February 2023 - Kernel - mailweb.openeuler.org

[PATCH openEuler-1.0-LTS 1/2] scsi: iscsi_tcp: Fix UAF during logout when accessing the shost ipaddress
by Yongqiang Liu 28 Feb '23

28 Feb '23

From: Mike Christie <michael.christie(a)oracle.com> mainline inclusion from mainline-v6.2-rc6~31 commit 6f1d64b13097e85abda0f91b5638000afc5f9a06 category: bugfix bugzilla: 188443, https://gitee.com/openeuler/kernel/issues/I6I8YD CVE: NA ---------------------------------------- Bug report and analysis from Ding Hui. During iSCSI session logout, if another task accesses the shost ipaddress attr, we can get a KASAN UAF report like this: [ 276.942144] BUG: KASAN: use-after-free in _raw_spin_lock_bh+0x78/0xe0 [ 276.942535] Write of size 4 at addr ffff8881053b45b8 by task cat/4088 [ 276.943511] CPU: 2 PID: 4088 Comm: cat Tainted: G E 6.1.0-rc8+ #3 [ 276.943997] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 11/12/2020 [ 276.944470] Call Trace: [ 276.944943] <TASK> [ 276.945397] dump_stack_lvl+0x34/0x48 [ 276.945887] print_address_description.constprop.0+0x86/0x1e7 [ 276.946421] print_report+0x36/0x4f [ 276.947358] kasan_report+0xad/0x130 [ 276.948234] kasan_check_range+0x35/0x1c0 [ 276.948674] _raw_spin_lock_bh+0x78/0xe0 [ 276.949989] iscsi_sw_tcp_host_get_param+0xad/0x2e0 [iscsi_tcp] [ 276.951765] show_host_param_ISCSI_HOST_PARAM_IPADDRESS+0xe9/0x130 [scsi_transport_iscsi] [ 276.952185] dev_attr_show+0x3f/0x80 [ 276.953005] sysfs_kf_seq_show+0x1fb/0x3e0 [ 276.953401] seq_read_iter+0x402/0x1020 [ 276.954260] vfs_read+0x532/0x7b0 [ 276.955113] ksys_read+0xed/0x1c0 [ 276.955952] do_syscall_64+0x38/0x90 [ 276.956347] entry_SYSCALL_64_after_hwframe+0x63/0xcd [ 276.956769] RIP: 0033:0x7f5d3a679222 [ 276.957161] Code: c0 e9 b2 fe ff ff 50 48 8d 3d 32 c0 0b 00 e8 a5 fe 01 00 0f 1f 44 00 00 f3 0f 1e fa 64 8b 04 25 18 00 00 00 85 c0 75 10 0f 05 <48> 3d 00 f0 ff ff 77 56 c3 0f 1f 44 00 00 48 83 ec 28 48 89 54 24 [ 276.958009] RSP: 002b:00007ffc864d16a8 EFLAGS: 00000246 ORIG_RAX: 0000000000000000 [ 276.958431] RAX: ffffffffffffffda RBX: 0000000000020000 RCX: 00007f5d3a679222 [ 276.958857] RDX: 0000000000020000 RSI: 00007f5d3a4fe000 RDI: 0000000000000003 [ 276.959281] RBP: 00007f5d3a4fe000 R08: 00000000ffffffff R09: 0000000000000000 [ 276.959682] R10: 0000000000000022 R11: 0000000000000246 R12: 0000000000020000 [ 276.960126] R13: 0000000000000003 R14: 0000000000000000 R15: 0000557a26dada58 [ 276.960536] </TASK> [ 276.961357] Allocated by task 2209: [ 276.961756] kasan_save_stack+0x1e/0x40 [ 276.962170] kasan_set_track+0x21/0x30 [ 276.962557] __kasan_kmalloc+0x7e/0x90 [ 276.962923] __kmalloc+0x5b/0x140 [ 276.963308] iscsi_alloc_session+0x28/0x840 [scsi_transport_iscsi] [ 276.963712] iscsi_session_setup+0xda/0xba0 [libiscsi] [ 276.964078] iscsi_sw_tcp_session_create+0x1fd/0x330 [iscsi_tcp] [ 276.964431] iscsi_if_create_session.isra.0+0x50/0x260 [scsi_transport_iscsi] [ 276.964793] iscsi_if_recv_msg+0xc5a/0x2660 [scsi_transport_iscsi] [ 276.965153] iscsi_if_rx+0x198/0x4b0 [scsi_transport_iscsi] [ 276.965546] netlink_unicast+0x4d5/0x7b0 [ 276.965905] netlink_sendmsg+0x78d/0xc30 [ 276.966236] sock_sendmsg+0xe5/0x120 [ 276.966576] ____sys_sendmsg+0x5fe/0x860 [ 276.966923] ___sys_sendmsg+0xe0/0x170 [ 276.967300] __sys_sendmsg+0xc8/0x170 [ 276.967666] do_syscall_64+0x38/0x90 [ 276.968028] entry_SYSCALL_64_after_hwframe+0x63/0xcd [ 276.968773] Freed by task 2209: [ 276.969111] kasan_save_stack+0x1e/0x40 [ 276.969449] kasan_set_track+0x21/0x30 [ 276.969789] kasan_save_free_info+0x2a/0x50 [ 276.970146] __kasan_slab_free+0x106/0x190 [ 276.970470] __kmem_cache_free+0x133/0x270 [ 276.970816] device_release+0x98/0x210 [ 276.971145] kobject_cleanup+0x101/0x360 [ 276.971462] iscsi_session_teardown+0x3fb/0x530 [libiscsi] [ 276.971775] iscsi_sw_tcp_session_destroy+0xd8/0x130 [iscsi_tcp] [ 276.972143] iscsi_if_recv_msg+0x1bf1/0x2660 [scsi_transport_iscsi] [ 276.972485] iscsi_if_rx+0x198/0x4b0 [scsi_transport_iscsi] [ 276.972808] netlink_unicast+0x4d5/0x7b0 [ 276.973201] netlink_sendmsg+0x78d/0xc30 [ 276.973544] sock_sendmsg+0xe5/0x120 [ 276.973864] ____sys_sendmsg+0x5fe/0x860 [ 276.974248] ___sys_sendmsg+0xe0/0x170 [ 276.974583] __sys_sendmsg+0xc8/0x170 [ 276.974891] do_syscall_64+0x38/0x90 [ 276.975216] entry_SYSCALL_64_after_hwframe+0x63/0xcd We can easily reproduce by two tasks: 1. while :; do iscsiadm -m node --login; iscsiadm -m node --logout; done 2. while :; do cat \ /sys/devices/platform/host*/iscsi_host/host*/ipaddress; done iscsid | cat --------------------------------+--------------------------------------- |- iscsi_sw_tcp_session_destroy | |- iscsi_session_teardown | |- device_release | |- iscsi_session_release ||- dev_attr_show |- kfree | |- show_host_param_ | ISCSI_HOST_PARAM_IPADDRESS | |- iscsi_sw_tcp_host_get_param | |- r/w tcp_sw_host->session (UAF) |- iscsi_host_remove | |- iscsi_host_free | Fix the above bug by splitting the session removal into 2 parts: 1. removal from iSCSI class which includes sysfs and removal from host tracking. 2. freeing of session. During iscsi_tcp host and session removal we can remove the session from sysfs then remove the host from sysfs. At this point we know userspace is not accessing the kernel via sysfs so we can free the session and host. Link: https://lore.kernel.org/r/20230117193937.21244-2-michael.christie@oracle.com Signed-off-by: Mike Christie <michael.christie(a)oracle.com> Reviewed-by: Lee Duncan <lduncan(a)suse.com> Acked-by: Ding Hui <dinghui(a)sangfor.com.cn> Signed-off-by: Martin K. Petersen <martin.petersen(a)oracle.com> Signed-off-by: Wenchao Hao <haowenchao2(a)huawei.com> Signed-off-by: Zhong Jinghua <zhongjinghua(a)huawei.com> Reviewed-by: Hou Tao <houtao1(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- drivers/scsi/iscsi_tcp.c | 11 +++++++++-- drivers/scsi/libiscsi.c | 39 +++++++++++++++++++++++++++++++-------- include/scsi/libiscsi.h | 2 ++ 3 files changed, 42 insertions(+), 10 deletions(-) diff --git a/drivers/scsi/iscsi_tcp.c b/drivers/scsi/iscsi_tcp.c index 241b1a310519..a5259fd39d6d 100644 --- a/drivers/scsi/iscsi_tcp.c +++ b/drivers/scsi/iscsi_tcp.c @@ -910,10 +910,17 @@ static void iscsi_sw_tcp_session_destroy(struct iscsi_cls_session *cls_session) if (WARN_ON_ONCE(session->leadconn)) return; + iscsi_session_remove(cls_session); + /* + * Our get_host_param needs to access the session, so remove the + * host from sysfs before freeing the session to make sure userspace + * is no longer accessing the callout. + */ + iscsi_host_remove(shost); + iscsi_tcp_r2tpool_free(cls_session->dd_data); - iscsi_session_teardown(cls_session); - iscsi_host_remove(shost); + iscsi_session_free(cls_session); iscsi_host_free(shost); } diff --git a/drivers/scsi/libiscsi.c b/drivers/scsi/libiscsi.c index 9f625a4d53c0..72463874d7b4 100644 --- a/drivers/scsi/libiscsi.c +++ b/drivers/scsi/libiscsi.c @@ -3018,20 +3018,34 @@ iscsi_session_setup(struct iscsi_transport *iscsit, struct Scsi_Host *shost, } EXPORT_SYMBOL_GPL(iscsi_session_setup); +/* + * issi_session_remove - Remove session from iSCSI class. + */ +void iscsi_session_remove(struct iscsi_cls_session *cls_session) +{ + struct iscsi_session *session = cls_session->dd_data; + struct Scsi_Host *shost = session->host; + + iscsi_remove_session(cls_session); + /* + * host removal only has to wait for its children to be removed from + * sysfs, and iscsi_tcp needs to do iscsi_host_remove before freeing + * the session, so drop the session count here. + */ + iscsi_host_dec_session_cnt(shost); +} +EXPORT_SYMBOL_GPL(iscsi_session_remove); + /** - * iscsi_session_teardown - destroy session, host, and cls_session + * iscsi_session_free - Free iscsi session and it's resources * @cls_session: iscsi session */ -void iscsi_session_teardown(struct iscsi_cls_session *cls_session) +void iscsi_session_free(struct iscsi_cls_session *cls_session) { struct iscsi_session *session = cls_session->dd_data; struct module *owner = cls_session->transport->owner; - struct Scsi_Host *shost = session->host; iscsi_pool_free(&session->cmdpool); - - iscsi_remove_session(cls_session); - kfree(session->password); kfree(session->password_in); kfree(session->username); @@ -3047,10 +3061,19 @@ void iscsi_session_teardown(struct iscsi_cls_session *cls_session) kfree(session->discovery_parent_type); iscsi_free_session(cls_session); - - iscsi_host_dec_session_cnt(shost); module_put(owner); } +EXPORT_SYMBOL_GPL(iscsi_session_free); + +/** + * iscsi_session_teardown - destroy session and cls_session + * @cls_session: iscsi session + */ +void iscsi_session_teardown(struct iscsi_cls_session *cls_session) +{ + iscsi_session_remove(cls_session); + iscsi_session_free(cls_session); +} EXPORT_SYMBOL_GPL(iscsi_session_teardown); /** diff --git a/include/scsi/libiscsi.h b/include/scsi/libiscsi.h index 254e72b46d10..2a8d1de70290 100644 --- a/include/scsi/libiscsi.h +++ b/include/scsi/libiscsi.h @@ -425,6 +425,8 @@ extern int iscsi_host_get_max_scsi_cmds(struct Scsi_Host *shost, extern struct iscsi_cls_session * iscsi_session_setup(struct iscsi_transport *, struct Scsi_Host *shost, uint16_t, int, int, uint32_t, unsigned int); +void iscsi_session_remove(struct iscsi_cls_session *cls_session); +void iscsi_session_free(struct iscsi_cls_session *cls_session); extern void iscsi_session_teardown(struct iscsi_cls_session *); extern void iscsi_session_recovery_timedout(struct iscsi_cls_session *); extern int iscsi_set_param(struct iscsi_cls_conn *cls_conn, -- 2.25.1

1 1

[PATCH openEuler-1.0-LTS] pciehp: fix the problem that the slot is powered on again after being powered off
by jiazhenyuan＠uniontech.com 28 Feb '23

28 Feb '23

From: jiazhenyuan <jiazhenyuan(a)uniontech.com> mainline inclusion from mainline-4.19-lts commit 32a8cef274feacd00b748a4f13b84d60aa6d82ff category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6ICDL CVE: NA ---------------------------------------- The DISABLE_SLOT event is lost when the slot is powered off. Signed-off-by: jiazhenyuan <jiazhenyuan(a)uniontech.com> --- drivers/pci/hotplug/pciehp_ctrl.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/pci/hotplug/pciehp_ctrl.c b/drivers/pci/hotplug/pciehp_ctrl.c index 2d549c97ac42..dd67dc540279 100644 --- a/drivers/pci/hotplug/pciehp_ctrl.c +++ b/drivers/pci/hotplug/pciehp_ctrl.c @@ -463,6 +463,7 @@ int pciehp_sysfs_disable_slot(struct slot *p_slot) mutex_unlock(&p_slot->lock); pci_config_pm_runtime_get(pdev); down_read(&ctrl->reset_lock); + atomic_or(DISABLE_SLOT, &ctrl->pending_events); pciehp_handle_disable_request(p_slot); up_read(&ctrl->reset_lock); pci_config_pm_runtime_put(pdev); -- 2.27.0

1 0

[PATCH openEuler-1.0-LTS] net: mpls: fix stale pointer if allocation fails during device rename
by Yongqiang Liu 28 Feb '23

28 Feb '23

From: Jakub Kicinski <kuba(a)kernel.org> stable inclusion from stable-v4.19.273 commit aa07c86e43ed8780d610ecfb2ce13da326729201 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6HZHU CVE: CVE-2023-26545 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit fda6c89fe3d9aca073495a664e1d5aea28cd4377 upstream. lianhui reports that when MPLS fails to register the sysctl table under new location (during device rename) the old pointers won't get overwritten and may be freed again (double free). Handle this gracefully. The best option would be unregistering the MPLS from the device completely on failure, but unfortunately mpls_ifdown() can fail. So failing fully is also unreliable. Another option is to register the new table first then only remove old one if the new one succeeds. That requires more code, changes order of notifications and two tables may be visible at the same time. sysctl point is not used in the rest of the code - set to NULL on failures and skip unregister if already NULL. Reported-by: lianhui tang <bluetlh(a)gmail.com> Fixes: 0fae3bf018d9 ("mpls: handle device renames for per-device sysctls") Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Zhengchao Shao <shaozhengchao(a)huawei.com> Reviewed-by: Liu Jian <liujian56(a)huawei.com> Reviewed-by: Wang Weiyang <wangweiyang2(a)huawei.com> Reviewed-by: Yue Haibing <yuehaibing(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- net/mpls/af_mpls.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 7623d9aec636..c7fd387baa61 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -1375,6 +1375,7 @@ static int mpls_dev_sysctl_register(struct net_device *dev, free: kfree(table); out: + mdev->sysctl = NULL; return -ENOBUFS; } @@ -1384,6 +1385,9 @@ static void mpls_dev_sysctl_unregister(struct net_device *dev, struct net *net = dev_net(dev); struct ctl_table *table; + if (!mdev->sysctl) + return; + table = mdev->sysctl->ctl_table_arg; unregister_net_sysctl_table(mdev->sysctl); kfree(table); -- 2.25.1

1 0

[PATCH openEuler-5.10-LTS-SP1] net: mpls: fix stale pointer if allocation fails during device rename
by Jialin Zhang 28 Feb '23

28 Feb '23

From: Jakub Kicinski <kuba(a)kernel.org> stable inclusion from stable-v5.10.169 commit 7ff0fdba82298d1f456c685e24930da89703c0fb category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6HZHU CVE: CVE-2023-26545 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit fda6c89fe3d9aca073495a664e1d5aea28cd4377 upstream. lianhui reports that when MPLS fails to register the sysctl table under new location (during device rename) the old pointers won't get overwritten and may be freed again (double free). Handle this gracefully. The best option would be unregistering the MPLS from the device completely on failure, but unfortunately mpls_ifdown() can fail. So failing fully is also unreliable. Another option is to register the new table first then only remove old one if the new one succeeds. That requires more code, changes order of notifications and two tables may be visible at the same time. sysctl point is not used in the rest of the code - set to NULL on failures and skip unregister if already NULL. Reported-by: lianhui tang <bluetlh(a)gmail.com> Fixes: 0fae3bf018d9 ("mpls: handle device renames for per-device sysctls") Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Zhengchao Shao <shaozhengchao(a)huawei.com> Reviewed-by: Liu Jian <liujian56(a)huawei.com> Reviewed-by: Wang Weiyang <wangweiyang2(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- net/mpls/af_mpls.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 72398149e4d4..1dcbdab9319b 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -1427,6 +1427,7 @@ static int mpls_dev_sysctl_register(struct net_device *dev, free: kfree(table); out: + mdev->sysctl = NULL; return -ENOBUFS; } @@ -1436,6 +1437,9 @@ static void mpls_dev_sysctl_unregister(struct net_device *dev, struct net *net = dev_net(dev); struct ctl_table *table; + if (!mdev->sysctl) + return; + table = mdev->sysctl->ctl_table_arg; unregister_net_sysctl_table(mdev->sysctl); kfree(table); -- 2.25.1

1 0

[PATCH openEuler-5.10-LTS 01/50] Revert "[Huawei] io_uring:drop identity before creating a private one"
by Jialin Zhang 28 Feb '23

28 Feb '23

From: Li Lingfeng <lilingfeng3(a)huawei.com> Offering: HULK hulk inclusion category: feature bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6BTWC ------------------------------- This reverts commit ab45921343c63dc9d461740e23003098697e0333. We need to apply patch 788d0824269bef (io_uring: import 5.15-stable io_uring) to move io_uring to separate directory and solve the problem of CVE-2023-0240. This patch fix a uaf problem of io_identity, and it can be reverted since io_identity is removed in patch 788d0824269bef. Signed-off-by: Li Lingfeng <lilingfeng3(a)huawei.com> Reviewed-by: Zhang Yi <yi.zhang(a)huawei.com> Reviewed-by: Wang Weiyang <wangweiyang2(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- fs/io_uring.c | 42 ------------------------------------------ 1 file changed, 42 deletions(-) diff --git a/fs/io_uring.c b/fs/io_uring.c index 4ace89ae4832..2397c2a1d919 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -1364,47 +1364,6 @@ static bool io_identity_cow(struct io_kiocb *req) return true; } -static void io_drop_identity(struct io_kiocb *req) -{ - struct io_identity *id = req->work.identity; - - if (req->work.flags & IO_WQ_WORK_MM) { - mmdrop(id->mm); - req->work.flags &= ~IO_WQ_WORK_MM; - } -#ifdef CONFIG_BLK_CGROUP - if (req->work.flags & IO_WQ_WORK_BLKCG) { - css_put(id->blkcg_css); - req->work.flags &= ~IO_WQ_WORK_BLKCG; - } -#endif - if (req->work.flags & IO_WQ_WORK_CREDS) { - put_cred(id->creds); - req->work.flags &= ~IO_WQ_WORK_CREDS; - } - if (req->work.flags & IO_WQ_WORK_FILES) { - put_files_struct(req->work.identity->files); - put_nsproxy(req->work.identity->nsproxy); - req->work.flags &= ~IO_WQ_WORK_FILES; - } - if (req->work.flags & IO_WQ_WORK_CANCEL) - req->work.flags &= ~IO_WQ_WORK_CANCEL; - if (req->work.flags & IO_WQ_WORK_FS) { - struct fs_struct *fs = id->fs; - - spin_lock(&id->fs->lock); - if (--fs->users) - fs = NULL; - spin_unlock(&id->fs->lock); - - if (fs) - free_fs_struct(fs); - req->work.flags &= ~IO_WQ_WORK_FS; - } - if (req->work.flags & IO_WQ_WORK_FSIZE) - req->work.flags &= ~IO_WQ_WORK_FSIZE; -} - static bool io_grab_identity(struct io_kiocb *req) { const struct io_op_def *def = &io_op_defs[req->opcode]; @@ -1510,7 +1469,6 @@ static void io_prep_async_work(struct io_kiocb *req) if (io_grab_identity(req)) return; - io_drop_identity(req); if (!io_identity_cow(req)) return; -- 2.25.1

1 49

[PATCH openEuler-5.10-LTS-SP1 01/52] Revert "[Huawei] io_uring:drop identity before creating a private one"
by Jialin Zhang 28 Feb '23

28 Feb '23

From: Li Lingfeng <lilingfeng3(a)huawei.com> Offering: HULK hulk inclusion category: feature bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6BTWC ------------------------------- This reverts commit 6d9aaec1edb197ca7e05873c1930c418b7ecaa1e. We need to apply patch 788d0824269bef (io_uring: import 5.15-stable io_uring) to move io_uring to separate directory and solve the problem of CVE-2023-0240. This patch fix a uaf problem of io_identity, and it can be reverted since io_identity is removed in patch 788d0824269bef. Signed-off-by: Li Lingfeng <lilingfeng3(a)huawei.com> Reviewed-by: Zhang Yi <yi.zhang(a)huawei.com> Reviewed-by: Wang Weiyang <wangweiyang2(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- fs/io_uring.c | 42 ------------------------------------------ 1 file changed, 42 deletions(-) diff --git a/fs/io_uring.c b/fs/io_uring.c index 21adb7ff7b2e..adb8fcef738a 100644 --- a/fs/io_uring.c +++ b/fs/io_uring.c @@ -1364,47 +1364,6 @@ static bool io_identity_cow(struct io_kiocb *req) return true; } -static void io_drop_identity(struct io_kiocb *req) -{ - struct io_identity *id = req->work.identity; - - if (req->work.flags & IO_WQ_WORK_MM) { - mmdrop(id->mm); - req->work.flags &= ~IO_WQ_WORK_MM; - } -#ifdef CONFIG_BLK_CGROUP - if (req->work.flags & IO_WQ_WORK_BLKCG) { - css_put(id->blkcg_css); - req->work.flags &= ~IO_WQ_WORK_BLKCG; - } -#endif - if (req->work.flags & IO_WQ_WORK_CREDS) { - put_cred(id->creds); - req->work.flags &= ~IO_WQ_WORK_CREDS; - } - if (req->work.flags & IO_WQ_WORK_FILES) { - put_files_struct(req->work.identity->files); - put_nsproxy(req->work.identity->nsproxy); - req->work.flags &= ~IO_WQ_WORK_FILES; - } - if (req->work.flags & IO_WQ_WORK_CANCEL) - req->work.flags &= ~IO_WQ_WORK_CANCEL; - if (req->work.flags & IO_WQ_WORK_FS) { - struct fs_struct *fs = id->fs; - - spin_lock(&id->fs->lock); - if (--fs->users) - fs = NULL; - spin_unlock(&id->fs->lock); - - if (fs) - free_fs_struct(fs); - req->work.flags &= ~IO_WQ_WORK_FS; - } - if (req->work.flags & IO_WQ_WORK_FSIZE) - req->work.flags &= ~IO_WQ_WORK_FSIZE; -} - static bool io_grab_identity(struct io_kiocb *req) { const struct io_op_def *def = &io_op_defs[req->opcode]; @@ -1510,7 +1469,6 @@ static void io_prep_async_work(struct io_kiocb *req) if (io_grab_identity(req)) return; - io_drop_identity(req); if (!io_identity_cow(req)) return; -- 2.25.1

1 51

[PATCH openEuler-1.0-LTS 1/3] selinux: further adjust init order for file_alloc_security hook
by Yongqiang Liu 27 Feb '23

27 Feb '23

From: "GONG, Ruiqi" <gongruiqi1(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6DRJ1 CVE: NA ---------------------------------------- After backporting commit cfff75d8973a ("selinux: reorder hooks to make runtime disable less broken") to the 4.19 kernel of openEuler-1.0-LTS, another kernel panic was triggered by running the POC of the aforementioned commit: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 PGD 800000001840b067 P4D 800000001840b067 PUD 1840c067 PMD 0 Oops: 0002 [#1] SMP PTI CPU: 7 PID: 273 Comm: exe Tainted: G OE 4.19.90+ #6 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.15.0-0-g2dd4b9b3f840-prebuilt.qemu.org 04/01/2014 RIP: 0010:selinux_file_open+0x49/0xf0 Code: 00 00 00 65 48 8b 04 25 28 00 00 00 48 89 44 24 20 31 c0 4c 89 e7 e8 a6 ec ff ff 49 8b 44 24 38 48 c7 c7 e0 a5 13 97 8b 40 1c <89> 45 08 e8 6f 80 ff ff ba 02 00 00 00 89 45 0c 8b 43 44 8b 73 40 RSP: 0018:ffffbb7300867ba0 EFLAGS: 00010246 RAX: 0000000000000003 RBX: ffff9dc301961400 RCX: 00000000000081ed RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffffffff9713a5e0 RBP: 0000000000000000 R08: 0000000000000001 R09: ffff9dc301fedcb0 R10: 0000000000000007 R11: 7fffffffffffffff R12: ffff9dc30204fd70 R13: 0000000000000000 R14: ffff9dc301961410 R15: ffffbb7300867c70 FS: 0000000000d258c0(0000) GS:ffff9dc33e9c0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000008 CR3: 00000000022bc000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: ? generic_permission+0x10a/0x190 security_file_open+0x26/0x90 do_dentry_open+0xd9/0x380 do_last+0x197/0x8d0 path_openat+0x89/0x280 do_filp_open+0x91/0x100 do_open_execat+0x79/0x180 __do_execve_file.isra.0+0x6dd/0x8b0 __x64_sys_execve+0x35/0x40 do_syscall_64+0x63/0x250 ? async_page_fault+0x8/0x30 entry_SYSCALL_64_after_hwframe+0x44/0xa9 RIP: 0033:0x49a5db Code: 41 89 01 eb da 66 2e 0f 1f 84 00 00 00 00 00 f7 d8 64 41 89 01 eb d6 0f 1f 84 00 00 00 00 00 f3 0f 1e fa b8 3b 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 c7 c1 e0 ff ff ff f7 d8 64 89 01 48 RSP: 002b:00007ffe7b1cebd8 EFLAGS: 00000246 ORIG_RAX: 000000000000003b RAX: ffffffffffffffda RBX: 0000000000d27ee0 RCX: 000000000049a5db RDX: 0000000000d27f08 RSI: 0000000000d27ee0 RDI: 0000000000d27f48 RBP: 0000000000d27f48 R08: fefefefefefefeff R09: fefefeff666d686f R10: 0000000000d25b90 R11: 0000000000000246 R12: 0000000000d27f08 R13: 0000000000655894 R14: 0000000000d27f08 R15: 0000000000d26ed0 Modules linked in: e1000(OE) CR2: 0000000000000008 ---[ end trace e4eb884974c22e2d ]--- RIP: 0010:selinux_file_open+0x49/0xf0 Code: 00 00 00 65 48 8b 04 25 28 00 00 00 48 89 44 24 20 31 c0 4c 89 e7 e8 a6 ec ff ff 49 8b 44 24 38 48 c7 c7 e0 a5 13 97 8b 40 1c <89> 45 08 e8 6f 80 ff ff ba 02 00 00 00 89 45 0c 8b 43 44 8b 73 40 RSP: 0018:ffffbb7300867ba0 EFLAGS: 00010246 RAX: 0000000000000003 RBX: ffff9dc301961400 RCX: 00000000000081ed RDX: 0000000000000001 RSI: 0000000000000000 RDI: ffffffff9713a5e0 RBP: 0000000000000000 R08: 0000000000000001 R09: ffff9dc301fedcb0 R10: 0000000000000007 R11: 7fffffffffffffff R12: ffff9dc30204fd70 R13: 0000000000000000 R14: ffff9dc301961410 R15: ffffbb7300867c70 FS: 0000000000d258c0(0000) GS:ffff9dc33e9c0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000008 CR3: 00000000022bc000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Kernel panic - not syncing: Fatal exception Kernel Offset: 0x14400000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff) ---[ end Kernel panic - not syncing: Fatal exception ]--- The problem was caused by selinux_file_open() accessing a file's fsec being NULL, which indicated that the file_alloc_security hook should be deleted later (at least after the file_open hook) when disabling SELinux at runtime. Here I put it into the "allocating" part. Signed-off-by: GONG, Ruiqi <gongruiqi1(a)huawei.com> Reviewed-by: Wang Weiyang <wangweiyang2(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- security/selinux/hooks.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/security/selinux/hooks.c b/security/selinux/hooks.c index cd26c1199353..85ac12d6b2f4 100644 --- a/security/selinux/hooks.c +++ b/security/selinux/hooks.c @@ -7040,7 +7040,6 @@ static struct security_hook_list selinux_hooks[] __lsm_ro_after_init = { LSM_HOOK_INIT(inode_copy_up_xattr, selinux_inode_copy_up_xattr), LSM_HOOK_INIT(file_permission, selinux_file_permission), - LSM_HOOK_INIT(file_alloc_security, selinux_file_alloc_security), LSM_HOOK_INIT(file_free_security, selinux_file_free_security), LSM_HOOK_INIT(file_ioctl, selinux_file_ioctl), LSM_HOOK_INIT(mmap_file, selinux_mmap_file), @@ -7207,6 +7206,7 @@ static struct security_hook_list selinux_hooks[] __lsm_ro_after_init = { LSM_HOOK_INIT(shm_alloc_security, selinux_shm_alloc_security), LSM_HOOK_INIT(sb_alloc_security, selinux_sb_alloc_security), LSM_HOOK_INIT(inode_alloc_security, selinux_inode_alloc_security), + LSM_HOOK_INIT(file_alloc_security, selinux_file_alloc_security), LSM_HOOK_INIT(cred_alloc_blank, selinux_cred_alloc_blank), LSM_HOOK_INIT(sem_alloc_security, selinux_sem_alloc_security), LSM_HOOK_INIT(secid_to_secctx, selinux_secid_to_secctx), -- 2.25.1

1 2

[PATCH openEuler-1.0-LTS 1/1] nbd: fix assignment error for first_minor in nbd_dev_add
by Yongqiang Liu 25 Feb '23

25 Feb '23

From: Zhong Jinghua <zhongjinghua(a)huawei.com> hulk inclusion category: bugfix bugzilla: 188413, https://gitee.com/openeuler/kernel/issues/I6GWYG CVE: NA ---------------------------------------- A panic error is like below: nbd_genl_connect nbd_dev_add first_minor = index << part_shift; // index =-1 ... __device_add_disk blk_alloc_devt *devt = MKDEV(disk->major, disk->first_minor + part->partno); // part->partno = 0， first_minor = 11...110000 major is covered There, index < 0 will reassign an index, but here disk->first_minor is assigned -1 << part_shift. This causes to the creation of the device with the same major and minor device numbers each time the incoming index<0, and this will lead to creation of kobject failed: Warning: kobject_add_internal failed for 4095:1048544 with -EEXIST, don't try to register things with the same name in the same directory. Fix it by moving the first_minor assignment down to after getting the new index. Fixes: 01f7594e62e9 ("nbd: Fix use-after-free in blk_mq_free_rqs") Signed-off-by: Zhong Jinghua <zhongjinghua(a)huawei.com> Reviewed-by: Yu Kuai <yukuai3(a)huawei.com> Reviewed-by: Hou Tao <houtao1(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- drivers/block/nbd.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/block/nbd.c b/drivers/block/nbd.c index 8dbbd676d275..551b2f7da8ef 100644 --- a/drivers/block/nbd.c +++ b/drivers/block/nbd.c @@ -1706,7 +1706,6 @@ static int nbd_dev_add(int index) struct gendisk *disk; struct request_queue *q; int err = -ENOMEM; - int first_minor = index << part_shift; nbd = kzalloc(sizeof(struct nbd_device), GFP_KERNEL); if (!nbd) @@ -1770,7 +1769,7 @@ static int nbd_dev_add(int index) refcount_set(&nbd->refs, 1); INIT_LIST_HEAD(&nbd->list); disk->major = NBD_MAJOR; - disk->first_minor = first_minor; + disk->first_minor = index << part_shift; disk->fops = &nbd_fops; disk->private_data = nbd; sprintf(disk->disk_name, "nbd%d", index); -- 2.25.1

2 1

openEuler缺陷issue处理流程_修正
by zhangjialin (F) 23 Feb '23

23 Feb '23

1 0

openEuler缺陷issue处理流程图
by zhangjialin (F) 23 Feb '23

23 Feb '23

1 0

[PATCH openEuler-1.0-LTS 00/13] mainline bugfix backport
by Yongqiang Liu 23 Feb '23

23 Feb '23

*** BLURB HERE *** Andy Shevchenko (1): ipmi: use %*ph to print small buffer Dan Carpenter (1): evm: Fix a small race in init_desc() Eric Biggers (1): crypto: rsa-pkcs1pad - fix buffer overread in pkcs1pad_verify_complete() Greg Kroah-Hartman (1): iommu: Properly export iommu_group_get_for_dev() Herbert Xu (2): crypto: algif_skcipher - EBUSY on aio should be an error crypto: algif_skcipher - Use chunksize instead of blocksize Lubomir Rintel (1): component: do not dereference opaque pointer in debugfs Nishka Dasgupta (2): of: unittest: Add of_node_put() before return of: resolver: Add of_node_put() before return and break Ondrej Mosnacek (1): selinux: reorder hooks to make runtime disable less broken Roberto Sassu (1): evm: Check also if *tfm is an error pointer in init_desc() Will Deacon (2): drivers/iommu: Export core IOMMU API symbols to permit modular drivers drivers/iommu: Allow IOMMU bus ops to be unregistered crypto/algif_skcipher.c | 4 +- crypto/rsa-pkcs1pad.c | 2 + drivers/base/component.c | 8 +-- drivers/char/ipmi/ipmi_msghandler.c | 27 ++------- drivers/iommu/iommu-sysfs.c | 5 ++ drivers/iommu/iommu.c | 12 ++++ drivers/of/resolver.c | 12 +++- drivers/of/unittest.c | 4 +- security/integrity/evm/evm_crypto.c | 45 +++++++------- security/selinux/hooks.c | 93 ++++++++++++++++++++--------- 10 files changed, 129 insertions(+), 83 deletions(-) -- 2.25.1

1 13

openEuler缺陷issue处理流程
by zhangjialin (F) 23 Feb '23

23 Feb '23

1 0

议题征集//答复: openEuler Kernel SIG双周例会
by liaotao (C) 23 Feb '23

23 Feb '23

当前例会议题：议题一：进展update（10min） -- 张伽琳 & 郑增凯议题二：openeuler 实时内核（PREEMPT_RT）规划讨论 --丁翔欢迎大家继续申报~ -----原始约会----- 发件人: openEuler conference <public(a)openeuler.org> 发送时间: 2023年2月21日 10:06 收件人: dev@openeuler.org,kernel-discuss@openeuler.org,kernel@openeuler.org 主题: openEuler Kernel SIG双周例会时间: 2023年2月24日星期五 14:00-15:30(UTC+08:00) 北京，重庆，香港特别行政区，乌鲁木齐。地点: 您好！ Kernel SIG 邀请您参加 2023-02-24 14:00 召开的Zoom会议(自动录制) 会议主题：openEuler Kernel SIG双周例会会议内容：欢迎您参加Kernel SIG双周例会，当前议题： 1. 进展update 2. 议题征集中欢迎大家积极申报议题（新增议题可以直接回复邮件，或录入会议看板）会议链接：https://us06web.zoom.us/j/82111503728?pwd=TWRGa2REc0V2ZmlGZ2ZreUF1OXA3dz09 会议纪要：https://etherpad.openeuler.org/p/Kernel-meetings 温馨提醒：建议接入会议后修改参会人的姓名，也可以使用您在gitee.com的ID 更多资讯尽在：https://openeuler.org/zh/ Hello! openEuler Kernel SIG invites you to attend the Zoom conference(auto recording) will be held at 2023-02-24 14:00, The subject of the conference is openEuler Kernel SIG双周例会, Summary: 欢迎您参加Kernel SIG双周例会，当前议题： 1. 进展update 2. 议题征集中欢迎大家积极申报议题（新增议题可以直接回复邮件，或录入会议看板） You can join the meeting at https://us06web.zoom.us/j/82111503728?pwd=TWRGa2REc0V2ZmlGZ2ZreUF1OXA3dz09. Add topics at https://etherpad.openeuler.org/p/Kernel-meetings. Note: You are advised to change the participant name after joining the conference or use your ID at gitee.com. More information: https://openeuler.org/en/

1 0

[PATCH openEuler-1.0-LTS] dhugetlb: isolate hwpoison hugepage when release
by Yongqiang Liu 23 Feb '23

23 Feb '23

From: Liu Shixin <liushixin2(a)huawei.com> hulk inclusion category: bugfix bugzilla: 46904, https://gitee.com/openeuler/kernel/issues/I6GSKP CVE: NA -------------------------------- For a hwpoison hugetlb page, the page will be freed firstly. If succeed, it will be dissolved and released to buddy system, then isolate the hwpoison page. For a hwpoison hugepage belong to dynamic hugetlb, we isolate the hugepage without dissolve it. Add a check in free_huge_page_to_dhugetlb_pool() to isolate the hwpoison hugepage directly. And keep HUGETLB_PAGE_DTOR after free to ensure the PageHuge() check return true in dissolve_free_huge_page(). Fixes: 0f0535e57da("dhugetlb: skip dissolve hugepage belonging to dynamic hugetlb") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- mm/hugetlb.c | 4 +++- 1 file changed, 3 insertions(+), 1 deletion(-) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index ed89df6fc5de..9f974270c84d 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -1348,8 +1348,9 @@ static void free_huge_page_to_dhugetlb_pool(struct page *page, } spin_lock(&hpool->lock); + if (PageHWPoison(page)) + goto out; ClearPagePool(page); - set_compound_page_dtor(page, NULL_COMPOUND_DTOR); if (!hstate_is_gigantic(h)) { list_add(&page->lru, &hpool->dhugetlb_2M_freelists); hpool->free_reserved_2M++; @@ -1375,6 +1376,7 @@ static void free_huge_page_to_dhugetlb_pool(struct page *page, trace_dhugetlb_alloc_free(hpool, page, hpool->free_reserved_1G, DHUGETLB_FREE_1G); } +out: spin_unlock(&hpool->lock); dhugetlb_pool_put(hpool); } -- 2.25.1

1 0

[PATCH openEuler-5.10-LTS-SP1 01/29] media: vivid: fix compose size exceed boundary
by Jialin Zhang 22 Feb '23

