- Kernel - mailweb.openeuler.org

[PATCH OLK-5.10 v5 00/21] xfs: atomic writes for xfs
by Long Li 14 Jun '24

14 Jun '24

THis patch set support xfs atomic writes feature. Alan Adamson (1): nvme: Atomic write support Darrick J. Wong (3): fs: xfs: Introduce FORCEALIGN inode flag fs: xfs: Enable file data forcealign feature fs: xfs: Make file data allocations observe the 'forcealign' flag John Garry (11): block: Add core atomic write support fs: xfs: Do not free EOF blocks for forcealign fs: iomap: Sub-extent zeroing fs: xfs: iomap: Sub-extent zeroing fs: Add FS_XFLAG_ATOMICWRITES flag fs: iomap: Atomic write support fs: xfs: Support FS_XFLAG_ATOMICWRITES for forcealign fs: xfs: Validate atomic writes fs: xfs: Support setting FMODE_CAN_ATOMIC_WRITE xfs: Update xfs_is_falloc_aligned() mask for forcealign xfs: Only free full extents for forcealign Long Li (5): block: get atomic write max sectors xfs: support atomic write ioctl xfs: fix set xflags fail when inode has extent hit xfs: make bunmapi observe forcealigin block: fix kabi in struct queue_limits Prasad Singamsetty (1): fs: Initial atomic write support Documentation/ABI/testing/sysfs-block | 52 ++++++++++++++ block/blk-core.c | 39 ++++++++++- block/blk-merge.c | 95 +++++++++++++++++++++++++- block/blk-mq-debugfs.c | 1 + block/blk-settings.c | 57 ++++++++++++++++ block/blk-sysfs.c | 34 ++++++++++ block/blk.h | 1 + drivers/md/dm-table.c | 5 +- drivers/md/dm.c | 4 +- drivers/nvme/host/core.c | 56 +++++++++++++++ fs/aio.c | 8 +-- fs/iomap/direct-io.c | 32 +++++++-- fs/read_write.c | 2 +- fs/xfs/libxfs/xfs_bmap.c | 24 ++++++- fs/xfs/libxfs/xfs_format.h | 15 +++- fs/xfs/libxfs/xfs_inode_buf.c | 40 +++++++++++ fs/xfs/libxfs/xfs_inode_buf.h | 3 + fs/xfs/libxfs/xfs_sb.c | 4 ++ fs/xfs/xfs_bmap_util.c | 10 ++- fs/xfs/xfs_file.c | 38 ++++++++++- fs/xfs/xfs_inode.c | 30 ++++++++ fs/xfs/xfs_inode.h | 11 +++ fs/xfs/xfs_ioctl.c | 98 ++++++++++++++++++++++++++- fs/xfs/xfs_iomap.c | 19 +++++- fs/xfs/xfs_mount.h | 5 ++ fs/xfs/xfs_super.c | 9 +++ include/linux/blk_types.h | 7 ++ include/linux/blkdev.h | 75 +++++++++++++++++++- include/linux/fs.h | 30 +++++++- include/linux/iomap.h | 1 + include/uapi/linux/fs.h | 9 ++- io_uring/io_uring.c | 2 +- 32 files changed, 776 insertions(+), 40 deletions(-) -- 2.31.1

2 22

[PATCH OLK-5.10] Bluetooth: L2CAP: Fix not validating setsockopt user input
by Kaixiong Yu 14 Jun '24

