- Kernel - mailweb.openeuler.org

[PATCH OLK-5.10] i2c: core: fix adapter deregistration race
by Zhang Yuwei 12 Aug '26

12 Aug '26

From: Johan Hovold <johan(a)kernel.org> mainline inclusion from mainline-v7.2-rc1 commit b1a58ed9eab146b36f41a55db8f5d7ce9fdedf3f category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16577 CVE: CVE-2026-64279 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Adapters can be looked up by their id using i2c_get_adapter() which takes a reference to the embedded struct device. Remove the adapter from the IDR before tearing it down during deregistration (and on registration failure) to make sure its resources are not accessed after having been freed (e.g. the device name). Fixes: 35fc37f81881 ("i2c: Limit core locking to the necessary sections") Cc: stable(a)vger.kernel.org # 2.6.31 Cc: Jean Delvare <khali(a)linux-fr.org> Signed-off-by: Johan Hovold <johan(a)kernel.org> Signed-off-by: Wolfram Sang <wsa+renesas(a)sang-engineering.com> Conflicts: drivers/i2c/i2c-core-base.c [contxt conflict] Signed-off-by: Zhang Yuwei <zhangyuwei20(a)huawei.com> --- drivers/i2c/i2c-core-base.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/i2c/i2c-core-base.c b/drivers/i2c/i2c-core-base.c index ae9682826156..45cde9cd9417 100644 --- a/drivers/i2c/i2c-core-base.c +++ b/drivers/i2c/i2c-core-base.c @@ -1496,6 +1496,9 @@ static int i2c_register_adapter(struct i2c_adapter *adap) res = device_register(&adap->dev); if (res) { pr_err("adapter '%s': can't register device (%d)\n", adap->name, res); + mutex_lock(&core_lock); + idr_replace(&i2c_adapter_idr, NULL, adap->nr); + mutex_unlock(&core_lock); goto out_list; } @@ -1537,6 +1540,9 @@ static int i2c_register_adapter(struct i2c_adapter *adap) return 0; out_reg: + mutex_lock(&core_lock); + idr_replace(&i2c_adapter_idr, NULL, adap->nr); + mutex_unlock(&core_lock); init_completion(&adap->dev_released); device_unregister(&adap->dev); wait_for_completion(&adap->dev_released); @@ -1693,6 +1699,8 @@ void i2c_del_adapter(struct i2c_adapter *adap) /* First make sure that this adapter was ever added */ mutex_lock(&core_lock); found = idr_find(&i2c_adapter_idr, adap->nr); + if (found == adap) + idr_replace(&i2c_adapter_idr, NULL, adap->nr); mutex_unlock(&core_lock); if (found != adap) { pr_debug("attempting to delete unregistered adapter [%s]\n", adap->name); -- 2.22.0

2 1

[PATCH OLK-6.6] i2c: core: fix adapter deregistration race
by Zhang Yuwei 12 Aug '26

12 Aug '26

From: Johan Hovold <johan(a)kernel.org> mainline inclusion from mainline-v7.2-rc1 commit b1a58ed9eab146b36f41a55db8f5d7ce9fdedf3f category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16577 CVE: CVE-2026-64279 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- Adapters can be looked up by their id using i2c_get_adapter() which takes a reference to the embedded struct device. Remove the adapter from the IDR before tearing it down during deregistration (and on registration failure) to make sure its resources are not accessed after having been freed (e.g. the device name). Fixes: 35fc37f81881 ("i2c: Limit core locking to the necessary sections") Cc: stable(a)vger.kernel.org # 2.6.31 Cc: Jean Delvare <khali(a)linux-fr.org> Signed-off-by: Johan Hovold <johan(a)kernel.org> Signed-off-by: Wolfram Sang <wsa+renesas(a)sang-engineering.com> Conflicts: drivers/i2c/i2c-core-base.c [contxt conflict] Signed-off-by: Zhang Yuwei <zhangyuwei20(a)huawei.com> --- drivers/i2c/i2c-core-base.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/i2c/i2c-core-base.c b/drivers/i2c/i2c-core-base.c index d87e99c0448c..4127cf8b7e49 100644 --- a/drivers/i2c/i2c-core-base.c +++ b/drivers/i2c/i2c-core-base.c @@ -1525,6 +1525,9 @@ static int i2c_register_adapter(struct i2c_adapter *adap) res = device_register(&adap->dev); if (res) { pr_err("adapter '%s': can't register device (%d)\n", adap->name, res); + mutex_lock(&core_lock); + idr_replace(&i2c_adapter_idr, NULL, adap->nr); + mutex_unlock(&core_lock); goto err_remove_irq_domain; } @@ -1567,6 +1570,9 @@ static int i2c_register_adapter(struct i2c_adapter *adap) return 0; out_reg: + mutex_lock(&core_lock); + idr_replace(&i2c_adapter_idr, NULL, adap->nr); + mutex_unlock(&core_lock); init_completion(&adap->dev_released); device_unregister(&adap->dev); wait_for_completion(&adap->dev_released); @@ -1725,6 +1731,8 @@ void i2c_del_adapter(struct i2c_adapter *adap) /* First make sure that this adapter was ever added */ mutex_lock(&core_lock); found = idr_find(&i2c_adapter_idr, adap->nr); + if (found == adap) + idr_replace(&i2c_adapter_idr, NULL, adap->nr); mutex_unlock(&core_lock); if (found != adap) { pr_debug("attempting to delete unregistered adapter [%s]\n", adap->name); -- 2.22.0

2 1

[PATCH OLK-5.10 0/2] *** proc: fix ptrace access locking and error paths ***
by Zhang Yuwei 12 Aug '26

12 Aug '26

*** BLURB HERE *** Jann Horn (2): proc: protect ptrace_may_access() with exec_update_lock (part 1) proc: Fix broken error paths for namespace links fs/proc/array.c | 6 ++++++ fs/proc/base.c | 41 ++++++++++++++++++++++++----------------- fs/proc/namespaces.c | 16 +++++++++++++++- 3 files changed, 45 insertions(+), 18 deletions(-) -- 2.22.0

2 3

[PATCH openEuler-1.0-LTS] RDMA/core: Prefer NLA_NUL_STRING
by Xia Fukun 11 Aug '26

11 Aug '26

From: Florian Westphal <fw(a)strlen.de> mainline inclusion from mainline-v7.1-rc1 commit 6ed3d14fc45d3da6025e7fe4a6a09066856698e2 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16189 CVE: CVE-2026-63860 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- These attributes are evaluated as c-string (passed to strcmp), but NLA_STRING doesn't check for the presence of a \0 terminator. Either this needs to switch to nla_strcmp() and needs to adjust printf fmt specifier to not use plain %s, or this needs to use NLA_NUL_STRING. As the code has been this way for long time, it seems to me that userspace does include the terminating nul, even tough its not enforced so far, and thus NLA_NUL_STRING use is the simpler solution. Fixes: 30dc5e63d6a5 ("RDMA/core: Add support for iWARP Port Mapper user space service") Link: https://patch.msgid.link/r/20260330122742.13315-1-fw@strlen.de Signed-off-by: Florian Westphal <fw(a)strlen.de> Signed-off-by: Jason Gunthorpe <jgg(a)nvidia.com> Conflicts: drivers/infiniband/core/iwpm_msg.c [Context differences, unrelated to the fix patch.] Signed-off-by: Xia Fukun <xiafukun(a)huawei.com> --- drivers/infiniband/core/iwpm_msg.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) diff --git a/drivers/infiniband/core/iwpm_msg.c b/drivers/infiniband/core/iwpm_msg.c index 8861c052155a..37abfd24073e 100644 --- a/drivers/infiniband/core/iwpm_msg.c +++ b/drivers/infiniband/core/iwpm_msg.c @@ -336,9 +336,9 @@ int iwpm_remove_mapping(struct sockaddr_storage *local_addr, u8 nl_client) /* netlink attribute policy for the received response to register pid request */ static const struct nla_policy resp_reg_policy[IWPM_NLA_RREG_PID_MAX] = { [IWPM_NLA_RREG_PID_SEQ] = { .type = NLA_U32 }, - [IWPM_NLA_RREG_IBDEV_NAME] = { .type = NLA_STRING, + [IWPM_NLA_RREG_IBDEV_NAME] = { .type = NLA_NUL_STRING, .len = IWPM_DEVNAME_SIZE - 1 }, - [IWPM_NLA_RREG_ULIB_NAME] = { .type = NLA_STRING, + [IWPM_NLA_RREG_ULIB_NAME] = { .type = NLA_NUL_STRING, .len = IWPM_ULIBNAME_SIZE - 1 }, [IWPM_NLA_RREG_ULIB_VER] = { .type = NLA_U16 }, [IWPM_NLA_RREG_PID_ERR] = { .type = NLA_U16 } @@ -630,7 +630,7 @@ int iwpm_remote_info_cb(struct sk_buff *skb, struct netlink_callback *cb) /* netlink attribute policy for the received request for mapping info */ static const struct nla_policy resp_mapinfo_policy[IWPM_NLA_MAPINFO_REQ_MAX] = { - [IWPM_NLA_MAPINFO_ULIB_NAME] = { .type = NLA_STRING, + [IWPM_NLA_MAPINFO_ULIB_NAME] = { .type = NLA_NUL_STRING, .len = IWPM_ULIBNAME_SIZE - 1 }, [IWPM_NLA_MAPINFO_ULIB_VER] = { .type = NLA_U16 } }; -- 2.34.1