22 Feb '23

From: Liu Shixin <liushixin2(a)huawei.com> stable inclusion from stable-v5.10.163 commit f9d19f3a044ca651b0be52a4bf951ffe74259b9f category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6CIF8 CVE: CVE-2023-0615 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 94a7ad9283464b75b12516c5512541d467cefcf8 ] syzkaller found a bug: BUG: unable to handle page fault for address: ffffc9000a3b1000 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 100000067 P4D 100000067 PUD 10015f067 PMD 1121ca067 PTE 0 Oops: 0002 [#1] PREEMPT SMP CPU: 0 PID: 23489 Comm: vivid-000-vid-c Not tainted 6.1.0-rc1+ #512 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014 RIP: 0010:memcpy_erms+0x6/0x10 [...] Call Trace: <TASK> ? tpg_fill_plane_buffer+0x856/0x15b0 vivid_fillbuff+0x8ac/0x1110 vivid_thread_vid_cap_tick+0x361/0xc90 vivid_thread_vid_cap+0x21a/0x3a0 kthread+0x143/0x180 ret_from_fork+0x1f/0x30 </TASK> This is because we forget to check boundary after adjust compose->height int V4L2_SEL_TGT_CROP case. Add v4l2_rect_map_inside() to fix this problem for this case. Fixes: ef834f7836ec ("[media] vivid: add the video capture and output parts") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> Signed-off-by: Hans Verkuil <hverkuil-cisco(a)xs4all.nl> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Wang Hai <wanghai38(a)huawei.com> Signed-off-by: Longlong Xia <xialonglong1(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- drivers/media/test-drivers/vivid/vivid-vid-cap.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/media/test-drivers/vivid/vivid-vid-cap.c b/drivers/media/test-drivers/vivid/vivid-vid-cap.c index eadf28ab1e39..eeb0aeb62f79 100644 --- a/drivers/media/test-drivers/vivid/vivid-vid-cap.c +++ b/drivers/media/test-drivers/vivid/vivid-vid-cap.c @@ -953,6 +953,7 @@ int vivid_vid_cap_s_selection(struct file *file, void *fh, struct v4l2_selection if (dev->has_compose_cap) { v4l2_rect_set_min_size(compose, &min_rect); v4l2_rect_set_max_size(compose, &max_rect); + v4l2_rect_map_inside(compose, &fmt); } dev->fmt_cap_rect = fmt; tpg_s_buf_height(&dev->tpg, fmt.height); -- 2.25.1

1 28

[PATCH openEuler-5.10-LTS 01/18] media: vivid: fix compose size exceed boundary
by Jialin Zhang 22 Feb '23

22 Feb '23

From: Liu Shixin <liushixin2(a)huawei.com> stable inclusion from stable-v5.10.163 commit f9d19f3a044ca651b0be52a4bf951ffe74259b9f category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6CIF8 CVE: CVE-2023-0615 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 94a7ad9283464b75b12516c5512541d467cefcf8 ] syzkaller found a bug: BUG: unable to handle page fault for address: ffffc9000a3b1000 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 100000067 P4D 100000067 PUD 10015f067 PMD 1121ca067 PTE 0 Oops: 0002 [#1] PREEMPT SMP CPU: 0 PID: 23489 Comm: vivid-000-vid-c Not tainted 6.1.0-rc1+ #512 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014 RIP: 0010:memcpy_erms+0x6/0x10 [...] Call Trace: <TASK> ? tpg_fill_plane_buffer+0x856/0x15b0 vivid_fillbuff+0x8ac/0x1110 vivid_thread_vid_cap_tick+0x361/0xc90 vivid_thread_vid_cap+0x21a/0x3a0 kthread+0x143/0x180 ret_from_fork+0x1f/0x30 </TASK> This is because we forget to check boundary after adjust compose->height int V4L2_SEL_TGT_CROP case. Add v4l2_rect_map_inside() to fix this problem for this case. Fixes: ef834f7836ec ("[media] vivid: add the video capture and output parts") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> Signed-off-by: Hans Verkuil <hverkuil-cisco(a)xs4all.nl> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Wang Hai <wanghai38(a)huawei.com> Signed-off-by: Longlong Xia <xialonglong1(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- drivers/media/test-drivers/vivid/vivid-vid-cap.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/media/test-drivers/vivid/vivid-vid-cap.c b/drivers/media/test-drivers/vivid/vivid-vid-cap.c index eadf28ab1e39..eeb0aeb62f79 100644 --- a/drivers/media/test-drivers/vivid/vivid-vid-cap.c +++ b/drivers/media/test-drivers/vivid/vivid-vid-cap.c @@ -953,6 +953,7 @@ int vivid_vid_cap_s_selection(struct file *file, void *fh, struct v4l2_selection if (dev->has_compose_cap) { v4l2_rect_set_min_size(compose, &min_rect); v4l2_rect_set_max_size(compose, &max_rect); + v4l2_rect_map_inside(compose, &fmt); } dev->fmt_cap_rect = fmt; tpg_s_buf_height(&dev->tpg, fmt.height); -- 2.25.1

1 17

[PATCH openEuler-5.10-LTS 01/18] media: vivid: fix compose size exceed boundary
by Jialin Zhang 22 Feb '23

22 Feb '23

From: Liu Shixin <liushixin2(a)huawei.com> stable inclusion from stable-v5.10.163 commit f9d19f3a044ca651b0be52a4bf951ffe74259b9f category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6CIF8 CVE: CVE-2023-0615 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 94a7ad9283464b75b12516c5512541d467cefcf8 ] syzkaller found a bug: BUG: unable to handle page fault for address: ffffc9000a3b1000 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 100000067 P4D 100000067 PUD 10015f067 PMD 1121ca067 PTE 0 Oops: 0002 [#1] PREEMPT SMP CPU: 0 PID: 23489 Comm: vivid-000-vid-c Not tainted 6.1.0-rc1+ #512 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014 RIP: 0010:memcpy_erms+0x6/0x10 [...] Call Trace: <TASK> ? tpg_fill_plane_buffer+0x856/0x15b0 vivid_fillbuff+0x8ac/0x1110 vivid_thread_vid_cap_tick+0x361/0xc90 vivid_thread_vid_cap+0x21a/0x3a0 kthread+0x143/0x180 ret_from_fork+0x1f/0x30 </TASK> This is because we forget to check boundary after adjust compose->height int V4L2_SEL_TGT_CROP case. Add v4l2_rect_map_inside() to fix this problem for this case. Fixes: ef834f7836ec ("[media] vivid: add the video capture and output parts") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> Signed-off-by: Hans Verkuil <hverkuil-cisco(a)xs4all.nl> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Wang Hai <wanghai38(a)huawei.com> Signed-off-by: Longlong Xia <xialonglong1(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- drivers/media/test-drivers/vivid/vivid-vid-cap.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/media/test-drivers/vivid/vivid-vid-cap.c b/drivers/media/test-drivers/vivid/vivid-vid-cap.c index eadf28ab1e39..eeb0aeb62f79 100644 --- a/drivers/media/test-drivers/vivid/vivid-vid-cap.c +++ b/drivers/media/test-drivers/vivid/vivid-vid-cap.c @@ -953,6 +953,7 @@ int vivid_vid_cap_s_selection(struct file *file, void *fh, struct v4l2_selection if (dev->has_compose_cap) { v4l2_rect_set_min_size(compose, &min_rect); v4l2_rect_set_max_size(compose, &max_rect); + v4l2_rect_map_inside(compose, &fmt); } dev->fmt_cap_rect = fmt; tpg_s_buf_height(&dev->tpg, fmt.height); -- 2.25.1

1 17

[PATCH openEuler-5.10-LTS-SP1 01/29] media: vivid: fix compose size exceed boundary
by Jialin Zhang 22 Feb '23

22 Feb '23

From: Liu Shixin <liushixin2(a)huawei.com> stable inclusion from stable-v5.10.163 commit f9d19f3a044ca651b0be52a4bf951ffe74259b9f category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6CIF8 CVE: CVE-2023-0615 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 94a7ad9283464b75b12516c5512541d467cefcf8 ] syzkaller found a bug: BUG: unable to handle page fault for address: ffffc9000a3b1000 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 100000067 P4D 100000067 PUD 10015f067 PMD 1121ca067 PTE 0 Oops: 0002 [#1] PREEMPT SMP CPU: 0 PID: 23489 Comm: vivid-000-vid-c Not tainted 6.1.0-rc1+ #512 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014 RIP: 0010:memcpy_erms+0x6/0x10 [...] Call Trace: <TASK> ? tpg_fill_plane_buffer+0x856/0x15b0 vivid_fillbuff+0x8ac/0x1110 vivid_thread_vid_cap_tick+0x361/0xc90 vivid_thread_vid_cap+0x21a/0x3a0 kthread+0x143/0x180 ret_from_fork+0x1f/0x30 </TASK> This is because we forget to check boundary after adjust compose->height int V4L2_SEL_TGT_CROP case. Add v4l2_rect_map_inside() to fix this problem for this case. Fixes: ef834f7836ec ("[media] vivid: add the video capture and output parts") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> Signed-off-by: Hans Verkuil <hverkuil-cisco(a)xs4all.nl> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Wang Hai <wanghai38(a)huawei.com> Signed-off-by: Longlong Xia <xialonglong1(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- drivers/media/test-drivers/vivid/vivid-vid-cap.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/media/test-drivers/vivid/vivid-vid-cap.c b/drivers/media/test-drivers/vivid/vivid-vid-cap.c index eadf28ab1e39..eeb0aeb62f79 100644 --- a/drivers/media/test-drivers/vivid/vivid-vid-cap.c +++ b/drivers/media/test-drivers/vivid/vivid-vid-cap.c @@ -953,6 +953,7 @@ int vivid_vid_cap_s_selection(struct file *file, void *fh, struct v4l2_selection if (dev->has_compose_cap) { v4l2_rect_set_min_size(compose, &min_rect); v4l2_rect_set_max_size(compose, &max_rect); + v4l2_rect_map_inside(compose, &fmt); } dev->fmt_cap_rect = fmt; tpg_s_buf_height(&dev->tpg, fmt.height); -- 2.25.1

1 28

[PATCH openEuler-1.0-LTS] mm/sharepool: Fix null-pointer-deference in sp_free_area
by Yongqiang Liu 22 Feb '23

22 Feb '23

From: Wang Wensheng <wangwensheng4(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6G76L CVE: NA ---------------------------------------------- When a process is deleted from a group, the process does not apply for memory from the shared group. Otherwise, the UAF problem occurs. We checked this, but it didn't do a good job of preventing sp_alloc and del_task concurrency. The process applies for memory after passing the check, which violates our requirements and causes problems. The solution is to place the checked code in the critical area to ensure that no memory can be allocated after the check is passed. [ T7596] Unable to handle kernel NULL pointer dereference at virtual address 0000000000000098 [ T7596] Mem abort info: [ T7596] ESR = 0x96000004 [ T7596] EC = 0x25: DABT (current EL), IL = 32 bits [ T7596] SET = 0, FnV = 0 [ T7596] EA = 0, S1PTW = 0 [ T7596] Data abort info: [ T7596] ISV = 0, ISS = 0x00000004 [ T7596] CM = 0, WnR = 0 [ T7596] user pgtable: 4k pages, 48-bit VAs, pgdp=00000001040a3000 [ T7596] [0000000000000098] pgd=0000000000000000, p4d=0000000000000000 [ T7596] Internal error: Oops: 96000004 [#1] SMP [ T7596] Modules linked in: sharepool_dev(OE) [last unloaded: demo] [ T7596] CPU: 1 PID: 7596 Comm: test_sp_group_d Tainted: G OE 5.10.0+ #8 [ T7596] Hardware name: linux,dummy-virt (DT) [ T7596] pstate: 20000005 (nzCv daif -PAN -UAO -TCO BTYPE=--) [ T7596] pc : sp_free_area+0x34/0x120 [ T7596] lr : sp_free_area+0x30/0x120 [ T7596] sp : ffff80001c6a3b20 [ T7596] x29: ffff80001c6a3b20 x28: 0000000000000009 [ T7596] x27: 0000000000000000 x26: ffff800011c49d20 [ T7596] x25: ffff0000c227f6c0 x24: 0000000000000008 [ T7596] x23: ffff0000c0cf0ce8 x22: 0000000000000001 [ T7596] x21: ffff0000c4082b30 x20: 0000000000000000 [ T7596] x19: ffff0000c4082b00 x18: 0000000000000000 [ T7596] x17: 0000000000000000 x16: 0000000000000000 [ T7596] x15: 0000000000000000 x14: 0000000000000000 [ T7596] x13: 0000000000000000 x12: ffff0005fffe12c0 [ T7596] x11: 0000000000000008 x10: ffff0005fffe12c0 [ T7596] x9 : ffff8000103eb690 x8 : 0000000000000001 [ T7596] x7 : 0000000000210d00 x6 : 0000000000000000 [ T7596] x5 : ffff8000123edea0 x4 : 0000000000000030 [ T7596] x3 : ffffeff000000000 x2 : 0000eff000000000 [ T7596] x1 : 0000e80000000000 x0 : 0000000000000000 [ T7596] Call trace: [ T7596] sp_free_area+0x34/0x120 [ T7596] __sp_area_drop_locked+0x3c/0x60 [ T7596] sp_area_drop+0x80/0xbc [ T7596] remove_vma+0x54/0x70 [ T7596] exit_mmap+0x114/0x1d0 [ T7596] mmput+0x90/0x1ec [ T7596] exit_mm+0x1d0/0x2f0 [ T7596] do_exit+0x180/0x400 [ T7596] do_group_exit+0x40/0x114 [ T7596] get_signal+0x1e8/0x720 [ T7596] do_signal+0x11c/0x1e4 [ T7596] do_notify_resume+0x15c/0x250 [ T7596] work_pending+0xc/0x6d8 [ T7596] Code: f9400001 f9402c00 97fff0e5 aa0003f4 (f9404c00) [ T7596] ---[ end trace 3c8368d77e758ebd ]--- Signed-off-by: Wang Wensheng <wangwensheng4(a)huawei.com> Reviewed-by: Weilong Chen <chenweilong(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- mm/share_pool.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) diff --git a/mm/share_pool.c b/mm/share_pool.c index 5d8344fe805a..a136c70b7588 100644 --- a/mm/share_pool.c +++ b/mm/share_pool.c @@ -1712,14 +1712,6 @@ int mg_sp_group_del_task(int pid, int spg_id) ret = -EINVAL; goto out; } - - if (!list_empty(&spg->spa_list)) { - up_write(&sp_group_sem); - pr_err_ratelimited("spa is not empty"); - ret = -EINVAL; - goto out; - } - ret = get_task(pid, &tsk); if (ret) { up_write(&sp_group_sem); @@ -1743,6 +1735,15 @@ int mg_sp_group_del_task(int pid, int spg_id) } down_write(&spg->rw_lock); + + if (!list_empty(&spg->spa_list)) { + up_write(&spg->rw_lock); + up_write(&sp_group_sem); + pr_err_ratelimited("spa is not empty"); + ret = -EINVAL; + goto out_put_mm; + } + if (list_is_singular(&spg->procs)) is_alive = spg->is_alive = false; spg->proc_num--; -- 2.25.1

1 0

openEuler Kernel SIG双周例会
by openEuler conference 21 Feb '23

21 Feb '23

您好！ Kernel SIG 邀请您参加 2023-02-24 14:00 召开的Zoom会议(自动录制) 会议主题：openEuler Kernel SIG双周例会会议内容：欢迎您参加Kernel SIG双周例会，当前议题： 1. 进展update 2. 议题征集中欢迎大家积极申报议题（新增议题可以直接回复邮件，或录入会议看板）会议链接：https://us06web.zoom.us/j/82111503728?pwd=TWRGa2REc0V2ZmlGZ2ZreUF1OXA3dz09 会议纪要：https://etherpad.openeuler.org/p/Kernel-meetings 温馨提醒：建议接入会议后修改参会人的姓名，也可以使用您在gitee.com的ID 更多资讯尽在：https://openeuler.org/zh/ Hello! openEuler Kernel SIG invites you to attend the Zoom conference(auto recording) will be held at 2023-02-24 14:00, The subject of the conference is openEuler Kernel SIG双周例会, Summary: 欢迎您参加Kernel SIG双周例会，当前议题： 1. 进展update 2. 议题征集中欢迎大家积极申报议题（新增议题可以直接回复邮件，或录入会议看板） You can join the meeting at https://us06web.zoom.us/j/82111503728?pwd=TWRGa2REc0V2ZmlGZ2ZreUF1OXA3dz09. Add topics at https://etherpad.openeuler.org/p/Kernel-meetings. Note: You are advised to change the participant name after joining the conference or use your ID at gitee.com. More information: https://openeuler.org/en/

1 0

Intel Arch SIG 例会
by openEuler conference 20 Feb '23

20 Feb '23

您好！ sig-Intel-Arch SIG 邀请您参加 2023-02-21 10:00 召开的Zoom会议会议主题：Intel Arch SIG 例会会议内容： 1. Intel几个复杂特性的支持计划 2. QEMU和kernel选型对未来lntel平台支持和release安排讨论 3. openEuler在SPR上的用户诉求和反馈会议链接：https://us06web.zoom.us/j/87058792984?pwd=UFlzSzA4OHNyMEIvRkpCdWFwdEdRQT09 会议纪要：https://etherpad.openeuler.org/p/sig-Intel-Arch-meetings 温馨提醒：建议接入会议后修改参会人的姓名，也可以使用您在gitee.com的ID 更多资讯尽在：https://openeuler.org/zh/ Hello! openEuler sig-Intel-Arch SIG invites you to attend the Zoom conference will be held at 2023-02-21 10:00, The subject of the conference is Intel Arch SIG 例会, Summary: 1. Intel几个复杂特性的支持计划 2. QEMU和kernel选型对未来lntel平台支持和release安排讨论 3. openEuler在SPR上的用户诉求和反馈 You can join the meeting at https://us06web.zoom.us/j/87058792984?pwd=UFlzSzA4OHNyMEIvRkpCdWFwdEdRQT09. Add topics at https://etherpad.openeuler.org/p/sig-Intel-Arch-meetings. Note: You are advised to change the participant name after joining the conference or use your ID at gitee.com. More information: https://openeuler.org/en/

1 0

None
by guofeng.yang＠amicro.com.cn 20 Feb '23

20 Feb '23

kernel-leave(a)openeuler.org Best Regards 嵌入式软件开发部杨国锋珠海一微半导体股份有限公司 / Amicro Semiconductor Co., Ltd. 横琴总部中心：珠海市横琴新区ICC横琴国际商务中心2座27楼香洲研发中心：香洲区红山路26号阳光大厦13楼深圳办公中心：深圳市南山区国家工程实验室大楼A座1102室电话：0756-2666456 邮编：519000 Phone：18814181955 E-Mail：guofeng.yang(a)amicro.com.cn This email may contain confidential and/or privileged information from Amicro,and is intended solely for the attention and use of the person(s) named above. If you are not the intended recipient (or have received this email in error), please notify the sender immediately and destroy this email. Any unauthorized copying, disclosure or distribution of the material in this email is strictly forbidden. The content provided in this email can not be guaranteed and assured to be accurate, appropriate for all, and complete by Amicro, and Amicro can not be held responsible forany error or negligence derived therefrom. (此电子邮件包含来自一微半导体的信息，而且是机密的或者专用的信息。这些信息是供所有以上列出的个人或者团体使用的。如果您不是此邮件的预期收件人，请勿阅读、复制、转发或存储此邮件。如果已误收此邮件，请通知发件人。本公司不担保本电子邮件中信息的准确性、适当性或完整性，并且对此产生的任何错误或疏忽不承担任何责任.)

1 0

[PATCH openEuler-1.0-LTS 1/5] mm/memcg_memfs_info: fix potential oom_lock recursion deadlock
by Yongqiang Liu 16 Feb '23

16 Feb '23

From: Liu Shixin <liushixin2(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6AXGS CVE: NA -------------------------------- syzbot is reporting GFP_KERNEL allocation with oom_lock held when reporting memcg OOM [1]. If this allocation triggers the global OOM situation then the system can livelock because the GFP_KERNEL allocation with oom_lock held cannot trigger the global OOM killer because __alloc_pages_may_oom() fails to hold oom_lock. The problem mentioned above has been fixed by patch[2]. The is the same problem in memcg_memfs_info feature too. Refer to the patch[2], fix it by removing the allocation from mem_cgroup_print_memfs_info() completely, and pass static buffer when calling from memcg OOM path. Link: https://syzkaller.appspot.com/bug?extid=2d2aeadc6ce1e1f11d45 [1] Link: https://lkml.kernel.org/r/86afb39f-8c65-bec2-6cfc-c5e3cd600c0b@I-love.SAKUR… [2] Fixes: 6b1d4d3a3713 ("mm/memcg_memfs_info: show files that having pages charged in mem_cgroup") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- include/linux/memcg_memfs_info.h | 4 +++- mm/memcg_memfs_info.c | 20 ++++++++++---------- mm/memcontrol.c | 3 ++- 3 files changed, 15 insertions(+), 12 deletions(-) diff --git a/include/linux/memcg_memfs_info.h b/include/linux/memcg_memfs_info.h index 658a91e22bd7..b5e3709baa9e 100644 --- a/include/linux/memcg_memfs_info.h +++ b/include/linux/memcg_memfs_info.h @@ -6,11 +6,13 @@ #include <linux/seq_file.h> #ifdef CONFIG_MEMCG_MEMFS_INFO -void mem_cgroup_print_memfs_info(struct mem_cgroup *memcg, struct seq_file *m); +void mem_cgroup_print_memfs_info(struct mem_cgroup *memcg, char *pathbuf, + struct seq_file *m); int mem_cgroup_memfs_files_show(struct seq_file *m, void *v); void mem_cgroup_memfs_info_init(void); #else static inline void mem_cgroup_print_memfs_info(struct mem_cgroup *memcg, + char *pathbuf, struct seq_file *m) { } diff --git a/mm/memcg_memfs_info.c b/mm/memcg_memfs_info.c index 346175026cae..632e03da673b 100644 --- a/mm/memcg_memfs_info.c +++ b/mm/memcg_memfs_info.c @@ -162,7 +162,8 @@ static void memfs_show_files_in_mem_cgroup(struct super_block *sb, void *data) mntput(pfc->vfsmnt); } -void mem_cgroup_print_memfs_info(struct mem_cgroup *memcg, struct seq_file *m) +void mem_cgroup_print_memfs_info(struct mem_cgroup *memcg, char *pathbuf, + struct seq_file *m) { struct print_files_control pfc = { .memcg = memcg, @@ -170,17 +171,11 @@ void mem_cgroup_print_memfs_info(struct mem_cgroup *memcg, struct seq_file *m) .max_print_files = memfs_max_print_files, .size_threshold = memfs_size_threshold, }; - char *pathbuf; int i; if (!memfs_enable || !memcg) return; - pathbuf = kmalloc(PATH_MAX, GFP_KERNEL); - if (!pathbuf) { - SEQ_printf(m, "Show memfs failed due to OOM\n"); - return; - } pfc.pathbuf = pathbuf; pfc.pathbuf_size = PATH_MAX; @@ -197,15 +192,20 @@ void mem_cgroup_print_memfs_info(struct mem_cgroup *memcg, struct seq_file *m) SEQ_printf(m, "total files: %lu, total memory-size: %lukB\n", pfc.total_print_files, pfc.total_files_size >> 10); } - - kfree(pfc.pathbuf); } int mem_cgroup_memfs_files_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); + char *pathbuf; - mem_cgroup_print_memfs_info(memcg, m); + pathbuf = kmalloc(PATH_MAX, GFP_KERNEL); + if (!pathbuf) { + SEQ_printf(m, "Show memfs abort: failed to allocate memory\n"); + return 0; + } + mem_cgroup_print_memfs_info(memcg, pathbuf, m); + kfree(pathbuf); return 0; } diff --git a/mm/memcontrol.c b/mm/memcontrol.c index bdc90e6fc082..fd40fef49e45 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -1490,6 +1490,7 @@ void mem_cgroup_print_oom_meminfo(struct mem_cgroup *memcg) { struct mem_cgroup *iter; unsigned int i; + static char pathbuf[PATH_MAX]; pr_info("memory: usage %llukB, limit %llukB, failcnt %lu\n", K((u64)page_counter_read(&memcg->memory)), @@ -1522,7 +1523,7 @@ void mem_cgroup_print_oom_meminfo(struct mem_cgroup *memcg) pr_cont("\n"); } - mem_cgroup_print_memfs_info(memcg, NULL); + mem_cgroup_print_memfs_info(memcg, pathbuf, NULL); } /* -- 2.25.1

1 4

[PATCH openEuler-5.10-LTS-SP1 001/263] KVM: PPC: Book3S HV: Context tracking exit guest context before enabling irqs
by Jialin Zhang 15 Feb '23

15 Feb '23

From: Nicholas Piggin <npiggin(a)gmail.com> stable inclusion from stable-v5.10.145 commit 6bae47548188ae957578e6d92d4b8753dec435e4 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6D0VF Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 112665286d08c87e66d699e7cba43c1497ad165f ] Interrupts that occur in kernel mode expect that context tracking is set to kernel. Enabling local irqs before context tracking switches from guest to host means interrupts can come in and trigger warnings about wrong context, and possibly worse. Signed-off-by: Nicholas Piggin <npiggin(a)gmail.com> Signed-off-by: Michael Ellerman <mpe(a)ellerman.id.au> Link: https://lore.kernel.org/r/20210130130852.2952424-3-npiggin@gmail.com Stable-dep-of: 235cee162459 ("KVM: PPC: Tick accounting should defer vtime accounting 'til after IRQ handling") Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> Reviewed-by: Zheng Zengkai <zhengzengkai(a)huawei.com> --- arch/powerpc/kvm/book3s_hv.c | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) diff --git a/arch/powerpc/kvm/book3s_hv.c b/arch/powerpc/kvm/book3s_hv.c index 38b7a3491aac..d6c4e27f7ed9 100644 --- a/arch/powerpc/kvm/book3s_hv.c +++ b/arch/powerpc/kvm/book3s_hv.c @@ -3399,8 +3399,9 @@ static noinline void kvmppc_run_core(struct kvmppc_vcore *vc) kvmppc_set_host_core(pcpu); + guest_exit_irqoff(); + local_irq_enable(); - guest_exit(); /* Let secondaries go back to the offline loop */ for (i = 0; i < controlled_threads; ++i) { @@ -4235,8 +4236,9 @@ int kvmhv_run_single_vcpu(struct kvm_vcpu *vcpu, u64 time_limit, kvmppc_set_host_core(pcpu); + guest_exit_irqoff(); + local_irq_enable(); - guest_exit(); cpumask_clear_cpu(pcpu, &kvm->arch.cpu_in_guest); -- 2.25.1

1 262

[PATCH openEuler-1.0-LTS 1/2] net: bridge: mcast: add and enforce startup query interval minimum
by Yongqiang Liu 15 Feb '23

15 Feb '23

From: Nikolay Aleksandrov <nikolay(a)nvidia.com> mainline inclusion from mainline-v5.16-rc8 commit f83a112bd91a494cdee671aec74e777470fb4a07 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6DKZJ CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- As reported[1] if startup query interval is set too low in combination with large number of startup queries and we have multiple bridges or even a single bridge with multiple querier vlans configured we can crash the machine. Add a 1 second minimum which must be enforced by overwriting the value if set lower (i.e. without returning an error) to avoid breaking user-space. If that happens a log message is emitted to let the admin know that the startup interval has been set to the minimum. It doesn't make sense to make the startup interval lower than the normal query interval so use the same value of 1 second. The issue has been present since these intervals could be user-controlled. [1] https://lore.kernel.org/netdev/e8b9ce41-57b9-b6e2-a46a-ff9c791cf0ba@gmail.c… Fixes: d902eee43f19 ("bridge: Add multicast count/interval sysfs entries") Reported-by: Eric Dumazet <eric.dumazet(a)gmail.com> Signed-off-by: Nikolay Aleksandrov <nikolay(a)nvidia.com> Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Conflicts: net/bridge/br_multicast.c net/bridge/br_netlink.c net/bridge/br_private.h net/bridge/br_sysfs_br.c Signed-off-by: Zhengchao Shao <shaozhengchao(a)huawei.com> Reviewed-by: Yue Haibing <yuehaibing(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- net/bridge/br_multicast.c | 16 ++++++++++++++++ net/bridge/br_netlink.c | 2 +- net/bridge/br_private.h | 5 +++++ net/bridge/br_sysfs_br.c | 2 +- 4 files changed, 23 insertions(+), 2 deletions(-) diff --git a/net/bridge/br_multicast.c b/net/bridge/br_multicast.c index 3504c3acd38f..ed67a43efdef 100644 --- a/net/bridge/br_multicast.c +++ b/net/bridge/br_multicast.c @@ -2351,6 +2351,22 @@ int br_multicast_set_mld_version(struct net_bridge *br, unsigned long val) } #endif +void br_multicast_set_startup_query_intvl(struct net_bridge *br, + unsigned long val) +{ + unsigned long intvl_jiffies = clock_t_to_jiffies(val); + + if (intvl_jiffies < BR_MULTICAST_STARTUP_QUERY_INTVL_MIN) { + br_info(br, + "trying to set multicast startup query interval below minimum, setting to %lu (%ums)\n", + jiffies_to_clock_t(BR_MULTICAST_STARTUP_QUERY_INTVL_MIN), + jiffies_to_msecs(BR_MULTICAST_STARTUP_QUERY_INTVL_MIN)); + intvl_jiffies = BR_MULTICAST_STARTUP_QUERY_INTVL_MIN; + } + + br->multicast_startup_query_interval = intvl_jiffies; +} + /** * br_multicast_list_adjacent - Returns snooped multicast addresses * @dev: The bridge port adjacent to which to retrieve addresses diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c index c00cb376263a..b3bcfb77077a 100644 --- a/net/bridge/br_netlink.c +++ b/net/bridge/br_netlink.c @@ -1237,7 +1237,7 @@ static int br_changelink(struct net_device *brdev, struct nlattr *tb[], if (data[IFLA_BR_MCAST_STARTUP_QUERY_INTVL]) { u64 val = nla_get_u64(data[IFLA_BR_MCAST_STARTUP_QUERY_INTVL]); - br->multicast_startup_query_interval = clock_t_to_jiffies(val); + br_multicast_set_startup_query_intvl(br, val); } if (data[IFLA_BR_MCAST_STATS_ENABLED]) { diff --git a/net/bridge/br_private.h b/net/bridge/br_private.h index 23fed55942fa..877fbc65a2fe 100644 --- a/net/bridge/br_private.h +++ b/net/bridge/br_private.h @@ -31,6 +31,9 @@ #define BR_PORT_BITS 10 #define BR_MAX_PORTS (1<<BR_PORT_BITS) +#define BR_MULTICAST_QUERY_INTVL_MIN msecs_to_jiffies(1000) +#define BR_MULTICAST_STARTUP_QUERY_INTVL_MIN BR_MULTICAST_QUERY_INTVL_MIN + #define BR_VERSION "2.3" /* Control of forwarding link local multicast */ @@ -1200,4 +1203,6 @@ void br_do_proxy_suppress_arp(struct sk_buff *skb, struct net_bridge *br, void br_do_suppress_nd(struct sk_buff *skb, struct net_bridge *br, u16 vid, struct net_bridge_port *p, struct nd_msg *msg); struct nd_msg *br_is_nd_neigh_msg(struct sk_buff *skb, struct nd_msg *m); +void br_multicast_set_startup_query_intvl(struct net_bridge *br, + unsigned long val); #endif diff --git a/net/bridge/br_sysfs_br.c b/net/bridge/br_sysfs_br.c index 0318a69888d4..bfbffc5088ac 100644 --- a/net/bridge/br_sysfs_br.c +++ b/net/bridge/br_sysfs_br.c @@ -618,7 +618,7 @@ static ssize_t multicast_startup_query_interval_show( static int set_startup_query_interval(struct net_bridge *br, unsigned long val) { - br->multicast_startup_query_interval = clock_t_to_jiffies(val); + br_multicast_set_startup_query_intvl(br, val); return 0; } -- 2.25.1

1 1

[PATCH openEuler-5.10-LTS 1/6] cifs: sanitize multiple delimiters in prepath
by Jialin Zhang 14 Feb '23

14 Feb '23

From: Thiago Rafael Becker <trbecker(a)gmail.com> mainline inclusion from mainline-v5.16-rc6 commit a31080899d5fdafcccf7f39dd214a814a2c82626 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6D6RL CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- mount.cifs can pass a device with multiple delimiters in it. This will cause rename(2) to fail with ENOENT. V2: - Make sanitize_path more readable. - Fix multiple delimiters between UNC and prepath. - Avoid a memory leak if a bad user starts putting a lot of delimiters in the path on purpose. BugLink: https://bugzilla.redhat.com/show_bug.cgi?id=2031200 Fixes: 24e0a1eff9e2 ("cifs: switch to new mount api") Cc: stable(a)vger.kernel.org # 5.11+ Acked-by: Ronnie Sahlberg <lsahlber(a)redhat.com> Signed-off-by: Thiago Rafael Becker <trbecker(a)gmail.com> Signed-off-by: Steve French <stfrench(a)microsoft.com> conflicts: fs/cifs/fs_context.c Signed-off-by: ZhaoLong Wang <wangzhaolong1(a)huawei.com> Reviewed-by: Zhang Xiaoxu <zhangxiaoxu5(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- fs/cifs/connect.c | 38 +++++++++++++++++++++++++++++++++++++- 1 file changed, 37 insertions(+), 1 deletion(-) diff --git a/fs/cifs/connect.c b/fs/cifs/connect.c index 7f5d173760cf..0a2eced46e35 100644 --- a/fs/cifs/connect.c +++ b/fs/cifs/connect.c @@ -1316,6 +1316,42 @@ static int get_option_gid(substring_t args[], kgid_t *result) return 0; } +/* + * Remove duplicate path delimiters. Windows is supposed to do that + * but there are some bugs that prevent rename from working if there are + * multiple delimiters. + * + * Returns a sanitized duplicate of @path. The caller is responsible for + * cleaning up the original. + */ +#define IS_DELIM(c) ((c) == '/' || (c) == '\\') +static char *sanitize_path(char *path) +{ + char *cursor1 = path, *cursor2 = path; + + /* skip all prepended delimiters */ + while (IS_DELIM(*cursor1)) + cursor1++; + + /* copy the first letter */ + *cursor2 = *cursor1; + + /* copy the remainder... */ + while (*(cursor1++)) { + /* ... skipping all duplicated delimiters */ + if (IS_DELIM(*cursor1) && IS_DELIM(*cursor2)) + continue; + *(++cursor2) = *cursor1; + } + + /* if the last character is a delimiter, skip it */ + if (IS_DELIM(*(cursor2 - 1))) + cursor2--; + + *(cursor2) = '\0'; + return kstrdup(path, GFP_KERNEL); +} + /* * Parse a devname into substrings and populate the vol->UNC and vol->prepath * fields with the result. Returns 0 on success and an error otherwise. @@ -1364,7 +1400,7 @@ cifs_parse_devname(const char *devname, struct smb_vol *vol) if (!*pos) return 0; - vol->prepath = kstrdup(pos, GFP_KERNEL); + vol->prepath = sanitize_path(pos); if (!vol->prepath) return -ENOMEM; -- 2.25.1

1 5

[PATCH openEuler-5.10-LTS-SP1 1/8] bpf, sockmap: Fix sk->sk_forward_alloc warn_on in sk_stream_kill_queues
by Jialin Zhang 14 Feb '23

14 Feb '23

From: Wang Yufen <wangyufen(a)huawei.com> mainline inclusion from mainline-v6.0-rc1 commit d8616ee2affcff37c5d315310da557a694a3303d category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6BRTY CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- During TCP sockmap redirect pressure test, the following warning is triggered: WARNING: CPU: 3 PID: 2145 at net/core/stream.c:205 sk_stream_kill_queues+0xbc/0xd0 CPU: 3 PID: 2145 Comm: iperf Kdump: loaded Tainted: G W 5.10.0+ #9 Call Trace: inet_csk_destroy_sock+0x55/0x110 inet_csk_listen_stop+0xbb/0x380 tcp_close+0x41b/0x480 inet_release+0x42/0x80 __sock_release+0x3d/0xa0 sock_close+0x11/0x20 __fput+0x9d/0x240 task_work_run+0x62/0x90 exit_to_user_mode_prepare+0x110/0x120 syscall_exit_to_user_mode+0x27/0x190 entry_SYSCALL_64_after_hwframe+0x44/0xa9 The reason we observed is that: When the listener is closing, a connection may have completed the three-way handshake but not accepted, and the client has sent some packets. The child sks in accept queue release by inet_child_forget()->inet_csk_destroy_sock(), but psocks of child sks have not released. To fix, add sock_map_destroy to release psocks. Signed-off-by: Wang Yufen <wangyufen(a)huawei.com> Signed-off-by: Daniel Borkmann <daniel(a)iogearbox.net> Signed-off-by: Andrii Nakryiko <andrii(a)kernel.org> Acked-by: Jakub Sitnicki <jakub(a)cloudflare.com> Acked-by: John Fastabend <john.fastabend(a)gmail.com> Link: https://lore.kernel.org/bpf/20220524075311.649153-1-wangyufen@huawei.com Signed-off-by: Liu Jian <liujian56(a)huawei.com> Conflicts: include/linux/bpf.h Reviewed-by: Wang Yufen <wangyufen(a)huawei.com> Reviewed-by: Yue Haibing <yuehaibing(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- include/linux/bpf.h | 1 + include/linux/skmsg.h | 1 + net/core/skmsg.c | 1 + net/core/sock_map.c | 23 +++++++++++++++++++++++ net/ipv4/tcp_bpf.c | 1 + 5 files changed, 27 insertions(+) diff --git a/include/linux/bpf.h b/include/linux/bpf.h index 8d95f4c66275..79fb7c1be8fd 100644 --- a/include/linux/bpf.h +++ b/include/linux/bpf.h @@ -1845,6 +1845,7 @@ static inline bool bpf_map_is_dev_bound(struct bpf_map *map) struct bpf_map *bpf_map_offload_map_alloc(union bpf_attr *attr); void bpf_map_offload_map_free(struct bpf_map *map); +void sock_map_destroy(struct sock *sk); #else static inline int bpf_prog_offload_init(struct bpf_prog *prog, union bpf_attr *attr) diff --git a/include/linux/skmsg.h b/include/linux/skmsg.h index 1899a1c8421c..a83885c5bb86 100644 --- a/include/linux/skmsg.h +++ b/include/linux/skmsg.h @@ -105,6 +105,7 @@ struct sk_psock { spinlock_t link_lock; refcount_t refcnt; void (*saved_unhash)(struct sock *sk); + void (*saved_destroy)(struct sock *sk); void (*saved_close)(struct sock *sk, long timeout); void (*saved_write_space)(struct sock *sk); struct proto *sk_proto; diff --git a/net/core/skmsg.c b/net/core/skmsg.c index 622d93d56953..925863fab5bd 100644 --- a/net/core/skmsg.c +++ b/net/core/skmsg.c @@ -616,6 +616,7 @@ struct sk_psock *sk_psock_init(struct sock *sk, int node) psock->eval = __SK_NONE; psock->sk_proto = prot; psock->saved_unhash = prot->unhash; + psock->saved_destroy = prot->destroy; psock->saved_close = prot->close; psock->saved_write_space = sk->sk_write_space; diff --git a/net/core/sock_map.c b/net/core/sock_map.c index 28e518da4bcb..85df06298c98 100644 --- a/net/core/sock_map.c +++ b/net/core/sock_map.c @@ -1560,6 +1560,29 @@ void sock_map_unhash(struct sock *sk) saved_unhash(sk); } +void sock_map_destroy(struct sock *sk) +{ + void (*saved_destroy)(struct sock *sk); + struct sk_psock *psock; + + rcu_read_lock(); + psock = sk_psock_get(sk); + if (unlikely(!psock)) { + rcu_read_unlock(); + if (sk->sk_prot->destroy) + sk->sk_prot->destroy(sk); + return; + } + + saved_destroy = psock->saved_destroy; + sock_map_remove_links(sk, psock); + rcu_read_unlock(); + sk_psock_stop(psock, true); + sk_psock_put(sk, psock); + saved_destroy(sk); +} +EXPORT_SYMBOL_GPL(sock_map_destroy); + void sock_map_close(struct sock *sk, long timeout) { void (*saved_close)(struct sock *sk, long timeout); diff --git a/net/ipv4/tcp_bpf.c b/net/ipv4/tcp_bpf.c index afeaf35194de..1a8f129fb309 100644 --- a/net/ipv4/tcp_bpf.c +++ b/net/ipv4/tcp_bpf.c @@ -577,6 +577,7 @@ static void tcp_bpf_rebuild_protos(struct proto prot[TCP_BPF_NUM_CFGS], struct proto *base) { prot[TCP_BPF_BASE] = *base; + prot[TCP_BPF_BASE].destroy = sock_map_destroy; prot[TCP_BPF_BASE].close = sock_map_close; prot[TCP_BPF_BASE].recvmsg = tcp_bpf_recvmsg; prot[TCP_BPF_BASE].stream_memory_read = tcp_bpf_stream_read; -- 2.25.1

