January 2025 - Kernel - mailweb.openeuler.org

[PATCH OLK-5.10] quota: flush quota_release_work upon quota writeback
by Baokun Li 15 Jan '25

15 Jan '25

From: Ojaswin Mujoo <ojaswin(a)linux.ibm.com> stable inclusion from stable-v5.10.231 commit 6f3821acd7c3143145999248087de5fb4b48cf26 category: bugfix bugzilla: https://gitee.com/src-openeuler/kernel/issues/IBHBWQ CVE: CVE-2024-56780 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- [ Upstream commit ac6f420291b3fee1113f21d612fa88b628afab5b ] One of the paths quota writeback is called from is: freeze_super() sync_filesystem() ext4_sync_fs() dquot_writeback_dquots() Since we currently don't always flush the quota_release_work queue in this path, we can end up with the following race: 1. dquot are added to releasing_dquots list during regular operations. 2. FS Freeze starts, however, this does not flush the quota_release_work queue. 3. Freeze completes. 4. Kernel eventually tries to flush the workqueue while FS is frozen which hits a WARN_ON since transaction gets started during frozen state: ext4_journal_check_start+0x28/0x110 [ext4] (unreliable) __ext4_journal_start_sb+0x64/0x1c0 [ext4] ext4_release_dquot+0x90/0x1d0 [ext4] quota_release_workfn+0x43c/0x4d0 Which is the following line: WARN_ON(sb->s_writers.frozen == SB_FREEZE_COMPLETE); Which ultimately results in generic/390 failing due to dmesg noise. This was detected on powerpc machine 15 cores. To avoid this, make sure to flush the workqueue during dquot_writeback_dquots() so we dont have any pending workitems after freeze. Reported-by: Disha Goel <disgoel(a)linux.ibm.com> CC: stable(a)vger.kernel.org Fixes: dabc8b207566 ("quota: fix dqput() to follow the guarantees dquot_srcu should provide") Reviewed-by: Baokun Li <libaokun1(a)huawei.com> Signed-off-by: Ojaswin Mujoo <ojaswin(a)linux.ibm.com> Signed-off-by: Jan Kara <jack(a)suse.cz> Link: https://patch.msgid.link/20241121123855.645335-2-ojaswin@linux.ibm.com Signed-off-by: Sasha Levin <sashal(a)kernel.org> Signed-off-by: Baokun Li <libaokun1(a)huawei.com> --- fs/quota/dquot.c | 2 ++ 1 file changed, 2 insertions(+) diff --git a/fs/quota/dquot.c b/fs/quota/dquot.c index 001482fac9fd..94fa70e38087 100644 --- a/fs/quota/dquot.c +++ b/fs/quota/dquot.c @@ -687,6 +687,8 @@ int dquot_writeback_dquots(struct super_block *sb, int type) WARN_ON_ONCE(!rwsem_is_locked(&sb->s_umount)); + flush_delayed_work(&quota_release_work); + for (cnt = 0; cnt < MAXQUOTAS; cnt++) { if (type != -1 && cnt != type) continue; -- 2.46.1

2 1

[PATCH OLK-6.6 0/1] CVE-2024-56724
by Liao Chen 15 Jan '25

15 Jan '25

fix CVE-2024-56724 Andy Shevchenko (1): mfd: intel_soc_pmic_bxtwc: Use IRQ domain for TMU device drivers/mfd/intel_soc_pmic_bxtwc.c | 31 ++++++++++++++------------ drivers/platform/x86/intel/bxtwc_tmu.c | 22 +++++------------- 2 files changed, 23 insertions(+), 30 deletions(-) -- 2.34.1

2 2

[PATCH OLK-6.6 0/2] *** jffs2: Prevent rtime decompress memory corruption ***
by Yongjian Sun 15 Jan '25

15 Jan '25

*** BLURB HERE *** Kinsey Moore (1): [Backport] jffs2: Prevent rtime decompress memory corruption Richard Weinberger (1): [Backport] jffs2: Fix rtime decompressor fs/jffs2/compr_rtime.c | 3 +++ 1 file changed, 3 insertions(+) -- 2.39.2