2 1

[PATCH v2 OLK-6.6] mm/memory: fix spurious warning when unmapping device-private/exclusive pages
by Ze Zuo 11 Aug '26

11 Aug '26

From: Alistair Popple <apopple(a)nvidia.com> stable inclusion from stable-v6.6.143 commit e7af1b15c884ed12bb69da11aec095045d861ee8 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16387 CVE: CVE-2026-64131 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit be3f38d05cc5a7c3f13e51994c5dd043ab604d28 ] Device private and exclusive entries are only supported for anonymous folios. This condition is tested in __migrate_device_pages() and make_device_exclusive() using folio_test_anon(). However the unmap path tests this assumption using vma_is_anonymous(). This is wrong because whilst anonymous VMAs can only contain folios where folio_test_anon() is true the opposite relation does not hold. A folio for which folio_test_anon() is true does not imply vma_is_anonymous() is true. Such a condition can occur if for example a folio is part of a private filebacked mapping. In this case vma_is_anonymous() is false as the mapping is filebacked, but folio_test_anon() may be true, thus permitting devices to migrate the folio to device private memory. This can lead to the following spurious warnings during process teardown: [ 772.737706] ------------[ cut here ]------------ [ 772.739201] WARNING: mm/memory.c:1754 at unmap_page_range.cold+0x26/0x18a, CPU#17: hmm-tests/2041 [ 772.742050] Modules linked in: test_hmm nvidia_uvm(O) nvidia(O) [ 772.743959] CPU: 17 UID: 0 PID: 2041 Comm: hmm-tests Tainted: G W O 7.0.0+ #387 PREEMPT(full) [ 772.747104] Tainted: [W]=WARN, [O]=OOT_MODULE [ 772.748509] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.17.0-0-gb52ca86e094d-prebuilt.qemu.org 04/01/2014 [ 772.752117] RIP: 0010:unmap_page_range.cold+0x26/0x18a [ 772.753780] Code: 7e fe ff ff 48 89 4c 24 78 4c 89 44 24 38 e8 f2 ff b1 00 48 8b 4c 24 78 4c 8b 44 24 38 48 8b 44 24 18 48 83 78 48 00 74 04 90 <0f> 0b 90 48 89 ca b8 ff ff 37 00 48 c1 ea 03 48 c1 e0 2a 80 3c 02 [ 772.759602] RSP: 0018:ffff888112607550 EFLAGS: 00010286 [ 772.761310] RAX: ffff88811bbf4dc0 RBX: dffffc0000000000 RCX: ffffea03e9bfffd8 [ 772.763583] RDX: 1ffff1102377e9c1 RSI: 0000000000000008 RDI: ffff88811bbf4e08 [ 772.765914] RBP: 0000000000000006 R08: ffff8881059f7448 R09: ffffed10224c0e68 [ 772.768184] R10: ffff888112607347 R11: 0000000000000001 R12: 0000000000000001 [ 772.770461] R13: ffffea03e9bfffc0 R14: ffff888112607908 R15: ffffea03e9bfffc0 [ 772.772782] FS: 00007f327caa2780(0000) GS:ffff888427b7d000(0000) knlGS:0000000000000000 [ 772.775328] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 772.777187] CR2: 00007f327ca89000 CR3: 00000001994d5000 CR4: 00000000000006f0 [ 772.779135] Call Trace: [ 772.779792] <TASK> [ 772.780317] ? dmirror_interval_invalidate+0x1a3/0x290 [test_hmm] [ 772.781873] ? vm_normal_page_pud+0x2b0/0x2b0 [ 772.782992] ? __rwlock_init+0x150/0x150 [ 772.784006] ? lock_release+0x216/0x2b0 [ 772.785008] ? __mmu_notifier_invalidate_range_start+0x505/0x6e0 [ 772.786522] ? lock_release+0x216/0x2b0 [ 772.787498] ? unmap_single_vma+0xb6/0x210 [ 772.788573] unmap_vmas+0x27d/0x520 [ 772.789506] ? unmap_single_vma+0x210/0x210 [ 772.790607] ? mas_update_gap.part.0+0x620/0x620 [ 772.791834] unmap_region+0x19e/0x350 [ 772.792769] ? remove_vma+0x130/0x130 [ 772.793684] ? mas_alloc_nodes+0x1f2/0x300 [ 772.794730] vms_complete_munmap_vmas+0x8c1/0xe20 [ 772.795926] ? unmap_region+0x350/0x350 [ 772.796917] do_vmi_align_munmap+0x36a/0x4e0 [ 772.798018] ? lock_release+0x216/0x2b0 [ 772.799024] ? vma_shrink+0x620/0x620 [ 772.799983] do_vmi_munmap+0x150/0x2c0 [ 772.800939] __vm_munmap+0x161/0x2c0 [ 772.801872] ? expand_downwards+0xd60/0xd60 [ 772.802948] ? clockevents_program_event+0x1ef/0x540 [ 772.804217] ? lock_release+0x216/0x2b0 [ 772.805158] __x64_sys_munmap+0x59/0x80 [ 772.805776] do_syscall_64+0xfc/0x670 [ 772.806336] ? irqentry_exit+0xda/0x580 [ 772.806976] entry_SYSCALL_64_after_hwframe+0x4b/0x53 [ 772.807772] RIP: 0033:0x7f327cbb2717 [ 772.808323] Code: 73 01 c3 48 8b 0d f9 76 0d 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 b8 0b 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d c9 76 0d 00 f7 d8 64 89 01 48 [ 772.811337] RSP: 002b:00007ffde7f57d38 EFLAGS: 00000202 ORIG_RAX: 000000000000000b [ 772.812564] RAX: ffffffffffffffda RBX: 00007f327cc9c000 RCX: 00007f327cbb2717 [ 772.813733] RDX: 0000000000000000 RSI: 0000000000400000 RDI: 00007f327c289000 [ 772.814867] RBP: 0000000000421360 R08: 000000000000001a R09: 0000000000000000 [ 772.815991] R10: 0000000000000003 R11: 0000000000000202 R12: 00007ffde7f57d74 [ 772.817121] R13: 00007f327c689010 R14: 0000000000100000 R15: 00007f327c289000 [ 772.818272] </TASK> [ 772.818614] irq event stamp: 0 [ 772.819159] hardirqs last enabled at (0): [<0000000000000000>] 0x0 [ 772.820174] hardirqs last disabled at (0): [<ffffffff82a57ab3>] copy_process+0x19f3/0x6440 [ 772.821511] softirqs last enabled at (0): [<ffffffff82a57b00>] copy_process+0x1a40/0x6440 [ 772.822869] softirqs last disabled at (0): [<0000000000000000>] 0x0 [ 772.823871] ---[ end trace 0000000000000000 ]--- Fix this by using the same check for folio_test_anon() in zap_nonpresent_ptes(). Also add a hmm-test case for this. Link: https://lore.kernel.org/20260501065116.2057242-1-apopple@nvidia.com Fixes: 999dad824c39 ("mm/shmem: persist uffd-wp bit across zapping for file-backed") Signed-off-by: Alistair Popple <apopple(a)nvidia.com> Reported-by: Arsen Arsenović <aarsenovic(a)baylibre.com> Reviewed-by: Balbir Singh <balbirs(a)nvidia.com> Cc: David Hildenbrand <david(a)kernel.org> Cc: Jason Gunthorpe <jgg(a)ziepe.ca> Cc: John Hubbard <jhubbard(a)nvidia.com> Cc: Leon Romanovsky <leon(a)kernel.org> Cc: Liam R. Howlett <liam(a)infradead.org> Cc: Lorenzo Stoakes <ljs(a)kernel.org> Cc: Peter Xu <peterx(a)redhat.com> Cc: Matthew Brost <matthew.brost(a)intel.com> Cc: Michal Hocko <mhocko(a)suse.com> Cc: Mike Rapoport <rppt(a)kernel.org> Cc: Shuah Khan <shuah(a)kernel.org> Cc: Suren Baghdasaryan <surenb(a)google.com> Cc: Thomas Hellström <thomas.hellstrom(a)linux.intel.com> Cc: Vlastimil Babka <vbabka(a)kernel.org> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> [ adapted `folio_test_anon(folio)` to `PageAnon(page)` ] Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Conflicts: mm/memory.c [Conflict due to merge commit a23f517b0e15 ("mm: convert mm_counter() to take a folio"), context conflict.] Signed-off-by: Ze Zuo <zuoze1(a)huawei.com> --- mm/memory.c | 2 +- tools/testing/selftests/mm/hmm-tests.c | 50 ++++++++++++++++++++++++++ 2 files changed, 51 insertions(+), 1 deletion(-) diff --git a/mm/memory.c b/mm/memory.c index 0b56a022c3f8..789b4b301415 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1635,7 +1635,7 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb, * consider uffd-wp bit when zap. For more information, * see zap_install_uffd_wp_if_needed(). */ - WARN_ON_ONCE(!vma_is_anonymous(vma)); + WARN_ON_ONCE(!folio_test_anon(folio)); rss[mm_counter(folio)]--; if (is_device_private_entry(entry)) folio_remove_rmap_pte(folio, page, vma); diff --git a/tools/testing/selftests/mm/hmm-tests.c b/tools/testing/selftests/mm/hmm-tests.c index 0cc1056c8fd7..5286e400ce8b 100644 --- a/tools/testing/selftests/mm/hmm-tests.c +++ b/tools/testing/selftests/mm/hmm-tests.c @@ -998,6 +998,56 @@ TEST_F(hmm, migrate) hmm_buffer_free(buffer); } +/* + * Migrate private file memory to device private memory. + */ +TEST_F(hmm, migrate_file_private) +{ + struct hmm_buffer *buffer; + unsigned long npages; + unsigned long size; + unsigned long i; + int *ptr; + int ret; + int fd; + + npages = ALIGN(HMM_BUFFER_SIZE, self->page_size) >> self->page_shift; + ASSERT_NE(npages, 0); + size = npages << self->page_shift; + + fd = hmm_create_file(size); + ASSERT_GE(fd, 0); + + buffer = malloc(sizeof(*buffer)); + ASSERT_NE(buffer, NULL); + + buffer->fd = fd; + buffer->size = size; + buffer->mirror = malloc(size); + ASSERT_NE(buffer->mirror, NULL); + + buffer->ptr = mmap(NULL, size, + PROT_READ | PROT_WRITE, + MAP_PRIVATE, + buffer->fd, 0); + ASSERT_NE(buffer->ptr, MAP_FAILED); + + /* Initialize buffer in system memory. */ + for (i = 0, ptr = buffer->ptr; i < size / sizeof(*ptr); ++i) + ptr[i] = i; + + /* Migrate memory to device. */ + ret = hmm_migrate_sys_to_dev(self->fd, buffer, npages); + ASSERT_EQ(ret, 0); + ASSERT_EQ(buffer->cpages, npages); + + /* Check what the device read. */ + for (i = 0, ptr = buffer->mirror; i < size / sizeof(*ptr); ++i) + ASSERT_EQ(ptr[i], i); + + hmm_buffer_free(buffer); +} + /* * Migrate anonymous memory to device private memory and fault some of it back * to system memory, then try migrating the resulting mix of system and device -- 2.25.1