1 7

[PATCH openEuler-1.0-LTS 1/2] x86/bugs: Flush IBP in ib_prctl_set()
by Yongqiang Liu 14 Feb '23

14 Feb '23

From: Rodrigo Branco <bsdaemon(a)google.com> stable inclusion from stable-v4.19.270 commit 940ede60d74d2fc7291b96cb38072d705333c8e0 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6CU98 CVE: CVE-2023-0045 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=… -------------------------------- commit a664ec9158eeddd75121d39c9a0758016097fa96 upstream. We missed the window between the TIF flag update and the next reschedule. Signed-off-by: Rodrigo Branco <bsdaemon(a)google.com> Reviewed-by: Borislav Petkov (AMD) <bp(a)alien8.de> Signed-off-by: Ingo Molnar <mingo(a)kernel.org> Cc: <stable(a)vger.kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Yuyao Lin <linyuyao1(a)huawei.com> Reviewed-by: Wei Li <liwei391(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- arch/x86/kernel/cpu/bugs.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/arch/x86/kernel/cpu/bugs.c b/arch/x86/kernel/cpu/bugs.c index 3ba948bbd77d..d4d114d62e2e 100644 --- a/arch/x86/kernel/cpu/bugs.c +++ b/arch/x86/kernel/cpu/bugs.c @@ -1547,6 +1547,8 @@ static int ib_prctl_set(struct task_struct *task, unsigned long ctrl) if (ctrl == PR_SPEC_FORCE_DISABLE) task_set_spec_ib_force_disable(task); task_update_spec_tif(task); + if (task == current) + indirect_branch_prediction_barrier(); break; default: return -ERANGE; -- 2.25.1

1 1

[PATCH] vfio/migration: Modify the vf id acquisition method
by Longfang Liu 14 Feb '23

14 Feb '23

driver inclusion category:feature bugzilla: https://gitee.com/openeuler/kernel/issues/I6DRLU CVE: NA When using the old PCI_FUNC method to obtain the vf id, if the number of vf big than 8, the obtained id number will be wrong, and the new method needs to be used to correct it. Signed-off-by: Longfang Liu <liulongfang(a)huawei.com> Signed-off-by: JiangShui Yang <yangjiangshui(a)h-partners.com> --- .../hisilicon/migration/acc_vf_migration.c | 17 ++++++++++-- .../hisilicon/migration/acc_vf_migration.h | 26 +++++++++++++++++++ 2 files changed, 41 insertions(+), 2 deletions(-) diff --git a/drivers/crypto/hisilicon/migration/acc_vf_migration.c b/drivers/crypto/hisilicon/migration/acc_vf_migration.c index 920f19916fea..7c5fc4eb02ac 100644 --- a/drivers/crypto/hisilicon/migration/acc_vf_migration.c +++ b/drivers/crypto/hisilicon/migration/acc_vf_migration.c @@ -1610,6 +1610,19 @@ static int acc_vf_dev_init(struct pci_dev *pdev, struct hisi_qm *pf_qm, return -ENOMEM; } +static int hisi_acc_get_vf_id(struct pci_dev *dev) +{ + struct pci_dev *pf; + + if (!dev->is_virtfn) + return -EINVAL; + + pf = pci_physfn(dev); + return (((dev->bus->number << 8) + dev->devfn) - + ((pf->bus->number << 8) + pf->devfn + pf->sriov->offset)) / + pf->sriov->stride; +} + static void *acc_vf_probe(struct pci_dev *pdev) { struct acc_vf_migration *acc_vf_dev; @@ -1635,7 +1648,7 @@ static void *acc_vf_probe(struct pci_dev *pdev) return ERR_PTR(-EINVAL); } - vf_id = PCI_FUNC(vf_dev->devfn); + vf_id = hisi_acc_get_vf_id(vf_dev); if (vf_id < 0) { dev_info(&pdev->dev, "vf device: %s, vf id: %d\n", pf_qm->dev_name, vf_id); @@ -1652,7 +1665,7 @@ static void *acc_vf_probe(struct pci_dev *pdev) return ERR_PTR(-ENOMEM); } - acc_vf_dev->vf_id = vf_id; + acc_vf_dev->vf_id = vf_id + 1; acc_vf_dev->vf_vendor = pdev->vendor; acc_vf_dev->vf_device = pdev->device; acc_vf_dev->pf_dev = pf_dev; diff --git a/drivers/crypto/hisilicon/migration/acc_vf_migration.h b/drivers/crypto/hisilicon/migration/acc_vf_migration.h index 2f459eecb8f0..8b38b9943d35 100644 --- a/drivers/crypto/hisilicon/migration/acc_vf_migration.h +++ b/drivers/crypto/hisilicon/migration/acc_vf_migration.h @@ -215,6 +215,32 @@ struct acc_vf_region_ops { struct vfio_info_cap *caps); }; +/* Single Root I/O Virtualization */ +struct pci_sriov { + int pos; /* Capability position */ + int nres; /* Number of resources */ + u32 cap; /* SR-IOV Capabilities */ + u16 ctrl; /* SR-IOV Control */ + u16 total_VFs; /* Total VFs associated with the PF */ + u16 initial_VFs; /* Initial VFs associated with the PF */ + u16 num_VFs; /* Number of VFs available */ + u16 offset; /* First VF Routing ID offset */ + u16 stride; /* Following VF stride */ + u16 vf_device; /* VF device ID */ + u32 pgsz; /* Page size for BAR alignment */ + u8 link; /* Function Dependency Link */ + u8 max_VF_buses; /* Max buses consumed by VFs */ + u16 driver_max_VFs; /* Max num VFs driver supports */ + struct pci_dev *dev; /* Lowest numbered PF */ + struct pci_dev *self; /* This PF */ + u32 class; /* VF device */ + u8 hdr_type; /* VF header type */ + u16 subsystem_vendor; /* VF subsystem vendor */ + u16 subsystem_device; /* VF subsystem device */ + resource_size_t barsz[PCI_SRIOV_NUM_BARS]; /* VF BAR size */ + bool drivers_autoprobe; /* Auto probing of VFs by driver */ +}; + struct acc_vf_region { u32 type; u32 subtype; -- 2.33.0

1 0

Intel Arch SIG 例会
by openEuler conference 14 Feb '23

14 Feb '23

您好！ sig-Intel-Arch SIG 邀请您参加 2023-02-21 10:00 召开的Zoom会议会议主题：Intel Arch SIG 例会会议内容： 1. Intel几个复杂特性的支持计划 2. QEMU和kernel选型对未来lntel平台支持和release安排讨论 3. openEuler在SPR上的用户诉求和反馈会议链接：https://us06web.zoom.us/j/87058792984?pwd=UFlzSzA4OHNyMEIvRkpCdWFwdEdRQT09 会议纪要：https://etherpad.openeuler.org/p/sig-Intel-Arch-meetings 温馨提醒：建议接入会议后修改参会人的姓名，也可以使用您在gitee.com的ID 更多资讯尽在：https://openeuler.org/zh/ Hello! openEuler sig-Intel-Arch SIG invites you to attend the Zoom conference will be held at 2023-02-21 10:00, The subject of the conference is Intel Arch SIG 例会, Summary: 1. Intel几个复杂特性的支持计划 2. QEMU和kernel选型对未来lntel平台支持和release安排讨论 3. openEuler在SPR上的用户诉求和反馈 You can join the meeting at https://us06web.zoom.us/j/87058792984?pwd=UFlzSzA4OHNyMEIvRkpCdWFwdEdRQT09. Add topics at https://etherpad.openeuler.org/p/sig-Intel-Arch-meetings. Note: You are advised to change the participant name after joining the conference or use your ID at gitee.com. More information: https://openeuler.org/en/

1 0

[Cancel] Intel Arch SIG 例会
by openEuler conference 14 Feb '23

14 Feb '23

Sorry! The Zoom meeting will be held at 2023-02-20 10:00 scheduled by openEuler sig-Intel-Arch SIG has been cancelled.

1 0

Intel Arch SIG 例会
by openEuler conference 14 Feb '23

14 Feb '23

您好！ sig-Intel-Arch SIG 邀请您参加 2023-02-20 10:00 召开的Zoom会议会议主题：Intel Arch SIG 例会会议内容： 1. Intel几个复杂特性的支持计划 2. QEMU和kernel选型对未来lntel平台支持和release安排讨论 3. openEuler在SPR上的用户诉求和反馈会议链接：https://us06web.zoom.us/j/87427026722?pwd=eE9lQXBNck9KS0kzek9Da2VYV2VRdz09 会议纪要：https://etherpad.openeuler.org/p/sig-Intel-Arch-meetings 温馨提醒：建议接入会议后修改参会人的姓名，也可以使用您在gitee.com的ID 更多资讯尽在：https://openeuler.org/zh/ Hello! openEuler sig-Intel-Arch SIG invites you to attend the Zoom conference will be held at 2023-02-20 10:00, The subject of the conference is Intel Arch SIG 例会, Summary: 1. Intel几个复杂特性的支持计划 2. QEMU和kernel选型对未来lntel平台支持和release安排讨论 3. openEuler在SPR上的用户诉求和反馈 You can join the meeting at https://us06web.zoom.us/j/87427026722?pwd=eE9lQXBNck9KS0kzek9Da2VYV2VRdz09. Add topics at https://etherpad.openeuler.org/p/sig-Intel-Arch-meetings. Note: You are advised to change the participant name after joining the conference or use your ID at gitee.com. More information: https://openeuler.org/en/

1 0

[PATCH openEuler-1.0-LTS] media: vivid: fix compose size exceed boundary
by Yongqiang Liu 13 Feb '23

13 Feb '23

From: Liu Shixin <liushixin2(a)huawei.com> stable inclusion from stable-v4.19.270 commit 9c7fba9503b826f0c061d136f8f0c9f953ed18b9 category: bugfix bugzilla: https://gitee.com/src-openeuler/risc-v-kernel/issues/I6CIGU CVE: CVE-2023-0615 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 94a7ad9283464b75b12516c5512541d467cefcf8 ] syzkaller found a bug: BUG: unable to handle page fault for address: ffffc9000a3b1000 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 100000067 P4D 100000067 PUD 10015f067 PMD 1121ca067 PTE 0 Oops: 0002 [#1] PREEMPT SMP CPU: 0 PID: 23489 Comm: vivid-000-vid-c Not tainted 6.1.0-rc1+ #512 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014 RIP: 0010:memcpy_erms+0x6/0x10 [...] Call Trace: <TASK> ? tpg_fill_plane_buffer+0x856/0x15b0 vivid_fillbuff+0x8ac/0x1110 vivid_thread_vid_cap_tick+0x361/0xc90 vivid_thread_vid_cap+0x21a/0x3a0 kthread+0x143/0x180 ret_from_fork+0x1f/0x30 </TASK> This is because we forget to check boundary after adjust compose->height int V4L2_SEL_TGT_CROP case. Add v4l2_rect_map_inside() to fix this problem for this case. Fixes: ef834f7836ec ("[media] vivid: add the video capture and output parts") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> Signed-off-by: Hans Verkuil <hverkuil-cisco(a)xs4all.nl> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Longlong Xia <xialonglong1(a)huawei.com> Reviewed-by: Nanyong Sun <sunnanyong(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- drivers/media/platform/vivid/vivid-vid-cap.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/media/platform/vivid/vivid-vid-cap.c b/drivers/media/platform/vivid/vivid-vid-cap.c index c58ae489f39c..9eaac497ec34 100644 --- a/drivers/media/platform/vivid/vivid-vid-cap.c +++ b/drivers/media/platform/vivid/vivid-vid-cap.c @@ -915,6 +915,7 @@ int vivid_vid_cap_s_selection(struct file *file, void *fh, struct v4l2_selection if (dev->has_compose_cap) { v4l2_rect_set_min_size(compose, &min_rect); v4l2_rect_set_max_size(compose, &max_rect); + v4l2_rect_map_inside(compose, &fmt); } dev->fmt_cap_rect = fmt; tpg_s_buf_height(&dev->tpg, fmt.height); -- 2.25.1

1 0

[PATCH openEuler-5.10-LTS-SP1 01/87] NFSD: Fix verifier returned in stable WRITEs
by Jialin Zhang 13 Feb '23

13 Feb '23

From: Chuck Lever <chuck.lever(a)oracle.com> stable inclusion from stable-v5.10.143 commit d3d885507b52b28f54770a42d650251876d433a3 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6D0U6 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit f11ad7aa653130b71e2e89bed207f387718216d5 upstream. RFC 8881 explains the purpose of the write verifier this way: > The final portion of the result is the field writeverf. This field > is the write verifier and is a cookie that the client can use to > determine whether a server has changed instance state (e.g., server > restart) between a call to WRITE and a subsequent call to either > WRITE or COMMIT. But then it says: > This cookie MUST be unchanged during a single instance of the > NFSv4.1 server and MUST be unique between instances of the NFSv4.1 > server. If the cookie changes, then the client MUST assume that > any data written with an UNSTABLE4 value for committed and an old > writeverf in the reply has been lost and will need to be > recovered. RFC 1813 has similar language for NFSv3. NFSv2 does not have a write verifier since it doesn't implement the COMMIT procedure. Since commit 19e0663ff9bc ("nfsd: Ensure sampling of the write verifier is atomic with the write"), the Linux NFS server has returned a boot-time-based verifier for UNSTABLE WRITEs, but a zero verifier for FILE_SYNC and DATA_SYNC WRITEs. FILE_SYNC and DATA_SYNC WRITEs are not followed up with a COMMIT, so there's no need for clients to compare verifiers for stable writes. However, by returning a different verifier for stable and unstable writes, the above commit puts the Linux NFS server a step farther out of compliance with the first MUST above. At least one NFS client (FreeBSD) noticed the difference, making this a potential regression. [Removed down_write to fix the conflict in the cherry-pick. The down_write functionality was no longer needed there. Upstream commit 555dbf1a9aac6d3150c8b52fa35f768a692f4eeb titled nfsd: Replace use of rwsem with errseq_t removed those and replace it with new functionality that was more scalable. This commit is already backported onto 5.10 and so removing down_write ensures consistency with that change. Tested by compiling and booting successfully. - kochera] Reported-by: Rick Macklem <rmacklem(a)uoguelph.ca> Link: https://lore.kernel.org/linux-nfs/YQXPR0101MB096857EEACF04A6DF1FC6D9BDD749@… Fixes: 19e0663ff9bc ("nfsd: Ensure sampling of the write verifier is atomic with the write") Signed-off-by: Chuck Lever <chuck.lever(a)oracle.com> Signed-off-by: Michael Kochera <kochera(a)google.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> Reviewed-by: Zheng Zengkai <zhengzengkai(a)huawei.com> --- fs/nfsd/vfs.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c index c852bb5ff212..a4ae1fcd2ab1 100644 --- a/fs/nfsd/vfs.c +++ b/fs/nfsd/vfs.c @@ -1014,6 +1014,10 @@ nfsd_vfs_write(struct svc_rqst *rqstp, struct svc_fh *fhp, struct nfsd_file *nf, iov_iter_kvec(&iter, WRITE, vec, vlen, *cnt); since = READ_ONCE(file->f_wb_err); if (flags & RWF_SYNC) { + if (verf) + nfsd_copy_boot_verifier(verf, + net_generic(SVC_NET(rqstp), + nfsd_net_id)); host_err = vfs_iter_write(file, &iter, &pos, flags); if (host_err < 0) nfsd_reset_boot_verifier(net_generic(SVC_NET(rqstp), -- 2.25.1

1 86

openEuler Kernel SIG双周例会
by openEuler conference 10 Feb '23

10 Feb '23

您好！ Kernel SIG 邀请您参加 2023-02-10 14:00 召开的Zoom会议(自动录制) 会议主题：openEuler Kernel SIG双周例会会议内容：欢迎您参加年后第一次Kernel SIG 双周例会： 1. 进展update 2. 议题征集中欢迎大家积极申报议题（新增议题可以直接回复邮件，或录入会议看板）会议链接：https://us06web.zoom.us/j/89146789515?pwd=NVN5RmdCSFh2Y1g4SXUwK2NCRnFEZz09 会议纪要：https://etherpad.openeuler.org/p/Kernel-meetings 温馨提醒：建议接入会议后修改参会人的姓名，也可以使用您在gitee.com的ID 更多资讯尽在：https://openeuler.org/zh/ Hello! openEuler Kernel SIG invites you to attend the Zoom conference(auto recording) will be held at 2023-02-10 14:00, The subject of the conference is openEuler Kernel SIG双周例会, Summary: 欢迎您参加年后第一次Kernel SIG 双周例会： 1. 进展update 2. 议题征集中欢迎大家积极申报议题（新增议题可以直接回复邮件，或录入会议看板） You can join the meeting at https://us06web.zoom.us/j/89146789515?pwd=NVN5RmdCSFh2Y1g4SXUwK2NCRnFEZz09. Add topics at https://etherpad.openeuler.org/p/Kernel-meetings. Note: You are advised to change the participant name after joining the conference or use your ID at gitee.com. More information: https://openeuler.org/en/

4 3

[PATCH openEuler-5.10-LTS 1/5] crypto: hisilicon/qm - add functions for releasing resources
by Zheng Zengkai 09 Feb '23

09 Feb '23

From: Weili Qian <qianweili(a)huawei.com> mainline inclusion from mainline-v6.0-rc1 commit bf081d6fa8e9 category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I6DRLU CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… ---------------------------------------------------------------------- The resources allocated by hisi_qm_memory_init() are released by hisi_qm_uninit(). Add hisi_qm_memory_uninit() to release resources, no functional change. conficts: drivers/crypto/hisilicon/qm.c Signed-off-by: Weili Qian <qianweili(a)huawei.com> Signed-off-by: JiangShui Yang <yangjiangshui(a)h-partners.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> --- drivers/crypto/hisilicon/qm.c | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c index 57de083e8967..ef09e45bdc3e 100644 --- a/drivers/crypto/hisilicon/qm.c +++ b/drivers/crypto/hisilicon/qm.c @@ -3480,6 +3480,21 @@ static void hisi_qm_pci_uninit(struct hisi_qm *qm) pci_disable_device(pdev); } +static void hisi_qm_memory_uninit(struct hisi_qm *qm) +{ + struct device *dev = &qm->pdev->dev; + + hisi_qp_memory_uninit(qm, qm->qp_num); + if (qm->qdma.va) { + hisi_qm_cache_wb(qm); + dma_free_coherent(dev, qm->qdma.size, + qm->qdma.va, qm->qdma.dma); + } + + idr_destroy(&qm->qp_idr); + kfree(qm->factor); +} + /** * hisi_qm_uninit() - Uninitialize qm. * @qm: The qm needed uninit. @@ -3488,26 +3503,15 @@ static void hisi_qm_pci_uninit(struct hisi_qm *qm) */ void hisi_qm_uninit(struct hisi_qm *qm) { - struct pci_dev *pdev = qm->pdev; - struct device *dev = &pdev->dev; - qm_cmd_uninit(qm); - kfree(qm->factor); - down_write(&qm->qps_lock); + down_write(&qm->qps_lock); if (!qm_avail_state(qm, QM_CLOSE)) { up_write(&qm->qps_lock); return; } - hisi_qp_memory_uninit(qm, qm->qp_num); - idr_destroy(&qm->qp_idr); - - if (qm->qdma.va) { - hisi_qm_cache_wb(qm); - dma_free_coherent(dev, qm->qdma.size, - qm->qdma.va, qm->qdma.dma); - } + hisi_qm_memory_uninit(qm); up_write(&qm->qps_lock); hisi_qm_set_state(qm, VF_NOT_READY); -- 2.20.1

1 4

[PATCH openEuler-1.0-LTS 00/49] linux-4.19.y inclusion
by Yongqiang Liu 09 Feb '23

09 Feb '23

Chen Zhongjin (1): perf: Fix possible memleak in pmu_dev_alloc() Christian Brauner (1): pnode: terminate at peers of source Chuck Lever (1): SUNRPC: Don't leak netobj memory when gss_read_proxy_verf() fails Dan Carpenter (2): bonding: uninitialized variable in bond_miimon_inspect() ipmi: fix use after free in _ipmi_destroy_user() Enzo Matsumiya (1): cifs: do not include page data when checking signature Eric Dumazet (1): net: stream: purge sk_error_queue in sk_stream_kill_queues() Greg Kroah-Hartman (1): prlimit: do_prlimit needs to have a speculation check Huaxin Lu (1): ima: Fix a potential NULL pointer access in ima_restore_measurement_list Isaac J. Manjarres (1): driver core: Fix bus_type.match() error handling in __driver_attach() Jakub Kicinski (2): bpf: pull before calling skb_postpull_rcsum() net: stream: don't purge sk_error_queue in sk_stream_kill_queues() Jan Kara (2): mbcache: automatically delete entries from cache on freeing ext4: fix deadlock due to mbcache entry corruption Jann Horn (2): mm/khugepaged: fix GUP-fast interaction by sending IPI mm/khugepaged: invoke MMU notifiers in shmem/file collapse paths Jiamei Xie (1): serial: amba-pl011: avoid SBSA UART accessing DMACR register Jiang Li (1): md/raid1: stop mdx_raid1 thread when raid1 array run failed Li Zetao (1): ACPICA: Fix use-after-free in acpi_ut_copy_ipackage_to_ipackage() Mark Rutland (1): arm64: cmpxchg_double*: hazard against entire exchange variable Michael S. Tsirkin (1): PCI: Fix pci_device_is_present() for VFs by checking PF Mikulas Patocka (1): md: fix a crash in mempool_free Paolo Abeni (1): net/ulp: prevent ULP without clone op from entering the LISTEN status Rafael J. Wysocki (1): ACPICA: Fix error code path in acpi_ds_call_control_method() Sascha Hauer (1): PCI/sysfs: Fix double free in error path Schspa Shi (1): mrp: introduce active flags to prevent UAF when applicant uninit Stanislav Fomichev (1): bpf: make sure skb->len != 0 when redirecting to a tunneling device Stephen Boyd (1): pstore: Avoid kcore oops by vmap()ing with VM_IOREMAP Subash Abhinov Kasiviswanathan (1): skbuff: Account for tail adjustment during pull operations Ulf Hansson (1): cpuidle: dt: Return the correct numbers of parsed idle states Volker Lendecke (1): cifs: Fix uninitialized memory read for smb311 posix symlink create Wang ShaoBo (1): SUNRPC: Fix missing release socket in rpc_sockname() Wang Weiyang (1): device_cgroup: Roll back to original exceptions after copy failure Wang Yufen (2): pstore/ram: Fix error return code in ramoops_probe() binfmt: Fix error return code in load_elf_fdpic_binary() Xiu Jianfeng (1): ima: Fix misuse of dereference of pointer in template_desc_init_fields() Yang Jihong (1): blktrace: Fix output non-blktrace event when blk_classic option enabled Yang Shi (1): mm: gup: fix the fast GUP race against THP collapse Yang Yingliang (2): class: fix possible memory leak in __class_register() chardev: fix error handling in cdev_device_add() Ye Bin (1): blk-mq: fix possible memleak when register 'hctx' failed Yuan Can (1): perf: arm_dsu: Fix hotplug callback leak in dsu_pmu_init() Zhang Tianci (1): ovl: Use ovl mounter's fsuid and fsgid in ovl_link() Zhang Yiqun (1): crypto: tcrypt - Fix multibuffer skcipher speed test mem leak Zhang Yuchen (2): ipmi: fix memleak when unload ipmi driver ipmi: fix long wait in unload when IPMI disconnect ZhangPeng (1): pinctrl: pinconf-generic: add missing of_node_put() delisun (1): serial: pl011: Do not clear RX FIFO & RX interrupt in unthrottle. minoura makoto (1): SUNRPC: ensure the matching upcall is in-flight upon downcall arch/arm64/include/asm/atomic_ll_sc.h | 2 +- arch/arm64/include/asm/atomic_lse.h | 2 +- block/blk-mq-sysfs.c | 11 ++- crypto/tcrypt.c | 9 -- drivers/acpi/acpica/dsmethod.c | 10 ++- drivers/acpi/acpica/utcopy.c | 7 -- drivers/base/class.c | 5 ++ drivers/base/dd.c | 8 +- drivers/char/ipmi/ipmi_msghandler.c | 12 ++- drivers/char/ipmi/ipmi_si_intf.c | 27 ++++-- drivers/cpuidle/dt_idle_states.c | 2 +- drivers/md/md.c | 9 +- drivers/md/raid1.c | 1 + drivers/net/bonding/bond_main.c | 2 +- drivers/pci/pci-sysfs.c | 13 ++- drivers/pci/pci.c | 2 + drivers/perf/arm_dsu_pmu.c | 6 +- drivers/pinctrl/pinconf-generic.c | 4 +- drivers/tty/serial/amba-pl011.c | 14 ++- fs/binfmt_elf_fdpic.c | 5 +- fs/char_dev.c | 2 +- fs/cifs/link.c | 1 + fs/cifs/smb2pdu.c | 15 ++-- fs/ext4/xattr.c | 4 +- fs/mbcache.c | 118 ++++++++++---------------- fs/overlayfs/dir.c | 46 ++++++---- fs/pnode.c | 2 +- fs/pstore/ram.c | 2 + fs/pstore/ram_core.c | 6 +- include/asm-generic/tlb.h | 6 ++ include/linux/mbcache.h | 33 ++++--- include/linux/sunrpc/rpc_pipe_fs.h | 5 ++ include/net/mrp.h | 1 + kernel/events/core.c | 8 +- kernel/sys.c | 2 + kernel/trace/blktrace.c | 3 +- mm/gup.c | 34 ++++++-- mm/khugepaged.c | 34 ++++++-- mm/memory.c | 1 + mm/mmu_gather.c | 5 ++ net/802/mrp.c | 18 ++-- net/core/filter.c | 11 ++- net/core/skbuff.c | 3 + net/core/stream.c | 7 +- net/ipv4/inet_connection_sock.c | 16 +++- net/sunrpc/auth_gss/auth_gss.c | 19 ++++- net/sunrpc/auth_gss/svcauth_gss.c | 4 +- net/sunrpc/clnt.c | 2 +- security/device_cgroup.c | 33 ++++++- security/integrity/ima/ima_template.c | 9 +- 50 files changed, 406 insertions(+), 195 deletions(-) -- 2.25.1

1 49

[PATCH openEuler-1.0-LTS 0/5] mainline backport
by Yongqiang Liu 09 Feb '23

09 Feb '23

Ard Biesheuvel (1): crypto: tcrypt - avoid signed overflow in byte count Guenter Roeck (1): lib/list_debug.c: Detect uninitialized lists Guoqing Jiang (1): md: protect md_unregister_thread from reentrancy Konstantin Khlebnikov (1): sched/rt: Optimize checking group RT scheduler constraints Sean Christopherson (1): hugetlbfs: fix off-by-one error in hugetlb_vmdelete_list() crypto/tcrypt.c | 20 ++++++++++---------- drivers/md/md.c | 15 ++++++++++----- fs/hugetlbfs/inode.c | 7 ++++--- kernel/sched/rt.c | 24 +++++++++++------------- lib/list_debug.c | 12 ++++++++++-- 5 files changed, 45 insertions(+), 33 deletions(-) -- 2.25.1

1 5

[PATCH openEuler-1.0-LTS] mm: sharepool: fix hugepage_rsvd count increase error
by Yongqiang Liu 09 Feb '23

09 Feb '23

From: Guo Mengqi <guomengqi3(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I5RO2H CVE: NA -------------------------------- When nr_hugepages is configured, sharepool allocates hugepages first from hugetlb pool, then from buddy system if the pool had been used up. Current page release function treat the buddy system hugepages as hugetlb pages, which caused HugePages_Rsvd to increase improperly. Add a check in page release function: if the page is temporary, do not call hugetlb_unreserve_pages. Signed-off-by: Guo Mengqi <guomengqi3(a)huawei.com> Reviewed-by: Weilong Chen <chenweilong(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- fs/hugetlbfs/inode.c | 20 +++++++++++++++++--- 1 file changed, 17 insertions(+), 3 deletions(-) diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index f29d6d2fb052..d70c8aaf5acc 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -448,6 +448,14 @@ hugetlb_vmdelete_list(struct rb_root_cached *root, pgoff_t start, pgoff_t end) } } +static inline bool PageHugeTemporary(struct page *page) +{ + if (!PageHuge(page)) + return false; + + return (unsigned long)page[2].mapping == -1U; +} + /* * remove_inode_hugepages handles two distinct cases: truncation and hole * punch. There are subtle differences in operation for each case. @@ -532,11 +540,17 @@ static void remove_inode_hugepages(struct inode *inode, loff_t lstart, */ VM_BUG_ON(PagePrivate(page)); remove_huge_page(page); - freed++; - if (!truncate_op) { - if (unlikely(hugetlb_unreserve_pages(inode, + /* + * if the page is from buddy system, do not add to freed. + * because freed is used for hugetlbfs reservation accounting. + */ + if (!PageHugeTemporary(page)) { + freed++; + if (!truncate_op) { + if (unlikely(hugetlb_unreserve_pages(inode, index, index + 1, 1))) hugetlb_fix_reserve_counts(inode); + } } unlock_page(page); -- 2.25.1

1 0

[PATCH openEuler-5.10-LTS 1/5] crypto: hisilicon/qm - add functions for releasing resources
by Zheng Zengkai 09 Feb '23

09 Feb '23

From: Weili Qian <qianweili(a)huawei.com> mainline inclusion from mainline-v6.0-rc1 commit bf081d6fa8e9 category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I6DRLU CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… ---------------------------------------------------------------------- The resources allocated by hisi_qm_memory_init() are released by hisi_qm_uninit(). Add hisi_qm_memory_uninit() to release resources, no functional change. conficts: drivers/crypto/hisilicon/qm.c Signed-off-by: Weili Qian <qianweili(a)huawei.com> Signed-off-by: JiangShui Yang <yangjiangshui(a)h-partners.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> --- drivers/crypto/hisilicon/qm.c | 30 +++++++++++++++++------------- 1 file changed, 17 insertions(+), 13 deletions(-) diff --git a/drivers/crypto/hisilicon/qm.c b/drivers/crypto/hisilicon/qm.c index 57de083e8967..ef09e45bdc3e 100644 --- a/drivers/crypto/hisilicon/qm.c +++ b/drivers/crypto/hisilicon/qm.c @@ -3480,6 +3480,21 @@ static void hisi_qm_pci_uninit(struct hisi_qm *qm) pci_disable_device(pdev); } +static void hisi_qm_memory_uninit(struct hisi_qm *qm) +{ + struct device *dev = &qm->pdev->dev; + + hisi_qp_memory_uninit(qm, qm->qp_num); + if (qm->qdma.va) { + hisi_qm_cache_wb(qm); + dma_free_coherent(dev, qm->qdma.size, + qm->qdma.va, qm->qdma.dma); + } + + idr_destroy(&qm->qp_idr); + kfree(qm->factor); +} + /** * hisi_qm_uninit() - Uninitialize qm. * @qm: The qm needed uninit. @@ -3488,26 +3503,15 @@ static void hisi_qm_pci_uninit(struct hisi_qm *qm) */ void hisi_qm_uninit(struct hisi_qm *qm) { - struct pci_dev *pdev = qm->pdev; - struct device *dev = &pdev->dev; - qm_cmd_uninit(qm); - kfree(qm->factor); - down_write(&qm->qps_lock); + down_write(&qm->qps_lock); if (!qm_avail_state(qm, QM_CLOSE)) { up_write(&qm->qps_lock); return; } - hisi_qp_memory_uninit(qm, qm->qp_num); - idr_destroy(&qm->qp_idr); - - if (qm->qdma.va) { - hisi_qm_cache_wb(qm); - dma_free_coherent(dev, qm->qdma.size, - qm->qdma.va, qm->qdma.dma); - } + hisi_qm_memory_uninit(qm); up_write(&qm->qps_lock); hisi_qm_set_state(qm, VF_NOT_READY); -- 2.20.1

1 4

[PATCH openEuler-1.0-LTS 1/3] irqchip/gic-v3-its: introduce CONFIG_GENERIC_PENDING_IRQ
by Yongqiang Liu 09 Feb '23

09 Feb '23

From: Yipeng Zou <zouyipeng(a)huawei.com> hulk inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I6BO2R CVE: NA -------------------------------- Now, There is some issues about LPI migration on ARM SMP platform. For example, NIC device generates MSI and sends LPI to CPU0 via ITS, meanwhile irqbalance running on CPU1 set irq affinty of NIC to CPU1, the next interrupt will be sent to CPU2, due to the state of irq is still in progress, kernel does not end up performing irq handler on CPU2, which results in some userland service timeouts, the sequence of events is shown as follows: NIC CPU0 CPU1 Generate IRQ#1 READ_IAR Lock irq_desc Set IRQD_IN_PROGRESS Unlock irq_desc Lock irq_desc Change LPI Affinity Unlock irq_desc Call irq_handler Generate IRQ#2 READ_IAR Lock irq_desc Check IRQD_IN_PROGRESS Unlock irq_desc Return from interrupt#2 Lock irq_desc Clear IRQD_IN_PROGRESS Unlock irq_desc return from interrupt#1 For this scenario, We can enable CONFIG_GENERIC_PENDING_IRQ to avoid this. The CONFIG_GENERIC_PENDING_IRQ will delay all action that modify LPI affinity until the next interrupt eoi handler. Signed-off-by: Yipeng Zou <zouyipeng(a)huawei.com> Reviewed-by: Liao Chang <liaochang1(a)huawei.com> Reviewed-by: Zhang Jianhua <chris.zjh(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- drivers/irqchip/irq-gic-v3-its.c | 8 +++++++- kernel/irq/Kconfig | 4 +++- kernel/irq/migration.c | 2 ++ 3 files changed, 12 insertions(+), 2 deletions(-) diff --git a/drivers/irqchip/irq-gic-v3-its.c b/drivers/irqchip/irq-gic-v3-its.c index 15a0292e8e61..7a0adf67761e 100644 --- a/drivers/irqchip/irq-gic-v3-its.c +++ b/drivers/irqchip/irq-gic-v3-its.c @@ -1162,6 +1162,12 @@ static void its_unmask_irq(struct irq_data *d) lpi_update_config(d, 0, LPI_PROP_ENABLED); } +static void its_irq_chip_eoi(struct irq_data *d) +{ + irq_move_irq(d); + irq_chip_eoi_parent(d); +} + static int its_set_affinity(struct irq_data *d, const struct cpumask *mask_val, bool force) { @@ -1485,7 +1491,7 @@ static struct irq_chip its_irq_chip = { .name = "ITS", .irq_mask = its_mask_irq, .irq_unmask = its_unmask_irq, - .irq_eoi = irq_chip_eoi_parent, + .irq_eoi = its_irq_chip_eoi, .irq_set_affinity = its_set_affinity, .irq_compose_msi_msg = its_irq_compose_msi_msg, .irq_set_irqchip_state = its_irq_set_irqchip_state, diff --git a/kernel/irq/Kconfig b/kernel/irq/Kconfig index 14a85d0161ea..eebbced84e44 100644 --- a/kernel/irq/Kconfig +++ b/kernel/irq/Kconfig @@ -32,7 +32,9 @@ config GENERIC_IRQ_LEGACY_ALLOC_HWIRQ # Support for delayed migration from interrupt context config GENERIC_PENDING_IRQ - bool + bool "Support for delayed migration from interrupt context" + depends on SMP + default n # Support for generic irq migrating off cpu before the cpu is offline. config GENERIC_IRQ_MIGRATION diff --git a/kernel/irq/migration.c b/kernel/irq/migration.c index def48589ea48..bcb61ee69c20 100644 --- a/kernel/irq/migration.c +++ b/kernel/irq/migration.c @@ -117,3 +117,5 @@ void __irq_move_irq(struct irq_data *idata) if (!masked) idata->chip->irq_unmask(idata); } + +void __weak irq_force_complete_move(struct irq_desc *desc) { } -- 2.25.1

1 2

[PATCH openEuler-1.0-LTS 1/3] jbd2: Fix data missing when reusing bh which is ready to be checkpointed
by Yongqiang Liu 09 Feb '23