2 3

[PATCH OLK-5.10 0/2] *** jffs2: Prevent rtime decompress memory corruption ***
by Yongjian Sun 15 Jan '25

15 Jan '25

*** BLURB HERE *** Kinsey Moore (1): [Backport] jffs2: Prevent rtime decompress memory corruption Richard Weinberger (1): [Backport] jffs2: Fix rtime decompressor fs/jffs2/compr_rtime.c | 3 +++ 1 file changed, 3 insertions(+) -- 2.39.2

2 3

[PATCH OLK-5.10] NFS: Fix up nfs_ctx_key_to_expire()
by Yang Erkun 15 Jan '25

15 Jan '25

From: Trond Myklebust <trond.myklebust(a)hammerspace.com> mainline inclusion from mainline-v5.16-rc1 commit ca05cbae2a0468e5d78e9b4605936a8bf5da328b category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/IBGMYF Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?… -------------------------------- If the cached credential exists but doesn't have any expiration callback then exit early. Fix up atomicity issues when replacing the credential with a new one since the existing code could lead to refcount leaks. Signed-off-by: Trond Myklebust <trond.myklebust(a)hammerspace.com> Signed-off-by: Yang Erkun <yangerkun(a)huawei.com> --- fs/nfs/inode.c | 4 ++-- fs/nfs/write.c | 41 ++++++++++++++++++++++++++++------------- include/linux/nfs_fs.h | 2 +- 3 files changed, 31 insertions(+), 16 deletions(-) diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c index cd7d7aa2b3c6..22d425f21087 100644 --- a/fs/nfs/inode.c +++ b/fs/nfs/inode.c @@ -998,7 +998,7 @@ struct nfs_open_context *alloc_nfs_open_context(struct dentry *dentry, ctx->cred = get_cred(filp->f_cred); else ctx->cred = get_current_cred(); - ctx->ll_cred = NULL; + rcu_assign_pointer(ctx->ll_cred, NULL); ctx->state = NULL; ctx->mode = f_mode; ctx->flags = 0; @@ -1037,7 +1037,7 @@ static void __put_nfs_open_context(struct nfs_open_context *ctx, int is_sync) put_cred(ctx->cred); dput(ctx->dentry); nfs_sb_deactive(sb); - put_rpccred(ctx->ll_cred); + put_rpccred(rcu_dereference_protected(ctx->ll_cred, 1)); kfree(ctx->mdsthreshold); kfree_rcu(ctx, rcu_head); } diff --git a/fs/nfs/write.c b/fs/nfs/write.c index 4cf060691979..92239a7571aa 100644 --- a/fs/nfs/write.c +++ b/fs/nfs/write.c @@ -1221,7 +1221,7 @@ nfs_key_timeout_notify(struct file *filp, struct inode *inode) struct nfs_open_context *ctx = nfs_file_open_context(filp); if (nfs_ctx_key_to_expire(ctx, inode) && - !ctx->ll_cred) + !rcu_access_pointer(ctx->ll_cred)) /* Already expired! */ return -EACCES; return 0; @@ -1233,23 +1233,38 @@ nfs_key_timeout_notify(struct file *filp, struct inode *inode) bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx, struct inode *inode) { struct rpc_auth *auth = NFS_SERVER(inode)->client->cl_auth; - struct rpc_cred *cred = ctx->ll_cred; + struct rpc_cred *cred, *new, *old = NULL; struct auth_cred acred = { .cred = ctx->cred, }; + bool ret = false; - if (cred && !cred->cr_ops->crmatch(&acred, cred, 0)) { - put_rpccred(cred); - ctx->ll_cred = NULL; - cred = NULL; - } - if (!cred) - cred = auth->au_ops->lookup_cred(auth, &acred, 0); - if (!cred || IS_ERR(cred)) + rcu_read_lock(); + cred = rcu_dereference(ctx->ll_cred); + if (cred && !(cred->cr_ops->crkey_timeout && + cred->cr_ops->crkey_timeout(cred))) + goto out; + rcu_read_unlock(); + + new = auth->au_ops->lookup_cred(auth, &acred, 0); + if (new == cred) { + put_rpccred(new); return true; - ctx->ll_cred = cred; - return !!(cred->cr_ops->crkey_timeout && - cred->cr_ops->crkey_timeout(cred)); + } + if (IS_ERR_OR_NULL(new)) { + new = NULL; + ret = true; + } else if (new->cr_ops->crkey_timeout && + new->cr_ops->crkey_timeout(new)) + ret = true; + + rcu_read_lock(); + old = rcu_dereference_protected(xchg(&ctx->ll_cred, + RCU_INITIALIZER(new)), 1); +out: + rcu_read_unlock(); + put_rpccred(old); + return ret; } /* diff --git a/include/linux/nfs_fs.h b/include/linux/nfs_fs.h index 7488864589a7..010176828f15 100644 --- a/include/linux/nfs_fs.h +++ b/include/linux/nfs_fs.h @@ -71,7 +71,7 @@ struct nfs_open_context { fl_owner_t flock_owner; struct dentry *dentry; const struct cred *cred; - struct rpc_cred *ll_cred; /* low-level cred - use to check for expiry */ + struct rpc_cred __rcu *ll_cred; /* low-level cred - use to check for expiry */ struct nfs4_state *state; fmode_t mode; -- 2.39.2