2 1

[PATCH OLK-5.10] USB: serial: mxuport: fix memory corruption with small endpoint
by Liu Mingrui 11 Aug '26

11 Aug '26

From: Johan Hovold <johan(a)kernel.org> stable inclusion from stable-v5.10.259 commit 086b858b5f5125bc9d967ea2bd825f83d9f8f29d category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16227 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit 4085f0dbb1ce2251c9a5938d693de6593f0ab2bd upstream. Make sure that the bulk-out endpoint max packet size is at least eight bytes to avoid user-controlled slab corruption should a malicious device report a smaller size. Fixes: ee467a1f2066 ("USB: serial: add Moxa UPORT 12XX/14XX/16XX driver") Cc: stable(a)vger.kernel.org # 3.14 Cc: Andrew Lunn <andrew(a)lunn.ch> Reviewed-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Johan Hovold <johan(a)kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Lin Yujun <linyujun809(a)h-partners.com> Signed-off-by: Mingrui Liu <liumingrui(a)huawei.com> --- drivers/usb/serial/mxuport.c | 8 ++++++++ 1 file changed, 8 insertions(+) diff --git a/drivers/usb/serial/mxuport.c b/drivers/usb/serial/mxuport.c index 5d38c2a0f590..217b114db6bf 100644 --- a/drivers/usb/serial/mxuport.c +++ b/drivers/usb/serial/mxuport.c @@ -969,6 +969,14 @@ static int mxuport_calc_num_ports(struct usb_serial *serial, */ BUILD_BUG_ON(ARRAY_SIZE(epds->bulk_out) < 16); + /* + * The bulk-out buffers must be large enough for the four-byte header + * (and following data), but assume anything smaller than eight bytes + * is broken. + */ + if (usb_endpoint_maxp(epds->bulk_out[0]) < 8) + return -EINVAL; + for (i = 1; i < num_ports; ++i) epds->bulk_out[i] = epds->bulk_out[0]; -- 2.34.1

2 1

[PATCH OLK-6.6] erofs: fix stack-out-of-bounds in erofs_put_metabuf
by Liu Mingrui 11 Aug '26