09 Feb '23

From: Zhihao Cheng <chengzhihao1(a)huawei.com> maillist inclusion category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I6C5HV CVE: NA Reference: https://lore.kernel.org/lkml/20230110015327.1181863-1-chengzhihao1@huawei.c… -------------------------------- Following process will make data lost and could lead to a filesystem corrupted problem: 1. jh(bh) is inserted into T1->t_checkpoint_list, bh is dirty, and jh->b_transaction = NULL 2. T1 is added into journal->j_checkpoint_transactions. 3. Get bh prepare to write while doing checkpoing: PA PB do_get_write_access jbd2_log_do_checkpoint spin_lock(&jh->b_state_lock) if (buffer_dirty(bh)) clear_buffer_dirty(bh) // clear buffer dirty set_buffer_jbddirty(bh) transaction = journal->j_checkpoint_transactions jh = transaction->t_checkpoint_list if (!buffer_dirty(bh)) __jbd2_journal_remove_checkpoint(jh) // bh won't be flushed jbd2_cleanup_journal_tail __jbd2_journal_file_buffer(jh, transaction, BJ_Reserved) 4. Aborting journal/Power-cut before writing latest bh on journal area. In this way we get a corrupted filesystem with bh's data lost. Fix it by moving the clearing of buffer_dirty bit just before the call to __jbd2_journal_file_buffer(), both bit clearing and jh->b_transaction assignment are under journal->j_list_lock locked, so that jbd2_log_do_checkpoint() will wait until jh's new transaction fininshed even bh is currently not dirty. And journal_shrink_one_cp_list() won't remove jh from checkpoint list if the buffer head is reused in do_get_write_access(). Fetch a reproducer in [Link]. Link: https://bugzilla.kernel.org/show_bug.cgi?id=216898 Cc: <stable(a)kernel.org> Signed-off-by: Zhihao Cheng <chengzhihao1(a)huawei.com> Signed-off-by: zhanchengbin <zhanchengbin1(a)huawei.com> Suggested-by: Jan Kara <jack(a)suse.cz> Conflicts: fs/jbd2/transaction.c [ 464170647b5648bb8("jbd2: Make state lock a spinlock") is not applied. ] Signed-off-by: Zhihao Cheng <chengzhihao1(a)huawei.com> Reviewed-by: Yang Erkun <yangerkun(a)huawei.com> Reviewed-by: Zhang Yi <yi.zhang(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- fs/jbd2/transaction.c | 50 +++++++++++++++++++++++++------------------ 1 file changed, 29 insertions(+), 21 deletions(-) diff --git a/fs/jbd2/transaction.c b/fs/jbd2/transaction.c index 3b31bd1f7b77..149190c2ac89 100644 --- a/fs/jbd2/transaction.c +++ b/fs/jbd2/transaction.c @@ -935,36 +935,28 @@ do_get_write_access(handle_t *handle, struct journal_head *jh, * ie. locked but not dirty) or tune2fs (which may actually have * the buffer dirtied, ugh.) */ - if (buffer_dirty(bh)) { + if (buffer_dirty(bh) && jh->b_transaction) { + warn_dirty_buffer(bh); /* - * First question: is this buffer already part of the current - * transaction or the existing committing transaction? - */ - if (jh->b_transaction) { - J_ASSERT_JH(jh, - jh->b_transaction == transaction || - jh->b_transaction == - journal->j_committing_transaction); - if (jh->b_next_transaction) - J_ASSERT_JH(jh, jh->b_next_transaction == - transaction); - warn_dirty_buffer(bh); - } - /* - * In any case we need to clean the dirty flag and we must - * do it under the buffer lock to be sure we don't race - * with running write-out. + * We need to clean the dirty flag and we must do it under the + * buffer lock to be sure we don't race with running write-out. */ JBUFFER_TRACE(jh, "Journalling dirty buffer"); clear_buffer_dirty(bh); + /* + * The buffer is going to be added to BJ_Reserved list now and + * nothing guarantees jbd2_journal_dirty_metadata() will be + * ever called for it. So we need to set jbddirty bit here to + * make sure the buffer is dirtied and written out when the + * journaling machinery is done with it. + */ set_buffer_jbddirty(bh); } - unlock_buffer(bh); - error = -EROFS; if (is_handle_aborted(handle)) { jbd_unlock_bh_state(bh); + unlock_buffer(bh); goto out; } error = 0; @@ -974,8 +966,10 @@ do_get_write_access(handle_t *handle, struct journal_head *jh, * b_next_transaction points to it */ if (jh->b_transaction == transaction || - jh->b_next_transaction == transaction) + jh->b_next_transaction == transaction) { + unlock_buffer(bh); goto done; + } /* * this is the first time this transaction is touching this buffer, @@ -999,10 +993,24 @@ do_get_write_access(handle_t *handle, struct journal_head *jh, */ smp_wmb(); spin_lock(&journal->j_list_lock); + if (test_clear_buffer_dirty(bh)) { + /* + * Execute buffer dirty clearing and jh->b_transaction + * assignment under journal->j_list_lock locked to + * prevent bh being removed from checkpoint list if + * the buffer is in an intermediate state (not dirty + * and jh->b_transaction is NULL). + */ + JBUFFER_TRACE(jh, "Journalling dirty buffer"); + set_buffer_jbddirty(bh); + } __jbd2_journal_file_buffer(jh, transaction, BJ_Reserved); spin_unlock(&journal->j_list_lock); + unlock_buffer(bh); goto done; } + unlock_buffer(bh); + /* * If there is already a copy-out version of this buffer, then we don't * need to make another one -- 2.25.1

1 2

[PATCH openEuler-1.0-LTS] x86/unwind: Fix orc entry for paravirt {save,restore}_fl
by Yongqiang Liu 08 Feb '23

08 Feb '23

From: Chen Zhongjin <chenzhongjin(a)huawei.com> hulk inclusion category: bugfix bugzilla: 187818, https://gitee.com/openeuler/kernel/issues/I6DK3O CVE: NA -------------------------------- When CONFIG_PARAVIRT is enabled, the code of {save,restore}_fl is defined as: ff 14 25 00 00 00 00 callq *0x0 which will be patched to call the xen paravirt function, or native implementation, in 'paravirt_patch_64.c': pushfq; popq %rax // for native_save_fl pushq %rdi; popfq // for native_restore_fl The orc metadata is generated with insn 'callq', so it can become inconsistent with the real insn 'push;pop'. This makes stacktrace on the 'pop' insn fail and incorrect stacktrace result can be returned. To prevent reliable stacktrace broken, check the insns when unwind pt_regs stack frame: When there are 'push;pop' combination and both insns don't change orc entry, it means the stack state is inconsistent with orc on pop. Add one slot to sp_offset for on original orc entry to get the correct orc entry. Signed-off-by: Chen Zhongjin <chenzhongjin(a)huawei.com> Reviewed-by: Xu Kuohai <xukuohai(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- arch/x86/kernel/unwind_orc.c | 43 ++++++++++++++++++++++++++++++++++++ 1 file changed, 43 insertions(+) diff --git a/arch/x86/kernel/unwind_orc.c b/arch/x86/kernel/unwind_orc.c index 0402bc0b5851..ec4a14e4f639 100644 --- a/arch/x86/kernel/unwind_orc.c +++ b/arch/x86/kernel/unwind_orc.c @@ -138,6 +138,34 @@ static struct orc_entry null_orc_entry = { .type = ORC_TYPE_CALL }; +#ifdef CONFIG_PARAVIRT +static bool check_paravirt(struct unwind_state *state, struct orc_entry *orc) +{ + u8 *ip = (u8 *)state->ip; + + /* + * In paravirt_patch_64.c, patched paravirt opcode should be: + * pushfq; popq %rax // 0x9c 0x58 + * pushq %rdi; popfq // 0x57 0x9d + * + * Error unwinding only happens when: + * 1. In irq or preempt context. + * 2. Current insn is popq, and it doesn't change orc. + * 3. Last insn doesn't change orc, checking it first to + * promise ip - 1 is valid. + * 4. Last byte fits pushf. + */ + if (state->regs && orc->type == ORC_TYPE_CALL && + (ip[0] == 0x58 || ip[0] == 0x9d) && + orc == orc_find((unsigned long)(ip + 1)) && + orc == orc_find((unsigned long)(ip - 1)) && + (ip[-1] == 0x9c || ip[-1] == 0x57)) + return true; + + return false; +} +#endif + static struct orc_entry *orc_find(unsigned long ip) { static struct orc_entry *orc; @@ -415,6 +443,9 @@ bool unwind_next_frame(struct unwind_state *state) enum stack_type prev_type = state->stack_info.type; struct orc_entry *orc; bool indirect = false; +#ifdef CONFIG_PARAVIRT + struct orc_entry para_orc; +#endif if (unwind_done(state)) return false; @@ -436,6 +467,18 @@ bool unwind_next_frame(struct unwind_state *state) if (!orc) goto err; +#ifdef CONFIG_PARAVIRT + /* + * When hitting paravirt POP insn, the orc entry should add + * one slot for PUSH insn. + */ + if (!state->error && check_paravirt(state, orc)) { + para_orc = *orc; + para_orc.sp_offset += sizeof(long); + orc = &para_orc; + } +#endif + /* End-of-stack check for kernel threads: */ if (orc->sp_reg == ORC_REG_UNDEFINED) { if (!orc->end) -- 2.25.1

1 0

[PATCH openEuler-1.0-LTS 1/2] drm/i915/gvt: fix double free bug in split_2MB_gtt_entry
by Yongqiang Liu 07 Feb '23

07 Feb '23

From: Zheng Wang <zyytlz.wz(a)163.com> mainline inclusion from mainline-v6.2-rc3 commit 4a61648af68f5ba4884f0e3b494ee1cabc4b6620 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I5XXFF CVE: CVE-2022-3707 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- If intel_gvt_dma_map_guest_page failed, it will call ppgtt_invalidate_spt, which will finally free the spt. But the caller function ppgtt_populate_spt_by_guest_entry does not notice that, it will free spt again in its error path. Fix this by canceling the mapping of DMA address and freeing sub_spt. Besides, leave the handle of spt destroy to caller function instead of callee function when error occurs. Fixes: b901b252b6cf ("drm/i915/gvt: Add 2M huge gtt support") Signed-off-by: Zheng Wang <zyytlz.wz(a)163.com> Reviewed-by: Zhenyu Wang <zhenyuw(a)linux.intel.com> Signed-off-by: Zhenyu Wang <zhenyuw(a)linux.intel.com> Link: http://patchwork.freedesktop.org/patch/msgid/20221229165641.1192455-1-zyytl… conflicts: drivers/gpu/drm/i915/gvt/gtt.c Signed-off-by: Wang ShaoBo <bobo.shaobowang(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Reviewed-by: Wei Li <liwei391(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- drivers/gpu/drm/i915/gvt/gtt.c | 17 +++++++++++++---- 1 file changed, 13 insertions(+), 4 deletions(-) diff --git a/drivers/gpu/drm/i915/gvt/gtt.c b/drivers/gpu/drm/i915/gvt/gtt.c index 40b32b4d1d98..afbc648befec 100644 --- a/drivers/gpu/drm/i915/gvt/gtt.c +++ b/drivers/gpu/drm/i915/gvt/gtt.c @@ -1155,10 +1155,8 @@ static int split_2MB_gtt_entry(struct intel_vgpu *vgpu, for_each_shadow_entry(sub_spt, &sub_se, sub_index) { ret = intel_gvt_hypervisor_dma_map_guest_page(vgpu, start_gfn + sub_index, PAGE_SIZE, &dma_addr); - if (ret) { - ppgtt_invalidate_spt(spt); - return ret; - } + if (ret) + goto err; sub_se.val64 = se->val64; /* Copy the PAT field from PDE. */ @@ -1177,6 +1175,17 @@ static int split_2MB_gtt_entry(struct intel_vgpu *vgpu, ops->set_pfn(se, sub_spt->shadow_page.mfn); ppgtt_set_shadow_entry(spt, se, index); return 0; +err: + /* Cancel the existing addess mappings of DMA addr. */ + for_each_present_shadow_entry(sub_spt, &sub_se, sub_index) { + gvt_vdbg_mm("invalidate 4K entry\n"); + ppgtt_invalidate_pte(sub_spt, &sub_se); + } + /* Release the new allocated spt. */ + trace_spt_change(sub_spt->vgpu->id, "release", sub_spt, + sub_spt->guest_page.gfn, sub_spt->shadow_page.type); + ppgtt_free_spt(sub_spt); + return ret; } static int split_64KB_gtt_entry(struct intel_vgpu *vgpu, -- 2.25.1

1 1

[PATCH openEuler-5.10-LTS 01/10] xfs: fix potential log item leak
by Jialin Zhang 07 Feb '23

07 Feb '23

From: Dave Chinner <dchinner(a)redhat.com> mainline inclusion from mainline-v5.19-rc1 commit c230a4a85bcdbfc1a7415deec6caf04e8fca1301 category: bugfix bugzilla: 187372, https://gitee.com/openeuler/kernel/issues/I5K0OM CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Ever since we added shadown format buffers to the log items, log items need to handle the item being released with shadow buffers attached. Due to the fact this requirement was added at the same time we added new rmap/reflink intents, we missed the cleanup of those items. In theory, this means shadow buffers can be leaked in a very small window when a shutdown is initiated. Testing with KASAN shows this leak does not happen in practice - we haven't identified a single leak in several years of shutdown testing since ~v4.8 kernels. However, the intent whiteout cleanup mechanism results in every cancelled intent in exactly the same state as this tiny race window creates and so if intents down clean up shadow buffers on final release we will leak the shadow buffer for just about every intent we create. Hence we start with this patch to close this condition off and ensure that when whiteouts start to be used we don't leak lots of memory. Signed-off-by: Dave Chinner <dchinner(a)redhat.com> Reviewed-by: Darrick J. Wong <djwong(a)kernel.org> Reviewed-by: Allison Henderson <allison.henderson(a)oracle.com> Signed-off-by: Dave Chinner <david(a)fromorbit.com> conflicts: fs/xfs/xfs_bmap_item.c fs/xfs/xfs_icreate_item.c fs/xfs/xfs_refcount_item.c fs/xfs/xfs_rmap_item.c Signed-off-by: Li Nan <linan122(a)huawei.com> Reviewed-by: Yang Erkun <yangerkun(a)huawei.com> Reviewed-by: Zhang Yi <yi.zhang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- fs/xfs/xfs_bmap_item.c | 2 ++ fs/xfs/xfs_icreate_item.c | 1 + fs/xfs/xfs_refcount_item.c | 2 ++ fs/xfs/xfs_rmap_item.c | 2 ++ 4 files changed, 7 insertions(+) diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c index 44ec0f2d5253..e6de8081451f 100644 --- a/fs/xfs/xfs_bmap_item.c +++ b/fs/xfs/xfs_bmap_item.c @@ -40,6 +40,7 @@ STATIC void xfs_bui_item_free( struct xfs_bui_log_item *buip) { + kmem_free(buip->bui_item.li_lv_shadow); kmem_cache_free(xfs_bui_zone, buip); } @@ -199,6 +200,7 @@ xfs_bud_item_release( struct xfs_bud_log_item *budp = BUD_ITEM(lip); xfs_bui_release(budp->bud_buip); + kmem_free(budp->bud_item.li_lv_shadow); kmem_cache_free(xfs_bud_zone, budp); } diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c index 9b3994b9c716..aa8c7c261d24 100644 --- a/fs/xfs/xfs_icreate_item.c +++ b/fs/xfs/xfs_icreate_item.c @@ -63,6 +63,7 @@ STATIC void xfs_icreate_item_release( struct xfs_log_item *lip) { + kmem_free(ICR_ITEM(lip)->ic_item.li_lv_shadow); kmem_cache_free(xfs_icreate_zone, ICR_ITEM(lip)); } diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c index 0dee316283a9..9f4ff45c7a93 100644 --- a/fs/xfs/xfs_refcount_item.c +++ b/fs/xfs/xfs_refcount_item.c @@ -35,6 +35,7 @@ STATIC void xfs_cui_item_free( struct xfs_cui_log_item *cuip) { + kmem_free(cuip->cui_item.li_lv_shadow); if (cuip->cui_format.cui_nextents > XFS_CUI_MAX_FAST_EXTENTS) kmem_free(cuip); else @@ -204,6 +205,7 @@ xfs_cud_item_release( struct xfs_cud_log_item *cudp = CUD_ITEM(lip); xfs_cui_release(cudp->cud_cuip); + kmem_free(cudp->cud_item.li_lv_shadow); kmem_cache_free(xfs_cud_zone, cudp); } diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index 20905953fe76..b5447ac7cb9b 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -35,6 +35,7 @@ STATIC void xfs_rui_item_free( struct xfs_rui_log_item *ruip) { + kmem_free(ruip->rui_item.li_lv_shadow); if (ruip->rui_format.rui_nextents > XFS_RUI_MAX_FAST_EXTENTS) kmem_free(ruip); else @@ -227,6 +228,7 @@ xfs_rud_item_release( struct xfs_rud_log_item *rudp = RUD_ITEM(lip); xfs_rui_release(rudp->rud_ruip); + kmem_free(rudp->rud_item.li_lv_shadow); kmem_cache_free(xfs_rud_zone, rudp); } -- 2.25.1

1 9

[PATCH openEuler-5.10-LTS-SP1 01/20] xfs: fix potential log item leak
by Jialin Zhang 07 Feb '23

07 Feb '23

From: Dave Chinner <dchinner(a)redhat.com> mainline inclusion from mainline-v5.19-rc1 commit c230a4a85bcdbfc1a7415deec6caf04e8fca1301 category: bugfix bugzilla: 187372, https://gitee.com/openeuler/kernel/issues/I5K0OM CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Ever since we added shadown format buffers to the log items, log items need to handle the item being released with shadow buffers attached. Due to the fact this requirement was added at the same time we added new rmap/reflink intents, we missed the cleanup of those items. In theory, this means shadow buffers can be leaked in a very small window when a shutdown is initiated. Testing with KASAN shows this leak does not happen in practice - we haven't identified a single leak in several years of shutdown testing since ~v4.8 kernels. However, the intent whiteout cleanup mechanism results in every cancelled intent in exactly the same state as this tiny race window creates and so if intents down clean up shadow buffers on final release we will leak the shadow buffer for just about every intent we create. Hence we start with this patch to close this condition off and ensure that when whiteouts start to be used we don't leak lots of memory. Signed-off-by: Dave Chinner <dchinner(a)redhat.com> Reviewed-by: Darrick J. Wong <djwong(a)kernel.org> Reviewed-by: Allison Henderson <allison.henderson(a)oracle.com> Signed-off-by: Dave Chinner <david(a)fromorbit.com> conflicts: fs/xfs/xfs_bmap_item.c fs/xfs/xfs_icreate_item.c fs/xfs/xfs_refcount_item.c fs/xfs/xfs_rmap_item.c Signed-off-by: Li Nan <linan122(a)huawei.com> Reviewed-by: Yang Erkun <yangerkun(a)huawei.com> Reviewed-by: Zhang Yi <yi.zhang(a)huawei.com> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> --- fs/xfs/xfs_bmap_item.c | 2 ++ fs/xfs/xfs_icreate_item.c | 1 + fs/xfs/xfs_refcount_item.c | 2 ++ fs/xfs/xfs_rmap_item.c | 2 ++ 4 files changed, 7 insertions(+) diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c index 44ec0f2d5253..e6de8081451f 100644 --- a/fs/xfs/xfs_bmap_item.c +++ b/fs/xfs/xfs_bmap_item.c @@ -40,6 +40,7 @@ STATIC void xfs_bui_item_free( struct xfs_bui_log_item *buip) { + kmem_free(buip->bui_item.li_lv_shadow); kmem_cache_free(xfs_bui_zone, buip); } @@ -199,6 +200,7 @@ xfs_bud_item_release( struct xfs_bud_log_item *budp = BUD_ITEM(lip); xfs_bui_release(budp->bud_buip); + kmem_free(budp->bud_item.li_lv_shadow); kmem_cache_free(xfs_bud_zone, budp); } diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c index 9b3994b9c716..aa8c7c261d24 100644 --- a/fs/xfs/xfs_icreate_item.c +++ b/fs/xfs/xfs_icreate_item.c @@ -63,6 +63,7 @@ STATIC void xfs_icreate_item_release( struct xfs_log_item *lip) { + kmem_free(ICR_ITEM(lip)->ic_item.li_lv_shadow); kmem_cache_free(xfs_icreate_zone, ICR_ITEM(lip)); } diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c index 0dee316283a9..9f4ff45c7a93 100644 --- a/fs/xfs/xfs_refcount_item.c +++ b/fs/xfs/xfs_refcount_item.c @@ -35,6 +35,7 @@ STATIC void xfs_cui_item_free( struct xfs_cui_log_item *cuip) { + kmem_free(cuip->cui_item.li_lv_shadow); if (cuip->cui_format.cui_nextents > XFS_CUI_MAX_FAST_EXTENTS) kmem_free(cuip); else @@ -204,6 +205,7 @@ xfs_cud_item_release( struct xfs_cud_log_item *cudp = CUD_ITEM(lip); xfs_cui_release(cudp->cud_cuip); + kmem_free(cudp->cud_item.li_lv_shadow); kmem_cache_free(xfs_cud_zone, cudp); } diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c index 20905953fe76..b5447ac7cb9b 100644 --- a/fs/xfs/xfs_rmap_item.c +++ b/fs/xfs/xfs_rmap_item.c @@ -35,6 +35,7 @@ STATIC void xfs_rui_item_free( struct xfs_rui_log_item *ruip) { + kmem_free(ruip->rui_item.li_lv_shadow); if (ruip->rui_format.rui_nextents > XFS_RUI_MAX_FAST_EXTENTS) kmem_free(ruip); else @@ -227,6 +228,7 @@ xfs_rud_item_release( struct xfs_rud_log_item *rudp = RUD_ITEM(lip); xfs_rui_release(rudp->rud_ruip); + kmem_free(rudp->rud_item.li_lv_shadow); kmem_cache_free(xfs_rud_zone, rudp); } -- 2.25.1

1 19

[PATCH openEuler-5.10-LTS-SP1 01/74] drm/msm/dsi: fix the inconsistent indenting
by Jialin Zhang 07 Feb '23

07 Feb '23

From: sunliming <sunliming(a)kylinos.cn> stable inclusion from stable-v5.10.142 commit 631fbefd877721d15c4f525cc71e851e0e588c8e category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6CSFH Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 2f25a1fb4ec516c5ad67afd754334b491b9f09a5 ] Fix the inconsistent indenting in function msm_dsi_dphy_timing_calc_v3(). Fix the following smatch warnings: drivers/gpu/drm/msm/dsi/phy/dsi_phy.c:350 msm_dsi_dphy_timing_calc_v3() warn: inconsistent indenting Fixes: f1fa7ff44056 ("drm/msm/dsi: implement auto PHY timing calculator for 10nm PHY") Reported-by: kernel test robot <lkp(a)intel.com> Signed-off-by: sunliming <sunliming(a)kylinos.cn> Reviewed-by: Abhinav Kumar <quic_abhinavk(a)quicinc.com> Patchwork: https://patchwork.freedesktop.org/patch/494662/ Link: https://lore.kernel.org/r/20220719015622.646718-1-sunliming@kylinos.cn Signed-off-by: Abhinav Kumar <quic_abhinavk(a)quicinc.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Jialin Zhang <zhangjialin11(a)huawei.com> Reviewed-by: Zheng Zengkai <zhengzengkai(a)huawei.com> --- drivers/gpu/drm/msm/dsi/phy/dsi_phy.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c index e07986ab52c2..2e0be85ec394 100644 --- a/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c +++ b/drivers/gpu/drm/msm/dsi/phy/dsi_phy.c @@ -345,7 +345,7 @@ int msm_dsi_dphy_timing_calc_v3(struct msm_dsi_dphy_timing *timing, } else { timing->shared_timings.clk_pre = linear_inter(tmax, tmin, pcnt2, 0, false); - timing->shared_timings.clk_pre_inc_by_2 = 0; + timing->shared_timings.clk_pre_inc_by_2 = 0; } timing->ta_go = 3; -- 2.25.1

1 73

[PATCH openEuler-1.0-LTS 1/2] block: don't allow a disk link holder to itself
by Yongqiang Liu 07 Feb '23

07 Feb '23

From: Yu Kuai <yukuai3(a)huawei.com> mainline inclusion from mainline-v6.2-rc1 commit 077a4033541fc96fb0a955985aab7d1f353da831 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6B4N7 CVE: NA -------------------------------- After creating a dm device, then user can reload such dm with itself, and dead loop will be triggered because dm keep looking up to itself. Test procedures: 1) dmsetup create test --table "xxx sda", assume dm-0 is created 2) dmsetup suspend test 3) dmsetup reload test --table "xxx dm-0" 4) dmsetup resume test Test result: BUG: TASK stack guard page was hit at 00000000736a261f (stack is 000000008d12c88d..00000000c8dd82d5) stack guard page: 0000 [#1] PREEMPT SMP CPU: 29 PID: 946 Comm: systemd-udevd Not tainted 6.1.0-rc3-next-20221101-00006-g17640ca3b0ee #1295 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS ?-20190727_073836-buildvm-ppc64le-16.ppc.fedoraproject.org-3.fc31 04/01/2014 RIP: 0010:dm_prepare_ioctl+0xf/0x1e0 Code: da 48 83 05 4a 7c 99 0b 01 41 89 c4 eb cd e8 b8 1f 40 00 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 57 48 83 05 a1 5a 99 0b 01 <41> 56 49 89 d6 41 55 4c 8d af 90 02 00 00 9 RSP: 0018:ffffc90002090000 EFLAGS: 00010206 RAX: ffff8881049d6800 RBX: ffff88817e589000 RCX: 0000000000000000 RDX: ffffc90002090010 RSI: ffffc9000209001c RDI: ffff88817e589000 RBP: 00000000484a101d R08: 0000000000000000 R09: 0000000000000007 R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000005331 R13: 0000000000005331 R14: 0000000000000000 R15: 0000000000000000 FS: 00007fddf9609200(0000) GS:ffff889fbfd40000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffc9000208fff8 CR3: 0000000179043000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: <TASK> dm_blk_ioctl+0x50/0x1c0 ? dm_prepare_ioctl+0xe0/0x1e0 dm_blk_ioctl+0x88/0x1c0 dm_blk_ioctl+0x88/0x1c0 ......(a lot of same lines) dm_blk_ioctl+0x88/0x1c0 dm_blk_ioctl+0x88/0x1c0 blkdev_ioctl+0x184/0x3e0 __x64_sys_ioctl+0xa3/0x110 do_syscall_64+0x35/0x80 entry_SYSCALL_64_after_hwframe+0x63/0xcd RIP: 0033:0x7fddf7306577 Code: b3 66 90 48 8b 05 11 89 2c 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d e1 88 8 RSP: 002b:00007ffd0b2ec318 EFLAGS: 00000246 ORIG_RAX: 0000000000000010 RAX: ffffffffffffffda RBX: 00005634ef478320 RCX: 00007fddf7306577 RDX: 0000000000000000 RSI: 0000000000005331 RDI: 0000000000000007 RBP: 0000000000000007 R08: 00005634ef4843e0 R09: 0000000000000080 R10: 00007fddf75cfb38 R11: 0000000000000246 R12: 00000000030d4000 R13: 0000000000000000 R14: 0000000000000000 R15: 00005634ef48b800 </TASK> Modules linked in: ---[ end trace 0000000000000000 ]--- RIP: 0010:dm_prepare_ioctl+0xf/0x1e0 Code: da 48 83 05 4a 7c 99 0b 01 41 89 c4 eb cd e8 b8 1f 40 00 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 41 57 48 83 05 a1 5a 99 0b 01 <41> 56 49 89 d6 41 55 4c 8d af 90 02 00 00 9 RSP: 0018:ffffc90002090000 EFLAGS: 00010206 RAX: ffff8881049d6800 RBX: ffff88817e589000 RCX: 0000000000000000 RDX: ffffc90002090010 RSI: ffffc9000209001c RDI: ffff88817e589000 RBP: 00000000484a101d R08: 0000000000000000 R09: 0000000000000007 R10: 0000000000000000 R11: 0000000000000000 R12: 0000000000005331 R13: 0000000000005331 R14: 0000000000000000 R15: 0000000000000000 FS: 00007fddf9609200(0000) GS:ffff889fbfd40000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: ffffc9000208fff8 CR3: 0000000179043000 CR4: 00000000000006e0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Kernel panic - not syncing: Fatal exception in interrupt Kernel Offset: disabled ---[ end Kernel panic - not syncing: Fatal exception in interrupt ]--- Fix the problem by forbidding a disk to create link to itself. Signed-off-by: Yu Kuai <yukuai3(a)huawei.com> Reviewed-by: Christoph Hellwig <hch(a)lst.de> Link: https://lore.kernel.org/r/20221115141054.1051801-11-yukuai1@huaweicloud.com Signed-off-by: Jens Axboe <axboe(a)kernel.dk> Signed-off-by: Li Lingfeng <lilingfeng3(a)huawei.com> Reviewed-by: Yang Erkun <yangerkun(a)huawei.com> Reviewed-by: Zhang Yi <yi.zhang(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- fs/block_dev.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/block_dev.c b/fs/block_dev.c index 2680092c022d..6ba91b97753f 100644 --- a/fs/block_dev.c +++ b/fs/block_dev.c @@ -1305,6 +1305,8 @@ int bd_link_disk_holder(struct block_device *bdev, struct gendisk *disk) struct bd_holder_disk *holder; int ret = 0; + if (bdev->bd_disk == disk) + return -EINVAL; /* * bdev could be deleted beneath us which would implicitly destroy * the holder directory. Hold on to it. -- 2.25.1

1 1

[PATCH openEuler-1.0-LTS 1/3] ipv6: raw: Deduct extension header length in rawv6_push_pending_frames
by Yongqiang Liu 06 Feb '23

06 Feb '23

From: Herbert Xu <herbert(a)gondor.apana.org.au> stable inclusion from stable-v4.19.270 commit f487d636e49bc1fdfbd8105bc1ab159164e2d8bd category: bugfix bugzilla: 188291, https://gitee.com/src-openeuler/kernel/issues/I6B1V2 CVE: CVE-2023-0394 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=… -------------------------------- commit cb3e9864cdbe35ff6378966660edbcbac955fe17 upstream. The total cork length created by ip6_append_data includes extension headers, so we must exclude them when comparing them against the IPV6_CHECKSUM offset which does not include extension headers. Reported-by: Kyle Zeng <zengyhkyle(a)gmail.com> Fixes: 357b40a18b04 ("[IPV6]: IPV6_CHECKSUM socket option can corrupt kernel memory") Signed-off-by: Herbert Xu <herbert(a)gondor.apana.org.au> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Lu Wei <luwei32(a)huawei.com> Reviewed-by: Yue Haibing <yuehaibing(a)huawei.com> Reviewed-by: Xiu Jianfeng <xiujianfeng(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- net/ipv6/raw.c | 4 ++++ 1 file changed, 4 insertions(+) diff --git a/net/ipv6/raw.c b/net/ipv6/raw.c index 412d4f60a408..8ce6414edd88 100644 --- a/net/ipv6/raw.c +++ b/net/ipv6/raw.c @@ -543,6 +543,7 @@ static int rawv6_recvmsg(struct sock *sk, struct msghdr *msg, size_t len, static int rawv6_push_pending_frames(struct sock *sk, struct flowi6 *fl6, struct raw6_sock *rp) { + struct ipv6_txoptions *opt; struct sk_buff *skb; int err = 0; int offset; @@ -560,6 +561,9 @@ static int rawv6_push_pending_frames(struct sock *sk, struct flowi6 *fl6, offset = rp->offset; total_len = inet_sk(sk)->cork.base.length; + opt = inet6_sk(sk)->cork.opt; + total_len -= opt ? opt->opt_flen : 0; + if (offset >= total_len - 1) { err = -EINVAL; ip6_flush_pending_frames(sk); -- 2.25.1

1 2

[openEuler-5.10 1/4] SUNRPC: Clean up the handling of page padding in rpc_prepare_reply_pages()
by Zheng Zengkai 06 Feb '23

06 Feb '23

From: Trond Myklebust <trond.myklebust(a)hammerspace.com> mainline inclusion from mainline-v5.11-rc1 commit 9ed5af268e88f6e5b65376be98d652b37cb20d7b bugzilla: https://gitee.com/openeuler/kernel/issues/I6D0MJ Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- rpc_prepare_reply_pages() currently expects the 'hdrsize' argument to contain the length of the data that we expect to want placed in the head kvec plus a count of 1 word of padding that is placed after the page data. This is very confusing when trying to read the code, and sometimes leads to callers adding an arbitrary value of '1' just in order to satisfy the requirement (whether or not the page data actually needs such padding). This patch aims to clarify the code by changing the 'hdrsize' argument to remove that 1 word of padding. This means we need to subtract the padding from all the existing callers. Fixes: 02ef04e432ba ("NFS: Account for XDR pad of buf->pages") Signed-off-by: Trond Myklebust <trond.myklebust(a)hammerspace.com> Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> --- fs/nfs/nfs2xdr.c | 19 ++++++++++--------- fs/nfs/nfs3xdr.c | 29 ++++++++++++++++------------- fs/nfs/nfs4xdr.c | 36 +++++++++++++++++++----------------- net/sunrpc/clnt.c | 5 +---- net/sunrpc/xdr.c | 3 --- 5 files changed, 46 insertions(+), 46 deletions(-) diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c index 5e6453e9b307..3d5ba43f44bb 100644 --- a/fs/nfs/nfs2xdr.c +++ b/fs/nfs/nfs2xdr.c @@ -34,6 +34,7 @@ * Declare the space requirements for NFS arguments and replies as * number of 32bit-words */ +#define NFS_pagepad_sz (1) /* Page padding */ #define NFS_fhandle_sz (8) #define NFS_sattr_sz (8) #define NFS_filename_sz (1+(NFS2_MAXNAMLEN>>2)) @@ -56,11 +57,11 @@ #define NFS_attrstat_sz (1+NFS_fattr_sz) #define NFS_diropres_sz (1+NFS_fhandle_sz+NFS_fattr_sz) -#define NFS_readlinkres_sz (2+1) -#define NFS_readres_sz (1+NFS_fattr_sz+1+1) +#define NFS_readlinkres_sz (2+NFS_pagepad_sz) +#define NFS_readres_sz (1+NFS_fattr_sz+1+NFS_pagepad_sz) #define NFS_writeres_sz (NFS_attrstat_sz) #define NFS_stat_sz (1) -#define NFS_readdirres_sz (1+1) +#define NFS_readdirres_sz (1+NFS_pagepad_sz) #define NFS_statfsres_sz (1+NFS_info_sz) static int nfs_stat_to_errno(enum nfs_stat); @@ -592,8 +593,8 @@ static void nfs2_xdr_enc_readlinkargs(struct rpc_rqst *req, const struct nfs_readlinkargs *args = data; encode_fhandle(xdr, args->fh); - rpc_prepare_reply_pages(req, args->pages, args->pgbase, - args->pglen, NFS_readlinkres_sz); + rpc_prepare_reply_pages(req, args->pages, args->pgbase, args->pglen, + NFS_readlinkres_sz - NFS_pagepad_sz); } /* @@ -628,8 +629,8 @@ static void nfs2_xdr_enc_readargs(struct rpc_rqst *req, const struct nfs_pgio_args *args = data; encode_readargs(xdr, args); - rpc_prepare_reply_pages(req, args->pages, args->pgbase, - args->count, NFS_readres_sz); + rpc_prepare_reply_pages(req, args->pages, args->pgbase, args->count, + NFS_readres_sz - NFS_pagepad_sz); req->rq_rcv_buf.flags |= XDRBUF_READ; } @@ -786,8 +787,8 @@ static void nfs2_xdr_enc_readdirargs(struct rpc_rqst *req, const struct nfs_readdirargs *args = data; encode_readdirargs(xdr, args); - rpc_prepare_reply_pages(req, args->pages, 0, - args->count, NFS_readdirres_sz); + rpc_prepare_reply_pages(req, args->pages, 0, args->count, + NFS_readdirres_sz - NFS_pagepad_sz); } /* diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c index b5a9379b1450..2c0cad7c51d6 100644 --- a/fs/nfs/nfs3xdr.c +++ b/fs/nfs/nfs3xdr.c @@ -33,6 +33,7 @@ * Declare the space requirements for NFS arguments and replies as * number of 32bit-words */ +#define NFS3_pagepad_sz (1) /* Page padding */ #define NFS3_fhandle_sz (1+16) #define NFS3_fh_sz (NFS3_fhandle_sz) /* shorthand */ #define NFS3_post_op_fh_sz (1+NFS3_fh_sz) @@ -70,13 +71,13 @@ #define NFS3_removeres_sz (NFS3_setattrres_sz) #define NFS3_lookupres_sz (1+NFS3_fh_sz+(2 * NFS3_post_op_attr_sz)) #define NFS3_accessres_sz (1+NFS3_post_op_attr_sz+1) -#define NFS3_readlinkres_sz (1+NFS3_post_op_attr_sz+1+1) -#define NFS3_readres_sz (1+NFS3_post_op_attr_sz+3+1) +#define NFS3_readlinkres_sz (1+NFS3_post_op_attr_sz+1+NFS3_pagepad_sz) +#define NFS3_readres_sz (1+NFS3_post_op_attr_sz+3+NFS3_pagepad_sz) #define NFS3_writeres_sz (1+NFS3_wcc_data_sz+4) #define NFS3_createres_sz (1+NFS3_post_op_fh_sz+NFS3_post_op_attr_sz+NFS3_wcc_data_sz) #define NFS3_renameres_sz (1+(2 * NFS3_wcc_data_sz)) #define NFS3_linkres_sz (1+NFS3_post_op_attr_sz+NFS3_wcc_data_sz) -#define NFS3_readdirres_sz (1+NFS3_post_op_attr_sz+2+1) +#define NFS3_readdirres_sz (1+NFS3_post_op_attr_sz+2+NFS3_pagepad_sz) #define NFS3_fsstatres_sz (1+NFS3_post_op_attr_sz+13) #define NFS3_fsinfores_sz (1+NFS3_post_op_attr_sz+12) #define NFS3_pathconfres_sz (1+NFS3_post_op_attr_sz+6) @@ -86,7 +87,8 @@ #define ACL3_setaclargs_sz (NFS3_fh_sz+1+ \ XDR_QUADLEN(NFS_ACL_INLINE_BUFSIZE)) #define ACL3_getaclres_sz (1+NFS3_post_op_attr_sz+1+ \ - XDR_QUADLEN(NFS_ACL_INLINE_BUFSIZE)+1) + XDR_QUADLEN(NFS_ACL_INLINE_BUFSIZE)+\ + NFS3_pagepad_sz) #define ACL3_setaclres_sz (1+NFS3_post_op_attr_sz) static int nfs3_stat_to_errno(enum nfs_stat); @@ -910,8 +912,8 @@ static void nfs3_xdr_enc_readlink3args(struct rpc_rqst *req, const struct nfs3_readlinkargs *args = data; encode_nfs_fh3(xdr, args->fh); - rpc_prepare_reply_pages(req, args->pages, args->pgbase, - args->pglen, NFS3_readlinkres_sz); + rpc_prepare_reply_pages(req, args->pages, args->pgbase, args->pglen, + NFS3_readlinkres_sz - NFS3_pagepad_sz); } /* @@ -940,7 +942,8 @@ static void nfs3_xdr_enc_read3args(struct rpc_rqst *req, const void *data) { const struct nfs_pgio_args *args = data; - unsigned int replen = args->replen ? args->replen : NFS3_readres_sz; + unsigned int replen = args->replen ? args->replen : + NFS3_readres_sz - NFS3_pagepad_sz; encode_read3args(xdr, args); rpc_prepare_reply_pages(req, args->pages, args->pgbase, @@ -1240,8 +1243,8 @@ static void nfs3_xdr_enc_readdir3args(struct rpc_rqst *req, const struct nfs3_readdirargs *args = data; encode_readdir3args(xdr, args); - rpc_prepare_reply_pages(req, args->pages, 0, - args->count, NFS3_readdirres_sz); + rpc_prepare_reply_pages(req, args->pages, 0, args->count, + NFS3_readdirres_sz - NFS3_pagepad_sz); } /* @@ -1282,8 +1285,8 @@ static void nfs3_xdr_enc_readdirplus3args(struct rpc_rqst *req, const struct nfs3_readdirargs *args = data; encode_readdirplus3args(xdr, args); - rpc_prepare_reply_pages(req, args->pages, 0, - args->count, NFS3_readdirres_sz); + rpc_prepare_reply_pages(req, args->pages, 0, args->count, + NFS3_readdirres_sz - NFS3_pagepad_sz); } /* @@ -1329,7 +1332,7 @@ static void nfs3_xdr_enc_getacl3args(struct rpc_rqst *req, if (args->mask & (NFS_ACL | NFS_DFACL)) { rpc_prepare_reply_pages(req, args->pages, 0, NFSACL_MAXPAGES << PAGE_SHIFT, - ACL3_getaclres_sz); + ACL3_getaclres_sz - NFS3_pagepad_sz); req->rq_rcv_buf.flags |= XDRBUF_SPARSE_PAGES; } } @@ -1649,7 +1652,7 @@ static int nfs3_xdr_dec_read3res(struct rpc_rqst *req, struct xdr_stream *xdr, result->op_status = status; if (status != NFS3_OK) goto out_status; - result->replen = 4 + ((xdr_stream_pos(xdr) - pos) >> 2); + result->replen = 3 + ((xdr_stream_pos(xdr) - pos) >> 2); error = decode_read3resok(xdr, result); out: return error; diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c index e2f0e3446e22..f64b28d5bc22 100644 --- a/fs/nfs/nfs4xdr.c +++ b/fs/nfs/nfs4xdr.c @@ -84,6 +84,7 @@ static int decode_layoutget(struct xdr_stream *xdr, struct rpc_rqst *req, /* lock,open owner id: * we currently use size 2 (u64) out of (NFS4_OPAQUE_LIMIT >> 2) */ +#define pagepad_maxsz (1) #define open_owner_id_maxsz (1 + 2 + 1 + 1 + 2) #define lock_owner_id_maxsz (1 + 1 + 4) #define decode_lockowner_maxsz (1 + XDR_QUADLEN(IDMAP_NAMESZ)) @@ -215,14 +216,14 @@ static int decode_layoutget(struct xdr_stream *xdr, struct rpc_rqst *req, nfs4_fattr_bitmap_maxsz) #define encode_read_maxsz (op_encode_hdr_maxsz + \ encode_stateid_maxsz + 3) -#define decode_read_maxsz (op_decode_hdr_maxsz + 2 + 1) +#define decode_read_maxsz (op_decode_hdr_maxsz + 2 + pagepad_maxsz) #define encode_readdir_maxsz (op_encode_hdr_maxsz + \ 2 + encode_verifier_maxsz + 5 + \ nfs4_label_maxsz) #define decode_readdir_maxsz (op_decode_hdr_maxsz + \ - decode_verifier_maxsz + 1) + decode_verifier_maxsz + pagepad_maxsz) #define encode_readlink_maxsz (op_encode_hdr_maxsz) -#define decode_readlink_maxsz (op_decode_hdr_maxsz + 1 + 1) +#define decode_readlink_maxsz (op_decode_hdr_maxsz + 1 + pagepad_maxsz) #define encode_write_maxsz (op_encode_hdr_maxsz + \ encode_stateid_maxsz + 4) #define decode_write_maxsz (op_decode_hdr_maxsz + \ @@ -284,14 +285,14 @@ static int decode_layoutget(struct xdr_stream *xdr, struct rpc_rqst *req, #define decode_delegreturn_maxsz (op_decode_hdr_maxsz) #define encode_getacl_maxsz (encode_getattr_maxsz) #define decode_getacl_maxsz (op_decode_hdr_maxsz + \ - nfs4_fattr_bitmap_maxsz + 1 + 1) + nfs4_fattr_bitmap_maxsz + 1 + pagepad_maxsz) #define encode_setacl_maxsz (op_encode_hdr_maxsz + \ encode_stateid_maxsz + 3) #define decode_setacl_maxsz (decode_setattr_maxsz) #define encode_fs_locations_maxsz \ (encode_getattr_maxsz) #define decode_fs_locations_maxsz \ - (1) + (pagepad_maxsz) #define encode_secinfo_maxsz (op_encode_hdr_maxsz + nfs4_name_maxsz) #define decode_secinfo_maxsz (op_decode_hdr_maxsz + 1 + ((NFS_MAX_SECFLAVORS * (16 + GSS_OID_MAX_LEN)) / 4)) @@ -393,12 +394,13 @@ static int decode_layoutget(struct xdr_stream *xdr, struct rpc_rqst *req, /* devaddr4 payload is read into page */ \ 1 /* notification bitmap length */ + \ 1 /* notification bitmap, word 0 */ + \ - 1 /* possible XDR padding */) + pagepad_maxsz /* possible XDR padding */) #define encode_layoutget_maxsz (op_encode_hdr_maxsz + 10 + \ encode_stateid_maxsz) #define decode_layoutget_maxsz (op_decode_hdr_maxsz + 8 + \ decode_stateid_maxsz + \ - XDR_QUADLEN(PNFS_LAYOUT_MAXSIZE) + 1) + XDR_QUADLEN(PNFS_LAYOUT_MAXSIZE) + \ + pagepad_maxsz) #define encode_layoutcommit_maxsz (op_encode_hdr_maxsz + \ 2 /* offset */ + \ 2 /* length */ + \ @@ -2342,7 +2344,7 @@ static void nfs4_xdr_enc_open(struct rpc_rqst *req, struct xdr_stream *xdr, encode_layoutget(xdr, args->lg_args, &hdr); rpc_prepare_reply_pages(req, args->lg_args->layout.pages, 0, args->lg_args->layout.pglen, - hdr.replen); + hdr.replen - pagepad_maxsz); } encode_nops(&hdr); } @@ -2388,7 +2390,7 @@ static void nfs4_xdr_enc_open_noattr(struct rpc_rqst *req, encode_layoutget(xdr, args->lg_args, &hdr); rpc_prepare_reply_pages(req, args->lg_args->layout.pages, 0, args->lg_args->layout.pglen, - hdr.replen); + hdr.replen - pagepad_maxsz); } encode_nops(&hdr); } @@ -2499,7 +2501,7 @@ static void nfs4_xdr_enc_readlink(struct rpc_rqst *req, struct xdr_stream *xdr, encode_readlink(xdr, args, req, &hdr); rpc_prepare_reply_pages(req, args->pages, args->pgbase, - args->pglen, hdr.replen); + args->pglen, hdr.replen - pagepad_maxsz); encode_nops(&hdr); } @@ -2520,7 +2522,7 @@ static void nfs4_xdr_enc_readdir(struct rpc_rqst *req, struct xdr_stream *xdr, encode_readdir(xdr, args, req, &hdr); rpc_prepare_reply_pages(req, args->pages, args->pgbase, - args->count, hdr.replen); + args->count, hdr.replen - pagepad_maxsz); encode_nops(&hdr); } @@ -2541,7 +2543,7 @@ static void nfs4_xdr_enc_read(struct rpc_rqst *req, struct xdr_stream *xdr, encode_read(xdr, args, &hdr); rpc_prepare_reply_pages(req, args->pages, args->pgbase, - args->count, hdr.replen); + args->count, hdr.replen - pagepad_maxsz); req->rq_rcv_buf.flags |= XDRBUF_READ; encode_nops(&hdr); } @@ -2588,7 +2590,7 @@ static void nfs4_xdr_enc_getacl(struct rpc_rqst *req, struct xdr_stream *xdr, ARRAY_SIZE(nfs4_acl_bitmap), &hdr); rpc_prepare_reply_pages(req, args->acl_pages, 0, - args->acl_len, replen + 1); + args->acl_len, replen); encode_nops(&hdr); } @@ -2810,7 +2812,7 @@ static void nfs4_xdr_enc_fs_locations(struct rpc_rqst *req, } rpc_prepare_reply_pages(req, (struct page **)&args->page, 0, - PAGE_SIZE, replen + 1); + PAGE_SIZE, replen); encode_nops(&hdr); } @@ -3014,14 +3016,14 @@ static void nfs4_xdr_enc_getdeviceinfo(struct rpc_rqst *req, encode_compound_hdr(xdr, req, &hdr); encode_sequence(xdr, &args->seq_args, &hdr); - replen = hdr.replen + op_decode_hdr_maxsz; + replen = hdr.replen + op_decode_hdr_maxsz + 2; encode_getdeviceinfo(xdr, args, &hdr); /* set up reply kvec. device_addr4 opaque data is read into the * pages */ rpc_prepare_reply_pages(req, args->pdev->pages, args->pdev->pgbase, - args->pdev->pglen, replen + 2 + 1); + args->pdev->pglen, replen); encode_nops(&hdr); } @@ -3043,7 +3045,7 @@ static void nfs4_xdr_enc_layoutget(struct rpc_rqst *req, encode_layoutget(xdr, args, &hdr); rpc_prepare_reply_pages(req, args->layout.pages, 0, - args->layout.pglen, hdr.replen); + args->layout.pglen, hdr.replen - pagepad_maxsz); encode_nops(&hdr); } diff --git a/net/sunrpc/clnt.c b/net/sunrpc/clnt.c index 78c6648af782..88a4dff14666 100644 --- a/net/sunrpc/clnt.c +++ b/net/sunrpc/clnt.c @@ -1251,10 +1251,7 @@ void rpc_prepare_reply_pages(struct rpc_rqst *req, struct page **pages, unsigned int base, unsigned int len, unsigned int hdrsize) { - /* Subtract one to force an extra word of buffer space for the - * payload's XDR pad to fall into the rcv_buf's tail iovec. - */ - hdrsize += RPC_REPHDRSIZE + req->rq_cred->cr_auth->au_ralign - 1; + hdrsize += RPC_REPHDRSIZE + req->rq_cred->cr_auth->au_ralign; xdr_inline_pages(&req->rq_rcv_buf, hdrsize << 2, pages, base, len); trace_rpc_xdr_reply_pages(req->rq_task, &req->rq_rcv_buf); diff --git a/net/sunrpc/xdr.c b/net/sunrpc/xdr.c index d84bb5037bb5..b1684b19b6d5 100644 --- a/net/sunrpc/xdr.c +++ b/net/sunrpc/xdr.c @@ -193,9 +193,6 @@ xdr_inline_pages(struct xdr_buf *xdr, unsigned int offset, tail->iov_base = buf + offset; tail->iov_len = buflen - offset; - if ((xdr->page_len & 3) == 0) - tail->iov_len -= sizeof(__be32); - xdr->buflen += len; } EXPORT_SYMBOL_GPL(xdr_inline_pages); -- 2.20.1

