Skip to contentSkip to conversation

proc/task_mmu: remove unnecessary inlines in function definitions

PATCH In review Suren Baghdasaryan 2 participants 7 messages Sep 14

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 proc/task_mmu: remove unnecessary inlines in function definitions

Uninlined drop_rcu() and reacquire_rcu() functions are not used when CONFIG_PROC_PAGE_MONITOR=n. Exclure them in this configuration.

Fixes 7edb437cbbfc ("proc/task_mmu: remove unnecessary inlines in function definitions")
---
 fs/proc/task_mmu.c | 2 ++
 1 file changed, 2 insertions(+)
fs/proc/task_mmu.c +2 -0
index 24425e230895..f8790a7095ff 100644
@@ -184,6 +184,7 @@ static bool fallback_to_mmap_lock(struct proc_maps_private *priv,
 	return true;
 }
 
+#ifdef CONFIG_PROC_PAGE_MONITOR
 static void drop_rcu(struct proc_maps_private *priv)
 {
 	if (priv->lock_ctx.mmap_locked)
@@ -201,6 +202,7 @@ static void reacquire_rcu(struct proc_maps_private *priv)
 	/* Reinitialize the iterator. */
 	vma_iter_set(&priv->iter, priv->lock_ctx.locked_vma->vm_end);
 }
+#endif
 
 static struct vm_area_struct *proc_get_vma(struct seq_file *m, loff_t *ppos)
 {

base-commit: baa8de2f3448d1466a888a805c18d01c998fe052

Signature · 1 line
2.55.0.1032.g73a4cd73de-goog
#1 lore raw
Quoted 3 lines · Suren Baghdasaryan

Uninlined drop_rcu() and reacquire_rcu() functions are not used when CONFIG_PROC_PAGE_MONITOR=n. Exclure them in this configuration.

s/Exclure/Exclude

Rushing too much...

Andrew, could you please fold this fixup into the patch it fixes (7edb437cbbfc in the mm-new branch)?
Thanks!

Quoted 34 lines
Fixes 7edb437cbbfc ("proc/task_mmu: remove unnecessary inlines in function definitions")
---
 fs/proc/task_mmu.c | 2 ++
 1 file changed, 2 insertions(+)
fs/proc/task_mmu.c +2 -0
index 24425e230895..f8790a7095ff 100644
@@ -184,6 +184,7 @@ static bool fallback_to_mmap_lock(struct proc_maps_private *priv,
        return true;
 }
 
+#ifdef CONFIG_PROC_PAGE_MONITOR
 static void drop_rcu(struct proc_maps_private *priv)
 {
        if (priv->lock_ctx.mmap_locked)
@@ -201,6 +202,7 @@ static void reacquire_rcu(struct proc_maps_private *priv)
        /* Reinitialize the iterator. */
        vma_iter_set(&priv->iter, priv->lock_ctx.locked_vma->vm_end);
 }
+#endif
 
 static struct vm_area_struct *proc_get_vma(struct seq_file *m, loff_t *ppos)
 {

base-commit: baa8de2f3448d1466a888a805c18d01c998fe052 --
2.55.0.1032.g73a4cd73de-goog

maintainer
#2 lore raw
In reply to Suren Baghdasaryan
Quoted 3 lines · On Mon, 14 Sep 2026 12:06:31 -0700 Suren Baghdasaryan <[email protected]> wrote:

Uninlined drop_rcu() and reacquire_rcu() functions are not used when CONFIG_PROC_PAGE_MONITOR=n. Exclure them in this configuration.

But there's also

fs/proc/task_mmu.c:133:12: warning: 'lock_ctx_mm' defined but not used [-Wunused-function]
133 | static int lock_ctx_mm(struct proc_maps_locking_ctx *lock_ctx)

      |            ^~~~~~~~~~~
#3 lore raw
Quoted 7 lines · Andrew Morton
Quoted 3 lines · On Mon, 14 Sep 2026 12:06:31 -0700 Suren Baghdasaryan <[email protected]> wrote:

Uninlined drop_rcu() and reacquire_rcu() functions are not used when CONFIG_PROC_PAGE_MONITOR=n. Exclure them in this configuration.