11 Aug '26

From: Mingrui Liu <liumingrui(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16872 -------------------------------- Commit 1d4b040dd8ee moved erofs_put_metabuf() to out_unlock, but some error paths jump there before buf is initialized, unlike mainline where buf is always set. This triggers : BUG: KASAN: stack-out-of-bounds in erofs_put_metabuf+0x54/0x158 [erofs] Call Trace: dump_backtrace+0xa0/0x128 show_stack+0x20/0x38 dump_stack_lvl+0x78/0xc8 print_address_description.constprop.0+0x84/0x3c8 print_report+0xb0/0x280 kasan_report+0x84/0xd0 __asan_load8+0x9c/0xc0 erofs_put_metabuf+0x54/0x158 [erofs] erofs_init_inode_xattrs+0x194/0x560 [erofs] erofs_getxattr+0xc8/0x238 [erofs] erofs_get_acl+0x6c/0x150 [erofs] __get_acl.part.0+0x218/0x390 vfs_get_acl+0xd0/0x150 do_get_acl+0x38/0x2a0 do_getxattr+0x9c/0x250 getxattr+0xe0/0x148 path_getxattr+0xdc/0x150 __arm64_sys_getxattr+0x60/0x80 invoke_syscall+0x64/0x178 el0_svc_common.constprop.0+0x11c/0x150 do_el0_svc+0x38/0x50 el0_svc+0x44/0x228 el0t_64_sync_handler+0x100/0x130 el0t_64_sync+0x3c8/0x3d0 Fix by zero-initializing buf at declaration so erofs_put_metabuf() can safely handle it. Fixes: 6a89c63750cb ("erofs: fix metabuf leak in inode xattr initialization") Signed-off-by: Mingrui Liu <liumingrui(a)huawei.com> --- fs/erofs/xattr.c | 3 ++- 1 file changed, 2 insertions(+), 1 deletion(-) diff --git a/fs/erofs/xattr.c b/fs/erofs/xattr.c index b270226be379..abd23e990a3a 100644 --- a/fs/erofs/xattr.c +++ b/fs/erofs/xattr.c @@ -34,6 +34,8 @@ static int erofs_init_inode_xattrs(struct inode *inode) struct super_block *sb = inode->i_sb; int ret = 0; + it.buf = __EROFS_BUF_INITIALIZER; + /* the most case is that xattrs of this inode are initialized. */ if (test_bit(EROFS_I_EA_INITED_BIT, &vi->flags)) { /* @@ -76,7 +78,6 @@ static int erofs_init_inode_xattrs(struct inode *inode) goto out_unlock; } - it.buf = __EROFS_BUF_INITIALIZER; erofs_init_metabuf(&it.buf, sb); it.pos = erofs_iloc(inode) + vi->inode_isize; -- 2.34.1

2 1

[PATCH OLK-5.10] md/raid5: validate payload size before accessing journal metadata
by Yi Yang 11 Aug '26

11 Aug '26

From: Junrui Luo <moonafterrain(a)outlook.com> stable inclusion from stable-v5.10.258 commit c96c6f01d84b5c67db1bf1cc8591c0b7146826fc category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/15314 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit b0cc3ae97e893bf54bbce447f4e9fd2e0b88bff9 upstream. r5c_recovery_analyze_meta_block() and r5l_recovery_verify_data_checksum_for_mb() iterate over payloads in a journal metadata block using on-disk payload size fields without validating them against the remaining space in the metadata block. A corrupted journal contains payload sizes extending beyond the PAGE_SIZE boundary can cause out-of-bounds reads when accessing payload fields or computing offsets. Add bounds validation for each payload type to ensure the full payload fits within meta_size before processing. Fixes: b4c625c67362 ("md/r5cache: r5cache recovery: part 1") Cc: stable(a)vger.kernel.org Signed-off-by: Junrui Luo <moonafterrain(a)outlook.com> Link: https://lore.kernel.org/linux-raid/SYBPR01MB78815E78D829BB86CD7C8015AF5FA@S… Signed-off-by: Yu Kuai <yukuai(a)fnnas.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Yi Yang <yiyang13(a)huawei.com> --- drivers/md/raid5-cache.c | 48 +++++++++++++++++++++++++++------------- 1 file changed, 33 insertions(+), 15 deletions(-) diff --git a/drivers/md/raid5-cache.c b/drivers/md/raid5-cache.c index d879fc525f50..75b508c1b841 100644 --- a/drivers/md/raid5-cache.c +++ b/drivers/md/raid5-cache.c @@ -2018,15 +2018,27 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, return -ENOMEM; while (mb_offset < le32_to_cpu(mb->meta_size)) { + sector_t payload_len; + payload = (void *)mb + mb_offset; payload_flush = (void *)mb + mb_offset; if (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_DATA) { + payload_len = sizeof(struct r5l_payload_data_parity) + + (sector_t)sizeof(__le32) * + (le32_to_cpu(payload->size) >> (PAGE_SHIFT - 9)); + if (mb_offset + payload_len > le32_to_cpu(mb->meta_size)) + goto mismatch; if (r5l_recovery_verify_data_checksum( log, ctx, page, log_offset, payload->checksum[0]) < 0) goto mismatch; } else if (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_PARITY) { + payload_len = sizeof(struct r5l_payload_data_parity) + + (sector_t)sizeof(__le32) * + (le32_to_cpu(payload->size) >> (PAGE_SHIFT - 9)); + if (mb_offset + payload_len > le32_to_cpu(mb->meta_size)) + goto mismatch; if (r5l_recovery_verify_data_checksum( log, ctx, page, log_offset, payload->checksum[0]) < 0) @@ -2039,22 +2051,18 @@ r5l_recovery_verify_data_checksum_for_mb(struct r5l_log *log, payload->checksum[1]) < 0) goto mismatch; } else if (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_FLUSH) { - /* nothing to do for R5LOG_PAYLOAD_FLUSH here */ + payload_len = sizeof(struct r5l_payload_flush) + + (sector_t)le32_to_cpu(payload_flush->size); + if (mb_offset + payload_len > le32_to_cpu(mb->meta_size)) + goto mismatch; } else /* not R5LOG_PAYLOAD_DATA/PARITY/FLUSH */ goto mismatch; - if (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_FLUSH) { - mb_offset += sizeof(struct r5l_payload_flush) + - le32_to_cpu(payload_flush->size); - } else { - /* DATA or PARITY payload */ + if (le16_to_cpu(payload->header.type) != R5LOG_PAYLOAD_FLUSH) { log_offset = r5l_ring_add(log, log_offset, le32_to_cpu(payload->size)); - mb_offset += sizeof(struct r5l_payload_data_parity) + - sizeof(__le32) * - (le32_to_cpu(payload->size) >> (PAGE_SHIFT - 9)); } - + mb_offset += payload_len; } put_page(page); @@ -2105,6 +2113,7 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, log_offset = r5l_ring_add(log, ctx->pos, BLOCK_SECTORS); while (mb_offset < le32_to_cpu(mb->meta_size)) { + sector_t payload_len; int dd; payload = (void *)mb + mb_offset; @@ -2113,6 +2122,12 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, if (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_FLUSH) { int i, count; + payload_len = sizeof(struct r5l_payload_flush) + + (sector_t)le32_to_cpu(payload_flush->size); + if (mb_offset + payload_len > + le32_to_cpu(mb->meta_size)) + return -EINVAL; + count = le32_to_cpu(payload_flush->size) / sizeof(__le64); for (i = 0; i < count; ++i) { stripe_sect = le64_to_cpu(payload_flush->flush_stripes[i]); @@ -2126,12 +2141,17 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, } } - mb_offset += sizeof(struct r5l_payload_flush) + - le32_to_cpu(payload_flush->size); + mb_offset += payload_len; continue; } /* DATA or PARITY payload */ + payload_len = sizeof(struct r5l_payload_data_parity) + + (sector_t)sizeof(__le32) * + (le32_to_cpu(payload->size) >> (PAGE_SHIFT - 9)); + if (mb_offset + payload_len > le32_to_cpu(mb->meta_size)) + return -EINVAL; + stripe_sect = (le16_to_cpu(payload->header.type) == R5LOG_PAYLOAD_DATA) ? raid5_compute_sector( conf, le64_to_cpu(payload->location), 0, &dd, @@ -2196,9 +2216,7 @@ r5c_recovery_analyze_meta_block(struct r5l_log *log, log_offset = r5l_ring_add(log, log_offset, le32_to_cpu(payload->size)); - mb_offset += sizeof(struct r5l_payload_data_parity) + - sizeof(__le32) * - (le32_to_cpu(payload->size) >> (PAGE_SHIFT - 9)); + mb_offset += payload_len; } return 0; -- 2.25.1