1 3

[PATCH OLK-5.10] net: hns3: add support handling tx dhcp packets for ROH
by Ke Chen 06 Feb '23

06 Feb '23

driver inclusion category: feature bugzilla: https://gitee.com/openeuler/kernel/issues/I6BSMN ----------------------------------------------------------------------- For ROH distributed scenario, EID is allocated by DHCP mode. Driver needs to convert the origin MAC address to EID format, and updates the destination MAC, chaddr and client id(if exists) when transmit DHCP packets. Meantime, the chaddr field should follow the source mac address, in order to make the dhcp server reply to the right client. For the payload of dhcp packet changed, so the checksum of L4 should be calculated too. Signed-off-by: Jian Shen <shenjian15(a)huawei.com> Signed-off-by: Ke Chen <chenke54(a)huawei.com> --- .../net/ethernet/hisilicon/hns3/hns3_enet.c | 172 +++++++++++++++++- .../net/ethernet/hisilicon/hns3/hns3_enet.h | 50 +++++ .../hisilicon/hns3/hns3pf/hclge_main.c | 9 + 3 files changed, 226 insertions(+), 5 deletions(-) diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c index cf79cd69c766..460f9d217a18 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.c @@ -1165,6 +1165,142 @@ static void hns3_tx_spare_reclaim_cb(struct hns3_enet_ring *ring, } } +static struct hns3_dhcp_packet *hns3_get_dhcp_packet(struct sk_buff *skb, + int *dhcp_len) +{ + struct hns3_dhcp_packet *dhcp; + union l4_hdr_info l4; + int l4_payload_len; + + l4.hdr = skb_transport_header(skb); + if (l4.udp->dest != htons(HNS3_DHCP_CLIENT_PORT) || + l4.udp->source != htons(HNS3_DHCP_SERVER_PORT)) + return NULL; + + dhcp = (struct hns3_dhcp_packet *)(l4.hdr + sizeof(struct udphdr)); + l4_payload_len = ntohs(l4.udp->len) - sizeof(struct udphdr); + if (l4_payload_len < offsetof(struct hns3_dhcp_packet, options) || + dhcp->hlen != ETH_ALEN || + dhcp->cookie != htonl(HNS3_DHCP_MAGIC)) + return NULL; + + *dhcp_len = l4_payload_len; + return dhcp; +} + +static u8 *hns3_dhcp_option_scan(struct hns3_dhcp_packet *packet, + struct hns3_dhcp_opt_state *opt_state) +{ + int opt_len; + u8 *cur_opt; + + /* option bytes: [code][len][data0~data[len-1]] */ + while (opt_state->rem > 0) { + switch (opt_state->opt_ptr[DHCP_OPT_CODE]) { + /* option padding and end have no len and data byte. */ + case DHCP_OPT_PADDING: + opt_state->rem--; + opt_state->opt_ptr++; + break; + case DHCP_OPT_END: + if (DHCP_OVERLOAD_USE_FILE(opt_state->overload_flag)) { + opt_state->overload_flag |= + DHCP_OVERLOAD_FILE_USED; + opt_state->opt_ptr = packet->file; + opt_state->rem = sizeof(packet->file); + break; + } + if (DHCP_OVERLOAD_USE_SNAME(opt_state->overload_flag)) { + opt_state->overload_flag |= + DHCP_OVERLOAD_SNAME_USED; + opt_state->opt_ptr = packet->sname; + opt_state->rem = sizeof(packet->sname); + break; + } + return NULL; + default: + if (opt_state->rem <= DHCP_OPT_LEN) + return NULL; + /* opt_len includes code, len and data bytes */ + opt_len = opt_state->opt_ptr[DHCP_OPT_LEN] + + DHCP_OPT_DATA; + cur_opt = opt_state->opt_ptr; + if (opt_state->rem < opt_len) + return NULL; + + opt_state->opt_ptr += opt_len; + opt_state->rem -= opt_len; + if (cur_opt[DHCP_OPT_CODE] == DHCP_OPT_OVERLOAD) { + opt_state->overload_flag |= + cur_opt[DHCP_OPT_DATA]; + break; + } + return cur_opt; + } + } + + return NULL; +} + +static void hns3_dhcp_update_option61(struct hns3_nic_priv *priv, + struct hns3_dhcp_packet *packet, + int dhcp_len) +{ + struct hns3_dhcp_opt_state opt_state; + u8 *cur_opt; + + opt_state.opt_ptr = packet->options; + opt_state.rem = dhcp_len - offsetof(struct hns3_dhcp_packet, options); + opt_state.overload_flag = 0; + + cur_opt = hns3_dhcp_option_scan(packet, &opt_state); + while (cur_opt) { + if (cur_opt[DHCP_OPT_CODE] != DHCP_OPT_CLIENT_ID) { + cur_opt = hns3_dhcp_option_scan(packet, &opt_state); + continue; + } + if (cur_opt[DHCP_OPT_LEN] > ETH_ALEN) + ether_addr_copy(&cur_opt[DHCP_CLIENT_ID_MAC_OFT], + priv->roh_perm_mac); + break; + } +} + +static void hns3_dhcp_cal_l4_csum(struct sk_buff *skb) +{ + union l3_hdr_info l3; + union l4_hdr_info l4; + __wsum csum = 0; + int offset; + + if (skb->ip_summed == CHECKSUM_PARTIAL) + return; + + l3.hdr = skb_network_header(skb); + l4.hdr = skb_transport_header(skb); + offset = skb_transport_offset(skb); + l4.udp->check = 0; + csum = csum_partial(l4.udp, ntohs(l4.udp->len), 0); + l4.udp->check = csum_tcpudp_magic(l3.v4->saddr, l3.v4->daddr, + skb->len - offset, IPPROTO_UDP, csum); +} + +static void hns3_dhcp_packet_convert(struct hns3_nic_priv *priv, + struct sk_buff *skb, + struct hns3_dhcp_packet *dhcp, + int dhcp_len) +{ + struct ethhdr *l2hdr = eth_hdr(skb); + + if (!dhcp) + return; + + ether_addr_copy(dhcp->chaddr, l2hdr->h_source); + hns3_dhcp_update_option61(priv, dhcp, dhcp_len); + /* for l4 payload changed, need to re-calculate the csum */ + hns3_dhcp_cal_l4_csum(skb); +} + static int hns3_set_tso(struct sk_buff *skb, u32 *paylen_fdop_ol4cs, u16 *mss, u32 *type_cs_vlan_tso, u32 *send_bytes) { @@ -1716,7 +1852,20 @@ static int hns3_handle_csum_partial(struct hns3_enet_ring *ring, return 0; } -static int hns3_fill_skb_desc(struct hns3_enet_ring *ring, +static bool hns3_roh_check_udpv4(struct sk_buff *skb) +{ + union l3_hdr_info l3; + + l3.hdr = skb_network_header(skb); + if (skb->protocol != htons(ETH_P_IP) || + l3.v4->version != IP_VERSION_IPV4) + return false; + + return l3.v4->protocol == IPPROTO_UDP; +} + +static int hns3_fill_skb_desc(struct hns3_nic_priv *priv, + struct hns3_enet_ring *ring, struct sk_buff *skb, struct hns3_desc *desc, struct hns3_desc_cb *desc_cb) { @@ -1741,6 +1890,14 @@ static int hns3_fill_skb_desc(struct hns3_enet_ring *ring, hnae3_set_field(param.paylen_fdop_ol4cs, HNS3_TXD_FD_OP_M, HNS3_TXD_FD_OP_S, fd_op); + if (hns3_roh_check_udpv4(skb)) { + struct hns3_dhcp_packet *dhcp; + int dhcp_len; + + dhcp = hns3_get_dhcp_packet(skb, &dhcp_len); + hns3_dhcp_packet_convert(priv, skb, dhcp, dhcp_len); + } + /* Set txbd */ desc->tx.ol_type_vlan_len_msec = cpu_to_le32(param.ol_type_vlan_len_msec); @@ -2338,15 +2495,16 @@ static int hns3_handle_desc_filling(struct hns3_enet_ring *ring, return hns3_fill_skb_to_desc(ring, skb, DESC_TYPE_SKB); } -static int hns3_handle_skb_desc(struct hns3_enet_ring *ring, +static int hns3_handle_skb_desc(struct hns3_nic_priv *priv, + struct hns3_enet_ring *ring, struct sk_buff *skb, struct hns3_desc_cb *desc_cb, int next_to_use_head) { int ret; - ret = hns3_fill_skb_desc(ring, skb, &ring->desc[ring->next_to_use], - desc_cb); + ret = hns3_fill_skb_desc(priv, ring, skb, + &ring->desc[ring->next_to_use], desc_cb); if (unlikely(ret < 0)) goto fill_err; @@ -2395,7 +2553,7 @@ netdev_tx_t hns3_nic_net_xmit(struct sk_buff *skb, struct net_device *netdev) goto out_err_tx_ok; } - ret = hns3_handle_skb_desc(ring, skb, desc_cb, ring->next_to_use); + ret = hns3_handle_skb_desc(priv, ring, skb, desc_cb, ring->next_to_use); if (unlikely(ret <= 0)) goto out_err_tx_ok; @@ -5226,6 +5384,9 @@ static int hns3_init_mac_addr(struct net_device *netdev) return 0; } + if (is_zero_ether_addr(priv->roh_perm_mac)) + ether_addr_copy(priv->roh_perm_mac, netdev->dev_addr); + if (h->ae_algo->ops->set_mac_addr) ret = h->ae_algo->ops->set_mac_addr(h, netdev->dev_addr, true); @@ -5377,6 +5538,7 @@ static int hns3_client_init(struct hnae3_handle *handle) priv->tx_timeout_count = 0; priv->max_non_tso_bd_num = ae_dev->dev_specs.max_non_tso_bd_num; set_bit(HNS3_NIC_STATE_DOWN, &priv->state); + eth_zero_addr(priv->roh_perm_mac); handle->msg_enable = netif_msg_init(debug, DEFAULT_MSG_LEVEL); diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h index ccfd38b0028e..85c352fff83b 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h +++ b/drivers/net/ethernet/hisilicon/hns3/hns3_enet.h @@ -604,6 +604,56 @@ struct hns3_nic_priv { struct hns3_enet_coalesce rx_coal; u32 tx_copybreak; u32 rx_copybreak; + u8 roh_perm_mac[ETH_ALEN]; +}; + +#define HNS3_DHCP_SERVER_PORT 68 +#define HNS3_DHCP_CLIENT_PORT 67 +#define HNS3_DHCP_MAGIC 0x63825363 +#define DHCP_OPT_CODE 0 +#define DHCP_OPT_LEN 1 +#define DHCP_OPT_DATA 2 +#define DHCP_CLIENT_ID_LEN 7 +#define DHCP_CLIENT_ID_MAC_OFT 3 +#define DHCP_OVERLOAD_FILE 0x1 +#define DHCP_OVERLOAD_SNAME 0x2 +#define DHCP_OVERLOAD_FILE_USED 0x101 +#define DHCP_OVERLOAD_SNAME_USED 0x202 +#define DHCP_OVERLOAD_USE_FILE(x) \ + (((x) & DHCP_OVERLOAD_FILE_USED) == DHCP_OVERLOAD_FILE) +#define DHCP_OVERLOAD_USE_SNAME(x) \ + (((x) & DHCP_OVERLOAD_SNAME_USED) == DHCP_OVERLOAD_SNAME) + +enum DHCP_OPTION_CODES { + DHCP_OPT_PADDING = 0, + DHCP_OPT_OVERLOAD = 52, + DHCP_OPT_CLIENT_ID = 61, + DHCP_OPT_END = 255 +}; + +struct hns3_dhcp_packet { + u8 op; + u8 htype; + u8 hlen; + u8 hops; + u32 xid; + u16 secs; + u16 flags; + u32 ciaddr; + u32 yiaddr; + u32 siaddr_nip; + u32 gateway_nip; + u8 chaddr[16]; /* link-layer client hardware address (MAC) */ + u8 sname[64]; + u8 file[128]; + u32 cookie; /* DHCP magic bytes: 0x63825363 */ + u8 options[312]; +}; + +struct hns3_dhcp_opt_state { + u8 *opt_ptr; /* refer to current option item */ + int rem; /* remain bytes in options */ + u32 overload_flag; /* whether use file and sname field as options */ }; union l3_hdr_info { diff --git a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c index eea17548416b..5c8a821aa61a 100644 --- a/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c +++ b/drivers/net/ethernet/hisilicon/hns3/hns3pf/hclge_main.c @@ -2866,12 +2866,21 @@ static void hclge_get_fec(struct hnae3_handle *handle, u8 *fec_ability, if (fec_mode) *fec_mode = mac->fec_mode; } + +static void hclge_roh_convert_mac_addr(struct hclge_dev *hdev) +{ +#define HCLGE_ROH_EID_MASK_BYTE 3 + + memset(hdev->hw.mac.mac_addr, 0, HCLGE_ROH_EID_MASK_BYTE); +} + static int hclge_mac_init(struct hclge_dev *hdev) { struct hclge_mac *mac = &hdev->hw.mac; int ret; hclge_mac_type_init(hdev); + hclge_roh_convert_mac_addr(hdev); hdev->support_sfp_query = true; hdev->hw.mac.duplex = HCLGE_MAC_FULL; -- 2.30.0

1 0

[OLK-5.10 v2 1/3] net: Use u64_stats_fetch_begin_irq() for stats fetch.
by Wang Yufen 06 Feb '23

06 Feb '23

From: Sebastian Andrzej Siewior <bigeasy(a)linutronix.de> stable inclusion from stable-v5.10.142 commit d71a1c9fce184718d1b3a51a9e8a6e31cbbb45ce category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6D0ZE Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… ------------------------------------------------- commit 278d3ba61563ceed3cb248383ced19e14ec7bc1f upstream. On 32bit-UP u64_stats_fetch_begin() disables only preemption. If the reader is in preemptible context and the writer side (u64_stats_update_begin*()) runs in an interrupt context (IRQ or softirq) then the writer can update the stats during the read operation. This update remains undetected. Use u64_stats_fetch_begin_irq() to ensure the stats fetch on 32bit-UP are not interrupted by a writer. 32bit-SMP remains unaffected by this change. Cc: "David S. Miller" <davem(a)davemloft.net> Cc: Catherine Sullivan <csully(a)google.com> Cc: David Awogbemila <awogbemila(a)google.com> Cc: Dimitris Michailidis <dmichail(a)fungible.com> Cc: Eric Dumazet <edumazet(a)google.com> Cc: Hans Ulli Kroll <ulli.kroll(a)googlemail.com> Cc: Jakub Kicinski <kuba(a)kernel.org> Cc: Jeroen de Borst <jeroendb(a)google.com> Cc: Johannes Berg <johannes(a)sipsolutions.net> Cc: Linus Walleij <linus.walleij(a)linaro.org> Cc: Paolo Abeni <pabeni(a)redhat.com> Cc: Simon Horman <simon.horman(a)corigine.com> Cc: linux-arm-kernel(a)lists.infradead.org Cc: linux-wireless(a)vger.kernel.org Cc: netdev(a)vger.kernel.org Cc: oss-drivers(a)corigine.com Cc: stable(a)vger.kernel.org Signed-off-by: Sebastian Andrzej Siewior <bigeasy(a)linutronix.de> Reviewed-by: Simon Horman <simon.horman(a)corigine.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> (cherry picked from commit d71a1c9fce184718d1b3a51a9e8a6e31cbbb45ce) Signed-off-by: Wang Yufen <wangyufen(a)huawei.com> Conflicts: drivers/net/ethernet/huawei/hinic/hinic_rx.c drivers/net/ethernet/huawei/hinic/hinic_tx.c Signed-off-by: Wang Yufen <wangyufen(a)huawei.com> --- drivers/net/ethernet/cortina/gemini.c | 24 +++++++++++----------- drivers/net/ethernet/google/gve/gve_ethtool.c | 16 +++++++-------- drivers/net/ethernet/google/gve/gve_main.c | 12 +++++------ drivers/net/ethernet/huawei/hinic/hinic_rx.c | 4 ++-- drivers/net/ethernet/huawei/hinic/hinic_tx.c | 4 ++-- .../net/ethernet/netronome/nfp/nfp_net_common.c | 8 ++++---- .../net/ethernet/netronome/nfp/nfp_net_ethtool.c | 8 ++++---- drivers/net/netdevsim/netdev.c | 4 ++-- net/mac80211/sta_info.c | 8 ++++---- net/mpls/af_mpls.c | 4 ++-- 10 files changed, 46 insertions(+), 46 deletions(-) diff --git a/drivers/net/ethernet/cortina/gemini.c b/drivers/net/ethernet/cortina/gemini.c index 3685878..b22ea40 100644 --- a/drivers/net/ethernet/cortina/gemini.c +++ b/drivers/net/ethernet/cortina/gemini.c @@ -1920,7 +1920,7 @@ static void gmac_get_stats64(struct net_device *netdev, /* Racing with RX NAPI */ do { - start = u64_stats_fetch_begin(&port->rx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->rx_stats_syncp); stats->rx_packets = port->stats.rx_packets; stats->rx_bytes = port->stats.rx_bytes; @@ -1932,11 +1932,11 @@ static void gmac_get_stats64(struct net_device *netdev, stats->rx_crc_errors = port->stats.rx_crc_errors; stats->rx_frame_errors = port->stats.rx_frame_errors; - } while (u64_stats_fetch_retry(&port->rx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->rx_stats_syncp, start)); /* Racing with MIB and TX completion interrupts */ do { - start = u64_stats_fetch_begin(&port->ir_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->ir_stats_syncp); stats->tx_errors = port->stats.tx_errors; stats->tx_packets = port->stats.tx_packets; @@ -1946,15 +1946,15 @@ static void gmac_get_stats64(struct net_device *netdev, stats->rx_missed_errors = port->stats.rx_missed_errors; stats->rx_fifo_errors = port->stats.rx_fifo_errors; - } while (u64_stats_fetch_retry(&port->ir_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->ir_stats_syncp, start)); /* Racing with hard_start_xmit */ do { - start = u64_stats_fetch_begin(&port->tx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->tx_stats_syncp); stats->tx_dropped = port->stats.tx_dropped; - } while (u64_stats_fetch_retry(&port->tx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->tx_stats_syncp, start)); stats->rx_dropped += stats->rx_missed_errors; } @@ -2032,18 +2032,18 @@ static void gmac_get_ethtool_stats(struct net_device *netdev, /* Racing with MIB interrupt */ do { p = values; - start = u64_stats_fetch_begin(&port->ir_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->ir_stats_syncp); for (i = 0; i < RX_STATS_NUM; i++) *p++ = port->hw_stats[i]; - } while (u64_stats_fetch_retry(&port->ir_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->ir_stats_syncp, start)); values = p; /* Racing with RX NAPI */ do { p = values; - start = u64_stats_fetch_begin(&port->rx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->rx_stats_syncp); for (i = 0; i < RX_STATUS_NUM; i++) *p++ = port->rx_stats[i]; @@ -2051,13 +2051,13 @@ static void gmac_get_ethtool_stats(struct net_device *netdev, *p++ = port->rx_csum_stats[i]; *p++ = port->rx_napi_exits; - } while (u64_stats_fetch_retry(&port->rx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->rx_stats_syncp, start)); values = p; /* Racing with TX start_xmit */ do { p = values; - start = u64_stats_fetch_begin(&port->tx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->tx_stats_syncp); for (i = 0; i < TX_MAX_FRAGS; i++) { *values++ = port->tx_frag_stats[i]; @@ -2066,7 +2066,7 @@ static void gmac_get_ethtool_stats(struct net_device *netdev, *values++ = port->tx_frags_linearized; *values++ = port->tx_hw_csummed; - } while (u64_stats_fetch_retry(&port->tx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->tx_stats_syncp, start)); } static int gmac_get_ksettings(struct net_device *netdev, diff --git a/drivers/net/ethernet/google/gve/gve_ethtool.c b/drivers/net/ethernet/google/gve/gve_ethtool.c index 66f9b37..80a8c0c 100644 --- a/drivers/net/ethernet/google/gve/gve_ethtool.c +++ b/drivers/net/ethernet/google/gve/gve_ethtool.c @@ -172,14 +172,14 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) struct gve_rx_ring *rx = &priv->rx[ring]; start = - u64_stats_fetch_begin(&priv->rx[ring].statss); + u64_stats_fetch_begin_irq(&priv->rx[ring].statss); tmp_rx_pkts = rx->rpackets; tmp_rx_bytes = rx->rbytes; tmp_rx_skb_alloc_fail = rx->rx_skb_alloc_fail; tmp_rx_buf_alloc_fail = rx->rx_buf_alloc_fail; tmp_rx_desc_err_dropped_pkt = rx->rx_desc_err_dropped_pkt; - } while (u64_stats_fetch_retry(&priv->rx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->rx[ring].statss, start)); rx_pkts += tmp_rx_pkts; rx_bytes += tmp_rx_bytes; @@ -193,10 +193,10 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) if (priv->tx) { do { start = - u64_stats_fetch_begin(&priv->tx[ring].statss); + u64_stats_fetch_begin_irq(&priv->tx[ring].statss); tmp_tx_pkts = priv->tx[ring].pkt_done; tmp_tx_bytes = priv->tx[ring].bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->tx[ring].statss, start)); tx_pkts += tmp_tx_pkts; tx_bytes += tmp_tx_bytes; @@ -254,13 +254,13 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) data[i++] = rx->cnt; do { start = - u64_stats_fetch_begin(&priv->rx[ring].statss); + u64_stats_fetch_begin_irq(&priv->rx[ring].statss); tmp_rx_bytes = rx->rbytes; tmp_rx_skb_alloc_fail = rx->rx_skb_alloc_fail; tmp_rx_buf_alloc_fail = rx->rx_buf_alloc_fail; tmp_rx_desc_err_dropped_pkt = rx->rx_desc_err_dropped_pkt; - } while (u64_stats_fetch_retry(&priv->rx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->rx[ring].statss, start)); data[i++] = tmp_rx_bytes; /* rx dropped packets */ @@ -313,9 +313,9 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) data[i++] = tx->done; do { start = - u64_stats_fetch_begin(&priv->tx[ring].statss); + u64_stats_fetch_begin_irq(&priv->tx[ring].statss); tmp_tx_bytes = tx->bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->tx[ring].statss, start)); data[i++] = tmp_tx_bytes; data[i++] = tx->wake_queue; diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index 6cb75bb..f0c1e6c8 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -40,10 +40,10 @@ static void gve_get_stats(struct net_device *dev, struct rtnl_link_stats64 *s) for (ring = 0; ring < priv->rx_cfg.num_queues; ring++) { do { start = - u64_stats_fetch_begin(&priv->rx[ring].statss); + u64_stats_fetch_begin_irq(&priv->rx[ring].statss); packets = priv->rx[ring].rpackets; bytes = priv->rx[ring].rbytes; - } while (u64_stats_fetch_retry(&priv->rx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->rx[ring].statss, start)); s->rx_packets += packets; s->rx_bytes += bytes; @@ -53,10 +53,10 @@ static void gve_get_stats(struct net_device *dev, struct rtnl_link_stats64 *s) for (ring = 0; ring < priv->tx_cfg.num_queues; ring++) { do { start = - u64_stats_fetch_begin(&priv->tx[ring].statss); + u64_stats_fetch_begin_irq(&priv->tx[ring].statss); packets = priv->tx[ring].pkt_done; bytes = priv->tx[ring].bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->tx[ring].statss, start)); s->tx_packets += packets; s->tx_bytes += bytes; @@ -1041,9 +1041,9 @@ void gve_handle_report_stats(struct gve_priv *priv) if (priv->tx) { for (idx = 0; idx < priv->tx_cfg.num_queues; idx++) { do { - start = u64_stats_fetch_begin(&priv->tx[idx].statss); + start = u64_stats_fetch_begin_irq(&priv->tx[idx].statss); tx_bytes = priv->tx[idx].bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[idx].statss, start)); + } while (u64_stats_fetch_retry_irq(&priv->tx[idx].statss, start)); stats[stats_idx++] = (struct stats) { .stat_name = cpu_to_be32(TX_WAKE_CNT), .value = cpu_to_be64(priv->tx[idx].wake_queue), diff --git a/drivers/net/ethernet/huawei/hinic/hinic_rx.c b/drivers/net/ethernet/huawei/hinic/hinic_rx.c index 57d5d79..1b57b67 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_rx.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_rx.c @@ -375,7 +375,7 @@ void hinic_rxq_get_stats(struct hinic_rxq *rxq, u64_stats_update_begin(&stats->syncp); do { - start = u64_stats_fetch_begin(&rxq_stats->syncp); + start = u64_stats_fetch_begin_irq(&rxq_stats->syncp); stats->bytes = rxq_stats->bytes; stats->packets = rxq_stats->packets; stats->errors = rxq_stats->csum_errors + @@ -384,7 +384,7 @@ void hinic_rxq_get_stats(struct hinic_rxq *rxq, stats->other_errors = rxq_stats->other_errors; stats->dropped = rxq_stats->dropped; stats->rx_buf_empty = rxq_stats->rx_buf_empty; - } while (u64_stats_fetch_retry(&rxq_stats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&rxq_stats->syncp, start)); u64_stats_update_end(&stats->syncp); } diff --git a/drivers/net/ethernet/huawei/hinic/hinic_tx.c b/drivers/net/ethernet/huawei/hinic/hinic_tx.c index 75fa344..ff37b6f 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_tx.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_tx.c @@ -61,7 +61,7 @@ void hinic_txq_get_stats(struct hinic_txq *txq, u64_stats_update_begin(&stats->syncp); do { - start = u64_stats_fetch_begin(&txq_stats->syncp); + start = u64_stats_fetch_begin_irq(&txq_stats->syncp); stats->bytes = txq_stats->bytes; stats->packets = txq_stats->packets; stats->busy = txq_stats->busy; @@ -69,7 +69,7 @@ void hinic_txq_get_stats(struct hinic_txq *txq, stats->dropped = txq_stats->dropped; stats->big_frags_pkts = txq_stats->big_frags_pkts; stats->big_udp_pkts = txq_stats->big_udp_pkts; - } while (u64_stats_fetch_retry(&txq_stats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&txq_stats->syncp, start)); u64_stats_update_end(&stats->syncp); } diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c index dfc1f32..5ab230aa 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c @@ -3373,21 +3373,21 @@ static void nfp_net_stat64(struct net_device *netdev, unsigned int start; do { - start = u64_stats_fetch_begin(&r_vec->rx_sync); + start = u64_stats_fetch_begin_irq(&r_vec->rx_sync); data[0] = r_vec->rx_pkts; data[1] = r_vec->rx_bytes; data[2] = r_vec->rx_drops; - } while (u64_stats_fetch_retry(&r_vec->rx_sync, start)); + } while (u64_stats_fetch_retry_irq(&r_vec->rx_sync, start)); stats->rx_packets += data[0]; stats->rx_bytes += data[1]; stats->rx_dropped += data[2]; do { - start = u64_stats_fetch_begin(&r_vec->tx_sync); + start = u64_stats_fetch_begin_irq(&r_vec->tx_sync); data[0] = r_vec->tx_pkts; data[1] = r_vec->tx_bytes; data[2] = r_vec->tx_errors; - } while (u64_stats_fetch_retry(&r_vec->tx_sync, start)); + } while (u64_stats_fetch_retry_irq(&r_vec->tx_sync, start)); stats->tx_packets += data[0]; stats->tx_bytes += data[1]; stats->tx_errors += data[2]; diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c index bfcd90f..d4136d3 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c @@ -498,7 +498,7 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) unsigned int start; do { - start = u64_stats_fetch_begin(&nn->r_vecs[i].rx_sync); + start = u64_stats_fetch_begin_irq(&nn->r_vecs[i].rx_sync); data[0] = nn->r_vecs[i].rx_pkts; tmp[0] = nn->r_vecs[i].hw_csum_rx_ok; tmp[1] = nn->r_vecs[i].hw_csum_rx_inner_ok; @@ -506,10 +506,10 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) tmp[3] = nn->r_vecs[i].hw_csum_rx_error; tmp[4] = nn->r_vecs[i].rx_replace_buf_alloc_fail; tmp[5] = nn->r_vecs[i].hw_tls_rx; - } while (u64_stats_fetch_retry(&nn->r_vecs[i].rx_sync, start)); + } while (u64_stats_fetch_retry_irq(&nn->r_vecs[i].rx_sync, start)); do { - start = u64_stats_fetch_begin(&nn->r_vecs[i].tx_sync); + start = u64_stats_fetch_begin_irq(&nn->r_vecs[i].tx_sync); data[1] = nn->r_vecs[i].tx_pkts; data[2] = nn->r_vecs[i].tx_busy; tmp[6] = nn->r_vecs[i].hw_csum_tx; @@ -519,7 +519,7 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) tmp[10] = nn->r_vecs[i].hw_tls_tx; tmp[11] = nn->r_vecs[i].tls_tx_fallback; tmp[12] = nn->r_vecs[i].tls_tx_no_fallback; - } while (u64_stats_fetch_retry(&nn->r_vecs[i].tx_sync, start)); + } while (u64_stats_fetch_retry_irq(&nn->r_vecs[i].tx_sync, start)); data += NN_RVEC_PER_Q_STATS; diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c index ad6dbf01..4fb0638 100644 --- a/drivers/net/netdevsim/netdev.c +++ b/drivers/net/netdevsim/netdev.c @@ -67,10 +67,10 @@ static int nsim_change_mtu(struct net_device *dev, int new_mtu) unsigned int start; do { - start = u64_stats_fetch_begin(&ns->syncp); + start = u64_stats_fetch_begin_irq(&ns->syncp); stats->tx_bytes = ns->tx_bytes; stats->tx_packets = ns->tx_packets; - } while (u64_stats_fetch_retry(&ns->syncp, start)); + } while (u64_stats_fetch_retry_irq(&ns->syncp, start)); } static int diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index 461c037..cee39ae 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -2175,9 +2175,9 @@ static inline u64 sta_get_tidstats_msdu(struct ieee80211_sta_rx_stats *rxstats, u64 value; do { - start = u64_stats_fetch_begin(&rxstats->syncp); + start = u64_stats_fetch_begin_irq(&rxstats->syncp); value = rxstats->msdu[tid]; - } while (u64_stats_fetch_retry(&rxstats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&rxstats->syncp, start)); return value; } @@ -2241,9 +2241,9 @@ static inline u64 sta_get_stats_bytes(struct ieee80211_sta_rx_stats *rxstats) u64 value; do { - start = u64_stats_fetch_begin(&rxstats->syncp); + start = u64_stats_fetch_begin_irq(&rxstats->syncp); value = rxstats->bytes; - } while (u64_stats_fetch_retry(&rxstats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&rxstats->syncp, start)); return value; } diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 9c047c1..7239814 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -1078,9 +1078,9 @@ static void mpls_get_stats(struct mpls_dev *mdev, p = per_cpu_ptr(mdev->stats, i); do { - start = u64_stats_fetch_begin(&p->syncp); + start = u64_stats_fetch_begin_irq(&p->syncp); local = p->stats; - } while (u64_stats_fetch_retry(&p->syncp, start)); + } while (u64_stats_fetch_retry_irq(&p->syncp, start)); stats->rx_packets += local.rx_packets; stats->rx_bytes += local.rx_bytes; -- 1.8.3.1

