* Re: [PATCH] ALSA: ump: Fix double-free of ump->out_cvts on legacy rawmidi attach failure
[not found] <20260729001509.11164-1-kartikey406@gmail.com>
@ 2026-07-29 0:15 ` syzbot
0 siblings, 0 replies; 6+ messages in thread
From: syzbot @ 2026-07-29 0:15 UTC (permalink / raw)
To: kartikey406; +Cc: kartikey406, linux-kernel, syzkaller-bugs
> #syz test: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
This crash does not have a reproducer. I cannot test it.
>
>
> If snd_rawmidi_new() fails inside snd_ump_attach_legacy_rawmidi(), the
> error path frees ump->out_cvts but leaves the pointer dangling. Since
> ump->out_cvts is a field of the long-lived struct snd_ump_endpoint
> (not the rawmidi device that failed to be created), it gets freed a
> second time later during normal endpoint teardown, in
> snd_ump_endpoint_free(), invoked via snd_rawmidi_free()'s
> private_free callback when the sound card is released. This results
> in a KASAN double-free/invalid-free.
>
> Clear ump->out_cvts to NULL after freeing it on the error path, so
> the later unconditional kfree() in snd_ump_endpoint_free() becomes a
> harmless no-op.
>
> Reported-by: syzbot+b6cab840e6a85641c7ad@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=b6cab840e6a85641c7ad
> Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
> ---
> sound/core/ump.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/sound/core/ump.c b/sound/core/ump.c
> index 70520c7ca293..632c13baf21e 100644
> --- a/sound/core/ump.c
> +++ b/sound/core/ump.c
> @@ -1365,6 +1365,7 @@ int snd_ump_attach_legacy_rawmidi(struct snd_ump_endpoint *ump,
> &rmidi);
> if (err < 0) {
> kfree(ump->out_cvts);
> + ump->out_cvts = NULL;
> return err;
> }
>
> --
> 2.43.0
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] ALSA: ump: Fix double-free of ump->out_cvts on legacy rawmidi attach failure
[not found] <20260729001642.11188-1-kartikey406@gmail.com>
@ 2026-07-29 0:16 ` syzbot
0 siblings, 0 replies; 6+ messages in thread
From: syzbot @ 2026-07-29 0:16 UTC (permalink / raw)
To: kartikey406; +Cc: kartikey406, linux-kernel, syzkaller-bugs
> #syz test: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
This crash does not have a reproducer. I cannot test it.
>
>
> If snd_rawmidi_new() fails inside snd_ump_attach_legacy_rawmidi(), the
> error path frees ump->out_cvts but leaves the pointer dangling. Since
> ump->out_cvts is a field of the long-lived struct snd_ump_endpoint
> (not the rawmidi device that failed to be created), it gets freed a
> second time later during normal endpoint teardown, in
> snd_ump_endpoint_free(), invoked via snd_rawmidi_free()'s
> private_free callback when the sound card is released. This results
> in a KASAN double-free/invalid-free.
>
> Clear ump->out_cvts to NULL after freeing it on the error path, so
> the later unconditional kfree() in snd_ump_endpoint_free() becomes a
> harmless no-op.
>
> Reported-by: syzbot+b6cab840e6a85641c7ad@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=b6cab840e6a85641c7ad
> Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
> ---
> sound/core/ump.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/sound/core/ump.c b/sound/core/ump.c
> index 70520c7ca293..632c13baf21e 100644
> --- a/sound/core/ump.c
> +++ b/sound/core/ump.c
> @@ -1365,6 +1365,7 @@ int snd_ump_attach_legacy_rawmidi(struct snd_ump_endpoint *ump,
> &rmidi);
> if (err < 0) {
> kfree(ump->out_cvts);
> + ump->out_cvts = NULL;
> return err;
> }
>
> --
> 2.43.0
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] ALSA: ump: Fix double-free of ump->out_cvts on legacy rawmidi attach failure
[not found] <20260729001845.11372-1-kartikey406@gmail.com>
@ 2026-07-29 0:18 ` syzbot
0 siblings, 0 replies; 6+ messages in thread
From: syzbot @ 2026-07-29 0:18 UTC (permalink / raw)
To: kartikey406; +Cc: kartikey406, linux-kernel, syzkaller-bugs
> #syz test: git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git master
This crash does not have a reproducer. I cannot test it.