2 1

[PATCH OLK-6.6] mm/memory: fix spurious warning when unmapping device-private/exclusive pages
by Ze Zuo 11 Aug '26

11 Aug '26

From: Alistair Popple <apopple(a)nvidia.com> stable inclusion from stable-v6.6.143 commit e7af1b15c884ed12bb69da11aec095045d861ee8 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/16387 CVE : CVE-2026-64131 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit be3f38d05cc5a7c3f13e51994c5dd043ab604d28 ] Device private and exclusive entries are only supported for anonymous folios. This condition is tested in __migrate_device_pages() and make_device_exclusive() using folio_test_anon(). However the unmap path tests this assumption using vma_is_anonymous(). This is wrong because whilst anonymous VMAs can only contain folios where folio_test_anon() is true the opposite relation does not hold. A folio for which folio_test_anon() is true does not imply vma_is_anonymous() is true. Such a condition can occur if for example a folio is part of a private filebacked mapping. In this case vma_is_anonymous() is false as the mapping is filebacked, but folio_test_anon() may be true, thus permitting devices to migrate the folio to device private memory. This can lead to the following spurious warnings during process teardown: [ 772.737706] ------------[ cut here ]------------ [ 772.739201] WARNING: mm/memory.c:1754 at unmap_page_range.cold+0x26/0x18a, CPU#17: hmm-tests/2041 [ 772.742050] Modules linked in: test_hmm nvidia_uvm(O) nvidia(O) [ 772.743959] CPU: 17 UID: 0 PID: 2041 Comm: hmm-tests Tainted: G W O 7.0.0+ #387 PREEMPT(full) [ 772.747104] Tainted: [W]=WARN, [O]=OOT_MODULE [ 772.748509] Hardware name: QEMU Standard PC (Q35 + ICH9, 2009), BIOS rel-1.17.0-0-gb52ca86e094d-prebuilt.qemu.org 04/01/2014 [ 772.752117] RIP: 0010:unmap_page_range.cold+0x26/0x18a [ 772.753780] Code: 7e fe ff ff 48 89 4c 24 78 4c 89 44 24 38 e8 f2 ff b1 00 48 8b 4c 24 78 4c 8b 44 24 38 48 8b 44 24 18 48 83 78 48 00 74 04 90 <0f> 0b 90 48 89 ca b8 ff ff 37 00 48 c1 ea 03 48 c1 e0 2a 80 3c 02 [ 772.759602] RSP: 0018:ffff888112607550 EFLAGS: 00010286 [ 772.761310] RAX: ffff88811bbf4dc0 RBX: dffffc0000000000 RCX: ffffea03e9bfffd8 [ 772.763583] RDX: 1ffff1102377e9c1 RSI: 0000000000000008 RDI: ffff88811bbf4e08 [ 772.765914] RBP: 0000000000000006 R08: ffff8881059f7448 R09: ffffed10224c0e68 [ 772.768184] R10: ffff888112607347 R11: 0000000000000001 R12: 0000000000000001 [ 772.770461] R13: ffffea03e9bfffc0 R14: ffff888112607908 R15: ffffea03e9bfffc0 [ 772.772782] FS: 00007f327caa2780(0000) GS:ffff888427b7d000(0000) knlGS:0000000000000000 [ 772.775328] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [ 772.777187] CR2: 00007f327ca89000 CR3: 00000001994d5000 CR4: 00000000000006f0 [ 772.779135] Call Trace: [ 772.779792] <TASK> [ 772.780317] ? dmirror_interval_invalidate+0x1a3/0x290 [test_hmm] [ 772.781873] ? vm_normal_page_pud+0x2b0/0x2b0 [ 772.782992] ? __rwlock_init+0x150/0x150 [ 772.784006] ? lock_release+0x216/0x2b0 [ 772.785008] ? __mmu_notifier_invalidate_range_start+0x505/0x6e0 [ 772.786522] ? lock_release+0x216/0x2b0 [ 772.787498] ? unmap_single_vma+0xb6/0x210 [ 772.788573] unmap_vmas+0x27d/0x520 [ 772.789506] ? unmap_single_vma+0x210/0x210 [ 772.790607] ? mas_update_gap.part.0+0x620/0x620 [ 772.791834] unmap_region+0x19e/0x350 [ 772.792769] ? remove_vma+0x130/0x130 [ 772.793684] ? mas_alloc_nodes+0x1f2/0x300 [ 772.794730] vms_complete_munmap_vmas+0x8c1/0xe20 [ 772.795926] ? unmap_region+0x350/0x350 [ 772.796917] do_vmi_align_munmap+0x36a/0x4e0 [ 772.798018] ? lock_release+0x216/0x2b0 [ 772.799024] ? vma_shrink+0x620/0x620 [ 772.799983] do_vmi_munmap+0x150/0x2c0 [ 772.800939] __vm_munmap+0x161/0x2c0 [ 772.801872] ? expand_downwards+0xd60/0xd60 [ 772.802948] ? clockevents_program_event+0x1ef/0x540 [ 772.804217] ? lock_release+0x216/0x2b0 [ 772.805158] __x64_sys_munmap+0x59/0x80 [ 772.805776] do_syscall_64+0xfc/0x670 [ 772.806336] ? irqentry_exit+0xda/0x580 [ 772.806976] entry_SYSCALL_64_after_hwframe+0x4b/0x53 [ 772.807772] RIP: 0033:0x7f327cbb2717 [ 772.808323] Code: 73 01 c3 48 8b 0d f9 76 0d 00 f7 d8 64 89 01 48 83 c8 ff c3 66 2e 0f 1f 84 00 00 00 00 00 0f 1f 44 00 00 b8 0b 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d c9 76 0d 00 f7 d8 64 89 01 48 [ 772.811337] RSP: 002b:00007ffde7f57d38 EFLAGS: 00000202 ORIG_RAX: 000000000000000b [ 772.812564] RAX: ffffffffffffffda RBX: 00007f327cc9c000 RCX: 00007f327cbb2717 [ 772.813733] RDX: 0000000000000000 RSI: 0000000000400000 RDI: 00007f327c289000 [ 772.814867] RBP: 0000000000421360 R08: 000000000000001a R09: 0000000000000000 [ 772.815991] R10: 0000000000000003 R11: 0000000000000202 R12: 00007ffde7f57d74 [ 772.817121] R13: 00007f327c689010 R14: 0000000000100000 R15: 00007f327c289000 [ 772.818272] </TASK> [ 772.818614] irq event stamp: 0 [ 772.819159] hardirqs last enabled at (0): [<0000000000000000>] 0x0 [ 772.820174] hardirqs last disabled at (0): [<ffffffff82a57ab3>] copy_process+0x19f3/0x6440 [ 772.821511] softirqs last enabled at (0): [<ffffffff82a57b00>] copy_process+0x1a40/0x6440 [ 772.822869] softirqs last disabled at (0): [<0000000000000000>] 0x0 [ 772.823871] ---[ end trace 0000000000000000 ]--- Fix this by using the same check for folio_test_anon() in zap_nonpresent_ptes(). Also add a hmm-test case for this. Link: https://lore.kernel.org/20260501065116.2057242-1-apopple@nvidia.com Fixes: 999dad824c39 ("mm/shmem: persist uffd-wp bit across zapping for file-backed") Signed-off-by: Alistair Popple <apopple(a)nvidia.com> Reported-by: Arsen Arsenović <aarsenovic(a)baylibre.com> Reviewed-by: Balbir Singh <balbirs(a)nvidia.com> Cc: David Hildenbrand <david(a)kernel.org> Cc: Jason Gunthorpe <jgg(a)ziepe.ca> Cc: John Hubbard <jhubbard(a)nvidia.com> Cc: Leon Romanovsky <leon(a)kernel.org> Cc: Liam R. Howlett <liam(a)infradead.org> Cc: Lorenzo Stoakes <ljs(a)kernel.org> Cc: Peter Xu <peterx(a)redhat.com> Cc: Matthew Brost <matthew.brost(a)intel.com> Cc: Michal Hocko <mhocko(a)suse.com> Cc: Mike Rapoport <rppt(a)kernel.org> Cc: Shuah Khan <shuah(a)kernel.org> Cc: Suren Baghdasaryan <surenb(a)google.com> Cc: Thomas Hellström <thomas.hellstrom(a)linux.intel.com> Cc: Vlastimil Babka <vbabka(a)kernel.org> Cc: <stable(a)vger.kernel.org> Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org> [ adapted `folio_test_anon(folio)` to `PageAnon(page)` ] Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Conflicts: mm/memory.c [Conflict due to merge commit a23f517b0e15 ("mm: convert mm_counter() to take a folio"), context conflict.] Signed-off-by: Ze Zuo <zuoze1(a)huawei.com> --- mm/memory.c | 2 +- tools/testing/selftests/mm/hmm-tests.c | 50 ++++++++++++++++++++++++++ 2 files changed, 51 insertions(+), 1 deletion(-) diff --git a/mm/memory.c b/mm/memory.c index 0b56a022c3f8..789b4b301415 100644 --- a/mm/memory.c +++ b/mm/memory.c @@ -1635,7 +1635,7 @@ static unsigned long zap_pte_range(struct mmu_gather *tlb, * consider uffd-wp bit when zap. For more information, * see zap_install_uffd_wp_if_needed(). */ - WARN_ON_ONCE(!vma_is_anonymous(vma)); + WARN_ON_ONCE(!folio_test_anon(folio)); rss[mm_counter(folio)]--; if (is_device_private_entry(entry)) folio_remove_rmap_pte(folio, page, vma); diff --git a/tools/testing/selftests/mm/hmm-tests.c b/tools/testing/selftests/mm/hmm-tests.c index 0cc1056c8fd7..5286e400ce8b 100644 --- a/tools/testing/selftests/mm/hmm-tests.c +++ b/tools/testing/selftests/mm/hmm-tests.c @@ -998,6 +998,56 @@ TEST_F(hmm, migrate) hmm_buffer_free(buffer); } +/* + * Migrate private file memory to device private memory. + */ +TEST_F(hmm, migrate_file_private) +{ + struct hmm_buffer *buffer; + unsigned long npages; + unsigned long size; + unsigned long i; + int *ptr; + int ret; + int fd; + + npages = ALIGN(HMM_BUFFER_SIZE, self->page_size) >> self->page_shift; + ASSERT_NE(npages, 0); + size = npages << self->page_shift; + + fd = hmm_create_file(size); + ASSERT_GE(fd, 0); + + buffer = malloc(sizeof(*buffer)); + ASSERT_NE(buffer, NULL); + + buffer->fd = fd; + buffer->size = size; + buffer->mirror = malloc(size); + ASSERT_NE(buffer->mirror, NULL); + + buffer->ptr = mmap(NULL, size, + PROT_READ | PROT_WRITE, + MAP_PRIVATE, + buffer->fd, 0); + ASSERT_NE(buffer->ptr, MAP_FAILED); + + /* Initialize buffer in system memory. */ + for (i = 0, ptr = buffer->ptr; i < size / sizeof(*ptr); ++i) + ptr[i] = i; + + /* Migrate memory to device. */ + ret = hmm_migrate_sys_to_dev(self->fd, buffer, npages); + ASSERT_EQ(ret, 0); + ASSERT_EQ(buffer->cpages, npages); + + /* Check what the device read. */ + for (i = 0, ptr = buffer->mirror; i < size / sizeof(*ptr); ++i) + ASSERT_EQ(ptr[i], i); + + hmm_buffer_free(buffer); +} + /* * Migrate anonymous memory to device private memory and fault some of it back * to system memory, then try migrating the resulting mix of system and device -- 2.25.1