1 2

[PATCH openEuler-1.0-LTS 1/3] mm: hwpoison: refactor refcount check handling
by Yongqiang Liu 04 Feb '23

04 Feb '23

From: Yang Shi <shy828301(a)gmail.com> stable inclusion from stable-v5.15.86 commit a62b1bc603a1ded739e7cf543da29a3eb93cc534 category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/I6AR36 CVE: NA Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit dd0f230a0a80ff396c7ce587f16429f2a8131344 upstream. Memory failure will report failure if the page still has extra pinned refcount other than from hwpoison after the handler is done. Actually the check is not necessary for all handlers, so move the check into specific handlers. This would make the following keeping shmem page in page cache patch easier. There may be expected extra pin for some cases, for example, when the page is dirty and in swapcache. Link: https://lkml.kernel.org/r/20211020210755.23964-5-shy828301@gmail.com Signed-off-by: Yang Shi <shy828301(a)gmail.com> Signed-off-by: Naoya Horiguchi <naoya.horiguchi(a)nec.com> Suggested-by: Naoya Horiguchi <naoya.horiguchi(a)nec.com> Cc: Hugh Dickins <hughd(a)google.com> Cc: Kirill A. Shutemov <kirill.shutemov(a)linux.intel.com> Cc: Matthew Wilcox <willy(a)infradead.org> Cc: Oscar Salvador <osalvador(a)suse.de> Cc: Peter Xu <peterx(a)redhat.com> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds(a)linux-foundation.org> Cc: Naoya Horiguchi <naoya.horiguchi(a)linux.dev> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Ze Zuo <zuoze1(a)huawei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- mm/memory-failure.c | 93 +++++++++++++++++++++++++++++++-------------- 1 file changed, 64 insertions(+), 29 deletions(-) diff --git a/mm/memory-failure.c b/mm/memory-failure.c index 9a816fdf812d..b653637d5a00 100644 --- a/mm/memory-failure.c +++ b/mm/memory-failure.c @@ -655,12 +655,44 @@ static int truncate_error_page(struct page *p, unsigned long pfn, return ret; } +struct page_state { + unsigned long mask; + unsigned long res; + enum mf_action_page_type type; + + /* Callback ->action() has to unlock the relevant page inside it. */ + int (*action)(struct page_state *ps, struct page *p); +}; + +/* + * Return true if page is still referenced by others, otherwise return + * false. + * + * The extra_pins is true when one extra refcount is expected. + */ +static bool has_extra_refcount(struct page_state *ps, struct page *p, + bool extra_pins) +{ + int count = page_count(p) - 1; + + if (extra_pins) + count -= 1; + + if (count > 0) { + pr_err("Memory failure: %#lx: %s still referenced by %d users\n", + page_to_pfn(p), action_page_types[ps->type], count); + return true; + } + + return false; +} + /* * Error hit kernel page. * Do nothing, try to be lucky and not touch this instead. For a few cases we * could be more sophisticated. */ -static int me_kernel(struct page *p, unsigned long pfn) +static int me_kernel(struct page_state *ps, struct page *p) { unlock_page(p); return MF_IGNORED; @@ -669,9 +701,9 @@ static int me_kernel(struct page *p, unsigned long pfn) /* * Page in unknown state. Do nothing. */ -static int me_unknown(struct page *p, unsigned long pfn) +static int me_unknown(struct page_state *ps, struct page *p) { - pr_err("Memory failure: %#lx: Unknown page state\n", pfn); + pr_err("Memory failure: %#lx: Unknown page state\n", page_to_pfn(p)); unlock_page(p); return MF_FAILED; } @@ -679,7 +711,7 @@ static int me_unknown(struct page *p, unsigned long pfn) /* * Clean (or cleaned) page cache page. */ -static int me_pagecache_clean(struct page *p, unsigned long pfn) +static int me_pagecache_clean(struct page_state *ps, struct page *p) { int ret; struct address_space *mapping; @@ -716,9 +748,13 @@ static int me_pagecache_clean(struct page *p, unsigned long pfn) * * Open: to take i_mutex or not for this? Right now we don't. */ - ret = truncate_error_page(p, pfn, mapping); + ret = truncate_error_page(p, page_to_pfn(p), mapping); out: unlock_page(p); + + if (has_extra_refcount(ps, p, false)) + ret = MF_FAILED; + return ret; } @@ -727,7 +763,7 @@ static int me_pagecache_clean(struct page *p, unsigned long pfn) * Issues: when the error hit a hole page the error is not properly * propagated. */ -static int me_pagecache_dirty(struct page *p, unsigned long pfn) +static int me_pagecache_dirty(struct page_state *ps, struct page *p) { struct address_space *mapping = page_mapping(p); @@ -771,7 +807,7 @@ static int me_pagecache_dirty(struct page *p, unsigned long pfn) mapping_set_error(mapping, -EIO); } - return me_pagecache_clean(p, pfn); + return me_pagecache_clean(ps, p); } /* @@ -793,9 +829,10 @@ static int me_pagecache_dirty(struct page *p, unsigned long pfn) * Clean swap cache pages can be directly isolated. A later page fault will * bring in the known good data from disk. */ -static int me_swapcache_dirty(struct page *p, unsigned long pfn) +static int me_swapcache_dirty(struct page_state *ps, struct page *p) { int ret; + bool extra_pins = false; ClearPageDirty(p); /* Trigger EIO in shmem: */ @@ -803,10 +840,17 @@ static int me_swapcache_dirty(struct page *p, unsigned long pfn) ret = delete_from_lru_cache(p) ? MF_FAILED : MF_DELAYED; unlock_page(p); + + if (ret == MF_DELAYED) + extra_pins = true; + + if (has_extra_refcount(ps, p, extra_pins)) + ret = MF_FAILED; + return ret; } -static int me_swapcache_clean(struct page *p, unsigned long pfn) +static int me_swapcache_clean(struct page_state *ps, struct page *p) { int ret; @@ -814,6 +858,10 @@ static int me_swapcache_clean(struct page *p, unsigned long pfn) ret = delete_from_lru_cache(p) ? MF_FAILED : MF_RECOVERED; unlock_page(p); + + if (has_extra_refcount(ps, p, false)) + ret = MF_FAILED; + return ret; } @@ -823,7 +871,7 @@ static int me_swapcache_clean(struct page *p, unsigned long pfn) * - Error on hugepage is contained in hugepage unit (not in raw page unit.) * To narrow down kill region to one page, we need to break up pmd. */ -static int me_huge_page(struct page *p, unsigned long pfn) +static int me_huge_page(struct page_state *ps, struct page *p) { int res; struct page *hpage = compound_head(p); @@ -834,7 +882,7 @@ static int me_huge_page(struct page *p, unsigned long pfn) mapping = page_mapping(hpage); if (mapping) { - res = truncate_error_page(hpage, pfn, mapping); + res = truncate_error_page(hpage, page_to_pfn(p), mapping); unlock_page(hpage); } else { res = MF_FAILED; @@ -852,6 +900,9 @@ static int me_huge_page(struct page *p, unsigned long pfn) } } + if (has_extra_refcount(ps, p, false)) + res = MF_FAILED; + return res; } @@ -878,14 +929,7 @@ static int me_huge_page(struct page *p, unsigned long pfn) #define slab (1UL << PG_slab) #define reserved (1UL << PG_reserved) -static struct page_state { - unsigned long mask; - unsigned long res; - enum mf_action_page_type type; - - /* Callback ->action() has to unlock the relevant page inside it. */ - int (*action)(struct page *p, unsigned long pfn); -} error_states[] = { +static struct page_state error_states[] = { { reserved, reserved, MF_MSG_KERNEL, me_kernel }, /* * free pages are specially detected outside this table: @@ -946,19 +990,10 @@ static int page_action(struct page_state *ps, struct page *p, unsigned long pfn) { int result; - int count; /* page p should be unlocked after returning from ps->action(). */ - result = ps->action(p, pfn); + result = ps->action(ps, p); - count = page_count(p) - 1; - if (ps->action == me_swapcache_dirty && result == MF_DELAYED) - count--; - if (count > 0) { - pr_err("Memory failure: %#lx: %s still referenced by %d users\n", - pfn, action_page_types[ps->type], count); - result = MF_FAILED; - } action_result(pfn, ps->type, result); /* Could do more checks here if page looks ok */ -- 2.25.1

1 2

[PATCH openEuler-1.0-LTS 1/6] dhugetlb: backport dynamic hugetlb feature
by Yongqiang Liu 04 Feb '23