14 Jun '24

From: Luiz Augusto von Dentz <luiz.von.dentz(a)intel.com> stable inclusion from stable-v6.8.7 commit 8ee0c132a61df9723813c40e742dc5321824daa9 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I9QRQE CVE: CVE-2024-35965 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 4f3951242ace5efc7131932e2e01e6ac6baed846 ] Check user input length before copying data. Fixes: 33575df7be67 ("Bluetooth: move l2cap_sock_setsockopt() to l2cap_sock.c") Fixes: 3ee7b7cd8390 ("Bluetooth: Add BT_MODE socket option") Signed-off-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz(a)intel.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: net/bluetooth/l2cap_sock.c [The conflict occurs because the commit 54db3630deff566224de6cfb0767d2d398e68ed5 ("Bluetooth: Remove BT_HS"), which is from stable inclusion-v6.8.2, is not merged] Signed-off-by: Kaixiong Yu <yukaixiong(a)huawei.com> --- net/bluetooth/l2cap_sock.c | 54 +++++++++++++++----------------------- 1 file changed, 21 insertions(+), 33 deletions(-) diff --git a/net/bluetooth/l2cap_sock.c b/net/bluetooth/l2cap_sock.c index 756523e5402a..13afdc599b58 100644 --- a/net/bluetooth/l2cap_sock.c +++ b/net/bluetooth/l2cap_sock.c @@ -744,7 +744,7 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, struct sock *sk = sock->sk; struct l2cap_chan *chan = l2cap_pi(sk)->chan; struct l2cap_options opts; - int len, err = 0; + int err = 0; u32 opt; BT_DBG("sk %p", sk); @@ -771,11 +771,9 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, opts.max_tx = chan->max_tx; opts.txwin_size = chan->tx_win; - len = min_t(unsigned int, sizeof(opts), optlen); - if (copy_from_sockptr(&opts, optval, len)) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opts, sizeof(opts), optval, optlen); + if (err) break; - } if (opts.txwin_size > L2CAP_DEFAULT_EXT_WINDOW) { err = -EINVAL; @@ -818,10 +816,9 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, break; case L2CAP_LM: - if (copy_from_sockptr(&opt, optval, sizeof(u32))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) break; - } if (opt & L2CAP_LM_FIPS) { err = -EINVAL; @@ -902,7 +899,7 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, struct bt_security sec; struct bt_power pwr; struct l2cap_conn *conn; - int len, err = 0; + int err = 0; u32 opt; u16 mtu; u8 mode; @@ -928,11 +925,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, sec.level = BT_SECURITY_LOW; - len = min_t(unsigned int, sizeof(sec), optlen); - if (copy_from_sockptr(&sec, optval, len)) { - err = -EFAULT; + err = bt_copy_from_sockptr(&sec, sizeof(sec), optval, optlen); + if (err) break; - } if (sec.level < BT_SECURITY_LOW || sec.level > BT_SECURITY_FIPS) { @@ -977,10 +972,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; } - if (copy_from_sockptr(&opt, optval, sizeof(u32))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) break; - } if (opt) { set_bit(BT_SK_DEFER_SETUP, &bt_sk(sk)->flags); @@ -992,10 +986,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; case BT_FLUSHABLE: - if (copy_from_sockptr(&opt, optval, sizeof(u32))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) break; - } if (opt > BT_FLUSHABLE_ON) { err = -EINVAL; @@ -1027,11 +1020,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, pwr.force_active = BT_POWER_FORCE_ACTIVE_ON; - len = min_t(unsigned int, sizeof(pwr), optlen); - if (copy_from_sockptr(&pwr, optval, len)) { - err = -EFAULT; + err = bt_copy_from_sockptr(&pwr, sizeof(pwr), optval, optlen); + if (err) break; - } if (pwr.force_active) set_bit(FLAG_FORCE_ACTIVE, &chan->flags); @@ -1040,10 +1031,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; case BT_CHANNEL_POLICY: - if (copy_from_sockptr(&opt, optval, sizeof(u32))) { - err = -EFAULT; - break; - } + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) + break; if (opt > BT_CHANNEL_POLICY_AMP_PREFERRED) { err = -EINVAL; @@ -1088,10 +1078,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; } - if (copy_from_sockptr(&mtu, optval, sizeof(u16))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&mtu, sizeof(mtu), optval, optlen); + if (err) break; - } if (chan->mode == L2CAP_MODE_EXT_FLOWCTL && sk->sk_state == BT_CONNECTED) @@ -1119,10 +1108,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; } - if (copy_from_sockptr(&mode, optval, sizeof(u8))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&mode, sizeof(mode), optval, optlen); + if (err) break; - } BT_DBG("mode %u", mode); -- 2.25.1

2 1

[PATCH v2 OLK-5.10] arm64: mm: Replace global variable in pbha with static key
by Wupeng Ma 14 Jun '24