2 3

[PATCH OLK-6.6 V1] RDMA/rxe: Fix iova-to-va conversion for MR page sizes != PAGE_SIZE
by Cheng Yu 11 Aug '26

11 Aug '26

From: Li Zhijian <lizhijian(a)fujitsu.com> mainline inclusion from mainline-v7.0 commit 12985e5915a0b8354796efadaaeb201eed115377 category: bugfix bugzilla: https://atomgit.com/src-openeuler/kernel/issues/15591 CVE: CVE-2026-46325 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- The current implementation incorrectly handles memory regions (MRs) with page sizes different from the system PAGE_SIZE. The core issue is that rxe_set_page() is called with mr->page_size step increments, but the page_list stores individual struct page pointers, each representing PAGE_SIZE of memory. ib_sg_to_page() has ensured that when i>=1 either a) SG[i-1].dma_end and SG[i].dma_addr are contiguous or b) SG[i-1].dma_end and SG[i].dma_addr are mr->page_size aligned. This leads to incorrect iova-to-va conversion in scenarios: 1) page_size < PAGE_SIZE (e.g., MR: 4K, system: 64K): ibmr->iova = 0x181800 sg[0]: dma_addr=0x181800, len=0x800 sg[1]: dma_addr=0x173000, len=0x1000 Access iova = 0x181800 + 0x810 = 0x182010 Expected VA: 0x173010 (second SG, offset 0x10) Before fix: - index = (0x182010 >> 12) - (0x181800 >> 12) = 1 - page_offset = 0x182010 & 0xFFF = 0x10 - xarray[1] stores system page base 0x170000 - Resulting VA: 0x170000 + 0x10 = 0x170010 (wrong) 2) page_size > PAGE_SIZE (e.g., MR: 64K, system: 4K): ibmr->iova = 0x18f800 sg[0]: dma_addr=0x18f800, len=0x800 sg[1]: dma_addr=0x170000, len=0x1000 Access iova = 0x18f800 + 0x810 = 0x190010 Expected VA: 0x170010 (second SG, offset 0x10) Before fix: - index = (0x190010 >> 16) - (0x18f800 >> 16) = 1 - page_offset = 0x190010 & 0xFFFF = 0x10 - xarray[1] stores system page for dma_addr 0x170000 - Resulting VA: system page of 0x170000 + 0x10 = 0x170010 (wrong) Yi Zhang reported a kernel panic[1] years ago related to this defect. Solution: 1. Replace xarray with pre-allocated rxe_mr_page array for sequential indexing (all MR page indices are contiguous) 2. Each rxe_mr_page stores both struct page* and offset within the system page 3. Handle MR page_size != PAGE_SIZE relationships: - page_size > PAGE_SIZE: Split MR pages into multiple system pages - page_size <= PAGE_SIZE: Store offset within system page 4. Add boundary checks and compatibility validation This ensures correct iova-to-va conversion regardless of MR page size and system PAGE_SIZE relationship, while improving performance through array-based sequential access. Tests on 4K and 64K PAGE_SIZE hosts: - rdma-core/pytests $ ./build/bin/run_tests.py --dev eth0_rxe - blktest: $ TIMEOUT=30 QUICK_RUN=1 USE_RXE=1 NVMET_TRTYPES=rdma ./check nvme srp rnbd [1] https://lore.kernel.org/all/CAHj4cs9XRqE25jyVw9rj9YugffLn5+f=1znaBEnu1usLOc… Fixes: 592627ccbdff ("RDMA/rxe: Replace rxe_map and rxe_phys_buf by xarray") Signed-off-by: Li Zhijian <lizhijian(a)fujitsu.com> Link: https://patch.msgid.link/20260116032753.2574363-1-lizhijian@fujitsu.com Signed-off-by: Leon Romanovsky <leon(a)kernel.org> Conflicts: drivers/infiniband/sw/rxe/rxe_mr.c [Applied cleanly to: drivers/infiniband/sw/rxe/rxe_verbs.h. drivers/infiniband/sw/rxe/rxe_mr.c: The rxe_mr_copy_xarray function is renamed to rxe_mr_copy_page_info to match the new array-based access pattern; all xarray references are replaced with direct page_info array indexing. The upstream __alloc_mr_page_info and alloc_mr_page_info helpers are added with identical semantics. The rxe_mr_alloc function is removed as it's replaced by alloc_mr_page_info.] Signed-off-by: Cheng Yu <serein.chengyu(a)huawei.com> --- drivers/infiniband/sw/rxe/rxe_mr.c | 280 +++++++++++++++++--------- drivers/infiniband/sw/rxe/rxe_verbs.h | 10 +- 2 files changed, 193 insertions(+), 97 deletions(-) diff --git a/drivers/infiniband/sw/rxe/rxe_mr.c b/drivers/infiniband/sw/rxe/rxe_mr.c index bc81fde696ee..9b553028a55b 100644 --- a/drivers/infiniband/sw/rxe/rxe_mr.c +++ b/drivers/infiniband/sw/rxe/rxe_mr.c @@ -72,14 +72,46 @@ void rxe_mr_init_dma(int access, struct rxe_mr *mr) mr->ibmr.type = IB_MR_TYPE_DMA; } +/* + * Convert iova to page_info index. The page_info stores pages of size + * PAGE_SIZE, but MRs can have different page sizes. This function + * handles the conversion for all cases: + * + * 1. mr->page_size > PAGE_SIZE: + * The MR's iova may not be aligned to mr->page_size. We use the + * aligned base (iova & page_mask) as reference, then calculate + * which PAGE_SIZE sub-page the iova falls into. + * + * 2. mr->page_size <= PAGE_SIZE: + * Use simple shift arithmetic since each page_info entry corresponds + * to one or more MR pages. + */ static unsigned long rxe_mr_iova_to_index(struct rxe_mr *mr, u64 iova) { - return (iova >> mr->page_shift) - (mr->ibmr.iova >> mr->page_shift); + int idx; + + if (mr_page_size(mr) > PAGE_SIZE) + idx = (iova - (mr->ibmr.iova & mr->page_mask)) >> PAGE_SHIFT; + else + idx = (iova >> mr->page_shift) - + (mr->ibmr.iova >> mr->page_shift); + + WARN_ON(idx >= mr->nbuf); + return idx; } +/* + * Convert iova to offset within the page_info entry. + * + * For mr_page_size > PAGE_SIZE, the offset is within the system page. + * For mr_page_size <= PAGE_SIZE, the offset is within the MR page size. + */ static unsigned long rxe_mr_iova_to_page_offset(struct rxe_mr *mr, u64 iova) { - return iova & (mr_page_size(mr) - 1); + if (mr_page_size(mr) > PAGE_SIZE) + return iova & (PAGE_SIZE - 1); + else + return iova & (mr_page_size(mr) - 1); } static bool is_pmem_page(struct page *pg) @@ -93,37 +125,69 @@ static bool is_pmem_page(struct page *pg) static int rxe_mr_fill_pages_from_sgt(struct rxe_mr *mr, struct sg_table *sgt) { - XA_STATE(xas, &mr->page_list, 0); struct sg_page_iter sg_iter; struct page *page; bool persistent = !!