04 Feb '23

From: Liu Shixin <liushixin2(a)hauwei.com> hulk inclusion category: feature bugzilla: 46904, https://gitee.com/openeuler/kernel/issues/I6BDME CVE: NA -------------------------------- This feature has already beed supported on x86_64 and this is the origin description: Dynamic hugetlb which is based on Hugetlb, supports to be splited dynamically in a specified cgroup. We add a hugetlb_pool in a mem_cgroup to manage dynamic hugetlb for corresponding cgroup. After dynamic hugepages are allocated for a cgroup, these hugepages can be used as 1G/2M/4K pages by split/merge opreation. It is now supported on arm64. This feature will be limited to depends on ARM64_4K_PAGES and not support cont-bits hugepage. We merge the previous patches into one patch which is patch[1]. While merge the code ,we found some code can be isolated by config DYNAMIC_HUGETLB, so we add patch[2] to re-isolated them. In patch[3], we restrict the feature on mentioned limit. The patch[4] add skip of dissolve hugepage which may conflict with memory hotplug and memory failure. The patch[5] set DYNAMIC_HUGETLB to y in hulk_defconfig to enable by default. This patch includes all previous patches and the patches list is recorded in bugzilla. Signed-off-by: Liu Shixin <liushixin2(a)hauwei.com> Reviewed-by: Kefeng Wang <wangkefeng.wang(a)huawei.com> Signed-off-by: Yongqiang Liu <liuyongqiang13(a)huawei.com> --- fs/Kconfig | 9 + fs/hugetlbfs/inode.c | 4 + include/linux/gfp.h | 4 +- include/linux/hugetlb.h | 97 +++ include/linux/memcontrol.h | 15 + include/linux/page-flags.h | 3 + include/trace/events/dhugetlb.h | 123 ++++ include/trace/events/mmflags.h | 1 + kernel/cgroup/cgroup.c | 6 + mm/huge_memory.c | 16 +- mm/hugetlb.c | 1188 ++++++++++++++++++++++++++++++- mm/internal.h | 1 + mm/memcontrol.c | 391 ++++++++++ mm/page_alloc.c | 33 +- 14 files changed, 1862 insertions(+), 29 deletions(-) create mode 100644 include/trace/events/dhugetlb.h diff --git a/fs/Kconfig b/fs/Kconfig index 5921bfbebee4..e8800d8a73b3 100644 --- a/fs/Kconfig +++ b/fs/Kconfig @@ -211,6 +211,15 @@ config TMPFS_INODE64 If unsure, say N. +config DYNAMIC_HUGETLB + bool "Dynamic HugeTLB" + depends on HUGETLB_PAGE + depends on MEMCG + depends on CGROUP_HUGETLB + help + Dynamic hugepage are used in memcg and can be splited into small pages + automatically. The tasks in the memcg prefer to alloc dynamic hugepage. + config HUGETLBFS bool "HugeTLB file system support" depends on X86 || IA64 || SPARC64 || (S390 && 64BIT) || \ diff --git a/fs/hugetlbfs/inode.c b/fs/hugetlbfs/inode.c index 005e05c442c5..30a29936372c 100644 --- a/fs/hugetlbfs/inode.c +++ b/fs/hugetlbfs/inode.c @@ -1164,6 +1164,8 @@ static struct inode *hugetlbfs_alloc_inode(struct super_block *sb) * private inode. This simplifies hugetlbfs_destroy_inode. */ mpol_shared_policy_init(&p->policy, NULL); + /* Initialize hpool here in case of a quick call to destroy */ + p->hpool = get_dhugetlb_pool_from_task(current); return &p->vfs_inode; } @@ -1178,6 +1180,8 @@ static void hugetlbfs_destroy_inode(struct inode *inode) { hugetlbfs_inc_free_inodes(HUGETLBFS_SB(inode->i_sb)); mpol_free_shared_policy(&HUGETLBFS_I(inode)->policy); + dhugetlb_pool_put(HUGETLBFS_I(inode)->hpool); + HUGETLBFS_I(inode)->hpool = NULL; call_rcu(&inode->i_rcu, hugetlbfs_i_callback); } diff --git a/include/linux/gfp.h b/include/linux/gfp.h index 152cb9bdf436..74b0375d7d2b 100644 --- a/include/linux/gfp.h +++ b/include/linux/gfp.h @@ -501,7 +501,9 @@ static inline void arch_alloc_page(struct page *page, int order) { } struct page * __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, int preferred_nid, nodemask_t *nodemask); - +void prep_new_page(struct page *page, unsigned int order, gfp_t gfp_flags, + unsigned int alloc_flags); +bool free_pages_prepare(struct page *page, unsigned int order, bool check_free); static inline struct page * __alloc_pages(gfp_t gfp_mask, unsigned int order, int preferred_nid) { diff --git a/include/linux/hugetlb.h b/include/linux/hugetlb.h index 2d2b06b36bd0..3a82ea9283ec 100644 --- a/include/linux/hugetlb.h +++ b/include/linux/hugetlb.h @@ -289,6 +289,7 @@ struct hugetlbfs_inode_info { struct shared_policy policy; struct inode vfs_inode; unsigned int seals; + struct dhugetlb_pool *hpool; }; static inline struct hugetlbfs_inode_info *HUGETLBFS_I(struct inode *inode) @@ -655,6 +656,102 @@ static inline void set_huge_swap_pte_at(struct mm_struct *mm, unsigned long addr #endif /* CONFIG_HUGETLB_PAGE */ +#ifdef CONFIG_DYNAMIC_HUGETLB +/* The number of small_page_pool for a dhugetlb_pool */ +#define NR_SMPOOL num_possible_cpus() +/* The max page number in a small_page_pool */ +#define MAX_SMPOOL_PAGE 1024 +/* number to move between list */ +#define BATCH_SMPOOL_PAGE (MAX_SMPOOL_PAGE >> 2) +/* We don't need to try 5 times, or we can't migrate the pages. */ +#define HPOOL_RECLAIM_RETRIES 5 + +extern struct static_key_false dhugetlb_enabled_key; +#define dhugetlb_enabled (static_branch_unlikely(&dhugetlb_enabled_key)) + +#define DEFAULT_PAGESIZE 4096 +extern rwlock_t dhugetlb_pagelist_rwlock; +struct dhugetlb_pagelist { + unsigned long count; + struct dhugetlb_pool *hpool[0]; +}; +extern struct dhugetlb_pagelist *dhugetlb_pagelist_t; + +struct split_pages { + struct list_head list; + unsigned long start_pfn; + unsigned long free_pages; +}; + +struct small_page_pool { + spinlock_t lock; + unsigned long free_pages; + long used_pages; + struct list_head head_page; +}; + +struct dhugetlb_pool { + int nid; + spinlock_t lock; + spinlock_t reserved_lock; + atomic_t refcnt; + + struct mem_cgroup *attach_memcg; + + struct list_head dhugetlb_1G_freelists; + struct list_head dhugetlb_2M_freelists; + struct list_head dhugetlb_4K_freelists; + + struct list_head split_1G_freelists; + struct list_head split_2M_freelists; + + unsigned long total_nr_pages; + + unsigned long total_reserved_1G; + unsigned long free_reserved_1G; + unsigned long mmap_reserved_1G; + unsigned long used_1G; + unsigned long free_unreserved_1G; + unsigned long nr_split_1G; + + unsigned long total_reserved_2M; + unsigned long free_reserved_2M; + unsigned long mmap_reserved_2M; + unsigned long used_2M; + unsigned long free_unreserved_2M; + unsigned long nr_split_2M; + + unsigned long free_pages; + struct small_page_pool smpool[0]; +}; + +bool dhugetlb_pool_get(struct dhugetlb_pool *hpool); +void dhugetlb_pool_put(struct dhugetlb_pool *hpool); +struct dhugetlb_pool *hpool_alloc(unsigned long nid); +int alloc_hugepage_from_hugetlb(struct dhugetlb_pool *hpool, + unsigned long nid, unsigned long size); +bool free_dhugetlb_pool(struct dhugetlb_pool *hpool); +int update_dhugetlb_pagelist(unsigned long idx, struct dhugetlb_pool *hpool); +struct dhugetlb_pool *get_dhugetlb_pool_from_dhugetlb_pagelist( + struct page *page); +struct dhugetlb_pool *get_dhugetlb_pool_from_task(struct task_struct *tsk); +bool move_pages_from_hpool_to_smpool(struct dhugetlb_pool *hpool, + struct small_page_pool *smpool); +void move_pages_from_smpool_to_hpool(struct dhugetlb_pool *hpool, + struct small_page_pool *smpool); +void dhugetlb_reserve_hugepages(struct dhugetlb_pool *hpool, + unsigned long count, bool gigantic); +#else +#define dhugetlb_enabled 0 +struct dhugetlb_pool {}; +static inline struct dhugetlb_pool *get_dhugetlb_pool_from_task( + struct task_struct *tsk) +{ + return NULL; +} +static inline void dhugetlb_pool_put(struct dhugetlb_pool *hpool) { return; } +#endif /* CONFIG_DYNAMIC_HUGETLB */ + static inline spinlock_t *huge_pte_lock(struct hstate *h, struct mm_struct *mm, pte_t *pte) { diff --git a/include/linux/memcontrol.h b/include/linux/memcontrol.h index 4517d132d1e2..22f40d5e0e8b 100644 --- a/include/linux/memcontrol.h +++ b/include/linux/memcontrol.h @@ -326,6 +326,7 @@ struct mem_cgroup { }; struct mem_cgroup_extension { + struct dhugetlb_pool *hpool; #ifdef CONFIG_MEMCG_QOS /* Currently support 0 and -1. * in the future it can expand to other value. @@ -1406,4 +1407,18 @@ static inline void memcg_put_cache_ids(void) #endif /* CONFIG_MEMCG_KMEM */ +#ifdef CONFIG_DYNAMIC_HUGETLB +struct dhugetlb_pool *get_dhugetlb_pool_from_memcg(struct mem_cgroup *memcg); +struct page *alloc_page_from_dhugetlb_pool(gfp_t gfp_mask); +void free_page_to_dhugetlb_pool(struct page *page); +int dhugetlb_pool_force_empty(struct mem_cgroup *memcg); +bool dhugetlb_pool_is_free(struct cgroup_subsys_state *css); +#else +static inline struct page *alloc_page_from_dhugetlb_pool(gfp_t gfp_mask) +{ + return NULL; +} +static inline void free_page_to_dhugetlb_pool(struct page *page) {} +#endif + #endif /* _LINUX_MEMCONTROL_H */ diff --git a/include/linux/page-flags.h b/include/linux/page-flags.h index 0c5d1c4c71e6..fd6cd68e00a2 100644 --- a/include/linux/page-flags.h +++ b/include/linux/page-flags.h @@ -102,6 +102,7 @@ enum pageflags { PG_idle, #endif PG_percpu_ref, + PG_pool, __NR_PAGEFLAGS, /* Filesystems */ @@ -284,6 +285,7 @@ PAGEFLAG(Active, active, PF_HEAD) __CLEARPAGEFLAG(Active, active, PF_HEAD) __PAGEFLAG(Slab, slab, PF_NO_TAIL) __PAGEFLAG(SlobFree, slob_free, PF_NO_TAIL) PAGEFLAG(Checked, checked, PF_NO_COMPOUND) /* Used by some filesystems */ +PAGEFLAG(Pool, pool, PF_NO_TAIL) /* Xen */ PAGEFLAG(Pinned, pinned, PF_NO_COMPOUND) @@ -770,6 +772,7 @@ static inline void ClearPageSlabPfmemalloc(struct page *page) 1UL << PG_private | 1UL << PG_private_2 | \ 1UL << PG_writeback | 1UL << PG_reserved | \ 1UL << PG_slab | 1UL << PG_active | \ + 1UL << PG_pool | \ 1UL << PG_unevictable | __PG_MLOCKED) /* diff --git a/include/trace/events/dhugetlb.h b/include/trace/events/dhugetlb.h new file mode 100644 index 000000000000..20b3a54589d1 --- /dev/null +++ b/include/trace/events/dhugetlb.h @@ -0,0 +1,123 @@ +#undef TRACE_SYSTEM +#define TRACE_SYSTEM dhugetlb + +#if !defined(_TRACE_DHUGETLB_H) || defined(TRACE_HEADER_MULTI_READ) +#define _TRACE_DHUGETLB_H + +#include <linux/tracepoint.h> +#include <trace/events/mmflags.h> + +#define DHUGETLB_SPLIT_1G 0x01u +#define DHUGETLB_SPLIT_2M 0x02u +#define DHUGETLB_MERGE_4K 0x04u +#define DHUGETLB_MIGRATE_4K 0x08u +#define DHUGETLB_RESV_1G 0x10u +#define DHUGETLB_UNRESV_1G 0x20u +#define DHUGETLB_RESV_2M 0x40u +#define DHUGETLB_UNRESV_2M 0x80u +#define DHUGETLB_ALLOC_1G 0x100u +#define DHUGETLB_FREE_1G 0x200u +#define DHUGETLB_ALLOC_2M 0x400u +#define DHUGETLB_FREE_2M 0x800u + +#define __def_action_names \ + {(unsigned long)DHUGETLB_SPLIT_1G, "split_1G_to_2M"}, \ + {(unsigned long)DHUGETLB_SPLIT_2M, "split_2M_to_4K"}, \ + {(unsigned long)DHUGETLB_MERGE_4K, "merge_4K_to_2M"}, \ + {(unsigned long)DHUGETLB_MIGRATE_4K, "migrate_4K_to_2M"}, \ + {(unsigned long)DHUGETLB_RESV_1G, "resv_1G_page"}, \ + {(unsigned long)DHUGETLB_UNRESV_1G, "unresv_1G_page"}, \ + {(unsigned long)DHUGETLB_RESV_2M, "resv_2M_page"}, \ + {(unsigned long)DHUGETLB_UNRESV_2M, "unresv_2M_page"}, \ + {(unsigned long)DHUGETLB_ALLOC_1G, "alloc_1G_page"}, \ + {(unsigned long)DHUGETLB_FREE_1G, "free_1G_page"}, \ + {(unsigned long)DHUGETLB_ALLOC_2M, "alloc_2M_page"}, \ + {(unsigned long)DHUGETLB_FREE_2M, "free_2M_page"} + +#define show_action(action) \ + (action) ? __print_flags(action, "", \ + __def_action_names \ + ) : "none" + +TRACE_EVENT(dhugetlb_split_merge, + + TP_PROTO(const void *hpool, struct page *page, unsigned long action), + + TP_ARGS(hpool, page, action), + + TP_STRUCT__entry( + __field( const void *, hpool ) + __field( unsigned long, pfn ) + __field( unsigned long, action ) + ), + + TP_fast_assign( + __entry->hpool = hpool; + __entry->pfn = page ? page_to_pfn(page) : -1UL; + __entry->action = action; + ), + + TP_printk("hpool=%p page=%p pfn=%lu action=%s", + __entry->hpool, + __entry->pfn != -1UL ? pfn_to_page(__entry->pfn) : NULL, + __entry->pfn != -1UL ? __entry->pfn : 0, + show_action(__entry->action)) +); + +TRACE_EVENT(dhugetlb_acct_memory, + + TP_PROTO(const void *hpool, unsigned long count, unsigned long action), + + TP_ARGS(hpool, count, action), + + TP_STRUCT__entry( + __field( const void *, hpool ) + __field( unsigned long, count ) + __field( unsigned long, action ) + ), + + TP_fast_assign( + __entry->hpool = hpool; + __entry->count = count; + __entry->action = action; + ), + + TP_printk("hpool=%p action=%s, mmap_count=%lu", + __entry->hpool, + show_action(__entry->action), + __entry->count) +); + +TRACE_EVENT(dhugetlb_alloc_free, + + TP_PROTO(const void *hpool, struct page *page, unsigned long count, + unsigned long action), + + TP_ARGS(hpool, page, count, action), + + TP_STRUCT__entry( + __field( const void *, hpool ) + __field( unsigned long, pfn ) + __field( unsigned long, count ) + __field( unsigned long, action ) + ), + + TP_fast_assign( + __entry->hpool = hpool; + __entry->pfn = page ? page_to_pfn(page) : -1UL; + __entry->count = count; + __entry->action = action; + ), + + TP_printk("hpool=%p page=%p pfn=%lu action=%s free_count=%lu", + __entry->hpool, + __entry->pfn != -1UL ? pfn_to_page(__entry->pfn) : NULL, + __entry->pfn != -1UL ? __entry->pfn : 0, + show_action(__entry->action), + __entry->count) +); + +#endif /* _TRACE_DHUGETLB_H */ + +/* This part must be outside protection */ +#include <trace/define_trace.h> diff --git a/include/trace/events/mmflags.h b/include/trace/events/mmflags.h index b817bf1885a0..4d06b47129f3 100644 --- a/include/trace/events/mmflags.h +++ b/include/trace/events/mmflags.h @@ -81,6 +81,7 @@ #define __def_pageflag_names \ {1UL << PG_locked, "locked" }, \ + {1UL << PG_pool, "pool" }, \ {1UL << PG_waiters, "waiters" }, \ {1UL << PG_error, "error" }, \ {1UL << PG_referenced, "referenced" }, \ diff --git a/kernel/cgroup/cgroup.c b/kernel/cgroup/cgroup.c index 7456882e1a0f..b01490b71f32 100644 --- a/kernel/cgroup/cgroup.c +++ b/kernel/cgroup/cgroup.c @@ -65,6 +65,7 @@ /* let's not notify more than 100 times per second */ #define CGROUP_FILE_NOTIFY_MIN_INTV DIV_ROUND_UP(HZ, 100) +bool dhugetlb_pool_is_free(struct cgroup_subsys_state *css); /* * cgroup_mutex is the master lock. Any modification to cgroup or its * hierarchy must be performed while holding it. @@ -5280,6 +5281,11 @@ static int cgroup_destroy_locked(struct cgroup *cgrp) if (css_has_online_children(&cgrp->self)) return -EBUSY; +#ifdef CONFIG_MEMCG + /* If we use dynamic hugetlb, make sure dhugtlb_pool is free */ + if (!dhugetlb_pool_is_free(cgrp->subsys[memory_cgrp_id])) + return -EBUSY; +#endif /* * Mark @cgrp and the associated csets dead. The former prevents * further task migration and child creation by disabling diff --git a/mm/huge_memory.c b/mm/huge_memory.c index f8319265c1cf..484ffdbf5f45 100644 --- a/mm/huge_memory.c +++ b/mm/huge_memory.c @@ -396,6 +396,20 @@ static int __init hugepage_init(void) return -EINVAL; } + /* + * When we alloc some pages(order = 0), system may help us to alloc + * a page(order > 0) due to transparent hugepage. This result + * dynamic hugetlb to be skipped. Actually, using dynamic hugetlb + * means we have already optimized the program, so we should not + * use transparent hugepage in addition. + * (May result negative optimization) + */ + if (dhugetlb_enabled) { + transparent_hugepage_flags = 0; + pr_info("transparent hugepage is disabled due to confilct with dynamic hugetlb\n"); + return -EINVAL; + } + /* * hugepages can't be allocated by the buddy allocator */ @@ -2946,9 +2960,9 @@ static unsigned long deferred_split_count(struct shrinker *shrink, { struct pglist_data *pgdata = NODE_DATA(sc->nid); unsigned long *split_queue_len = &pgdata->split_queue_len; +#ifdef CONFIG_MEMCG struct mem_cgroup_extension *memcg_ext; -#ifdef CONFIG_MEMCG if (sc->memcg) { memcg_ext = container_of(sc->memcg, struct mem_cgroup_extension, memcg); split_queue_len = &memcg_ext->split_queue_len; diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 495d8b5b38fc..4c8c91acd6d5 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -27,6 +27,12 @@ #include <linux/jhash.h> #include <linux/mman.h> #include <linux/share_pool.h> +#include <linux/kthread.h> +#include <linux/cpuhotplug.h> +#include <linux/freezer.h> +#include <linux/delay.h> +#include <linux/migrate.h> +#include <linux/mm_inline.h> #include <asm/page.h> #include <asm/pgtable.h> @@ -39,8 +45,14 @@ #include <linux/userfaultfd_k.h> #include <linux/page_owner.h> #include <linux/share_pool.h> +#include <linux/memblock.h> #include "internal.h" +#if (defined CONFIG_DYNAMIC_HUGETLB) && (!defined __GENKSYMS__) +#define CREATE_TRACE_POINTS +#include <trace/events/dhugetlb.h> +#endif + int hugetlb_max_hstate __read_mostly; unsigned int default_hstate_idx; struct hstate hstates[HUGE_MAX_HSTATE]; @@ -89,7 +101,8 @@ static inline void ClearPageHugeFreed(struct page *head) } /* Forward declaration */ -static int hugetlb_acct_memory(struct hstate *h, long delta); +static int hugetlb_acct_memory(struct hstate *h, long delta, + struct dhugetlb_pool *hpool); static inline void unlock_or_release_subpool(struct hugepage_subpool *spool) { @@ -103,7 +116,7 @@ static inline void unlock_or_release_subpool(struct hugepage_subpool *spool) if (free) { if (spool->min_hpages != -1) hugetlb_acct_memory(spool->hstate, - -spool->min_hpages); + -spool->min_hpages, NULL); kfree(spool); } } @@ -123,7 +136,7 @@ struct hugepage_subpool *hugepage_new_subpool(struct hstate *h, long max_hpages, spool->hstate = h; spool->min_hpages = min_hpages; - if (min_hpages != -1 && hugetlb_acct_memory(h, min_hpages)) { + if (min_hpages != -1 && hugetlb_acct_memory(h, min_hpages, NULL)) { kfree(spool); return NULL; } @@ -149,13 +162,17 @@ void hugepage_put_subpool(struct hugepage_subpool *spool) * a subpool minimum size must be manitained. */ static long hugepage_subpool_get_pages(struct hugepage_subpool *spool, - long delta) + long delta, struct dhugetlb_pool *hpool) { long ret = delta; if (!spool) return ret; + /* Skip subpool when hugetlb file belongs to a hugetlb_pool */ + if (dhugetlb_enabled && hpool) + return ret; + spin_lock(&spool->lock); if (spool->max_hpages != -1) { /* maximum size accounting */ @@ -194,13 +211,17 @@ static long hugepage_subpool_get_pages(struct hugepage_subpool *spool, * in the case where a subpool minimum size must be maintained. */ static long hugepage_subpool_put_pages(struct hugepage_subpool *spool, - long delta) + long delta, struct dhugetlb_pool *hpool) { long ret = delta; if (!spool) return delta; + /* Skip subpool when hugetlb file belongs to a hugetlb_pool */ + if (dhugetlb_enabled && hpool) + return ret; + spin_lock(&spool->lock); if (spool->max_hpages != -1) /* maximum size accounting */ @@ -594,12 +615,13 @@ void hugetlb_fix_reserve_counts(struct inode *inode) struct hugepage_subpool *spool = subpool_inode(inode); long rsv_adjust; bool reserved = false; + struct dhugetlb_pool *hpool = HUGETLBFS_I(inode)->hpool; - rsv_adjust = hugepage_subpool_get_pages(spool, 1); + rsv_adjust = hugepage_subpool_get_pages(spool, 1, hpool); if (rsv_adjust > 0) { struct hstate *h = hstate_inode(inode); - if (!hugetlb_acct_memory(h, 1)) + if (!hugetlb_acct_memory(h, 1, hpool)) reserved = true; } else if (!rsv_adjust) { reserved = true; @@ -1300,6 +1322,56 @@ static inline void ClearPageHugeTemporary(struct page *page) page[2].mapping = NULL; } +#ifdef CONFIG_DYNAMIC_HUGETLB +static void free_huge_page_to_dhugetlb_pool(struct page *page, + bool restore_reserve) +{ + struct hstate *h = page_hstate(page); + struct dhugetlb_pool *hpool; + + hpool = get_dhugetlb_pool_from_dhugetlb_pagelist(page); + if (unlikely(!hpool)) { + pr_err("dhugetlb: free error: get hpool failed\n"); + return; + } + + spin_lock(&hpool->lock); + ClearPagePool(page); + set_compound_page_dtor(page, NULL_COMPOUND_DTOR); + if (!hstate_is_gigantic(h)) { + list_add(&page->lru, &hpool->dhugetlb_2M_freelists); + hpool->free_reserved_2M++; + hpool->used_2M--; + if (restore_reserve) { + hpool->mmap_reserved_2M++; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_2M, + DHUGETLB_RESV_2M); + } + trace_dhugetlb_alloc_free(hpool, page, hpool->free_reserved_2M, + DHUGETLB_FREE_2M); + } else { + list_add(&page->lru, &hpool->dhugetlb_1G_freelists); + hpool->free_reserved_1G++; + hpool->used_1G--; + if (restore_reserve) { + hpool->mmap_reserved_1G++; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_1G, + DHUGETLB_RESV_1G); + } + trace_dhugetlb_alloc_free(hpool, page, hpool->free_reserved_1G, + DHUGETLB_FREE_1G); + } + spin_unlock(&hpool->lock); + dhugetlb_pool_put(hpool); +} +#else +void free_huge_page_to_dhugetlb_pool(struct page *page, bool restore_reserve) +{ +} +#endif + void free_huge_page(struct page *page) { /* @@ -1320,6 +1392,17 @@ void free_huge_page(struct page *page) restore_reserve = PagePrivate(page); ClearPagePrivate(page); + if (dhugetlb_enabled && PagePool(page)) { + spin_lock(&hugetlb_lock); + clear_page_huge_active(page); + list_del(&page->lru); + hugetlb_cgroup_uncharge_page(hstate_index(h), + pages_per_huge_page(h), page); + spin_unlock(&hugetlb_lock); + free_huge_page_to_dhugetlb_pool(page, restore_reserve); + return; + } + /* * If PagePrivate() was set on page, page allocation consumed a * reservation. If the page was associated with a subpool, there @@ -1335,7 +1418,7 @@ void free_huge_page(struct page *page) * after page is free. Therefore, force restore_reserve * operation. */ - if (hugepage_subpool_put_pages(spool, 1) == 0) + if (hugepage_subpool_put_pages(spool, 1, NULL) == 0) restore_reserve = true; } @@ -2211,6 +2294,81 @@ static void restore_reserve_on_error(struct hstate *h, } } +#ifdef CONFIG_DYNAMIC_HUGETLB +static struct page *__alloc_huge_page_from_dhugetlb_pool( + struct dhugetlb_pool *hpool, int idx, bool need_unreserved) +{ + unsigned long flags; + struct page *page = NULL; + + spin_lock_irqsave(&hpool->lock, flags); + if (hstate_is_gigantic(&hstates[idx]) && hpool->free_reserved_1G) { + page = list_entry(hpool->dhugetlb_1G_freelists.next, + struct page, lru); + list_del(&page->lru); + hpool->free_reserved_1G--; + hpool->used_1G++; + if (need_unreserved) { + SetPagePrivate(page); + hpool->mmap_reserved_1G--; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_1G, + DHUGETLB_UNRESV_1G); + } + trace_dhugetlb_alloc_free(hpool, page, hpool->free_reserved_1G, + DHUGETLB_ALLOC_1G); + } else if (!hstate_is_gigantic(&hstates[idx]) && + hpool->free_reserved_2M) { + page = list_entry(hpool->dhugetlb_2M_freelists.next, + struct page, lru); + list_del(&page->lru); + hpool->free_reserved_2M--; + hpool->used_2M++; + if (need_unreserved) { + SetPagePrivate(page); + hpool->mmap_reserved_2M--; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_2M, + DHUGETLB_UNRESV_2M); + } + trace_dhugetlb_alloc_free(hpool, page, hpool->free_reserved_2M, + DHUGETLB_ALLOC_2M); + } + if (page) { + INIT_LIST_HEAD(&page->lru); + set_compound_page_dtor(page, HUGETLB_PAGE_DTOR); + set_page_refcounted(page); + SetPagePool(page); + } + spin_unlock_irqrestore(&hpool->lock, flags); + + return page; +} + +static struct page *alloc_huge_page_from_dhugetlb_pool( + struct vm_area_struct *vma, int idx, int avoid_reserve, + long gbl_chg, struct dhugetlb_pool *hpool) +{ + struct page *page; + bool need_unreserved = false; + + if (!avoid_reserve && vma_has_reserves(vma, gbl_chg)) + need_unreserved = true; + + page = __alloc_huge_page_from_dhugetlb_pool(hpool, idx, + need_unreserved); + + return page; +} +#else +static inline struct page *alloc_huge_page_from_dhugetlb_pool( + struct vm_area_struct *vma, int idx, int avoid_reserve, + long gbl_chg, struct dhugetlb_pool *hpool) +{ + return NULL; +} +#endif + struct page *alloc_huge_page(struct vm_area_struct *vma, unsigned long addr, int avoid_reserve) { @@ -2221,6 +2379,8 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, long gbl_chg; int ret, idx; struct hugetlb_cgroup *h_cg; + struct dhugetlb_pool *hpool = + HUGETLBFS_I(file_inode(vma->vm_file))->hpool; idx = hstate_index(h); /* @@ -2240,7 +2400,7 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, * checked against any subpool limit. */ if (map_chg || avoid_reserve) { - gbl_chg = hugepage_subpool_get_pages(spool, 1); + gbl_chg = hugepage_subpool_get_pages(spool, 1, hpool); if (gbl_chg < 0) { vma_end_reservation(h, vma, addr); return ERR_PTR(-ENOSPC); @@ -2262,6 +2422,26 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, if (ret) goto out_subpool_put; + if (dhugetlb_enabled && hpool) { + page = alloc_huge_page_from_dhugetlb_pool(vma, idx, + avoid_reserve, + gbl_chg, hpool); + if (page) { + /* + * Use hugetlb_lock to manage the account of + * hugetlb cgroup. + */ + spin_lock(&hugetlb_lock); + list_add(&page->lru, &h->hugepage_activelist); + hugetlb_cgroup_commit_charge(idx, + pages_per_huge_page(hstate_vma(vma)), + h_cg, page); + spin_unlock(&hugetlb_lock); + goto out; + } + goto out_uncharge_cgroup; + } + spin_lock(&hugetlb_lock); /* * glb_chg is passed to indicate whether or not a page must be taken @@ -2284,7 +2464,7 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, } hugetlb_cgroup_commit_charge(idx, pages_per_huge_page(h), h_cg, page); spin_unlock(&hugetlb_lock); - +out: set_page_private(page, (unsigned long)spool); map_commit = vma_commit_reservation(h, vma, addr); @@ -2300,8 +2480,8 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, */ long rsv_adjust; - rsv_adjust = hugepage_subpool_put_pages(spool, 1); - hugetlb_acct_memory(h, -rsv_adjust); + rsv_adjust = hugepage_subpool_put_pages(spool, 1, hpool); + hugetlb_acct_memory(h, -rsv_adjust, hpool); } return page; @@ -2309,7 +2489,7 @@ struct page *alloc_huge_page(struct vm_area_struct *vma, hugetlb_cgroup_uncharge_cgroup(idx, pages_per_huge_page(h), h_cg); out_subpool_put: if (map_chg || avoid_reserve) - hugepage_subpool_put_pages(spool, 1); + hugepage_subpool_put_pages(spool, 1, hpool); vma_end_reservation(h, vma, addr); return ERR_PTR(-ENOSPC); } @@ -3098,6 +3278,932 @@ static void hugetlb_register_all_nodes(void) { } #endif +#ifdef CONFIG_DYNAMIC_HUGETLB +static bool enable_dhugetlb; +DEFINE_STATIC_KEY_FALSE(dhugetlb_enabled_key); +DEFINE_RWLOCK(dhugetlb_pagelist_rwlock); +struct dhugetlb_pagelist *dhugetlb_pagelist_t; + +bool dhugetlb_pool_get(struct dhugetlb_pool *hpool) +{ + if (!hpool) + return false; + + return atomic_inc_not_zero(&hpool->refcnt); +} + +void dhugetlb_pool_put(struct dhugetlb_pool *hpool) +{ + if (!dhugetlb_enabled || !hpool) + return; + + if (atomic_dec_and_test(&hpool->refcnt)) { + css_put(&hpool->attach_memcg->css); + kfree(hpool); + } +} + +struct dhugetlb_pool *hpool_alloc(unsigned long nid) +{ + int i; + struct dhugetlb_pool *hpool; + + hpool = kzalloc(sizeof(struct dhugetlb_pool) + + NR_SMPOOL * sizeof(struct small_page_pool), GFP_KERNEL); + if (!hpool) + return NULL; + + spin_lock_init(&hpool->lock); + spin_lock_init(&hpool->reserved_lock); + hpool->nid = nid; + atomic_set(&hpool->refcnt, 1); + INIT_LIST_HEAD(&hpool->dhugetlb_1G_freelists); + INIT_LIST_HEAD(&hpool->dhugetlb_2M_freelists); + INIT_LIST_HEAD(&hpool->dhugetlb_4K_freelists); + INIT_LIST_HEAD(&hpool->split_1G_freelists); + INIT_LIST_HEAD(&hpool->split_2M_freelists); + + for (i = 0; i < NR_SMPOOL; i++) { + spin_lock_init(&hpool->smpool[i].lock); + INIT_LIST_HEAD(&hpool->smpool[i].head_page); + } + + return hpool; +} + +int alloc_hugepage_from_hugetlb(struct dhugetlb_pool *hpool, + unsigned long nid, unsigned long size) +{ + int ret; + struct page *page, *next; + unsigned long idx; + unsigned long i = 0; + struct hstate *h = size_to_hstate(PUD_SIZE); + + if (!h) + return -ENOMEM; + + spin_lock(&hpool->lock); + spin_lock(&hugetlb_lock); + if (h->free_huge_pages_node[nid] < size) { + ret = -ENOMEM; + goto out_unlock; + } + + list_for_each_entry_safe(page, next, &h->hugepage_freelists[nid], lru) { + idx = page_to_pfn(page) >> (PUD_SHIFT - PAGE_SHIFT); + ret = update_dhugetlb_pagelist(idx, hpool); + if (ret) + continue; + ClearPageHugeFreed(page); + list_move_tail(&page->lru, &hpool->dhugetlb_1G_freelists); + h->free_huge_pages--; + h->free_huge_pages_node[nid]--; + hpool->total_nr_pages++; + hpool->free_unreserved_1G++; + if (++i == size) + break; + } + ret = 0; +out_unlock: + spin_unlock(&hugetlb_lock); + spin_unlock(&hpool->lock); + return ret; +} + +/* + * When we assign a hugepage to dhugetlb_pool, we need to record it in + * dhugetlb_pagelist_t. In this situation, we just need read_lock because + * there is not conflit when write to dhugetlb_pagelist_t->hpool. + * + * If page's pfn is greater than dhugetlb_pagelist_t->count (which may + * occurs due to memory hotplug), we need to realloc enough memory so that + * pfn = dhugetlb_pagelist_t->count - 1 and then record it. + * In this situation, we need write_lock because while we are reallocating, + * the read request should wait. + */ +int update_dhugetlb_pagelist(unsigned long idx, struct dhugetlb_pool *hpool) +{ + read_lock(&dhugetlb_pagelist_rwlock); + if (idx >= dhugetlb_pagelist_t->count) { + unsigned long size; + struct dhugetlb_pagelist *tmp; + + read_unlock(&dhugetlb_pagelist_rwlock); + write_lock(&dhugetlb_pagelist_rwlock); + + size = sizeof(struct dhugetlb_pagelist) + + (idx + 1) * sizeof(struct dhugetlb_pool *); + tmp = krealloc(dhugetlb_pagelist_t, size, GFP_ATOMIC); + if (!tmp) { + write_unlock(&dhugetlb_pagelist_rwlock); + return -ENOMEM; + } + tmp->count = idx + 1; + dhugetlb_pagelist_t = tmp; + + write_unlock(&dhugetlb_pagelist_rwlock); + read_lock(&dhugetlb_pagelist_rwlock); + } + dhugetlb_pagelist_t->hpool[idx] = hpool; + read_unlock(&dhugetlb_pagelist_rwlock); + return 0; +} + +struct dhugetlb_pool *get_dhugetlb_pool_from_dhugetlb_pagelist( + struct page *page) +{ + struct dhugetlb_pool *hpool = NULL; + unsigned long idx = page_to_pfn(page) >> (PUD_SHIFT - PAGE_SHIFT); + + read_lock(&dhugetlb_pagelist_rwlock); + if (idx < dhugetlb_pagelist_t->count) + hpool = dhugetlb_pagelist_t->hpool[idx]; + read_unlock(&dhugetlb_pagelist_rwlock); + if (dhugetlb_pool_get(hpool)) + return hpool; + return NULL; +} + +struct dhugetlb_pool *get_dhugetlb_pool_from_task(struct task_struct *tsk) +{ + struct mem_cgroup *memcg; + struct dhugetlb_pool *hpool; + + if (!dhugetlb_enabled) + return NULL; + + rcu_read_lock(); + memcg = mem_cgroup_from_task(tsk); + rcu_read_unlock(); + + hpool = get_dhugetlb_pool_from_memcg(memcg); + + return hpool; +} + +static void add_new_huge_page_to_pool(struct dhugetlb_pool *hpool, + struct page *page, bool gigantic) +{ + lockdep_assert_held(&hpool->lock); + VM_BUG_ON_PAGE(page_mapcount(page), page); + INIT_LIST_HEAD(&page->lru); + + if (gigantic) { + prep_compound_gigantic_page(page, PUD_SHIFT - PAGE_SHIFT); + list_add_tail(&page->lru, &hpool->dhugetlb_1G_freelists); + hpool->free_unreserved_1G++; + } else { + prep_new_page(page, PMD_SHIFT - PAGE_SHIFT, __GFP_COMP, 0); + set_page_count(page, 0); + list_add_tail(&page->lru, &hpool->dhugetlb_2M_freelists); + hpool->free_unreserved_2M++; + } + set_page_private(page, 0); + page->mapping = NULL; + set_compound_page_dtor(page, HUGETLB_PAGE_DTOR); + set_hugetlb_cgroup(page, NULL); +} + +static void free_dhugetlb_pcpool(struct dhugetlb_pool *hpool) +{ + int i; + struct small_page_pool *smpool; + + for (i = 0; i < NR_SMPOOL; i++) { + smpool = &hpool->smpool[i]; + list_splice(&smpool->head_page, &hpool->dhugetlb_4K_freelists); + smpool->free_pages = 0; + smpool->used_pages = 0; + INIT_LIST_HEAD(&smpool->head_page); + } +} + +static void __free_dhugetlb_small_page(struct dhugetlb_pool *hpool) +{ + struct page *page, *next; + struct split_pages *split_huge, *split_next; + + if (list_empty(&hpool->dhugetlb_4K_freelists)) + return; + + list_for_each_entry_safe(page, next, + &hpool->dhugetlb_4K_freelists, lru) { + list_del(&page->lru); + add_new_huge_page_to_pool(hpool, page, false); + } + + list_for_each_entry_safe(split_huge, split_next, + &hpool->split_2M_freelists, list) { + list_del(&split_huge->list); + kfree(split_huge); + hpool->nr_split_2M--; + } + + hpool->free_pages = 0; + INIT_LIST_HEAD(&hpool->dhugetlb_4K_freelists); +} + +static void free_dhugetlb_small_page(struct dhugetlb_pool *hpool) +{ + struct page *page, *next; + unsigned long nr_pages = 1 << (PMD_SHIFT - PAGE_SHIFT); + + lockdep_assert_held(&hpool->lock); + if (list_empty(&hpool->dhugetlb_4K_freelists)) + return; + + list_for_each_entry_safe(page, next, + &hpool->dhugetlb_4K_freelists, lru) { + if (page_to_pfn(page) % nr_pages != 0) + list_del(&page->lru); + } + + __free_dhugetlb_small_page(hpool); +} + +static void __free_dhugetlb_huge_page(struct dhugetlb_pool *hpool) +{ + struct page *page, *next; + struct split_pages *split_giga, *split_next; + + if (list_empty(&hpool->dhugetlb_2M_freelists)) + return; + + list_for_each_entry_safe(page, next, + &hpool->dhugetlb_2M_freelists, lru) { + list_del(&page->lru); + add_new_huge_page_to_pool(hpool, page, true); + } + list_for_each_entry_safe(split_giga, split_next, + &hpool->split_1G_freelists, list) { + list_del(&split_giga->list); + kfree(split_giga); + hpool->nr_split_1G--; + } + + hpool->total_reserved_2M = 0; + hpool->free_reserved_2M = 0; + hpool->free_unreserved_2M = 0; + INIT_LIST_HEAD(&hpool->dhugetlb_2M_freelists); +} + +static void free_dhugetlb_huge_page(struct dhugetlb_pool *hpool) +{ + struct page *page, *next; + unsigned long nr_pages = 1 << (PUD_SHIFT - PAGE_SHIFT); + unsigned long block_size = 1 << (PMD_SHIFT - PAGE_SHIFT); + int i; + + lockdep_assert_held(&hpool->lock); + if (list_empty(&hpool->dhugetlb_2M_freelists)) + return; + + list_for_each_entry_safe(page, next, + &hpool->dhugetlb_2M_freelists, lru) { + set_compound_page_dtor(page, NULL_COMPOUND_DTOR); + atomic_set(compound_mapcount_ptr(page), 0); + for (i = 1; i < block_size; i++) + clear_compound_head(&page[i]); + set_compound_order(page, 0); + __ClearPageHead(page); + if (page_to_pfn(page) % nr_pages != 0) + list_del(&page->lru); + } + __free_dhugetlb_huge_page(hpool); +} + +static int try_migrate_page(struct page *page, unsigned long nid) +{ + unsigned long pfn = page_to_pfn(page); + int ret = 0; + + LIST_HEAD(source); + + if (!pfn_valid(pfn)) + return 0; + BUG_ON(PageHuge(page) || PageTransHuge(page)); + /* + * HWPoison pages have elevated reference counts so the migration + * would fail on them. It also doesn't make any sense to migrate them + * in the first place. Still try to unmap such a page in case it is + * still mapped(e.g. current hwpoison implementation doesn't unmap + * KSM pages but keep the unmap as the catch all safety net). + */ + if (PageHWPoison(page)) { + if (WARN_ON(PageLRU(page))) + isolate_lru_page(page); + if (page_mapped(page)) + try_to_unmap(page, + TTU_IGNORE_MLOCK | TTU_IGNORE_ACCESS); + return 0; + } + + if (!get_page_unless_zero(page)) + return 0; + /* + * We can skip free pages. And we can deal with pages on + * LRU and non-lru movable pages. + */ + if (PageLRU(page)) + ret = isolate_lru_page(page); + else + ret = isolate_movable_page(page, ISOLATE_UNEVICTABLE); + put_page(page); + if (ret) { + if (page_count(page)) + ret = -EBUSY; + return ret; + } + list_add_tail(&page->lru, &source); + if (!__PageMovable(page)) + inc_node_page_state(page, + NR_ISOLATED_ANON + page_is_file_cache(page)); + + ret = migrate_pages(&source, alloc_new_node_page, NULL, nid, + MIGRATE_SYNC_LIGHT, MR_COMPACTION); + if (ret) + putback_movable_pages(&source); + return ret; +} + +static void try_migrate_pages(struct dhugetlb_pool *hpool) +{ + int i, j; + unsigned long nr_free_pages; + struct split_pages *split_giga, *next; + unsigned int nr_pages = 1 << (PMD_SHIFT - PAGE_SHIFT); + struct page *page; + int sleep_interval = 100; /* wait for the migration */ + + spin_unlock(&hpool->lock); + for (i = NR_SMPOOL - 1; i >= 0; i--) + spin_unlock(&hpool->smpool[i].lock); + + msleep(sleep_interval); + dhugetlb_pool_force_empty(hpool->attach_memcg); + + spin_lock(&hpool->lock); + nr_free_pages = hpool->free_pages; + spin_unlock(&hpool->lock); + for (i = 0; i < NR_SMPOOL; i++) { + spin_lock(&hpool->smpool[i].lock); + nr_free_pages += hpool->smpool[i].free_pages; + spin_unlock(&hpool->smpool[i].lock); + } + + if (nr_free_pages >> HUGETLB_PAGE_ORDER < hpool->nr_split_2M) { + list_for_each_entry_safe(split_giga, next, + &hpool->split_1G_freelists, list) { + for (i = 0; i < nr_pages; i++) { + if (PageCompound(pfn_to_page( + split_giga->start_pfn + i * nr_pages))) + continue; + page = pfn_to_page(split_giga->start_pfn + + i * nr_pages); + for (j = 0; j < nr_pages; j++) { + if (PagePool(page + j)) + try_migrate_page(page + j, + hpool->nid); + } + } + } + } + + for (i = 0; i < NR_SMPOOL; i++) + spin_lock(&hpool->smpool[i].lock); + spin_lock(&hpool->lock); +} + +/* + * If there are some pages are still in use. We will try to reclaim/migrate it. + * After trying at most HPOOL_RECLAIM_RETRIES times, we may success. + * Or we will print the failed information and return false. + */ +static bool free_dhugetlb_pages(struct dhugetlb_pool *hpool) +{ + int i; + long used_pages; + int try_count = 0; + +retry: + used_pages = 0; + for (i = 0; i < NR_SMPOOL; i++) + used_pages += hpool->smpool[i].used_pages; + + if (try_count < HPOOL_RECLAIM_RETRIES && + (used_pages || hpool->used_2M || hpool->used_1G)) { + try_migrate_pages(hpool); + try_count++; + goto retry; + } + + if (used_pages) + pr_err("dhugetlb: some 4K pages not free, memcg: %s delete failed!\n", + hpool->attach_memcg->css.cgroup->kn->name); + else if (hpool->used_2M) + pr_err("dhugetlb: some 2M pages not free, memcg: %s delete failed!\n", + hpool->attach_memcg->css.cgroup->kn->name); + else if (hpool->used_1G) + pr_err("dhugetlb: some 1G pages not free, memcg: %s delete failed!\n", + hpool->attach_memcg->css.cgroup->kn->name); + else { + free_dhugetlb_pcpool(hpool); + free_dhugetlb_small_page(hpool); + free_dhugetlb_huge_page(hpool); + return true; + } + return false; +} + +static void free_back_hugetlb(struct dhugetlb_pool *hpool) +{ + int nid; + unsigned int nr_pages; + unsigned long pfn, idx; + struct page *page, *page_next, *p; + struct hstate *h = size_to_hstate(PUD_SIZE); + + if (!h) + return; + + spin_lock(&hugetlb_lock); + list_for_each_entry_safe(page, page_next, + &hpool->dhugetlb_1G_freelists, lru) { + nr_pages = 1 << huge_page_order(h); + pfn = page_to_pfn(page); + for (; nr_pages--; pfn++) { + p = pfn_to_page(pfn); + p->mapping = NULL; + } + SetPageHugeFreed(page); + set_compound_page_dtor(page, HUGETLB_PAGE_DTOR); + nid = page_to_nid(page); + BUG_ON(nid >= MAX_NUMNODES); + list_move(&page->lru, &h->hugepage_freelists[nid]); + h->free_huge_pages_node[nid]++; + read_lock(&dhugetlb_pagelist_rwlock); + idx = page_to_pfn(page) >> (PUD_SHIFT - PAGE_SHIFT); + if (idx < dhugetlb_pagelist_t->count) + dhugetlb_pagelist_t->hpool[idx] = NULL; + read_unlock(&dhugetlb_pagelist_rwlock); + } + h->free_huge_pages += hpool->total_nr_pages; + hpool->total_nr_pages = 0; + hpool->free_unreserved_1G = 0; + hpool->free_reserved_1G = 0; + hpool->total_reserved_1G = 0; + INIT_LIST_HEAD(&hpool->dhugetlb_1G_freelists); + spin_unlock(&hugetlb_lock); +} + +bool free_dhugetlb_pool(struct dhugetlb_pool *hpool) +{ + int i; + bool ret = false; + + for (i = 0; i < NR_SMPOOL; i++) + spin_lock(&hpool->smpool[i].lock); + spin_lock(&hpool->lock); + + ret = free_dhugetlb_pages(hpool); + if (!ret) + goto out_unlock; + + free_back_hugetlb(hpool); + +out_unlock: + spin_unlock(&hpool->lock); + for (i = NR_SMPOOL - 1; i >= 0; i--) + spin_unlock(&hpool->smpool[i].lock); + + if (ret) + dhugetlb_pool_put(hpool); + return ret; +} + +static void __split_free_huge_page(struct dhugetlb_pool *hpool, + struct page *page) +{ + int i; + int order_h = PUD_SHIFT - PAGE_SHIFT; + int order_m = PMD_SHIFT - PAGE_SHIFT; + int blocks = 1 << (order_h - order_m); + struct page *p = page + 1; + + lockdep_assert_held(&hpool->lock); + set_compound_page_dtor(page, NULL_COMPOUND_DTOR); + atomic_set(compound_mapcount_ptr(page), 0); + for (i = 1; i < (1 << order_h); i++, p = mem_map_next(p, page, i)) + clear_compound_head(p); + + set_compound_order(page, 0); + __ClearPageHead(page); + + /* make it be 2M huge pages and put it to huge pool */ + for (i = 0; i < blocks; i++, page += (1 << order_m)) + add_new_huge_page_to_pool(hpool, page, false); +} + +static void __split_free_small_page(struct dhugetlb_pool *hpool, + struct page *page) +{ + int i; + int nr_pages = 1 << (PMD_SHIFT - PAGE_SHIFT); + + lockdep_assert_held(&hpool->lock); + set_compound_page_dtor(page, NULL_COMPOUND_DTOR); + set_compound_order(page, 0); + for (i = 0; i < nr_pages; i++) { + if (i != 0) { + page[i].mapping = NULL; + clear_compound_head(&page[i]); + } else + __ClearPageHead(page); + + /* + * If a hugepage is mapped in private mode, the PG_uptodate bit + * will not be cleared when the hugepage freed. Clear the + * hugepage using free_pages_prepare() here. + */ + free_pages_prepare(&page[i], 0, false); + hpool->free_pages++; + list_add_tail(&page[i].lru, &hpool->dhugetlb_4K_freelists); + } +} + +static bool split_free_huge_page(struct dhugetlb_pool *hpool) +{ + struct page *page; + struct split_pages *split_page; + + lockdep_assert_held(&hpool->lock); + + if (!hpool->free_unreserved_1G) + return false; + + split_page = kzalloc(sizeof(struct split_pages), GFP_ATOMIC); + if (!split_page) + return false; + + page = list_entry(hpool->dhugetlb_1G_freelists.next, struct page, lru); + list_del(&page->lru); + hpool->free_unreserved_1G--; + + split_page->start_pfn = page_to_pfn(page); + list_add(&split_page->list, &hpool->split_1G_freelists); + hpool->nr_split_1G++; + + trace_dhugetlb_split_merge(hpool, page, DHUGETLB_SPLIT_1G); + + __split_free_huge_page(hpool, page); + return true; +} + +static bool split_free_small_page(struct dhugetlb_pool *hpool) +{ + struct page *page; + struct split_pages *split_page; + + lockdep_assert_held(&hpool->lock); + + if (!hpool->free_unreserved_2M && !split_free_huge_page(hpool)) + return false; + + split_page = kzalloc(sizeof(struct split_pages), GFP_ATOMIC); + if (!split_page) + return false; + + page = list_entry(hpool->dhugetlb_2M_freelists.next, struct page, lru); + list_del(&page->lru); + hpool->free_unreserved_2M--; + + split_page->start_pfn = page_to_pfn(page); + list_add(&split_page->list, &hpool->split_2M_freelists); + hpool->nr_split_2M++; + + trace_dhugetlb_split_merge(hpool, page, DHUGETLB_SPLIT_2M); + + __split_free_small_page(hpool, page); + return true; +} + +bool move_pages_from_hpool_to_smpool(struct dhugetlb_pool *hpool, + struct small_page_pool *smpool) +{ + int i = 0; + struct page *page, *next; + + if (!hpool->free_pages && !split_free_small_page(hpool)) + return false; + + list_for_each_entry_safe(page, next, + &hpool->dhugetlb_4K_freelists, lru) { + list_del(&page->lru); + hpool->free_pages--; + list_add_tail(&page->lru, &smpool->head_page); + smpool->free_pages++; + if (++i == BATCH_SMPOOL_PAGE) + break; + } + return true; +} + +void move_pages_from_smpool_to_hpool(struct dhugetlb_pool *hpool, + struct small_page_pool *smpool) +{ + int i = 0; + struct page *page, *next; + + list_for_each_entry_safe(page, next, &smpool->head_page, lru) { + list_del(&page->lru); + smpool->free_pages--; + list_add(&page->lru, &hpool->dhugetlb_4K_freelists); + hpool->free_pages++; + if (++i == BATCH_SMPOOL_PAGE) + break; + } +} + +static unsigned long list_len(struct list_head *head) +{ + unsigned long len = 0; + struct page *page; + + list_for_each_entry(page, head, lru) + len++; + + return len; +} + +static void hugetlb_migrate_pages(struct dhugetlb_pool *hpool, + unsigned long count) +{ + int i, try; + struct page *page; + struct split_pages *split_huge, *split_next; + unsigned long nr_pages = 1 << (PMD_SHIFT - PAGE_SHIFT); + LIST_HEAD(wait_page_list); + + list_for_each_entry_safe(split_huge, split_next, + &hpool->split_2M_freelists, list) { + /* + * Isolate free page first because we dont want them to be + * allocated. + */ + for (i = 0; i < nr_pages; i++) { + page = pfn_to_page(split_huge->start_pfn + i); + if (!PagePool(page)) + list_move(&page->lru, &wait_page_list); + } + + for (try = 0; try < HPOOL_RECLAIM_RETRIES; try++) { + /* + * Unlock and try migration, after migration we need + * to lock back. + */ + for (i = 0; i < NR_SMPOOL; i++) + hpool->smpool[i].free_pages = + list_len(&hpool->smpool[i].head_page); + hpool->free_pages = + list_len(&hpool->dhugetlb_4K_freelists); + spin_unlock(&hpool->lock); + for (i = NR_SMPOOL - 1; i >= 0; i--) + spin_unlock(&hpool->smpool[i].lock); + + for (i = 0; i < nr_pages; i++) { + page = pfn_to_page(split_huge->start_pfn + i); + if (PagePool(page)) + try_migrate_page(page, hpool->nid); + } + for (i = 0; i < NR_SMPOOL; i++) + spin_lock(&hpool->smpool[i].lock); + spin_lock(&hpool->lock); + + /* + * Isolate free page. If all page in the split_huge + * is free, return it. + */ + split_huge->free_pages = 0; + for (i = 0; i < nr_pages; i++) { + page = pfn_to_page(split_huge->start_pfn + i); + if (!PagePool(page)) { + list_move(&page->lru, &wait_page_list); + split_huge->free_pages++; + } + } + if (split_huge->free_pages == nr_pages) + break; + } + if (split_huge->free_pages == nr_pages) { + for (i = 0; i < nr_pages; i++) { + page = pfn_to_page(split_huge->start_pfn + i); + list_del(&page->lru); + } + INIT_LIST_HEAD(&wait_page_list); + page = pfn_to_page(split_huge->start_pfn); + add_new_huge_page_to_pool(hpool, page, false); + list_del(&split_huge->list); + kfree(split_huge); + hpool->nr_split_2M--; + + trace_dhugetlb_split_merge(hpool, page, + DHUGETLB_MIGRATE_4K); + + if (--count == 0) + return; + } else { + /* Failed, put back the isolate pages */ + list_splice(&wait_page_list, + &hpool->dhugetlb_4K_freelists); + INIT_LIST_HEAD(&wait_page_list); + } + } +} + +static unsigned long merge_free_split_huge(struct dhugetlb_pool *hpool, + unsigned long count) +{ + int i; + struct page *page; + struct split_pages *split_huge, *split_next; + unsigned long nr_pages = 1 << (PMD_SHIFT - PAGE_SHIFT); + + list_for_each_entry_safe(split_huge, split_next, + &hpool->split_2M_freelists, list) { + split_huge->free_pages = 0; + for (i = 0; i < nr_pages; i++) { + page = pfn_to_page(split_huge->start_pfn + i); + if (!PagePool(page)) + split_huge->free_pages++; + } + if (split_huge->free_pages == nr_pages) { + for (i = 0; i < nr_pages; i++) { + page = pfn_to_page(split_huge->start_pfn + i); + list_del(&page->lru); + } + page = pfn_to_page(split_huge->start_pfn); + add_new_huge_page_to_pool(hpool, page, false); + list_del(&split_huge->list); + kfree(split_huge); + hpool->nr_split_2M--; + + trace_dhugetlb_split_merge(hpool, page, + DHUGETLB_MERGE_4K); + + if (--count == 0) + return 0; + } + } + return count; +} + +static void merge_free_small_page(struct dhugetlb_pool *hpool, + unsigned long count) +{ + int i; + unsigned long need_migrate; + + if (!hpool->nr_split_2M) + return; + + need_migrate = merge_free_split_huge(hpool, count); + if (need_migrate) + hugetlb_migrate_pages(hpool, need_migrate); + + for (i = 0; i < NR_SMPOOL; i++) + hpool->smpool[i].free_pages = + list_len(&hpool->smpool[i].head_page); + hpool->free_pages = list_len(&hpool->dhugetlb_4K_freelists); +} + +static void dhugetlb_collect_2M_pages(struct dhugetlb_pool *hpool, + unsigned long count) +{ + int i; + + while (hpool->free_unreserved_1G && + count > hpool->free_unreserved_2M) + split_free_huge_page(hpool); + + /* + * If we try to merge 4K pages to 2M, we need to unlock hpool->lock + * first, and then try to lock every lock in order to avoid deadlock. + */ + if (count > hpool->free_unreserved_2M) { + spin_unlock(&hpool->lock); + for (i = 0; i < NR_SMPOOL; i++) + spin_lock(&hpool->smpool[i].lock); + spin_lock(&hpool->lock); + merge_free_small_page(hpool, count - hpool->free_unreserved_2M); + for (i = NR_SMPOOL - 1; i >= 0; i--) + spin_unlock(&hpool->smpool[i].lock); + } +} + +/* + * Parameter gigantic: true means reserve 1G pages and false means reserve + * 2M pages. When we want to reserve 2M pages more than + * hpool->free_unreserved_2M, we have to try split/merge. Still, we can't + * guarantee success. + */ +void dhugetlb_reserve_hugepages(struct dhugetlb_pool *hpool, + unsigned long count, bool gigantic) +{ + unsigned long delta; + + spin_lock(&hpool->lock); + if (gigantic) { + if (count > hpool->total_reserved_1G) { + delta = min(count - hpool->total_reserved_1G, + hpool->free_unreserved_1G); + hpool->total_reserved_1G += delta; + hpool->free_reserved_1G += delta; + hpool->free_unreserved_1G -= delta; + } else { + delta = min(hpool->total_reserved_1G - count, + hpool->free_reserved_1G - + hpool->mmap_reserved_1G); + hpool->total_reserved_1G -= delta; + hpool->free_reserved_1G -= delta; + hpool->free_unreserved_1G += delta; + } + } else { + if (count > hpool->total_reserved_2M) { + delta = count - hpool->total_reserved_2M; + if (delta > hpool->free_unreserved_2M) + dhugetlb_collect_2M_pages(hpool, delta); + delta = min(count - hpool->total_reserved_2M, + hpool->free_unreserved_2M); + hpool->total_reserved_2M += delta; + hpool->free_reserved_2M += delta; + hpool->free_unreserved_2M -= delta; + } else { + delta = min(hpool->total_reserved_2M - count, + hpool->free_reserved_2M - + hpool->mmap_reserved_2M); + hpool->total_reserved_2M -= delta; + hpool->free_reserved_2M -= delta; + hpool->free_unreserved_2M += delta; + } + } + spin_unlock(&hpool->lock); +} + +static int dhugetlb_acct_memory(struct hstate *h, long delta, + struct dhugetlb_pool *hpool) +{ + int ret = -ENOMEM; + + if (delta == 0) + return 0; + + spin_lock(&hpool->lock); + if (hstate_is_gigantic(h)) { + if (delta > 0 && delta <= hpool->free_reserved_1G - + hpool->mmap_reserved_1G) { + hpool->mmap_reserved_1G += delta; + ret = 0; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_1G, + DHUGETLB_RESV_1G); + } else if (delta < 0) { + hpool->mmap_reserved_1G -= (unsigned long)(-delta); + WARN_ON(hpool->mmap_reserved_1G < 0); + ret = 0; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_1G, + DHUGETLB_UNRESV_1G); + } + } else { + if (delta > 0 && delta <= hpool->free_reserved_2M - + hpool->mmap_reserved_2M) { + hpool->mmap_reserved_2M += delta; + ret = 0; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_2M, + DHUGETLB_RESV_2M); + } else if (delta < 0) { + hpool->mmap_reserved_2M -= (unsigned long)(-delta); + WARN_ON(hpool->mmap_reserved_2M < 0); + ret = 0; + trace_dhugetlb_acct_memory(hpool, + hpool->mmap_reserved_2M, + DHUGETLB_UNRESV_2M); + } + } + spin_unlock(&hpool->lock); + + return ret; +} +#else +static int dhugetlb_acct_memory(struct hstate *h, long delta, + struct dhugetlb_pool *hpool) +{ + return 0; +} +#endif /* CONFIG_DYNAMIC_HUGETLB */ + static int __init hugetlb_init(void) { int i; @@ -3134,6 +4240,23 @@ static int __init hugetlb_init(void) hugetlb_register_all_nodes(); hugetlb_cgroup_file_init(); +#ifdef CONFIG_DYNAMIC_HUGETLB + if (enable_dhugetlb) { + unsigned long count = max(max_pfn >> (PUD_SHIFT - PAGE_SHIFT), + (unsigned long)DEFAULT_PAGESIZE); + unsigned long size = sizeof(struct dhugetlb_pagelist) + + count * sizeof(struct dhugetlb_pool *); + dhugetlb_pagelist_t = kzalloc(size, GFP_KERNEL); + if (dhugetlb_pagelist_t) { + dhugetlb_pagelist_t->count = count; + static_branch_enable(&dhugetlb_enabled_key); + pr_info("Dynamic 1G hugepage enabled\n"); + } else + pr_info("Dynamic 1G hugepage disabled due to out of memory, need %lu\n", + size); + } +#endif + #ifdef CONFIG_SMP num_fault_mutexes = roundup_pow_of_two(8 * num_possible_cpus()); #else @@ -3270,6 +4393,16 @@ static int __init hugetlb_nrpages_setup(char *s) } __setup("hugepages=", hugetlb_nrpages_setup); +#ifdef CONFIG_DYNAMIC_HUGETLB +static int __init dhugetlb_setup(char *s) +{ + if (!strcmp(s, "on")) + enable_dhugetlb = true; + return 1; +} +__setup("dynamic_1G_hugepage=", dhugetlb_setup); +#endif + static int __init hugetlb_default_setup(char *s) { default_hstate_size = memparse(s, &s); @@ -3471,10 +4604,14 @@ unsigned long hugetlb_total_pages(void) return nr_total_pages; } -static int hugetlb_acct_memory(struct hstate *h, long delta) +static int hugetlb_acct_memory(struct hstate *h, long delta, + struct dhugetlb_pool *hpool) { int ret = -ENOMEM; + if (dhugetlb_enabled && hpool) + return dhugetlb_acct_memory(h, delta, hpool); + spin_lock(&hugetlb_lock); /* * When cpuset is configured, it breaks the strict hugetlb page @@ -3535,6 +4672,8 @@ static void hugetlb_vm_op_close(struct vm_area_struct *vma) struct hugepage_subpool *spool = subpool_vma(vma); unsigned long reserve, start, end; long gbl_reserve; + struct dhugetlb_pool *hpool = + HUGETLBFS_I(file_inode(vma->vm_file))->hpool; if (!resv || !is_vma_resv_set(vma, HPAGE_RESV_OWNER)) return; @@ -3551,8 +4690,8 @@ static void hugetlb_vm_op_close(struct vm_area_struct *vma) * Decrement reserve counts. The global reserve count may be * adjusted if the subpool has a minimum size. */ - gbl_reserve = hugepage_subpool_put_pages(spool, reserve); - hugetlb_acct_memory(h, -gbl_reserve); + gbl_reserve = hugepage_subpool_put_pages(spool, reserve, hpool); + hugetlb_acct_memory(h, -gbl_reserve, hpool); } } @@ -4934,6 +6073,7 @@ int hugetlb_reserve_pages(struct inode *inode, struct hugepage_subpool *spool = subpool_inode(inode); struct resv_map *resv_map; long gbl_reserve; + struct dhugetlb_pool *hpool = HUGETLBFS_I(inode)->hpool; /* This should never happen */ if (from > to) { @@ -4986,7 +6126,7 @@ int hugetlb_reserve_pages(struct inode *inode, * the subpool has a minimum size, there may be some global * reservations already in place (gbl_reserve). */ - gbl_reserve = hugepage_subpool_get_pages(spool, chg); + gbl_reserve = hugepage_subpool_get_pages(spool, chg, hpool); if (gbl_reserve < 0) { ret = -ENOSPC; goto out_err; @@ -4996,10 +6136,10 @@ int hugetlb_reserve_pages(struct inode *inode, * Check enough hugepages are available for the reservation. * Hand the pages back to the subpool if there are not */ - ret = hugetlb_acct_memory(h, gbl_reserve); + ret = hugetlb_acct_memory(h, gbl_reserve, hpool); if (ret < 0) { /* put back original number of pages, chg */ - (void)hugepage_subpool_put_pages(spool, chg); + (void)hugepage_subpool_put_pages(spool, chg, hpool); goto out_err; } @@ -5028,8 +6168,9 @@ int hugetlb_reserve_pages(struct inode *inode, long rsv_adjust; rsv_adjust = hugepage_subpool_put_pages(spool, - chg - add); - hugetlb_acct_memory(h, -rsv_adjust); + chg - add, + hpool); + hugetlb_acct_memory(h, -rsv_adjust, hpool); } } return 0; @@ -5051,6 +6192,7 @@ long hugetlb_unreserve_pages(struct inode *inode, long start, long end, long chg = 0; struct hugepage_subpool *spool = subpool_inode(inode); long gbl_reserve; + struct dhugetlb_pool *hpool = HUGETLBFS_I(inode)->hpool; /* * Since this routine can be called in the evict inode path for all @@ -5075,8 +6217,8 @@ long hugetlb_unreserve_pages(struct inode *inode, long start, long end, * If the subpool has a minimum size, the number of global * reservations to be released may be adjusted. */ - gbl_reserve = hugepage_subpool_put_pages(spool, (chg - freed)); - hugetlb_acct_memory(h, -gbl_reserve); + gbl_reserve = hugepage_subpool_put_pages(spool, (chg - freed), hpool); + hugetlb_acct_memory(h, -gbl_reserve, hpool); return 0; } diff --git a/mm/internal.h b/mm/internal.h index 1b861446c751..deffd247b010 100644 --- a/mm/internal.h +++ b/mm/internal.h @@ -182,6 +182,7 @@ extern void __putback_isolated_page(struct page *page, unsigned int order, int mt); extern void __free_pages_core(struct page *page, unsigned int order); extern void prep_compound_page(struct page *page, unsigned int order); +extern int check_new_page(struct page *page); extern void post_alloc_hook(struct page *page, unsigned int order, gfp_t gfp_flags); extern int user_min_free_kbytes; diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 63b91a030b02..bdc90e6fc082 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -997,6 +997,41 @@ static __always_inline struct mem_cgroup *get_mem_cgroup_from_current(void) return get_mem_cgroup_from_mm(current->mm); } +#ifdef CONFIG_DYNAMIC_HUGETLB +void free_page_to_dhugetlb_pool(struct page *page) +{ + struct dhugetlb_pool *hpool; + struct small_page_pool *smpool; + unsigned long flags; + + hpool = get_dhugetlb_pool_from_dhugetlb_pagelist(page); + if (unlikely(!hpool)) { + pr_err("dhugetlb: free error: get hpool failed\n"); + return; + } + + smpool = &hpool->smpool[smp_processor_id()]; + spin_lock_irqsave(&smpool->lock, flags); + + ClearPagePool(page); + if (!free_pages_prepare(page, 0, false)) { + SetPagePool(page); + goto out; + } + list_add(&page->lru, &smpool->head_page); + smpool->free_pages++; + smpool->used_pages--; + if (smpool->free_pages > MAX_SMPOOL_PAGE) { + spin_lock(&hpool->lock); + move_pages_from_smpool_to_hpool(hpool, smpool); + spin_unlock(&hpool->lock); + } +out: + spin_unlock_irqrestore(&smpool->lock, flags); + dhugetlb_pool_put(hpool); +} +#endif /* CONFIG_DYNAMIC_HUGETLB */ + /** * mem_cgroup_iter - iterate over memory cgroup hierarchy * @root: hierarchy root @@ -3118,6 +3153,31 @@ static int mem_cgroup_force_empty(struct mem_cgroup *memcg) return 0; } +#ifdef CONFIG_DYNAMIC_HUGETLB +int dhugetlb_pool_force_empty(struct mem_cgroup *memcg) +{ + lru_add_drain_all(); + + drain_all_stock(memcg); + + while (page_counter_read(&memcg->memory)) { + int progress; + + if (signal_pending(current)) + return -EINTR; + + progress = try_to_free_mem_cgroup_pages(memcg, 1, + GFP_HIGHUSER_MOVABLE, + false); + + if (!progress) { + congestion_wait(BLK_RW_ASYNC, HZ/10); + break; + } + } + return 0; +} +#endif static ssize_t mem_cgroup_force_empty_write(struct kernfs_open_file *of, char *buf, size_t nbytes, @@ -4652,6 +4712,305 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of, return ret; } +#ifdef CONFIG_DYNAMIC_HUGETLB +struct dhugetlb_pool *get_dhugetlb_pool_from_memcg(struct mem_cgroup *memcg) +{ + struct mem_cgroup_extension *memcg_ext; + + if (!memcg) + return NULL; + + memcg_ext = container_of(memcg, struct mem_cgroup_extension, memcg); + if (dhugetlb_pool_get(memcg_ext->hpool)) + return memcg_ext->hpool; + return NULL; +} + +static void set_dhugetlb_pool_to_memcg(struct mem_cgroup *memcg, + struct dhugetlb_pool *hpool) +{ + struct mem_cgroup_extension *memcg_ext; + + memcg_ext = container_of(memcg, struct mem_cgroup_extension, memcg); + + memcg_ext->hpool = hpool; +} + +static bool should_allocate_from_dhugetlb_pool(gfp_t gfp_mask) +{ + gfp_t gfp = gfp_mask & GFP_HIGHUSER_MOVABLE; + + if (current->flags & PF_KTHREAD) + return false; + + /* + * The cgroup only charges anonymous and file pages from usespage. + * some filesystem maybe has masked out the __GFP_IO | __GFP_FS + * to avoid recursive memory request. eg: loop device, xfs. + */ + if ((gfp | __GFP_IO | __GFP_FS) != GFP_HIGHUSER_MOVABLE) + return false; + + return true; +} + +static struct page *__alloc_page_from_dhugetlb_pool(void) +{ + bool ret; + struct dhugetlb_pool *hpool; + struct small_page_pool *smpool; + struct page *page = NULL; + unsigned long flags; + + hpool = get_dhugetlb_pool_from_task(current); + if (unlikely(!hpool)) + goto out; + + smpool = &hpool->smpool[smp_processor_id()]; + spin_lock_irqsave(&smpool->lock, flags); + + if (smpool->free_pages == 0) { + spin_lock(&hpool->lock); + ret = move_pages_from_hpool_to_smpool(hpool, smpool); + spin_unlock(&hpool->lock); + if (!ret) + goto unlock; + } + + page = list_entry(smpool->head_page.next, struct page, lru); + list_del(&page->lru); + smpool->free_pages--; + smpool->used_pages++; + check_new_page(page); + SetPagePool(page); +unlock: + spin_unlock_irqrestore(&smpool->lock, flags); +out: + dhugetlb_pool_put(hpool); + return page; +} + +struct page *alloc_page_from_dhugetlb_pool(gfp_t gfp_mask) +{ + struct page *page = NULL; + + if (should_allocate_from_dhugetlb_pool(gfp_mask)) + page = __alloc_page_from_dhugetlb_pool(); + + return page; +} + +static void assign_new_dhugetlb_pool(struct mem_cgroup *memcg, + unsigned long nid) +{ + struct dhugetlb_pool *hpool; + + hpool = hpool_alloc(nid); + if (!hpool) + return; + + hpool->attach_memcg = memcg; + css_get(&memcg->css); + set_dhugetlb_pool_to_memcg(memcg, hpool); +} + +static int update_dhugetlb_pool(struct mem_cgroup *memcg, + unsigned long nid, unsigned long size) +{ + int ret; + struct dhugetlb_pool *hpool = get_dhugetlb_pool_from_memcg(memcg); + + if (!hpool) { + if (memcg_has_children(memcg)) + return -EINVAL; + assign_new_dhugetlb_pool(memcg, nid); + hpool = get_dhugetlb_pool_from_memcg(memcg); + } + if (!hpool) + return -ENOMEM; + if (hpool->attach_memcg != memcg || hpool->nid != nid) { + dhugetlb_pool_put(hpool); + return -EINVAL; + } + + ret = alloc_hugepage_from_hugetlb(hpool, nid, size); + + dhugetlb_pool_put(hpool); + return ret; +} + +/* + * Test whether an process can allocate specified memory size. + * + * Input must be in format '<nid> <size>'. + * size is regarded as how many it does 1G huge page. + */ +static ssize_t memcg_write_dhugetlb(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) +{ + int ret; + unsigned long nid, size; + char *endp; + struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); + + if (!dhugetlb_enabled) + return -EINVAL; + + buf = strstrip(buf); + nid = memparse(buf, &endp); + if (*endp != ' ' || nid >= MAX_NUMNODES) + return -EINVAL; + + buf = endp + 1; + size = memparse(buf, &endp); + if (*endp != '\0' || size == 0) + return -EINVAL; + + ret = update_dhugetlb_pool(memcg, nid, size); + + return ret ?: nbytes; +} + +static int memcg_read_dhugetlb(struct seq_file *m, void *v) +{ + int i; + unsigned long free_pages; + long used_pages = 0; + struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); + struct dhugetlb_pool *hpool = get_dhugetlb_pool_from_memcg(memcg); + + if (!dhugetlb_enabled) + return 0; + if (!hpool) { + seq_printf(m, "Curent hierarchial have not memory pool.\n"); + return 0; + } + + for (i = 0; i < NR_SMPOOL; i++) + spin_lock(&hpool->smpool[i].lock); + spin_lock(&hpool->lock); + + free_pages = hpool->free_pages; + for (i = 0; i < NR_SMPOOL; i++) { + free_pages += hpool->smpool[i].free_pages; + used_pages += hpool->smpool[i].used_pages; + } + + seq_printf(m, "dhugetlb_total_pages %ld\n" + "1G_total_reserved_pages %ld\n" + "1G_free_reserved_pages %ld\n" + "1G_mmap_reserved_pages %ld\n" + "1G_used_pages %ld\n" + "1G_free_unreserved_pages %ld\n" + "2M_total_reserved_pages %ld\n" + "2M_free_reserved_pages %ld\n" + "2M_mmap_reserved_pages %ld\n" + "2M_used_pages %ld\n" + "2M_free_unreserved_pages %ld\n" + "4K_free_pages %ld\n" + "4K_used_pages %ld\n", + hpool->total_nr_pages, + hpool->total_reserved_1G, + hpool->free_reserved_1G, + hpool->mmap_reserved_1G, + hpool->used_1G, + hpool->free_unreserved_1G, + hpool->total_reserved_2M, + hpool->free_reserved_2M, + hpool->mmap_reserved_2M, + hpool->used_2M, + hpool->free_unreserved_2M, + free_pages, + used_pages); + + spin_unlock(&hpool->lock); + for (i = NR_SMPOOL - 1; i >= 0; i--) + spin_unlock(&hpool->smpool[i].lock); + dhugetlb_pool_put(hpool); + return 0; +} + +static int update_reserve_pages(struct kernfs_open_file *of, + char *buf, bool gigantic) +{ + unsigned long size; + char *endp; + struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); + struct dhugetlb_pool *hpool; + + if (!dhugetlb_enabled) + return -EINVAL; + + buf = strstrip(buf); + size = memparse(buf, &endp); + if (*endp != '\0') + return -EINVAL; + + hpool = get_dhugetlb_pool_from_memcg(memcg); + if (!hpool) + return -EINVAL; + spin_lock(&hpool->reserved_lock); + dhugetlb_reserve_hugepages(hpool, size, gigantic); + spin_unlock(&hpool->reserved_lock); + dhugetlb_pool_put(hpool); + return 0; +} + +static ssize_t dhugetlb_1G_reserve_write(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) +{ + return update_reserve_pages(of, buf, true) ?: nbytes; +} + +static ssize_t dhugetlb_2M_reserve_write(struct kernfs_open_file *of, + char *buf, size_t nbytes, loff_t off) +{ + return update_reserve_pages(of, buf, false) ?: nbytes; +} + +static void dhugetlb_pool_inherits(struct mem_cgroup *memcg, + struct mem_cgroup *parent) +{ + struct dhugetlb_pool *hpool; + + hpool = get_dhugetlb_pool_from_memcg(parent); + if (!hpool) + return; + + set_dhugetlb_pool_to_memcg(memcg, hpool); + dhugetlb_pool_put(hpool); +} + +static bool dhugetlb_pool_free(struct mem_cgroup *memcg) +{ + bool ret = true; + struct dhugetlb_pool *hpool; + + hpool = get_dhugetlb_pool_from_memcg(memcg); + if (hpool && hpool->attach_memcg == memcg) + ret = free_dhugetlb_pool(hpool); + dhugetlb_pool_put(hpool); + return ret; +} + +bool dhugetlb_pool_is_free(struct cgroup_subsys_state *css) +{ + if (dhugetlb_enabled) + return dhugetlb_pool_free(mem_cgroup_from_css(css)); + return true; +} +#else +static void dhugetlb_pool_inherits(struct mem_cgroup *memcg, + struct mem_cgroup *parent) +{ +} + +bool dhugetlb_pool_is_free(struct cgroup_subsys_state *css) +{ + return true; +} +#endif /* CONFIG_DYNAMIC_HUGETLB */ + static struct cftype mem_cgroup_legacy_files[] = { { .name = "usage_in_bytes", @@ -4700,6 +5059,27 @@ static struct cftype mem_cgroup_legacy_files[] = { .write = memcg_write_event_control, .flags = CFTYPE_NO_PREFIX | CFTYPE_WORLD_WRITABLE, }, +#ifdef CONFIG_DYNAMIC_HUGETLB + { + .name = "dhugetlb.nr_pages", + .write = memcg_write_dhugetlb, + .seq_show = memcg_read_dhugetlb, + .flags = CFTYPE_NO_PREFIX | CFTYPE_WORLD_WRITABLE | + CFTYPE_NOT_ON_ROOT, + }, + { + .name = "dhugetlb.1G.reserved_pages", + .write = dhugetlb_1G_reserve_write, + .flags = CFTYPE_NO_PREFIX | CFTYPE_WORLD_WRITABLE | + CFTYPE_NOT_ON_ROOT, + }, + { + .name = "dhugetlb.2M.reserved_pages", + .write = dhugetlb_2M_reserve_write, + .flags = CFTYPE_NO_PREFIX | CFTYPE_WORLD_WRITABLE | + CFTYPE_NOT_ON_ROOT, + }, +#endif { .name = "swappiness", .read_u64 = mem_cgroup_swappiness_read, @@ -5063,6 +5443,9 @@ mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) return &memcg->css; } + if (dhugetlb_enabled) + dhugetlb_pool_inherits(memcg, parent); + error = memcg_online_kmem(memcg); if (error) goto fail; @@ -5681,6 +6064,14 @@ static int mem_cgroup_can_attach(struct cgroup_taskset *tset) if (!p) return 0; + if (dhugetlb_enabled) { + struct dhugetlb_pool *hpool = get_dhugetlb_pool_from_task(p); + + if (hpool) { + dhugetlb_pool_put(hpool); + return -EPERM; + } + } /* * We are now commited to this value whatever it is. Changes in this * tunable will only affect upcoming migrations, not the current one. diff --git a/mm/page_alloc.c b/mm/page_alloc.c index a6a2f254f61f..e722d73a3724 100644 --- a/mm/page_alloc.c +++ b/mm/page_alloc.c @@ -1052,7 +1052,7 @@ static int free_tail_pages_check(struct page *head_page, struct page *page) return ret; } -static __always_inline bool free_pages_prepare(struct page *page, +__always_inline bool free_pages_prepare(struct page *page, unsigned int order, bool check_free) { int bad = 0; @@ -2012,7 +2012,7 @@ static void check_new_page_bad(struct page *page) /* * This page is about to be returned from the page allocator */ -static inline int check_new_page(struct page *page) +inline int check_new_page(struct page *page) { if (likely(page_expected_state(page, PAGE_FLAGS_CHECK_AT_PREP|__PG_HWPOISON))) @@ -2075,8 +2075,8 @@ inline void post_alloc_hook(struct page *page, unsigned int order, set_page_owner(page, order, gfp_flags); } -static void prep_new_page(struct page *page, unsigned int order, gfp_t gfp_flags, - unsigned int alloc_flags) +void prep_new_page(struct page *page, unsigned int order, gfp_t gfp_flags, + unsigned int alloc_flags) { int i; @@ -2955,6 +2955,12 @@ void free_unref_page(struct page *page) unsigned long flags; unsigned long pfn = page_to_pfn(page); + /* Free dynamic hugetlb page */ + if (dhugetlb_enabled && PagePool(page)) { + free_page_to_dhugetlb_pool(page); + return; + } + if (!free_unref_page_prepare(page, pfn)) return; @@ -2972,6 +2978,16 @@ void free_unref_page_list(struct list_head *list) unsigned long flags, pfn; int batch_count = 0; + /* Free dynamic hugetlb pages */ + if (dhugetlb_enabled) { + list_for_each_entry_safe(page, next, list, lru) { + if (PagePool(page)) { + list_del(&page->lru); + free_page_to_dhugetlb_pool(page); + } + } + } + /* Prepare pages for freeing */ list_for_each_entry_safe(page, next, list, lru) { pfn = page_to_pfn(page); @@ -4785,6 +4801,15 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, int preferred_nid, finalise_ac(gfp_mask, &ac); + /* Dynamic hugetlb allocation attemp */ + if (dhugetlb_enabled && likely(order == 0)) { + page = alloc_page_from_dhugetlb_pool(gfp_mask); + if (page) { + prep_new_page(page, order, gfp_mask, alloc_flags); + goto out; + } + } + /* First allocation attempt */ page = get_page_from_freelist(alloc_mask, order, alloc_flags, &ac); if (likely(page)) -- 2.25.1