Quoted 5 lines

But there's also

fs/proc/task_mmu.c:133:12: warning: 'lock_ctx_mm' defined but not used [-Wunused-function]
133 | static int lock_ctx_mm(struct proc_maps_locking_ctx *lock_ctx)

      |            ^~~~~~~~~~~

Ugh! This function is removed in later patches... Should I just repost v5 instead of fixing up in different places?

Quoted 1 line
maintainer
#4 lore raw
Quoted 9 lines · On Mon, 14 Sep 2026 13:11:52 -0700 Suren Baghdasaryan <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 3 lines · On Mon, 14 Sep 2026 12:06:31 -0700 Suren Baghdasaryan <[email protected]> wrote:

Uninlined drop_rcu() and reacquire_rcu() functions are not used when CONFIG_PROC_PAGE_MONITOR=n. Exclure them in this configuration.

Quoted 7 lines
Quoted 5 lines

But there's also

fs/proc/task_mmu.c:133:12: warning: 'lock_ctx_mm' defined but not used [-Wunused-function]
133 | static int lock_ctx_mm(struct proc_maps_locking_ctx *lock_ctx)

      |            ^~~~~~~~~~~

Ugh! This function is removed in later patches...

Yes. So it's a compile-time bisection hole, which is probably the worst kind of bisection hole.

Quoted 2 lines

Should I just repost
v5 instead of fixing up in different places?

That would be best. Leave it a few days, we'll survive.

#5 lore raw
Quoted 20 lines · Andrew Morton
Quoted 9 lines · On Mon, 14 Sep 2026 13:11:52 -0700 Suren Baghdasaryan <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 3 lines · On Mon, 14 Sep 2026 12:06:31 -0700 Suren Baghdasaryan <[email protected]> wrote:

Uninlined drop_rcu() and reacquire_rcu() functions are not used when CONFIG_PROC_PAGE_MONITOR=n. Exclure them in this configuration.

Quoted 7 lines
Quoted 5 lines

But there's also

fs/proc/task_mmu.c:133:12: warning: 'lock_ctx_mm' defined but not used [-Wunused-function]
133 | static int lock_ctx_mm(struct proc_maps_locking_ctx *lock_ctx)

      |            ^~~~~~~~~~~

Ugh! This function is removed in later patches...

Quoted 7 lines

Yes. So it's a compile-time bisection hole, which is probably the worst kind of bisection hole.

Quoted 2 lines

Should I just repost
v5 instead of fixing up in different places?

That would be best. Leave it a few days, we'll survive.

Ack. I'll post v5 on Thursday if nothing else blows up.

maintainer
#6 lore raw
Quoted 24 lines · On Mon, 14 Sep 2026 13:57:11 -0700 Suren Baghdasaryan <[email protected]> wrote:
Quoted 23 lines · Andrew Morton
Quoted 9 lines · On Mon, 14 Sep 2026 13:11:52 -0700 Suren Baghdasaryan <[email protected]> wrote:
Quoted 7 lines · Andrew Morton
Quoted 3 lines · On Mon, 14 Sep 2026 12:06:31 -0700 Suren Baghdasaryan <[email protected]> wrote:

Uninlined drop_rcu() and reacquire_rcu() functions are not used when CONFIG_PROC_PAGE_MONITOR=n. Exclure them in this configuration.

Quoted 7 lines
Quoted 5 lines

But there's also

fs/proc/task_mmu.c:133:12: warning: 'lock_ctx_mm' defined but not used [-Wunused-function]
133 | static int lock_ctx_mm(struct proc_maps_locking_ctx *lock_ctx)

      |            ^~~~~~~~~~~

Ugh! This function is removed in later patches...

Yes. So it's a compile-time bisection hole, which is probably the worst kind of bisection hole.

Quoted 6 lines
Quoted 4 lines
Quoted 2 lines

Should I just repost
v5 instead of fixing up in different places?

That would be best. Leave it a few days, we'll survive.

Ack. I'll post v5 on Thursday if nothing else blows up.

np. It's only in mm-new, which is performing its intended function. I'll keep this out of linux-next until we have the v2.