(mr->access & IB_ACCESS_FLUSH_PERSISTENT); + WARN_ON(mr_page_size(mr) != PAGE_SIZE); + __sg_page_iter_start(&sg_iter, sgt->sgl, sgt->orig_nents, 0); if (!__sg_page_iter_next(&sg_iter)) return 0; - do { - xas_lock(&xas); - while (true) { - page = sg_page_iter_page(&sg_iter); - - if (persistent && !is_pmem_page(page)) { - rxe_dbg_mr(mr, "Page can't be persistent\n"); - xas_set_err(&xas, -EINVAL); - break; - } + while (true) { + page = sg_page_iter_page(&sg_iter); - xas_store(&xas, page); - if (xas_error(&xas)) - break; - xas_next(&xas); - if (!__sg_page_iter_next(&sg_iter)) - break; + if (persistent && !is_pmem_page(page)) { + rxe_dbg_mr(mr, "Page can't be persistent\n"); + return -EINVAL; } - xas_unlock(&xas); - } while (xas_nomem(&xas, GFP_KERNEL)); - return xas_error(&xas); + mr->page_info[mr->nbuf].page = page; + mr->page_info[mr->nbuf].offset = 0; + mr->nbuf++; + + if (!__sg_page_iter_next(&sg_iter)) + break; + } + + return 0; +} + +static int __alloc_mr_page_info(struct rxe_mr *mr, int num_pages) +{ + mr->page_info = kcalloc(num_pages, sizeof(struct rxe_mr_page), + GFP_KERNEL); + if (!mr->page_info) + return -ENOMEM; + + mr->max_allowed_buf = num_pages; + mr->nbuf = 0; + + return 0; +} + +static int alloc_mr_page_info(struct rxe_mr *mr, int num_pages) +{ + int ret; + + WARN_ON(mr->num_buf); + ret = __alloc_mr_page_info(mr, num_pages); + if (ret) + return ret; + + mr->num_buf = num_pages; + + return 0; +} + +static void free_mr_page_info(struct rxe_mr *mr) +{ + if (!mr->page_info) + return; + + kfree(mr->page_info); + mr->page_info = NULL; } int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, @@ -134,8 +198,6 @@ int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, rxe_mr_init(access, mr); - xa_init(&mr->page_list); - umem = ib_umem_get(&rxe->ib_dev, start, length, access); if (IS_ERR(umem)) { rxe_dbg_mr(mr, "Unable to pin memory region err = %d\n", @@ -143,46 +205,24 @@ int rxe_mr_init_user(struct rxe_dev *rxe, u64 start, u64 length, u64 iova, return PTR_ERR(umem); } + err = alloc_mr_page_info(mr, ib_umem_num_pages(umem)); + if (err) + goto err2; + err = rxe_mr_fill_pages_from_sgt(mr, &umem->sgt_append.sgt); - if (err) { - ib_umem_release(umem); - return err; - } + if (err) + goto err1; mr->umem = umem; mr->ibmr.type = IB_MR_TYPE_USER; mr->state = RXE_MR_STATE_VALID; return 0; -} - -static int rxe_mr_alloc(struct rxe_mr *mr, int num_buf) -{ - XA_STATE(xas, &mr->page_list, 0); - int i = 0; - int err; - - xa_init(&mr->page_list); - - do { - xas_lock(&xas); - while (i != num_buf) { - xas_store(&xas, XA_ZERO_ENTRY); - if (xas_error(&xas)) - break; - xas_next(&xas); - i++; - } - xas_unlock(&xas); - } while (xas_nomem(&xas, GFP_KERNEL)); - - err = xas_error(&xas); - if (err) - return err; - - mr->num_buf = num_buf; - - return 0; +err1: + free_mr_page_info(mr); +err2: + ib_umem_release(umem); + return err; } int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr) @@ -192,7 +232,7 @@ int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr) /* always allow remote access for FMRs */ rxe_mr_init(RXE_ACCESS_REMOTE, mr); - err = rxe_mr_alloc(mr, max_pages); + err = alloc_mr_page_info(mr, max_pages); if (err) goto err1; @@ -205,26 +245,43 @@ int rxe_mr_init_fast(int max_pages, struct rxe_mr *mr) return err; } +/* + * I) MRs with page_size >= PAGE_SIZE, + * Split a large MR page (mr->page_size) into multiple PAGE_SIZE + * sub-pages and store them in page_info, offset is always 0. + * + * Called when mr->page_size > PAGE_SIZE. Each call to rxe_set_page() + * represents one mr->page_size region, which we must split into + * (mr->page_size >> PAGE_SHIFT) individual pages. + * + * II) MRs with page_size < PAGE_SIZE, + * Save each PAGE_SIZE page and its offset within the system page in page_info. + */ static int rxe_set_page(struct ib_mr *ibmr, u64 dma_addr) { struct rxe_mr *mr = to_rmr(ibmr); - struct page *page = ib_virt_dma_to_page(dma_addr); bool persistent = !!(mr->access & IB_ACCESS_FLUSH_PERSISTENT); - int err; + u32 i, pages_per_mr = mr_page_size(mr) >> PAGE_SHIFT; - if (persistent && !is_pmem_page(page)) { - rxe_dbg_mr(mr, "Page cannot be persistent\n"); - return -EINVAL; - } + pages_per_mr = MAX(1, pages_per_mr); - if (unlikely(mr->nbuf == mr->num_buf)) - return -ENOMEM; + for (i = 0; i < pages_per_mr; i++) { + u64 addr = dma_addr + i * PAGE_SIZE; + struct page *sub_page = ib_virt_dma_to_page(addr); - err = xa_err(xa_store(&mr->page_list, mr->nbuf, page, GFP_KERNEL)); - if (err) - return err; + if (unlikely(mr->nbuf >= mr->max_allowed_buf)) + return -ENOMEM; + + if (persistent && !is_pmem_page(sub_page)) { + rxe_dbg_mr(mr, "Page cannot be persistent\n"); + return -EINVAL; + } + + mr->page_info[mr->nbuf].page = sub_page; + mr->page_info[mr->nbuf].offset = addr & (PAGE_SIZE - 1); + mr->nbuf++; + } - mr->nbuf++; return 0; } @@ -234,6 +291,31 @@ int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sgl, struct