14 Jun '24

From: Ma Wupeng <mawupeng1(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/IA5L01 -------------------------------- During the lmbench test, it was observed that bw_mem experienced a degradation of 8% or more. To solve this issue, replacing the global variables with a static key. Fixes: 0d65528c8526 ("arm64: mm: Detect and enable PBHA bit0 at early startup") Signed-off-by: Ma Wupeng <mawupeng1(a)huawei.com> --- Changelog since v1: - add fix tag for bugfix --- drivers/soc/hisilicon/pbha.c | 17 +++++++++-------- include/linux/pbha.h | 13 ++++++++----- 2 files changed, 17 insertions(+), 13 deletions(-) diff --git a/drivers/soc/hisilicon/pbha.c b/drivers/soc/hisilicon/pbha.c index 57daa5c338b5..9e385c93bab5 100644 --- a/drivers/soc/hisilicon/pbha.c +++ b/drivers/soc/hisilicon/pbha.c @@ -17,8 +17,9 @@ #define HBM_MODE_CACHE 1 -bool __ro_after_init pbha_bit0_enabled; -bool __ro_after_init pbha_bit0_kernel_enabled; +__ro_after_init DEFINE_STATIC_KEY_FALSE(pbha_bit0_enabled); +__ro_after_init DEFINE_STATIC_KEY_FALSE(pbha_bit0_kernel_enabled); + static bool pbha_enabled_phase_1; void __init early_pbha_bit0_init(void) @@ -191,15 +192,15 @@ static int __init setup_pbha(char *str) return 0; if (strcmp(str, "enable") == 0) { - pbha_bit0_enabled = true; - pbha_bit0_kernel_enabled = true; + static_branch_enable(&pbha_bit0_enabled); + static_branch_enable(&pbha_bit0_kernel_enabled); } else if (strcmp(str, "user") == 0) { - pbha_bit0_enabled = true; + static_branch_enable(&pbha_bit0_enabled); } - if (pbha_bit0_enabled) - pr_info("pbha bit_0 enabled, kernel: %d\n", - pbha_bit0_kernel_enabled); + if (static_branch_likely(&pbha_bit0_enabled)) + pr_info("pbha bit_0 enabled, kernel: %ld\n", + static_branch_likely(&pbha_bit0_kernel_enabled)); return 0; } diff --git a/include/linux/pbha.h b/include/linux/pbha.h index 45a4b07778bf..a8ac441628bd 100644 --- a/include/linux/pbha.h +++ b/include/linux/pbha.h @@ -20,20 +20,22 @@ #define HBM_MODE_CACHE 1 #ifdef CONFIG_ARM64_PBHA -extern bool __ro_after_init pbha_bit0_enabled; -extern bool __ro_after_init pbha_bit0_kernel_enabled; +DECLARE_STATIC_KEY_FALSE(pbha_bit0_enabled); +DECLARE_STATIC_KEY_FALSE(pbha_bit0_kernel_enabled); + extern struct mm_walk_ops pbha_bit0_walk_ops; extern void __init early_pbha_bit0_init(void); extern int pbha_bit0_update_vma(struct mm_struct *mm, int val); static inline bool system_support_pbha_bit0(void) { - return pbha_bit0_enabled; + return static_branch_unlikely(&pbha_bit0_enabled); } static inline pgprot_t pgprot_pbha_bit0(pgprot_t prot) { - if (!system_support_pbha_bit0() || !pbha_bit0_kernel_enabled) + if (!system_support_pbha_bit0() || + !static_branch_unlikely(&pbha_bit0_kernel_enabled)) return prot; return pgprot_pbha(prot, PBHA_VAL_BIT0); @@ -45,7 +47,8 @@ static inline pte_t maybe_mk_pbha_bit0(pte_t pte, struct vm_area_struct *vma) return pte; /* global init task will update pbha bit0 iff kernel can do this */ - if (unlikely(is_global_init(current)) && pbha_bit0_kernel_enabled && + if (unlikely(is_global_init(current)) && + static_branch_unlikely(&pbha_bit0_kernel_enabled) && !(vma->vm_flags & VM_PBHA_BIT0)) vma->vm_flags |= VM_PBHA_BIT0; -- 2.25.1

2 1

[PATCH v2 OLK-5.10] arm64: mm: Replace global variable in pbha with static key
by Wupeng Ma 14 Jun '24