2 1

[PATCH OLK-5.10] fs/dirty_pages: Add exclusive open lock to prevent null pointer dereference
by Zizhi Wo 15 Jan '25

15 Jan '25

hulk inclusion category: bugfix bugzilla: https://gitee.com/openeuler/kernel/issues/IBHTNK -------------------------------- The current mechanism for printing dirty pages involves "/proc/dirty/buffer_size" and "/proc/dirty/dirty_list" being able to be opened simultaneously. If "buffer_size" is opened firstly and 1 is written to it, and then "dirty_list" is opened, followed by writing 0 to "buffer_size". Then variable "buf_dirty" will be freed and set to NULL. However, the user can still access "buf_dirty" through the already opened "dirty_list", which could lead to a NULL pointer dereference issue. To prevent the NULL pointer dereference problem, add an exclusive open lock. And now "/proc/dirty/buffer_size" and "/proc/dirty/dirty_list" cannot be opened at the same time unless "/proc/dirty/buffer_size" is opened in read-only mode. At the same time, "buff_used" and "buff_lock" in the original code are not needed, and the redundant code is deleted. Fixes: aeb96447becc ("fs/dirty_pages: dump the number of dirty pages for each inode") Signed-off-by: Zizhi Wo <wozizhi(a)huawei.com> --- fs/dirty_pages.c | 83 +++++++++++++++++++++++++++--------------------- 1 file changed, 47 insertions(+), 36 deletions(-) diff --git a/fs/dirty_pages.c b/fs/dirty_pages.c index f75a5158d94b..81bd52ad50f0 100644 --- a/fs/dirty_pages.c +++ b/fs/dirty_pages.c @@ -19,12 +19,10 @@ static char *buf_dirty; /* buffer to store number of dirty pages */ static unsigned long buf_size; /* size of buffer in bytes */ static long buff_num; /* size of buffer in number of pages */ static int buff_limit; /* filter threshold of dirty pages*/ +static unsigned long lock_word; /* for exclusive access to buffer */ static struct proc_dir_entry *dirty_dir; -static struct mutex buff_used; /* buffer is in used */ -static struct mutex buff_lock; /* lock when buffer is changed */ - /* proc root directory */ #define DIRTY_ROOT "dirty" /* proc file for buffer allocation and release */ @@ -40,6 +38,18 @@ static void seq_set_overflow(struct seq_file *m) m->count = m->size; } +static bool dirty_pages_lock(void) +{ + if (xchg(&lock_word, 1) == 1) + return false; + return true; +} + +static void dirty_pages_unlock(void) +{ + lock_word = 0; +} + static unsigned long dump_dirtypages_inode(struct inode *inode) { struct pagevec pvec; @@ -195,10 +205,6 @@ static ssize_t seq_read_dirty( size_t n; int err = 0; - if (!mutex_trylock(&buff_used)) - return -EBUSY; - /* don't allow buffer to change during read */ - mutex_lock(&buff_lock); if (m->count == 0) { memset(buf_dirty, 0, buf_size); if (!m->buf) { @@ -210,13 +216,6 @@ static ssize_t seq_read_dirty( goto done; } - /* if buffer changed somehow */ - if (m->size != buf_size) { - mutex_unlock(&buff_lock); - mutex_unlock(&buff_used); - return -EFAULT; - } - n = min(m->count - m->from, size); /* check if this is the last read */ if (n == 0) @@ -234,8 +233,6 @@ static ssize_t seq_read_dirty( else *ppos += copied; - mutex_unlock(&buff_lock); - mutex_unlock(&buff_used); return copied; } @@ -247,6 +244,7 @@ static void free_buf_dirty(void) buf_size = 0; } } + static ssize_t write_proc( struct file *filp, const char *buf, @@ -257,9 +255,6 @@ static ssize_t write_proc( int ret = 0; long old_buff_num; - if (!mutex_trylock(&buff_used)) - return -EBUSY; - if (count > PAGE_SIZE) { ret = -EINVAL; goto error; @@ -283,42 +278,44 @@ static ssize_t write_proc( goto free; } - mutex_lock(&buff_lock); - ret = count; if (buff_num == 0) { free_buf_dirty(); - goto out; + goto free; } if (buff_num == old_buff_num) - goto out; + goto free; free_buf_dirty(); buf_size = PAGE_SIZE * buff_num; buf_dirty = vzalloc(buf_size); - if (!buf_dirty) { + if (!buf_dirty) ret = -ENOMEM; - goto out; - } -out: - mutex_unlock(&buff_lock); free: kfree(msg); error: - mutex_unlock(&buff_used); return ret; } static int proc_dpages_open(struct inode *inode, struct file *filp) { + int ret; + + if (!dirty_pages_lock()) + return -EBUSY; + if (buf_dirty == NULL || buf_size == 0) { pr_warn("please allocate buffer before getting dirty pages\n"); + dirty_pages_unlock(); return -ENOMEM; } - return single_open(filp, proc_dpages_show, NULL); + ret = single_open(filp, proc_dpages_show, NULL); + if (ret) + dirty_pages_unlock(); + return ret; } static int seq_release_dirty(struct inode *inode, struct file *file) @@ -327,6 +324,7 @@ static int seq_release_dirty(struct inode *inode, struct file *file) /* we don't want to free the buf */ m->buf = NULL; + dirty_pages_unlock(); single_release(inode, file); return 0; } @@ -351,7 +349,16 @@ static int proc_limit_show(struct seq_file *m, void *v) static int proc_switch_open(struct inode *inode, struct file *filp) { - return single_open(filp, proc_switch_show, NULL); + int ret; + + if (((filp->f_flags & O_ACCMODE) != O_RDONLY) && !dirty_pages_lock()) + return -EBUSY; + + ret = single_open(filp, proc_switch_show, NULL); + if (ret && ((filp->f_flags & O_ACCMODE) != O_RDONLY)) + dirty_pages_unlock(); + + return ret; } static int proc_limit_open(struct inode *inode, struct file *filp) @@ -359,6 +366,14 @@ static int proc_limit_open(struct inode *inode, struct file *filp) return single_open(filp, proc_limit_show, NULL); } +static int proc_switch_release(struct inode *inode, struct file *filp) +{ + if ((filp->f_flags & O_ACCMODE) != O_RDONLY) + dirty_pages_unlock(); + + return single_release(inode, filp); +} + static ssize_t write_limit_proc( struct file *filp, const char *buf, @@ -405,7 +420,7 @@ static const struct proc_ops proc_switch_operations = { .proc_read = seq_read, .proc_write = write_proc, .proc_lseek = seq_lseek, - .proc_release = single_release, + .proc_release = proc_switch_release, }; static const struct proc_ops proc_limit_operations = { @@ -440,8 +455,6 @@ static int __init dpages_proc_init(void) if (!proc_file) goto fail_limit; - mutex_init(&buff_used); - mutex_init(&buff_lock); return 0; fail_limit: @@ -456,9 +469,7 @@ static int __init dpages_proc_init(void) static void dpages_proc_exit(void) { - mutex_lock(&buff_lock); free_buf_dirty(); - mutex_unlock(&buff_lock); remove_proc_entry(DIRTY_PAGES, dirty_dir); remove_proc_entry(DIRTY_SWITCH, dirty_dir); remove_proc_entry(DIRTY_LIMIT, dirty_dir); -- 2.39.2