>
> If snd_rawmidi_new() fails inside snd_ump_attach_legacy_rawmidi(), the
> error path frees ump->out_cvts but leaves the pointer dangling. Since
> ump->out_cvts is a field of the long-lived struct snd_ump_endpoint
> (not the rawmidi device that failed to be created), it gets freed a
> second time later during normal endpoint teardown, in
> snd_ump_endpoint_free(), invoked via snd_rawmidi_free()'s
> private_free callback when the sound card is released. This results
> in a KASAN double-free/invalid-free.
>
> Clear ump->out_cvts to NULL after freeing it on the error path, so
> the later unconditional kfree() in snd_ump_endpoint_free() becomes a
> harmless no-op.
>
> Reported-by: syzbot+b6cab840e6a85641c7ad@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=b6cab840e6a85641c7ad
> Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
> ---
> sound/core/ump.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/sound/core/ump.c b/sound/core/ump.c
> index 70520c7ca293..632c13baf21e 100644
> --- a/sound/core/ump.c
> +++ b/sound/core/ump.c
> @@ -1365,6 +1365,7 @@ int snd_ump_attach_legacy_rawmidi(struct snd_ump_endpoint *ump,
> &rmidi);
> if (err < 0) {
> kfree(ump->out_cvts);
> + ump->out_cvts = NULL;
> return err;
> }
>
> --
> 2.43.0
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [PATCH] ALSA: ump: Fix double-free of ump->out_cvts on legacy rawmidi attach failure
[not found] <20260729002134.11433-1-kartikey406@gmail.com>
@ 2026-07-29 0:21 ` syzbot
0 siblings, 0 replies; 6+ messages in thread
From: syzbot @ 2026-07-29 0:21 UTC (permalink / raw)
To: kartikey406; +Cc: kartikey406, linux-kernel, syzkaller-bugs
> #syz test: git://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git for-kernelci
This crash does not have a reproducer. I cannot test it.
>
> If snd_rawmidi_new() fails inside snd_ump_attach_legacy_rawmidi(), the
> error path frees ump->out_cvts but leaves the pointer dangling. Since
> ump->out_cvts is a field of the long-lived struct snd_ump_endpoint
> (not the rawmidi device that failed to be created), it gets freed a
> second time later during normal endpoint teardown, in
> snd_ump_endpoint_free(), invoked via snd_rawmidi_free()'s
> private_free callback when the sound card is released. This results
> in a KASAN double-free/invalid-free.
>
> Clear ump->out_cvts to NULL after freeing it on the error path, so
> the later unconditional kfree() in snd_ump_endpoint_free() becomes a
> harmless no-op.
>
> Reported-by: syzbot+b6cab840e6a85641c7ad@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=b6cab840e6a85641c7ad
> Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
> ---
> sound/core/ump.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/sound/core/ump.c b/sound/core/ump.c
> index 70520c7ca293..632c13baf21e 100644
> --- a/sound/core/ump.c
> +++ b/sound/core/ump.c
> @@ -1365,6 +1365,7 @@ int snd_ump_attach_legacy_rawmidi(struct snd_ump_endpoint *ump,
> &rmidi);
> if (err < 0) {
> kfree(ump->out_cvts);
> + ump->out_cvts = NULL;
> return err;
> }
>
> --
> 2.43.0
>
^ permalink raw reply [flat|nested] 6+ messages in thread
* [PATCH] ALSA: ump: Fix double-free of ump->out_cvts on legacy rawmidi attach failure
@ 2026-07-29 0:23 Deepanshu Kartikey
2026-07-29 6:33 ` Takashi Iwai
0 siblings, 1 reply; 6+ messages in thread
From: Deepanshu Kartikey @ 2026-07-29 0:23 UTC (permalink / raw)
To: perex, tiwai
Cc: kees, linux-sound, linux-kernel, Deepanshu Kartikey,
syzbot+b6cab840e6a85641c7ad
If snd_rawmidi_new() fails inside snd_ump_attach_legacy_rawmidi(), the
error path frees ump->out_cvts but leaves the pointer dangling. Since
ump->out_cvts is a field of the long-lived struct snd_ump_endpoint
(not the rawmidi device that failed to be created), it gets freed a
second time later during normal endpoint teardown, in
snd_ump_endpoint_free(), invoked via snd_rawmidi_free()'s
private_free callback when the sound card is released. This results
in a KASAN double-free/invalid-free.