14 Jun '24

From: Ma Wupeng <mawupeng1(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/IA5L01 -------------------------------- During the lmbench test, it was observed that bw_mem experienced a degradation of 8% or more. To solve this issue, replacing the global variables with a static key. Fixes: 0d65528c8526 ("arm64: mm: Detect and enable PBHA bit0 at early startup") Signed-off-by: Ma Wupeng <mawupeng1(a)huawei.com> --- Changelog since v1: - add fix tag for bugfix --- drivers/soc/hisilicon/pbha.c | 17 +++++++++-------- include/linux/pbha.h | 13 ++++++++----- 2 files changed, 17 insertions(+), 13 deletions(-) diff --git a/drivers/soc/hisilicon/pbha.c b/drivers/soc/hisilicon/pbha.c index 57daa5c338b5..9e385c93bab5 100644 --- a/drivers/soc/hisilicon/pbha.c +++ b/drivers/soc/hisilicon/pbha.c @@ -17,8 +17,9 @@ #define HBM_MODE_CACHE 1 -bool __ro_after_init pbha_bit0_enabled; -bool __ro_after_init pbha_bit0_kernel_enabled; +__ro_after_init DEFINE_STATIC_KEY_FALSE(pbha_bit0_enabled); +__ro_after_init DEFINE_STATIC_KEY_FALSE(pbha_bit0_kernel_enabled); + static bool pbha_enabled_phase_1; void __init early_pbha_bit0_init(void) @@ -191,15 +192,15 @@ static int __init setup_pbha(char *str) return 0; if (strcmp(str, "enable") == 0) { - pbha_bit0_enabled = true; - pbha_bit0_kernel_enabled = true; + static_branch_enable(&pbha_bit0_enabled); + static_branch_enable(&pbha_bit0_kernel_enabled); } else if (strcmp(str, "user") == 0) { - pbha_bit0_enabled = true; + static_branch_enable(&pbha_bit0_enabled); } - if (pbha_bit0_enabled) - pr_info("pbha bit_0 enabled, kernel: %d\n", - pbha_bit0_kernel_enabled); + if (static_branch_likely(&pbha_bit0_enabled)) + pr_info("pbha bit_0 enabled, kernel: %ld\n", + static_branch_likely(&pbha_bit0_kernel_enabled)); return 0; } diff --git a/include/linux/pbha.h b/include/linux/pbha.h index 45a4b07778bf..a8ac441628bd 100644 --- a/include/linux/pbha.h +++ b/include/linux/pbha.h @@ -20,20 +20,22 @@ #define HBM_MODE_CACHE 1 #ifdef CONFIG_ARM64_PBHA -extern bool __ro_after_init pbha_bit0_enabled; -extern bool __ro_after_init pbha_bit0_kernel_enabled; +DECLARE_STATIC_KEY_FALSE(pbha_bit0_enabled); +DECLARE_STATIC_KEY_FALSE(pbha_bit0_kernel_enabled); + extern struct mm_walk_ops pbha_bit0_walk_ops; extern void __init early_pbha_bit0_init(void); extern int pbha_bit0_update_vma(struct mm_struct *mm, int val); static inline bool system_support_pbha_bit0(void) { - return pbha_bit0_enabled; + return static_branch_unlikely(&pbha_bit0_enabled); } static inline pgprot_t pgprot_pbha_bit0(pgprot_t prot) { - if (!system_support_pbha_bit0() || !pbha_bit0_kernel_enabled) + if (!system_support_pbha_bit0() || + !static_branch_unlikely(&pbha_bit0_kernel_enabled)) return prot; return pgprot_pbha(prot, PBHA_VAL_BIT0); @@ -45,7 +47,8 @@ static inline pte_t maybe_mk_pbha_bit0(pte_t pte, struct vm_area_struct *vma) return pte; /* global init task will update pbha bit0 iff kernel can do this */ - if (unlikely(is_global_init(current)) && pbha_bit0_kernel_enabled && + if (unlikely(is_global_init(current)) && + static_branch_unlikely(&pbha_bit0_kernel_enabled) && !(vma->vm_flags & VM_PBHA_BIT0)) vma->vm_flags |= VM_PBHA_BIT0; -- 2.25.1

2 1

[PATCH openEuler-22.03-LTS-SP1] Bluetooth: L2CAP: Fix not validating setsockopt user input
by Kaixiong Yu 14 Jun '24

