[PATCH] uprobe: Move arch_uprobe_optimize right after handlers execution

Subsystems: performance events subsystem, the rest, uprobes

STALE341d

4 messages, 4 authors, 2025-10-03 · open the first message on its own page

[PATCH] uprobe: Move arch_uprobe_optimize right after handlers execution

From: Jiri Olsa <jolsa@kernel.org>
Date: 2025-10-01 13:25:03

It's less confusing to optimize uprobe right after handlers execution
and before we do the check for changed ip register to avoid situations
where changed ip register would skip uprobe optimization.

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 kernel/events/uprobes.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index 5dcf927310fd..c14ec27b976d 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c

@@ -2765,6 +2765,9 @@ static void handle_swbp(struct pt_regs *regs)
 
 	handler_chain(uprobe, regs);
 
+	/* Try to optimize after first hit. */
+	arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
+
 	/*
 	 * If user decided to take execution elsewhere, it makes little sense
 	 * to execute the original instruction, so let's skip it.

@@ -2772,9 +2775,6 @@ static void handle_swbp(struct pt_regs *regs)
 	if (instruction_pointer(regs) != bp_vaddr)
 		goto out;
 
-	/* Try to optimize after first hit. */
-	arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
-
 	if (arch_uprobe_skip_sstep(&uprobe->arch, regs))
 		goto out;

-- 
2.51.0

Re: [PATCH] uprobe: Move arch_uprobe_optimize right after handlers execution

From: Andrii Nakryiko <hidden>
Date: 2025-10-01 16:39:42

On Wed, Oct 1, 2025 at 6:25 AM Jiri Olsa [off-list ref] wrote:

It's less confusing to optimize uprobe right after handlers execution
and before we do the check for changed ip register to avoid situations
where changed ip register would skip uprobe optimization.

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>
---
 kernel/events/uprobes.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

makes sense

Acked-by: Andrii Nakryiko <andrii@kernel.org>

quoted hunk

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index 5dcf927310fd..c14ec27b976d 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c

@@ -2765,6 +2765,9 @@ static void handle_swbp(struct pt_regs *regs)

        handler_chain(uprobe, regs);

+       /* Try to optimize after first hit. */
+       arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
+
        /*
         * If user decided to take execution elsewhere, it makes little sense
         * to execute the original instruction, so let's skip it.

@@ -2772,9 +2775,6 @@ static void handle_swbp(struct pt_regs *regs)
        if (instruction_pointer(regs) != bp_vaddr)
                goto out;

-       /* Try to optimize after first hit. */
-       arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
-
        if (arch_uprobe_skip_sstep(&uprobe->arch, regs))
                goto out;

--

2.51.0

Re: [PATCH] uprobe: Move arch_uprobe_optimize right after handlers execution

From: Masami Hiramatsu (Google) <mhiramat@kernel.org>
Date: 2025-10-01 23:20:52

On Wed,  1 Oct 2025 15:24:49 +0200
Jiri Olsa [off-list ref] wrote:

It's less confusing to optimize uprobe right after handlers execution
and before we do the check for changed ip register to avoid situations
where changed ip register would skip uprobe optimization.

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Jiri Olsa <jolsa@kernel.org>

Looks good to me.

Reviewed-by: Masami Hiramatsu (Google) <mhiramat@kernel.org>


Thanks,

quoted hunk

---
 kernel/events/uprobes.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/kernel/events/uprobes.c b/kernel/events/uprobes.c
index 5dcf927310fd..c14ec27b976d 100644
--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c

@@ -2765,6 +2765,9 @@ static void handle_swbp(struct pt_regs *regs)
 
 	handler_chain(uprobe, regs);
 
+	/* Try to optimize after first hit. */
+	arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
+
 	/*
 	 * If user decided to take execution elsewhere, it makes little sense
 	 * to execute the original instruction, so let's skip it.

@@ -2772,9 +2775,6 @@ static void handle_swbp(struct pt_regs *regs)
 	if (instruction_pointer(regs) != bp_vaddr)
 		goto out;
 
-	/* Try to optimize after first hit. */
-	arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
-
 	if (arch_uprobe_skip_sstep(&uprobe->arch, regs))
 		goto out;

-- 
2.51.0


-- 
Masami Hiramatsu (Google) [off-list ref]

Re: [PATCH] uprobe: Move arch_uprobe_optimize right after handlers execution

From: Oleg Nesterov <oleg@redhat.com>
Date: 2025-10-03 11:49:52

On 10/01, Jiri Olsa wrote:

quoted hunk

--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c

@@ -2765,6 +2765,9 @@ static void handle_swbp(struct pt_regs *regs)
 
 	handler_chain(uprobe, regs);
 
+	/* Try to optimize after first hit. */
+	arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
+
 	/*
 	 * If user decided to take execution elsewhere, it makes little sense
 	 * to execute the original instruction, so let's skip it.

@@ -2772,9 +2775,6 @@ static void handle_swbp(struct pt_regs *regs)
 	if (instruction_pointer(regs) != bp_vaddr)
 		goto out;
 
-	/* Try to optimize after first hit. */
-	arch_uprobe_optimize(&uprobe->arch, bp_vaddr);
-
 	if (arch_uprobe_skip_sstep(&uprobe->arch, regs))
 		goto out;

Acked-by: Oleg Nesterov <oleg@redhat.com>

`h`	back out one level
`j`	next message in thread
`k`	previous message in thread
`l`	drill in
`Esc`	close help / fold thread tree
`?`	toggle this help