Clear ump->out_cvts to NULL after freeing it on the error path, so
the later unconditional kfree() in snd_ump_endpoint_free() becomes a
harmless no-op.
Reported-by: syzbot+b6cab840e6a85641c7ad@syzkaller.appspotmail.com
Closes: https://syzkaller.appspot.com/bug?extid=b6cab840e6a85641c7ad
Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
---
sound/core/ump.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/sound/core/ump.c b/sound/core/ump.c
index 70520c7ca293..632c13baf21e 100644
--- a/sound/core/ump.c
+++ b/sound/core/ump.c
@@ -1365,6 +1365,7 @@ int snd_ump_attach_legacy_rawmidi(struct snd_ump_endpoint *ump,
&rmidi);
if (err < 0) {
kfree(ump->out_cvts);
+ ump->out_cvts = NULL;
return err;
}
--
2.43.0
^ permalink raw reply related [flat|nested] 6+ messages in thread
* Re: [PATCH] ALSA: ump: Fix double-free of ump->out_cvts on legacy rawmidi attach failure
2026-07-29 0:23 Deepanshu Kartikey
@ 2026-07-29 6:33 ` Takashi Iwai
0 siblings, 0 replies; 6+ messages in thread
From: Takashi Iwai @ 2026-07-29 6:33 UTC (permalink / raw)
To: Deepanshu Kartikey
Cc: perex, tiwai, kees, linux-sound, linux-kernel,
syzbot+b6cab840e6a85641c7ad
On Wed, 29 Jul 2026 02:23:33 +0200,
Deepanshu Kartikey wrote:
>
> If snd_rawmidi_new() fails inside snd_ump_attach_legacy_rawmidi(), the
> error path frees ump->out_cvts but leaves the pointer dangling. Since
> ump->out_cvts is a field of the long-lived struct snd_ump_endpoint
> (not the rawmidi device that failed to be created), it gets freed a
> second time later during normal endpoint teardown, in
> snd_ump_endpoint_free(), invoked via snd_rawmidi_free()'s
> private_free callback when the sound card is released. This results
> in a KASAN double-free/invalid-free.
>
> Clear ump->out_cvts to NULL after freeing it on the error path, so
> the later unconditional kfree() in snd_ump_endpoint_free() becomes a
> harmless no-op.
>
> Reported-by: syzbot+b6cab840e6a85641c7ad@syzkaller.appspotmail.com
> Closes: https://syzkaller.appspot.com/bug?extid=b6cab840e6a85641c7ad
> Signed-off-by: Deepanshu Kartikey <kartikey406@gmail.com>
The very same fix has been already merged in commit
70c977815af0d997feb2d0c5d284d55689bf7051 on sound git tree.
thanks,
Takashi
^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2026-07-29 6:33 UTC | newest]
Thread overview: 6+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
[not found] <20260729001845.11372-1-kartikey406@gmail.com>
2026-07-29 0:18 ` [PATCH] ALSA: ump: Fix double-free of ump->out_cvts on legacy rawmidi attach failure syzbot
2026-07-29 0:23 Deepanshu Kartikey
2026-07-29 6:33 ` Takashi Iwai
[not found] <20260729002134.11433-1-kartikey406@gmail.com>
2026-07-29 0:21 ` syzbot
[not found] <20260729001642.11188-1-kartikey406@gmail.com>
2026-07-29 0:16 ` syzbot
[not found] <20260729001509.11164-1-kartikey406@gmail.com>
2026-07-29 0:15 ` syzbot
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.