14 Jun '24

From: Luiz Augusto von Dentz <luiz.von.dentz(a)intel.com> stable inclusion from stable-v6.8.7 commit 8ee0c132a61df9723813c40e742dc5321824daa9 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I9QRQE CVE: CVE-2024-35965 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit 4f3951242ace5efc7131932e2e01e6ac6baed846 ] Check user input length before copying data. Fixes: 33575df7be67 ("Bluetooth: move l2cap_sock_setsockopt() to l2cap_sock.c") Fixes: 3ee7b7cd8390 ("Bluetooth: Add BT_MODE socket option") Signed-off-by: Eric Dumazet <edumazet(a)google.com> Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz(a)intel.com> Signed-off-by: Sasha Levin <sashal(a)kernel.org> Conflicts: net/bluetooth/l2cap_sock.c [The conflict occurs because the commit 54db3630deff566224de6cfb0767d2d398e68ed5 ("Bluetooth: Remove BT_HS"), which is from stable inclusion-v6.8.2, is not merged.] Signed-off-by: Kaixiong Yu <yukaixiong(a)huawei.com> --- net/bluetooth/l2cap_sock.c | 52 +++++++++++++++----------------------- 1 file changed, 20 insertions(+), 32 deletions(-) diff --git a/net/bluetooth/l2cap_sock.c b/net/bluetooth/l2cap_sock.c index 63bfbb5918fe..c91f4c190bbc 100644 --- a/net/bluetooth/l2cap_sock.c +++ b/net/bluetooth/l2cap_sock.c @@ -744,7 +744,7 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, struct sock *sk = sock->sk; struct l2cap_chan *chan = l2cap_pi(sk)->chan; struct l2cap_options opts; - int len, err = 0; + int err = 0; u32 opt; BT_DBG("sk %p", sk); @@ -771,11 +771,9 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, opts.max_tx = chan->max_tx; opts.txwin_size = chan->tx_win; - len = min_t(unsigned int, sizeof(opts), optlen); - if (copy_from_sockptr(&opts, optval, len)) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opts, sizeof(opts), optval, optlen); + if (err) break; - } if (opts.txwin_size > L2CAP_DEFAULT_EXT_WINDOW) { err = -EINVAL; @@ -818,10 +816,9 @@ static int l2cap_sock_setsockopt_old(struct socket *sock, int optname, break; case L2CAP_LM: - if (copy_from_sockptr(&opt, optval, sizeof(u32))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) break; - } if (opt & L2CAP_LM_FIPS) { err = -EINVAL; @@ -902,7 +899,7 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, struct bt_security sec; struct bt_power pwr; struct l2cap_conn *conn; - int len, err = 0; + int err = 0; u32 opt; u16 mtu; u8 mode; @@ -928,11 +925,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, sec.level = BT_SECURITY_LOW; - len = min_t(unsigned int, sizeof(sec), optlen); - if (copy_from_sockptr(&sec, optval, len)) { - err = -EFAULT; + err = bt_copy_from_sockptr(&sec, sizeof(sec), optval, optlen); + if (err) break; - } if (sec.level < BT_SECURITY_LOW || sec.level > BT_SECURITY_FIPS) { @@ -977,10 +972,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; } - if (copy_from_sockptr(&opt, optval, sizeof(u32))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) break; - } if (opt) { set_bit(BT_SK_DEFER_SETUP, &bt_sk(sk)->flags); @@ -992,10 +986,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; case BT_FLUSHABLE: - if (copy_from_sockptr(&opt, optval, sizeof(u32))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) break; - } if (opt > BT_FLUSHABLE_ON) { err = -EINVAL; @@ -1027,11 +1020,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, pwr.force_active = BT_POWER_FORCE_ACTIVE_ON; - len = min_t(unsigned int, sizeof(pwr), optlen); - if (copy_from_sockptr(&pwr, optval, len)) { - err = -EFAULT; + err = bt_copy_from_sockptr(&pwr, sizeof(pwr), optval, optlen); + if (err) break; - } if (pwr.force_active) set_bit(FLAG_FORCE_ACTIVE, &chan->flags); @@ -1065,10 +1056,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; case BT_SNDMTU: - if (!bdaddr_type_is_le(chan->src_type)) { - err = -EINVAL; + err = bt_copy_from_sockptr(&opt, sizeof(opt), optval, optlen); + if (err) break; - } /* Setting is not supported as it's the remote side that * decides this. @@ -1088,10 +1078,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; } - if (copy_from_sockptr(&mtu, optval, sizeof(u16))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&mtu, sizeof(mtu), optval, optlen); + if (err) break; - } if (chan->mode == L2CAP_MODE_EXT_FLOWCTL && sk->sk_state == BT_CONNECTED) @@ -1119,10 +1108,9 @@ static int l2cap_sock_setsockopt(struct socket *sock, int level, int optname, break; } - if (copy_from_sockptr(&mode, optval, sizeof(u8))) { - err = -EFAULT; + err = bt_copy_from_sockptr(&mode, sizeof(mode), optval, optlen); + if (err) break; - } BT_DBG("mode %u", mode); -- 2.25.1

2 1

[PATCH OLK-5.10] arm64: mm: Replace global variable in pbha with static key
by Wupeng Ma 14 Jun '24