2 1

[PATCH openEuler-1.0-LTS] [Backport] drm/sti: avoid potential dereference of error pointers in sti_gdp_atomic_check
by Lin Ruifeng 15 Jan '25

15 Jan '25

From: Ma Ke <make24(a)iscas.ac.cn> stable inclusion from stable-v5.15.174 commit f5804567cf9605d6e5ec46c0bb786f7d50f18c13 bugzilla: https://gitee.com/src-openeuler/kernel/issues/IBHBX6 CVE: CVE-2024-56777 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit e965e771b069421c233d674c3c8cd8c7f7245f42 upstream. The return value of drm_atomic_get_crtc_state() needs to be checked. To avoid use of error pointer 'crtc_state' in case of the failure. Cc: stable(a)vger.kernel.org Fixes: dd86dc2f9ae1 ("drm/sti: implement atomic_check for the planes") Signed-off-by: Ma Ke <make24(a)iscas.ac.cn> Acked-by: Alain Volmat <alain.volmat(a)foss.st.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240909063359.1197065-1-make… Signed-off-by: Alain Volmat <alain.volmat(a)foss.st.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Conflicts: drivers/gpu/drm/sti/sti_gdp.c [Context conflicts] Signed-off-by: Lin Ruifeng <linruifeng4(a)huawei.com> --- drivers/gpu/drm/sti/sti_gdp.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/sti/sti_gdp.c b/drivers/gpu/drm/sti/sti_gdp.c index c32de6cbf061..0bd3230b2caa 100644 --- a/drivers/gpu/drm/sti/sti_gdp.c +++ b/drivers/gpu/drm/sti/sti_gdp.c @@ -629,6 +629,9 @@ static int sti_gdp_atomic_check(struct drm_plane *drm_plane, mixer = to_sti_mixer(crtc); crtc_state = drm_atomic_get_crtc_state(state->state, crtc); + if (IS_ERR(crtc_state)) + return PTR_ERR(crtc_state); + mode = &crtc_state->mode; dst_x = state->crtc_x; dst_y = state->crtc_y; -- 2.22.0

2 1

[PATCH OLK-6.6] [Backport] drm/sti: avoid potential dereference of error pointers in sti_gdp_atomic_check
by Lin Ruifeng 15 Jan '25

15 Jan '25

From: Ma Ke <make24(a)iscas.ac.cn> stable inclusion from stable-v6.6.64 commit 997b64c3f4c1827c5cfda8ae7f5d13f78d28b541 bugzilla: https://gitee.com/src-openeuler/kernel/issues/IBHBX6 CVE: CVE-2024-56777 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit e965e771b069421c233d674c3c8cd8c7f7245f42 upstream. The return value of drm_atomic_get_crtc_state() needs to be checked. To avoid use of error pointer 'crtc_state' in case of the failure. Cc: stable(a)vger.kernel.org Fixes: dd86dc2f9ae1 ("drm/sti: implement atomic_check for the planes") Signed-off-by: Ma Ke <make24(a)iscas.ac.cn> Acked-by: Alain Volmat <alain.volmat(a)foss.st.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240909063359.1197065-1-make… Signed-off-by: Alain Volmat <alain.volmat(a)foss.st.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Signed-off-by: Lin Ruifeng <linruifeng4(a)huawei.com> --- drivers/gpu/drm/sti/sti_gdp.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/sti/sti_gdp.c b/drivers/gpu/drm/sti/sti_gdp.c index 43c72c2604a0..f046f5f7ad25 100644 --- a/drivers/gpu/drm/sti/sti_gdp.c +++ b/drivers/gpu/drm/sti/sti_gdp.c @@ -638,6 +638,9 @@ static int sti_gdp_atomic_check(struct drm_plane *drm_plane, mixer = to_sti_mixer(crtc); crtc_state = drm_atomic_get_crtc_state(state, crtc); + if (IS_ERR(crtc_state)) + return PTR_ERR(crtc_state); + mode = &crtc_state->mode; dst_x = new_plane_state->crtc_x; dst_y = new_plane_state->crtc_y; -- 2.22.0

