fs/proc/task_mmu.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-)
The commit 65c66047259f ("proc: fix the issue of proc_mem_open returning NULL")
breaks `perf record -g -p PID` when profiling a kernel thread.
The strace of `perf record -g -p $(pgrep kswapd0)` shows:
openat(AT_FDCWD, "/proc/65/task/65/maps", O_RDONLY) = -1 ESRCH (No such process)
This patch partially reverts the commit to fix it.
Fixes: 65c66047259f ("proc: fix the issue of proc_mem_open returning NULL")
Signed-off-by: Jialin Wang <wjl.linux@gmail.com>
---
fs/proc/task_mmu.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 3d6d8a9f13fc..7a7ce26106ac 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -340,8 +340,8 @@ static int proc_maps_open(struct inode *inode, struct file *file,
priv->inode = inode;
priv->mm = proc_mem_open(inode, PTRACE_MODE_READ);
- if (IS_ERR_OR_NULL(priv->mm)) {
- int err = priv->mm ? PTR_ERR(priv->mm) : -ESRCH;
+ if (IS_ERR(priv->mm)) {
+ int err = PTR_ERR(priv->mm);
seq_release_private(inode, file);
return err;
--
2.50.0
On Fri, 8 Aug 2025 00:54:55 +0800 Jialin Wang <wjl.linux@gmail.com> wrote: > The commit 65c66047259f ("proc: fix the issue of proc_mem_open returning NULL") > breaks `perf record -g -p PID` when profiling a kernel thread. > > The strace of `perf record -g -p $(pgrep kswapd0)` shows: > > openat(AT_FDCWD, "/proc/65/task/65/maps", O_RDONLY) = -1 ESRCH (No such process) > > This patch partially reverts the commit to fix it. Thanks. But "breaks" is a rather thin description of the problem! Can you please describe the observed misbehavior fully? > Fixes: 65c66047259f ("proc: fix the issue of proc_mem_open returning NULL") Because we should backport this fix into 6.16.x -stable kernels. The -stable maintainers may wonder why we're requesting this. Also, any person who is having problems with their 6.16-based kernel will want such a description so they can decide whether this fix might address their problem. Thanks.
The commit 65c66047259f ("proc: fix the issue of proc_mem_open returning
NULL") caused proc_maps_open() to return -ESRCH when proc_mem_open()
returns NULL. This breaks legitimate /proc/<pid>/maps access for kernel
threads since kernel threads have NULL mm_struct.
The regression causes perf to fail and exit when profiling a kernel thread:
# perf record -v -g -p $(pgrep kswapd0)
...
couldn't open /proc/65/task/65/maps
This patch partially reverts the commit to fix it.
Fixes: 65c66047259f ("proc: fix the issue of proc_mem_open returning NULL")
Signed-off-by: Jialin Wang <wjl.linux@gmail.com>
---
Changes in v2 (Thanks to Andrew):
- Add more detailed misbehavior description in commit message
fs/proc/task_mmu.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/proc/task_mmu.c b/fs/proc/task_mmu.c
index 3d6d8a9f13fc..7a7ce26106ac 100644
--- a/fs/proc/task_mmu.c
+++ b/fs/proc/task_mmu.c
@@ -340,8 +340,8 @@ static int proc_maps_open(struct inode *inode, struct file *file,
priv->inode = inode;
priv->mm = proc_mem_open(inode, PTRACE_MODE_READ);
- if (IS_ERR_OR_NULL(priv->mm)) {
- int err = priv->mm ? PTR_ERR(priv->mm) : -ESRCH;
+ if (IS_ERR(priv->mm)) {
+ int err = PTR_ERR(priv->mm);
seq_release_private(inode, file);
return err;
--
2.50.0
© 2016 - 2025 Red Hat, Inc.