14 Jun '24

From: Ma Wupeng <mawupeng1(a)huawei.com> hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/IA5L01 -------------------------------- During the lmbench test, it was observed that bw_mem experienced a degradation of 8% or more. To solve this issue, replacing the global variables with a static key. Signed-off-by: Ma Wupeng <mawupeng1(a)huawei.com> --- drivers/soc/hisilicon/pbha.c | 17 +++++++++-------- include/linux/pbha.h | 13 ++++++++----- 2 files changed, 17 insertions(+), 13 deletions(-) diff --git a/drivers/soc/hisilicon/pbha.c b/drivers/soc/hisilicon/pbha.c index 57daa5c338b5..9e385c93bab5 100644 --- a/drivers/soc/hisilicon/pbha.c +++ b/drivers/soc/hisilicon/pbha.c @@ -17,8 +17,9 @@ #define HBM_MODE_CACHE 1 -bool __ro_after_init pbha_bit0_enabled; -bool __ro_after_init pbha_bit0_kernel_enabled; +__ro_after_init DEFINE_STATIC_KEY_FALSE(pbha_bit0_enabled); +__ro_after_init DEFINE_STATIC_KEY_FALSE(pbha_bit0_kernel_enabled); + static bool pbha_enabled_phase_1; void __init early_pbha_bit0_init(void) @@ -191,15 +192,15 @@ static int __init setup_pbha(char *str) return 0; if (strcmp(str, "enable") == 0) { - pbha_bit0_enabled = true; - pbha_bit0_kernel_enabled = true; + static_branch_enable(&pbha_bit0_enabled); + static_branch_enable(&pbha_bit0_kernel_enabled); } else if (strcmp(str, "user") == 0) { - pbha_bit0_enabled = true; + static_branch_enable(&pbha_bit0_enabled); } - if (pbha_bit0_enabled) - pr_info("pbha bit_0 enabled, kernel: %d\n", - pbha_bit0_kernel_enabled); + if (static_branch_likely(&pbha_bit0_enabled)) + pr_info("pbha bit_0 enabled, kernel: %ld\n", + static_branch_likely(&pbha_bit0_kernel_enabled)); return 0; } diff --git a/include/linux/pbha.h b/include/linux/pbha.h index 45a4b07778bf..a8ac441628bd 100644 --- a/include/linux/pbha.h +++ b/include/linux/pbha.h @@ -20,20 +20,22 @@ #define HBM_MODE_CACHE 1 #ifdef CONFIG_ARM64_PBHA -extern bool __ro_after_init pbha_bit0_enabled; -extern bool __ro_after_init pbha_bit0_kernel_enabled; +DECLARE_STATIC_KEY_FALSE(pbha_bit0_enabled); +DECLARE_STATIC_KEY_FALSE(pbha_bit0_kernel_enabled); + extern struct mm_walk_ops pbha_bit0_walk_ops; extern void __init early_pbha_bit0_init(void); extern int pbha_bit0_update_vma(struct mm_struct *mm, int val); static inline bool system_support_pbha_bit0(void) { - return pbha_bit0_enabled; + return static_branch_unlikely(&pbha_bit0_enabled); } static inline pgprot_t pgprot_pbha_bit0(pgprot_t prot) { - if (!system_support_pbha_bit0() || !pbha_bit0_kernel_enabled) + if (!system_support_pbha_bit0() || + !static_branch_unlikely(&pbha_bit0_kernel_enabled)) return prot; return pgprot_pbha(prot, PBHA_VAL_BIT0); @@ -45,7 +47,8 @@ static inline pte_t maybe_mk_pbha_bit0(pte_t pte, struct vm_area_struct *vma) return pte; /* global init task will update pbha bit0 iff kernel can do this */ - if (unlikely(is_global_init(current)) && pbha_bit0_kernel_enabled && + if (unlikely(is_global_init(current)) && + static_branch_unlikely(&pbha_bit0_kernel_enabled) && !(vma->vm_flags & VM_PBHA_BIT0)) vma->vm_flags |= VM_PBHA_BIT0; -- 2.25.1

2 1

[PATCH OLK-5.10 v4 00/26] xfs: atomic writes for xfs
by Long Li 14 Jun '24

14 Jun '24

THis patch set support xfs atomic writes feature. Alan Adamson (1): nvme: Atomic write support Darrick J. Wong (3): fs: xfs: Introduce FORCEALIGN inode flag fs: xfs: Enable file data forcealign feature fs: xfs: Make file data allocations observe the 'forcealign' flag John Garry (11): block: Add core atomic write support fs: xfs: Do not free EOF blocks for forcealign fs: iomap: Sub-extent zeroing fs: xfs: iomap: Sub-extent zeroing fs: Add FS_XFLAG_ATOMICWRITES flag fs: iomap: Atomic write support fs: xfs: Support FS_XFLAG_ATOMICWRITES for forcealign fs: xfs: Validate atomic writes fs: xfs: Support setting FMODE_CAN_ATOMIC_WRITE xfs: Update xfs_is_falloc_aligned() mask for forcealign xfs: Only free full extents for forcealign Long Li (6): block: get atomic write max sectors xfs: support atomic write ioctl xfs: atomic write file dio convert to mark IOCB_ATOMIC xfs: fix set xflags fail when inode has extent hit xfs: make bunmapi observe forcealigin block: fix kabi in struct queue_limits Prasad Singamsetty (1): fs: Initial atomic write support Zhang Yi (4): math64: add rem_u64() to just return the remainder iomap: pass blocksize to iomap_truncate_page() xfs: refactor the truncating order xfs: correct the truncate blocksize of realtime inode Documentation/ABI/testing/sysfs-block | 52 ++++++++++++ block/blk-core.c | 39 ++++++++- block/blk-merge.c | 95 ++++++++++++++++++++- block/blk-mq-debugfs.c | 1 + block/blk-settings.c | 57 +++++++++++++ block/blk-sysfs.c | 34 ++++++++ block/blk.h | 1 + drivers/md/dm-table.c | 5 +- drivers/md/dm.c | 4 +- drivers/nvme/host/core.c | 56 ++++++++++++ fs/aio.c | 8 +- fs/iomap/buffered-io.c | 8 +- fs/iomap/direct-io.c | 32 +++++-- fs/read_write.c | 2 +- fs/xfs/libxfs/xfs_bmap.c | 24 +++++- fs/xfs/libxfs/xfs_format.h | 15 +++- fs/xfs/libxfs/xfs_inode_buf.c | 40 +++++++++ fs/xfs/libxfs/xfs_inode_buf.h | 3 + fs/xfs/libxfs/xfs_sb.c | 4 + fs/xfs/xfs_bmap_util.c | 10 ++- fs/xfs/xfs_file.c | 40 ++++++++- fs/xfs/xfs_inode.c | 43 ++++++++++ fs/xfs/xfs_inode.h | 12 +++ fs/xfs/xfs_ioctl.c | 96 ++++++++++++++++++++- fs/xfs/xfs_iomap.c | 19 ++++- fs/xfs/xfs_iops.c | 118 +++++++++++++------------- fs/xfs/xfs_mount.h | 5 ++ fs/xfs/xfs_super.c | 9 ++ include/linux/blk_types.h | 7 ++ include/linux/blkdev.h | 77 ++++++++++++++++- include/linux/fs.h | 30 ++++++- include/linux/iomap.h | 5 +- include/linux/math64.h | 24 ++++++ include/uapi/linux/fs.h | 9 +- io_uring/io_uring.c | 2 +- 35 files changed, 884 insertions(+), 102 deletions(-) -- 2.31.1