2 1

[PATCH OLK-5.10] [Backport] drm/sti: avoid potential dereference of error pointers in sti_gdp_atomic_check
by Lin Ruifeng 15 Jan '25

15 Jan '25

From: Ma Ke <make24(a)iscas.ac.cn> stable inclusion from stable-v5.15.174 commit f5804567cf9605d6e5ec46c0bb786f7d50f18c13 bugzilla: https://gitee.com/src-openeuler/kernel/issues/IBHBX6 CVE: CVE-2024-56777 Reference: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id… -------------------------------- commit e965e771b069421c233d674c3c8cd8c7f7245f42 upstream. The return value of drm_atomic_get_crtc_state() needs to be checked. To avoid use of error pointer 'crtc_state' in case of the failure. Cc: stable(a)vger.kernel.org Fixes: dd86dc2f9ae1 ("drm/sti: implement atomic_check for the planes") Signed-off-by: Ma Ke <make24(a)iscas.ac.cn> Acked-by: Alain Volmat <alain.volmat(a)foss.st.com> Link: https://patchwork.freedesktop.org/patch/msgid/20240909063359.1197065-1-make… Signed-off-by: Alain Volmat <alain.volmat(a)foss.st.com> Signed-off-by: Greg Kroah-Hartman <gregkh(a)linuxfoundation.org> Conflicts: drivers/gpu/drm/sti/sti_gdp.c [Context conflicts] Signed-off-by: Lin Ruifeng <linruifeng4(a)huawei.com> --- drivers/gpu/drm/sti/sti_gdp.c | 3 +++ 1 file changed, 3 insertions(+) diff --git a/drivers/gpu/drm/sti/sti_gdp.c b/drivers/gpu/drm/sti/sti_gdp.c index 2d5a2b5b78b8..c268a2ef1431 100644 --- a/drivers/gpu/drm/sti/sti_gdp.c +++ b/drivers/gpu/drm/sti/sti_gdp.c @@ -634,6 +634,9 @@ static int sti_gdp_atomic_check(struct drm_plane *drm_plane, mixer = to_sti_mixer(crtc); crtc_state = drm_atomic_get_crtc_state(state->state, crtc); + if (IS_ERR(crtc_state)) + return PTR_ERR(crtc_state); + mode = &crtc_state->mode; dst_x = state->crtc_x; dst_y = state->crtc_y; -- 2.22.0

2 1

openEuler Kernel SIG双周例会
by openEuler conference 15 Jan '25

15 Jan '25

您好！ Kernel 邀请您参加 2025-01-17 14:00 召开的WeLink会议(自动录制) 会议主题：openEuler Kernel SIG双周例会会议内容： 1. 进展update 2. 议题征集中（新增议题可回复本邮件申报，也可直接填报至会议纪要看板）会议链接：https://meeting.huaweicloud.com:36443/#/j/988444011 会议纪要：https://etherpad.openeuler.org/p/Kernel-meetings 更多资讯尽在：https://www.openeuler.org/zh/ Hello! Kernel invites you to attend the WeLink conference(auto recording) will be held at 2025-01-17 14:00, The subject of the conference is openEuler Kernel SIG双周例会, Summary: 1. 进展update 2. 议题征集中（新增议题可回复本邮件申报，也可直接填报至会议纪要看板） You can join the meeting at https://meeting.huaweicloud.com:36443/#/j/988444011. Add topics at https://etherpad.openeuler.org/p/Kernel-meetings. More information: https://www.openeuler.org/en/

1 0