1 5

[OLK-5.10 1/3] net: Use u64_stats_fetch_begin_irq() for stats fetch.
by Wang Yufen 03 Feb '23

03 Feb '23

From: Sebastian Andrzej Siewior <bigeasy(a)linutronix.de> stable inclusion from stable-v5.10.142 commit d71a1c9fce184718d1b3a51a9e8a6e31cbbb45ce category: bugfix bugzilla: 188217 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… ------------------------------------------------- commit 278d3ba61563ceed3cb248383ced19e14ec7bc1f upstream. On 32bit-UP u64_stats_fetch_begin() disables only preemption. If the reader is in preemptible context and the writer side (u64_stats_update_begin*()) runs in an interrupt context (IRQ or softirq) then the writer can update the stats during the read operation. This update remains undetected. Use u64_stats_fetch_begin_irq() to ensure the stats fetch on 32bit-UP are not interrupted by a writer. 32bit-SMP remains unaffected by this change. Cc: "David S. Miller" <davem(a)davemloft.net> Cc: Catherine Sullivan <csully(a)google.com> Cc: David Awogbemila <awogbemila(a)google.com> Cc: Dimitris Michailidis <dmichail(a)fungible.com> Cc: Eric Dumazet <edumazet(a)google.com> Cc: Hans Ulli Kroll <ulli.kroll(a)googlemail.com> Cc: Jakub Kicinski <kuba(a)kernel.org> Cc: Jeroen de Borst <jeroendb(a)google.com> Cc: Johannes Berg <johannes(a)sipsolutions.net> Cc: Linus Walleij <linus.walleij(a)linaro.org> Cc: Paolo Abeni <pabeni(a)redhat.com> Cc: Simon Horman <simon.horman(a)corigine.com> Cc: linux-arm-kernel(a)lists.infradead.org Cc: linux-wireless(a)vger.kernel.org Cc: netdev(a)vger.kernel.org Cc: oss-drivers(a)corigine.com Cc: stable(a)vger.kernel.org Signed-off-by: Sebastian Andrzej Siewior <bigeasy(a)linutronix.de> Reviewed-by: Simon Horman <simon.horman(a)corigine.com> Signed-off-by: David S. Miller <davem(a)davemloft.net> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> (cherry picked from commit d71a1c9fce184718d1b3a51a9e8a6e31cbbb45ce) Signed-off-by: Wang Yufen <wangyufen(a)huawei.com> Conflicts: drivers/net/ethernet/huawei/hinic/hinic_rx.c drivers/net/ethernet/huawei/hinic/hinic_tx.c Signed-off-by: Wang Yufen <wangyufen(a)huawei.com> --- drivers/net/ethernet/cortina/gemini.c | 24 +++++++++++----------- drivers/net/ethernet/google/gve/gve_ethtool.c | 16 +++++++-------- drivers/net/ethernet/google/gve/gve_main.c | 12 +++++------ drivers/net/ethernet/huawei/hinic/hinic_rx.c | 4 ++-- drivers/net/ethernet/huawei/hinic/hinic_tx.c | 4 ++-- .../net/ethernet/netronome/nfp/nfp_net_common.c | 8 ++++---- .../net/ethernet/netronome/nfp/nfp_net_ethtool.c | 8 ++++---- drivers/net/netdevsim/netdev.c | 4 ++-- net/mac80211/sta_info.c | 8 ++++---- net/mpls/af_mpls.c | 4 ++-- 10 files changed, 46 insertions(+), 46 deletions(-) diff --git a/drivers/net/ethernet/cortina/gemini.c b/drivers/net/ethernet/cortina/gemini.c index 3685878..b22ea40 100644 --- a/drivers/net/ethernet/cortina/gemini.c +++ b/drivers/net/ethernet/cortina/gemini.c @@ -1920,7 +1920,7 @@ static void gmac_get_stats64(struct net_device *netdev, /* Racing with RX NAPI */ do { - start = u64_stats_fetch_begin(&port->rx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->rx_stats_syncp); stats->rx_packets = port->stats.rx_packets; stats->rx_bytes = port->stats.rx_bytes; @@ -1932,11 +1932,11 @@ static void gmac_get_stats64(struct net_device *netdev, stats->rx_crc_errors = port->stats.rx_crc_errors; stats->rx_frame_errors = port->stats.rx_frame_errors; - } while (u64_stats_fetch_retry(&port->rx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->rx_stats_syncp, start)); /* Racing with MIB and TX completion interrupts */ do { - start = u64_stats_fetch_begin(&port->ir_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->ir_stats_syncp); stats->tx_errors = port->stats.tx_errors; stats->tx_packets = port->stats.tx_packets; @@ -1946,15 +1946,15 @@ static void gmac_get_stats64(struct net_device *netdev, stats->rx_missed_errors = port->stats.rx_missed_errors; stats->rx_fifo_errors = port->stats.rx_fifo_errors; - } while (u64_stats_fetch_retry(&port->ir_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->ir_stats_syncp, start)); /* Racing with hard_start_xmit */ do { - start = u64_stats_fetch_begin(&port->tx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->tx_stats_syncp); stats->tx_dropped = port->stats.tx_dropped; - } while (u64_stats_fetch_retry(&port->tx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->tx_stats_syncp, start)); stats->rx_dropped += stats->rx_missed_errors; } @@ -2032,18 +2032,18 @@ static void gmac_get_ethtool_stats(struct net_device *netdev, /* Racing with MIB interrupt */ do { p = values; - start = u64_stats_fetch_begin(&port->ir_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->ir_stats_syncp); for (i = 0; i < RX_STATS_NUM; i++) *p++ = port->hw_stats[i]; - } while (u64_stats_fetch_retry(&port->ir_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->ir_stats_syncp, start)); values = p; /* Racing with RX NAPI */ do { p = values; - start = u64_stats_fetch_begin(&port->rx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->rx_stats_syncp); for (i = 0; i < RX_STATUS_NUM; i++) *p++ = port->rx_stats[i]; @@ -2051,13 +2051,13 @@ static void gmac_get_ethtool_stats(struct net_device *netdev, *p++ = port->rx_csum_stats[i]; *p++ = port->rx_napi_exits; - } while (u64_stats_fetch_retry(&port->rx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->rx_stats_syncp, start)); values = p; /* Racing with TX start_xmit */ do { p = values; - start = u64_stats_fetch_begin(&port->tx_stats_syncp); + start = u64_stats_fetch_begin_irq(&port->tx_stats_syncp); for (i = 0; i < TX_MAX_FRAGS; i++) { *values++ = port->tx_frag_stats[i]; @@ -2066,7 +2066,7 @@ static void gmac_get_ethtool_stats(struct net_device *netdev, *values++ = port->tx_frags_linearized; *values++ = port->tx_hw_csummed; - } while (u64_stats_fetch_retry(&port->tx_stats_syncp, start)); + } while (u64_stats_fetch_retry_irq(&port->tx_stats_syncp, start)); } static int gmac_get_ksettings(struct net_device *netdev, diff --git a/drivers/net/ethernet/google/gve/gve_ethtool.c b/drivers/net/ethernet/google/gve/gve_ethtool.c index 66f9b37..80a8c0c 100644 --- a/drivers/net/ethernet/google/gve/gve_ethtool.c +++ b/drivers/net/ethernet/google/gve/gve_ethtool.c @@ -172,14 +172,14 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) struct gve_rx_ring *rx = &priv->rx[ring]; start = - u64_stats_fetch_begin(&priv->rx[ring].statss); + u64_stats_fetch_begin_irq(&priv->rx[ring].statss); tmp_rx_pkts = rx->rpackets; tmp_rx_bytes = rx->rbytes; tmp_rx_skb_alloc_fail = rx->rx_skb_alloc_fail; tmp_rx_buf_alloc_fail = rx->rx_buf_alloc_fail; tmp_rx_desc_err_dropped_pkt = rx->rx_desc_err_dropped_pkt; - } while (u64_stats_fetch_retry(&priv->rx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->rx[ring].statss, start)); rx_pkts += tmp_rx_pkts; rx_bytes += tmp_rx_bytes; @@ -193,10 +193,10 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) if (priv->tx) { do { start = - u64_stats_fetch_begin(&priv->tx[ring].statss); + u64_stats_fetch_begin_irq(&priv->tx[ring].statss); tmp_tx_pkts = priv->tx[ring].pkt_done; tmp_tx_bytes = priv->tx[ring].bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->tx[ring].statss, start)); tx_pkts += tmp_tx_pkts; tx_bytes += tmp_tx_bytes; @@ -254,13 +254,13 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) data[i++] = rx->cnt; do { start = - u64_stats_fetch_begin(&priv->rx[ring].statss); + u64_stats_fetch_begin_irq(&priv->rx[ring].statss); tmp_rx_bytes = rx->rbytes; tmp_rx_skb_alloc_fail = rx->rx_skb_alloc_fail; tmp_rx_buf_alloc_fail = rx->rx_buf_alloc_fail; tmp_rx_desc_err_dropped_pkt = rx->rx_desc_err_dropped_pkt; - } while (u64_stats_fetch_retry(&priv->rx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->rx[ring].statss, start)); data[i++] = tmp_rx_bytes; /* rx dropped packets */ @@ -313,9 +313,9 @@ static int gve_get_sset_count(struct net_device *netdev, int sset) data[i++] = tx->done; do { start = - u64_stats_fetch_begin(&priv->tx[ring].statss); + u64_stats_fetch_begin_irq(&priv->tx[ring].statss); tmp_tx_bytes = tx->bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->tx[ring].statss, start)); data[i++] = tmp_tx_bytes; data[i++] = tx->wake_queue; diff --git a/drivers/net/ethernet/google/gve/gve_main.c b/drivers/net/ethernet/google/gve/gve_main.c index 6cb75bb..f0c1e6c8 100644 --- a/drivers/net/ethernet/google/gve/gve_main.c +++ b/drivers/net/ethernet/google/gve/gve_main.c @@ -40,10 +40,10 @@ static void gve_get_stats(struct net_device *dev, struct rtnl_link_stats64 *s) for (ring = 0; ring < priv->rx_cfg.num_queues; ring++) { do { start = - u64_stats_fetch_begin(&priv->rx[ring].statss); + u64_stats_fetch_begin_irq(&priv->rx[ring].statss); packets = priv->rx[ring].rpackets; bytes = priv->rx[ring].rbytes; - } while (u64_stats_fetch_retry(&priv->rx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->rx[ring].statss, start)); s->rx_packets += packets; s->rx_bytes += bytes; @@ -53,10 +53,10 @@ static void gve_get_stats(struct net_device *dev, struct rtnl_link_stats64 *s) for (ring = 0; ring < priv->tx_cfg.num_queues; ring++) { do { start = - u64_stats_fetch_begin(&priv->tx[ring].statss); + u64_stats_fetch_begin_irq(&priv->tx[ring].statss); packets = priv->tx[ring].pkt_done; bytes = priv->tx[ring].bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[ring].statss, + } while (u64_stats_fetch_retry_irq(&priv->tx[ring].statss, start)); s->tx_packets += packets; s->tx_bytes += bytes; @@ -1041,9 +1041,9 @@ void gve_handle_report_stats(struct gve_priv *priv) if (priv->tx) { for (idx = 0; idx < priv->tx_cfg.num_queues; idx++) { do { - start = u64_stats_fetch_begin(&priv->tx[idx].statss); + start = u64_stats_fetch_begin_irq(&priv->tx[idx].statss); tx_bytes = priv->tx[idx].bytes_done; - } while (u64_stats_fetch_retry(&priv->tx[idx].statss, start)); + } while (u64_stats_fetch_retry_irq(&priv->tx[idx].statss, start)); stats[stats_idx++] = (struct stats) { .stat_name = cpu_to_be32(TX_WAKE_CNT), .value = cpu_to_be64(priv->tx[idx].wake_queue), diff --git a/drivers/net/ethernet/huawei/hinic/hinic_rx.c b/drivers/net/ethernet/huawei/hinic/hinic_rx.c index 57d5d79..1b57b67 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_rx.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_rx.c @@ -375,7 +375,7 @@ void hinic_rxq_get_stats(struct hinic_rxq *rxq, u64_stats_update_begin(&stats->syncp); do { - start = u64_stats_fetch_begin(&rxq_stats->syncp); + start = u64_stats_fetch_begin_irq(&rxq_stats->syncp); stats->bytes = rxq_stats->bytes; stats->packets = rxq_stats->packets; stats->errors = rxq_stats->csum_errors + @@ -384,7 +384,7 @@ void hinic_rxq_get_stats(struct hinic_rxq *rxq, stats->other_errors = rxq_stats->other_errors; stats->dropped = rxq_stats->dropped; stats->rx_buf_empty = rxq_stats->rx_buf_empty; - } while (u64_stats_fetch_retry(&rxq_stats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&rxq_stats->syncp, start)); u64_stats_update_end(&stats->syncp); } diff --git a/drivers/net/ethernet/huawei/hinic/hinic_tx.c b/drivers/net/ethernet/huawei/hinic/hinic_tx.c index 75fa344..ff37b6f 100644 --- a/drivers/net/ethernet/huawei/hinic/hinic_tx.c +++ b/drivers/net/ethernet/huawei/hinic/hinic_tx.c @@ -61,7 +61,7 @@ void hinic_txq_get_stats(struct hinic_txq *txq, u64_stats_update_begin(&stats->syncp); do { - start = u64_stats_fetch_begin(&txq_stats->syncp); + start = u64_stats_fetch_begin_irq(&txq_stats->syncp); stats->bytes = txq_stats->bytes; stats->packets = txq_stats->packets; stats->busy = txq_stats->busy; @@ -69,7 +69,7 @@ void hinic_txq_get_stats(struct hinic_txq *txq, stats->dropped = txq_stats->dropped; stats->big_frags_pkts = txq_stats->big_frags_pkts; stats->big_udp_pkts = txq_stats->big_udp_pkts; - } while (u64_stats_fetch_retry(&txq_stats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&txq_stats->syncp, start)); u64_stats_update_end(&stats->syncp); } diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c index dfc1f32..5ab230aa 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_common.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_common.c @@ -3373,21 +3373,21 @@ static void nfp_net_stat64(struct net_device *netdev, unsigned int start; do { - start = u64_stats_fetch_begin(&r_vec->rx_sync); + start = u64_stats_fetch_begin_irq(&r_vec->rx_sync); data[0] = r_vec->rx_pkts; data[1] = r_vec->rx_bytes; data[2] = r_vec->rx_drops; - } while (u64_stats_fetch_retry(&r_vec->rx_sync, start)); + } while (u64_stats_fetch_retry_irq(&r_vec->rx_sync, start)); stats->rx_packets += data[0]; stats->rx_bytes += data[1]; stats->rx_dropped += data[2]; do { - start = u64_stats_fetch_begin(&r_vec->tx_sync); + start = u64_stats_fetch_begin_irq(&r_vec->tx_sync); data[0] = r_vec->tx_pkts; data[1] = r_vec->tx_bytes; data[2] = r_vec->tx_errors; - } while (u64_stats_fetch_retry(&r_vec->tx_sync, start)); + } while (u64_stats_fetch_retry_irq(&r_vec->tx_sync, start)); stats->tx_packets += data[0]; stats->tx_bytes += data[1]; stats->tx_errors += data[2]; diff --git a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c index bfcd90f..d4136d3 100644 --- a/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c +++ b/drivers/net/ethernet/netronome/nfp/nfp_net_ethtool.c @@ -498,7 +498,7 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) unsigned int start; do { - start = u64_stats_fetch_begin(&nn->r_vecs[i].rx_sync); + start = u64_stats_fetch_begin_irq(&nn->r_vecs[i].rx_sync); data[0] = nn->r_vecs[i].rx_pkts; tmp[0] = nn->r_vecs[i].hw_csum_rx_ok; tmp[1] = nn->r_vecs[i].hw_csum_rx_inner_ok; @@ -506,10 +506,10 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) tmp[3] = nn->r_vecs[i].hw_csum_rx_error; tmp[4] = nn->r_vecs[i].rx_replace_buf_alloc_fail; tmp[5] = nn->r_vecs[i].hw_tls_rx; - } while (u64_stats_fetch_retry(&nn->r_vecs[i].rx_sync, start)); + } while (u64_stats_fetch_retry_irq(&nn->r_vecs[i].rx_sync, start)); do { - start = u64_stats_fetch_begin(&nn->r_vecs[i].tx_sync); + start = u64_stats_fetch_begin_irq(&nn->r_vecs[i].tx_sync); data[1] = nn->r_vecs[i].tx_pkts; data[2] = nn->r_vecs[i].tx_busy; tmp[6] = nn->r_vecs[i].hw_csum_tx; @@ -519,7 +519,7 @@ static u64 *nfp_vnic_get_sw_stats(struct net_device *netdev, u64 *data) tmp[10] = nn->r_vecs[i].hw_tls_tx; tmp[11] = nn->r_vecs[i].tls_tx_fallback; tmp[12] = nn->r_vecs[i].tls_tx_no_fallback; - } while (u64_stats_fetch_retry(&nn->r_vecs[i].tx_sync, start)); + } while (u64_stats_fetch_retry_irq(&nn->r_vecs[i].tx_sync, start)); data += NN_RVEC_PER_Q_STATS; diff --git a/drivers/net/netdevsim/netdev.c b/drivers/net/netdevsim/netdev.c index ad6dbf01..4fb0638 100644 --- a/drivers/net/netdevsim/netdev.c +++ b/drivers/net/netdevsim/netdev.c @@ -67,10 +67,10 @@ static int nsim_change_mtu(struct net_device *dev, int new_mtu) unsigned int start; do { - start = u64_stats_fetch_begin(&ns->syncp); + start = u64_stats_fetch_begin_irq(&ns->syncp); stats->tx_bytes = ns->tx_bytes; stats->tx_packets = ns->tx_packets; - } while (u64_stats_fetch_retry(&ns->syncp, start)); + } while (u64_stats_fetch_retry_irq(&ns->syncp, start)); } static int diff --git a/net/mac80211/sta_info.c b/net/mac80211/sta_info.c index 461c037..cee39ae 100644 --- a/net/mac80211/sta_info.c +++ b/net/mac80211/sta_info.c @@ -2175,9 +2175,9 @@ static inline u64 sta_get_tidstats_msdu(struct ieee80211_sta_rx_stats *rxstats, u64 value; do { - start = u64_stats_fetch_begin(&rxstats->syncp); + start = u64_stats_fetch_begin_irq(&rxstats->syncp); value = rxstats->msdu[tid]; - } while (u64_stats_fetch_retry(&rxstats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&rxstats->syncp, start)); return value; } @@ -2241,9 +2241,9 @@ static inline u64 sta_get_stats_bytes(struct ieee80211_sta_rx_stats *rxstats) u64 value; do { - start = u64_stats_fetch_begin(&rxstats->syncp); + start = u64_stats_fetch_begin_irq(&rxstats->syncp); value = rxstats->bytes; - } while (u64_stats_fetch_retry(&rxstats->syncp, start)); + } while (u64_stats_fetch_retry_irq(&rxstats->syncp, start)); return value; } diff --git a/net/mpls/af_mpls.c b/net/mpls/af_mpls.c index 9c047c1..7239814 100644 --- a/net/mpls/af_mpls.c +++ b/net/mpls/af_mpls.c @@ -1078,9 +1078,9 @@ static void mpls_get_stats(struct mpls_dev *mdev, p = per_cpu_ptr(mdev->stats, i); do { - start = u64_stats_fetch_begin(&p->syncp); + start = u64_stats_fetch_begin_irq(&p->syncp); local = p->stats; - } while (u64_stats_fetch_retry(&p->syncp, start)); + } while (u64_stats_fetch_retry_irq(&p->syncp, start)); stats->rx_packets += local.rx_packets; stats->rx_bytes += local.rx_bytes; -- 1.8.3.1

1 2

[OLK-5.10 1/7] geneve: do not use RT_TOS for IPv6 flowlabel
by Wang Yufen 03 Feb '23

03 Feb '23

From: Matthias May <matthias.may(a)westermo.com> stable inclusion from stable-v5.10.138 commit 38b83883ce4e4efe8ff0a727192219cac2668d42 category: bugfix bugzilla: 188217 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… ------------------------------------------------- commit ca2bb69514a8bc7f83914122f0d596371352416c upstream. According to Guillaume Nault RT_TOS should never be used for IPv6. Quote: RT_TOS() is an old macro used to interprete IPv4 TOS as described in the obsolete RFC 1349. It's conceptually wrong to use it even in IPv4 code, although, given the current state of the code, most of the existing calls have no consequence. But using RT_TOS() in IPv6 code is always a bug: IPv6 never had a "TOS" field to be interpreted the RFC 1349 way. There's no historical compatibility to worry about. Fixes: 3a56f86f1be6 ("geneve: handle ipv6 priority like ipv4 tos") Acked-by: Guillaume Nault <gnault(a)redhat.com> Signed-off-by: Matthias May <matthias.may(a)westermo.com> Signed-off-by: Jakub Kicinski <kuba(a)kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> (cherry picked from commit 38b83883ce4e4efe8ff0a727192219cac2668d42) Signed-off-by: Wang Yufen <wangyufen(a)huawei.com> --- drivers/net/geneve.c | 3 +-- 1 file changed, 1 insertion(+), 2 deletions(-) diff --git a/drivers/net/geneve.c b/drivers/net/geneve.c index 5ddb2db..ba9947d 100644 --- a/drivers/net/geneve.c +++ b/drivers/net/geneve.c @@ -850,8 +850,7 @@ static struct dst_entry *geneve_get_v6_dst(struct sk_buff *skb, use_cache = false; } - fl6->flowlabel = ip6_make_flowinfo(RT_TOS(prio), - info->key.label); + fl6->flowlabel = ip6_make_flowinfo(prio, info->key.label); dst_cache = (struct dst_cache *)&info->dst_cache; if (use_cache) { dst = dst_cache_get_ip6(dst_cache, &fl6->saddr); -- 1.8.3.1

2 7

[PATCH openEuler-23.03] config: disable CONFIG_EFI_ZBOOT by default
by Xie XiuQi 01 Feb '23

01 Feb '23

CONFIG_EFI_ZBOOT is introduced to openEuler 22.03 LTS SP1 by this commit for loongarch and enabled by default: e46780727555("efi/libstub: implement generic EFI zboot"). However, if is enabled, the compiled version cannot be booted on openEuler 22.03 LTS. So, disable it on this version. Signed-off-by: Xie XiuQi <xiexiuqi(a)huawei.com> --- arch/arm64/configs/openeuler_defconfig | 14 +------------- 1 file changed, 1 insertion(+), 13 deletions(-) diff --git a/arch/arm64/configs/openeuler_defconfig b/arch/arm64/configs/openeuler_defconfig index fbb71f7520a8..63f2adbbc778 100644 --- a/arch/arm64/configs/openeuler_defconfig +++ b/arch/arm64/configs/openeuler_defconfig @@ -15,18 +15,6 @@ CONFIG_INIT_ENV_ARG_LIMIT=32 CONFIG_LOCALVERSION="" # CONFIG_LOCALVERSION_AUTO is not set CONFIG_BUILD_SALT="" -CONFIG_HAVE_KERNEL_GZIP=y -CONFIG_HAVE_KERNEL_LZMA=y -CONFIG_HAVE_KERNEL_XZ=y -CONFIG_HAVE_KERNEL_LZO=y -CONFIG_HAVE_KERNEL_LZ4=y -CONFIG_HAVE_KERNEL_ZSTD=y -CONFIG_KERNEL_GZIP=y -# CONFIG_KERNEL_LZMA is not set -# CONFIG_KERNEL_XZ is not set -# CONFIG_KERNEL_LZO is not set -# CONFIG_KERNEL_LZ4 is not set -# CONFIG_KERNEL_ZSTD is not set CONFIG_DEFAULT_INIT="" CONFIG_DEFAULT_HOSTNAME="(none)" CONFIG_SYSVIPC=y @@ -2008,7 +1996,7 @@ CONFIG_EFI_SOFT_RESERVE=y CONFIG_EFI_PARAMS_FROM_FDT=y CONFIG_EFI_RUNTIME_WRAPPERS=y CONFIG_EFI_GENERIC_STUB=y -CONFIG_EFI_ZBOOT=y +# CONFIG_EFI_ZBOOT is not set CONFIG_EFI_ARMSTUB_DTB_LOADER=y CONFIG_EFI_GENERIC_STUB_INITRD_CMDLINE_LOADER=y # CONFIG_EFI_BOOTLOADER_CONTROL is not set -- 2.33.0

1 0

[PATCH openEuler-5.10] Add relationship graph of openEuler kernel branches
by Zheng Zengkai 01 Feb '23

01 Feb '23

openEuler inclusion category: doc bugzilla: https://gitee.com/openeuler/kernel/issues/I6CAZ5 ----------------------------- Newcomers may get confused about the relationship of openEuler kernel branches, so add relationship graph of openEuler kernel branches to README in openEuler kernel master branch. Signed-off-by: Zheng Zengkai <zhengzengkai(a)huawei.com> --- README_openEuler | 45 +++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 45 insertions(+) create mode 100644 README_openEuler diff --git a/README_openEuler b/README_openEuler new file mode 100644 index 000000000000..7ec06f5b2538 --- /dev/null +++ b/README_openEuler @@ -0,0 +1,45 @@ +README for openEuler kernel branches + +Relationship of openEuler kernel 4.19 branches: + +------------------------------------------------------------------> Linux 4.19.y + | + | tag: v4.19.13 + | + -----------------------------------------------------------> openEuler-1.0-LTS + +Relationship of openEuler kernel 5.10 branches: + +------------------------------------------------------------------> Linux 5.10.y + | + | tag: v5.10 + | + --------------------------------------------------------------------> OLK-5.10 + | | | | | + | tag: 5.10.0-4.0.0 | + | | | | | + ------^-----^-----^-----^--------------> openEuler-21.03(Innovative version) + | | | | + | tag: 5.10.0-5.0.0 + | | | | + ------^-----^-----^--------------> openEuler-21.09(Innovative version) + | | | + | tag: 5.10.0-60.0.0 + | | | + ------^-----^------------------------------> openEuler-22.03-LTS + | | + | tag: 5.10.0-106.0.0 + | | + ------^--------------> openEuler-22.09(Innovative version) + | + | tag: 5.10.0-136.0.0 + | + ---------------------------> openEuler-22.03-LTS-SP1 + +Relationship of openEuler kernel 6.x branches: + +-------------------------------------------------------------------> Linux 6.1.y + | + | tag: v6.1 + | + -----------------------------------------------> devel-6.1(Innovative version) -- 2.20.1

1 0