2 27

[PATCH openEuler-1.0-LTS] dhugetlb: use rcu lock to protect hpool
by Liu Shixin 14 Jun '24

14 Jun '24

hulk inclusion category: bugfix bugzilla: 46904, https://gitee.com/openeuler/kernel/issues/IA5JXN -------------------------------- In freeing process, set memcg->hpool or dpool_global_hpool to NULL and then freeing it. To avoid UAF problem of hpool, we have to make sure the users that already hold pointer to hpool don't use the pointer after freeing hpool. The freeing of hpool should block until make sure all such users don't use hpool. Since anyone who want to use hpool have to increase its refcount at first. Use rcu_read_lock() and synchronize_rcu() to guarantee all such users failed to get hpool before freeing hpool. Fixes: 0bc0d0d57eda ("dhugetlb: backport dynamic hugetlb feature") Signed-off-by: Liu Shixin <liushixin2(a)huawei.com> --- mm/hugetlb.c | 1 + mm/memcontrol.c | 11 ++++++++--- 2 files changed, 9 insertions(+), 3 deletions(-) diff --git a/mm/hugetlb.c b/mm/hugetlb.c index 4be708677ce6..07b961d0a5ff 100644 --- a/mm/hugetlb.c +++ b/mm/hugetlb.c @@ -3360,6 +3360,7 @@ void dhugetlb_pool_put(struct dhugetlb_pool *hpool) if (atomic_dec_and_test(&hpool->refcnt)) { css_put(&hpool->attach_memcg->css); + synchronize_rcu(); kfree(hpool); } } diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 7dc68b1792ef..4ecb6f09b68b 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -4701,14 +4701,19 @@ static ssize_t memcg_write_event_control(struct kernfs_open_file *of, struct dhugetlb_pool *get_dhugetlb_pool_from_memcg(struct mem_cgroup *memcg) { struct mem_cgroup_extension *memcg_ext; + struct dhugetlb_pool *hpool; if (!memcg) return NULL; + rcu_read_lock(); memcg_ext = container_of(memcg, struct mem_cgroup_extension, memcg); - if (dhugetlb_pool_get(memcg_ext->hpool)) - return memcg_ext->hpool; - return NULL; + hpool = memcg_ext->hpool; + if (!dhugetlb_pool_get(hpool)) + hpool = NULL; + rcu_read_unlock(); + + return hpool; } static void set_dhugetlb_pool_to_memcg(struct mem_cgroup *memcg, -- 2.25.1

2 1

[PATCH openEuler-1.0-LTS] RDMA/cma: Ensure rdma_addr_cancel() happens before issuing more requests
by Zheng Zucheng 14 Jun '24