rxe_mr *mr = to_rmr(ibmr); unsigned int page_size = mr_page_size(mr); + /* + * Ensure page_size and PAGE_SIZE are compatible for mapping. + * We require one to be a multiple of the other for correct + * iova-to-page conversion. + */ + if (!IS_ALIGNED(page_size, PAGE_SIZE) && + !IS_ALIGNED(PAGE_SIZE, page_size)) { + rxe_dbg_mr(mr, "MR page size %u must be compatible with PAGE_SIZE %lu\n", + page_size, PAGE_SIZE); + return -EINVAL; + } + + if (mr_page_size(mr) > PAGE_SIZE) { + /* resize page_info if needed */ + u32 map_mr_pages = (page_size >> PAGE_SHIFT) * mr->num_buf; + + if (map_mr_pages > mr->max_allowed_buf) { + rxe_dbg_mr(mr, "requested pages %u exceed max %u\n", + map_mr_pages, mr->max_allowed_buf); + free_mr_page_info(mr); + if (__alloc_mr_page_info(mr, map_mr_pages)) + return -ENOMEM; + } + } + mr->nbuf = 0; mr->page_shift = ilog2(page_size); mr->page_mask = ~((u64)page_size - 1); @@ -245,30 +327,30 @@ int rxe_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sgl, static int rxe_mr_copy_xarray(struct rxe_mr *mr, u64 iova, void *addr, unsigned int length, enum rxe_mr_copy_dir dir) { - unsigned int page_offset = rxe_mr_iova_to_page_offset(mr, iova); - unsigned long index = rxe_mr_iova_to_index(mr, iova); unsigned int bytes; - struct page *page; - void *va; + u8 *va; while (length) { - page = xa_load(&mr->page_list, index); - if (!page) + unsigned long index = rxe_mr_iova_to_index(mr, iova); + struct rxe_mr_page *info = &mr->page_info[index]; + unsigned int page_offset = rxe_mr_iova_to_page_offset(mr, iova); + + if (!info->page) return -EFAULT; - bytes = min_t(unsigned int, length, - mr_page_size(mr) - page_offset); - va = kmap_local_page(page); + page_offset += info->offset; + bytes = min_t(unsigned int, length, PAGE_SIZE - page_offset); + va = kmap_local_page(info->page); + if (dir == RXE_FROM_MR_OBJ) memcpy(addr, va + page_offset, bytes); else memcpy(va + page_offset, addr, bytes); kunmap_local(va); - page_offset = 0; addr += bytes; + iova += bytes; length -= bytes; - index++; } return 0; @@ -423,9 +505,6 @@ int copy_data( int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, unsigned int length) { - unsigned int page_offset; - unsigned long index; - struct page *page; unsigned int bytes; int err; u8 *va; @@ -445,15 +524,16 @@ int rxe_flush_pmem_iova(struct rxe_mr *mr, u64 iova, unsigned int length) return err; while (length > 0) { - index = rxe_mr_iova_to_index(mr, iova); - page = xa_load(&mr->page_list, index); - page_offset = rxe_mr_iova_to_page_offset(mr, iova); - if (!page) + unsigned long index = rxe_mr_iova_to_index(mr, iova); + struct rxe_mr_page *info = &mr->page_info[index]; + unsigned int page_offset = rxe_mr_iova_to_page_offset(mr, iova); + + if (!info->page) return -EFAULT; - bytes = min_t(unsigned int, length, - mr_page_size(mr) - page_offset); + page_offset += info->offset; + bytes = min_t(unsigned int, length, PAGE_SIZE - page_offset); - va = kmap_local_page(page); + va = kmap_local_page(info->page); arch_wb_cache_pmem(va + page_offset, bytes); kunmap_local(va); @@ -487,6 +567,7 @@ int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, } else { unsigned long index; int err; + struct rxe_mr_page *info; err = mr_check_range(mr, iova, sizeof(value)); if (err) { @@ -495,9 +576,12 @@ int rxe_mr_do_atomic_op(struct rxe_mr *mr, u64 iova, int opcode, } page_offset = rxe_mr_iova_to_page_offset(mr, iova); index = rxe_mr_iova_to_index(mr, iova); - page = xa_load(&mr->page_list, index); - if (!page) + info = &mr->page_info[index]; + if (!info->page) return RESPST_ERR_RKEY_VIOLATION; + + page_offset += info->offset; + page = info->page; } if (unlikely(page_offset & 0x7)) { @@ -544,6 +628,7 @@ int rxe_mr_do_atomic_write(struct rxe_mr *mr, u64 iova, u64 value) } else { unsigned long index; int err; + struct rxe_mr_page *info; /* See IBA oA19-28 */ err = mr_check_range(mr, iova, sizeof(value)); @@ -553,9 +638,12 @@ int rxe_mr_do_atomic_write(struct rxe_mr *mr, u64 iova, u64 value) } page_offset = rxe_mr_iova_to_page_offset(mr, iova); index = rxe_mr_iova_to_index(mr, iova); - page = xa_load(&mr->page_list, index); - if (!page) + info = &mr->page_info[index]; + if (!info->page) return RESPST_ERR_RKEY_VIOLATION; + + page_offset += info->offset; + page = info->page; } /* See IBA A19.4.2 */ @@ -727,5 +815,5 @@ void rxe_mr_cleanup(struct rxe_pool_elem *elem) ib_umem_release(mr->umem); if (mr->ibmr.type != IB_MR_TYPE_DMA) - xa_destroy(&mr->page_list); + free_mr_page_info(mr); } diff --git a/drivers/infiniband/sw/rxe/rxe_verbs.h b/drivers/infiniband/sw/rxe/rxe_verbs.h index 4242b24a47da..3b26ae016abe 100644 --- a/drivers/infiniband/sw/rxe/rxe_verbs.h +++ b/drivers/infiniband/sw/rxe/rxe_verbs.h @@ -297,6 +297,11 @@ static inline int rkey_is_mw(u32 rkey) return (index >= RXE_MIN_MW_INDEX) && (index <= RXE_MAX_MW_INDEX); } +struct rxe_mr_page { + struct page *page; + unsigned int offset; /* offset in system page */ +}; + struct rxe_mr { struct rxe_pool_elem elem; struct ib_mr ibmr; @@ -313,10 +318,13 @@ struct rxe_mr { unsigned int page_shift; u64 page_mask; + /* size of page_info when mr allocated */ u32 num_buf; + /* real size of page_info */ + u32 max_allowed_buf; u32 nbuf; - struct xarray page_list; + struct rxe_mr_page *page_info; }; static inline unsigned int mr_page_size(struct rxe_mr *mr) -- 2.25.1

2 1