14 Jun '24

From: Jason Gunthorpe <jgg(a)nvidia.com> mainline inclusion from mainline-v5.15-rc4 commit 305d568b72f17f674155a2a8275f865f207b3808 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/I9R4OB CVE: CVE-2021-47391 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… --------------------------- The FSM can run in a circle allowing rdma_resolve_ip() to be called twice on the same id_priv. While this cannot happen without going through the work, it violates the invariant that the same address resolution background request cannot be active twice. CPU 1 CPU 2 rdma_resolve_addr(): RDMA_CM_IDLE -> RDMA_CM_ADDR_QUERY rdma_resolve_ip(addr_handler) #1 process_one_req(): for #1 addr_handler(): RDMA_CM_ADDR_QUERY -> RDMA_CM_ADDR_BOUND mutex_unlock(&id_priv->handler_mutex); [.. handler still running ..] rdma_resolve_addr(): RDMA_CM_ADDR_BOUND -> RDMA_CM_ADDR_QUERY rdma_resolve_ip(addr_handler) !! two requests are now on the req_list rdma_destroy_id(): destroy_id_handler_unlock(): _destroy_id(): cma_cancel_operation(): rdma_addr_cancel() // process_one_req() self removes it spin_lock_bh(&lock); cancel_delayed_work(&req->work); if (!list_empty(&req->list)) == true ! rdma_addr_cancel() returns after process_on_req #1 is done kfree(id_priv) process_one_req(): for #2 addr_handler(): mutex_lock(&id_priv->handler_mutex); !! Use after free on id_priv rdma_addr_cancel() expects there to be one req on the list and only cancels the first one. The self-removal behavior of the work only happens after the handler has returned. This yields a situations where the req_list can have two reqs for the same "handle" but rdma_addr_cancel() only cancels the first one. The second req remains active beyond rdma_destroy_id() and will use-after-free id_priv once it inevitably triggers. Fix this by remembering if the id_priv has called rdma_resolve_ip() and always cancel before calling it again. This ensures the req_list never gets more than one item in it and doesn't cost anything in the normal flow that never uses this strange error path. Link: https://lore.kernel.org/r/0-v1-3bc675b8006d+22-syz_cancel_uaf_jgg@nvidia.com Cc: stable(a)vger.kernel.org Fixes: e51060f08a61 ("IB: IP address based RDMA connection manager") Reported-by: syzbot+dc3dfba010d7671e05f5(a)syzkaller.appspotmail.com Signed-off-by: Jason Gunthorpe <jgg(a)nvidia.com> (cherry picked from commit 305d568b72f17f674155a2a8275f865f207b3808) Signed-off-by: Liu Jian <liujian56(a)huawei.com> Conflicts: drivers/infiniband/core/cma.c drivers/infiniband/core/cma_priv.h [Some contexts different. No functional impact.] Signed-off-by: Zheng Zucheng <zhengzucheng(a)huawei.com> --- drivers/infiniband/core/cma.c | 23 +++++++++++++++++++++++ drivers/infiniband/core/cma_priv.h | 1 + 2 files changed, 24 insertions(+) diff --git a/drivers/infiniband/core/cma.c b/drivers/infiniband/core/cma.c index 20a5501b740a..90cb1255ca64 100644 --- a/drivers/infiniband/core/cma.c +++ b/drivers/infiniband/core/cma.c @@ -1643,6 +1643,14 @@ static void cma_cancel_operation(struct rdma_id_private *id_priv, { switch (state) { case RDMA_CM_ADDR_QUERY: + /* + * We can avoid doing the rdma_addr_cancel() based on state, + * only RDMA_CM_ADDR_QUERY has a work that could still execute. + * Notice that the addr_handler work could still be exiting + * outside this state, however due to the interaction with the + * handler_mutex the work is guaranteed not to touch id_priv + * during exit. + */ rdma_addr_cancel(&id_priv->id.route.addr.dev_addr); break; case RDMA_CM_ROUTE_QUERY: @@ -3459,6 +3467,21 @@ int rdma_resolve_addr(struct rdma_cm_id *id, struct sockaddr *src_addr, if (dst_addr->sa_family == AF_IB) { ret = cma_resolve_ib_addr(id_priv); } else { + /* + * The FSM can return back to RDMA_CM_ADDR_BOUND after + * rdma_resolve_ip() is called, eg through the error + * path in addr_handler(). If this happens the existing + * request must be canceled before issuing a new one. + * Since canceling a request is a bit slow and this + * oddball path is rare, keep track once a request has + * been issued. The track turns out to be a permanent + * state since this is the only cancel as it is + * immediately before rdma_resolve_ip(). + */ + if (id_priv->used_resolve_ip) + rdma_addr_cancel(&id->route.addr.dev_addr); + else + id_priv->used_resolve_ip = 1; ret = rdma_resolve_ip(cma_src_addr(id_priv), dst_addr, &id->route.addr.dev_addr, timeout_ms, addr_handler, id_priv); diff --git a/drivers/infiniband/core/cma_priv.h b/drivers/infiniband/core/cma_priv.h index 194cfe78c447..e2e0bc89d847 100644 --- a/drivers/infiniband/core/cma_priv.h +++ b/drivers/infiniband/core/cma_priv.h @@ -87,6 +87,7 @@ struct rdma_id_private { bool tos_set; u8 reuseaddr; u8 afonly; + u8 used_resolve_ip; enum ib_gid_type gid_type; /* -- 2.34.1

2 1

[PATCH openEuler-22.03-LTS-SP1 0/2] Fix UAF problem of hpool
by Liu Shixin 14 Jun '24

14 Jun '24

Add rcu lock protection to avoid get refcount of hpool after it has been freed. Liu Shixin (2): mm/dynamic_hugetlb: simplify the refcount code mm/dynamic_hugetlb: use rcu lock to protect hpool mm/dynamic_hugetlb.c | 128 +++++++++++++++++++++++++++---------------- 1 file changed, 82 insertions(+), 46 deletions(-) -- 2.25.1

2 3

2024

2023

2022

2021

2020

2019

Kernel