[PATCH] kobject: fix out-of-bounds read in action parser

[PATCH] kobject: fix out-of-bounds read in action parser

[Jiacheng Xu]

kobject_action_type() uses the position of the first space in the input
as the length of the uevent action.  It then checks the byte at that
position in the corresponding action string.

An embedded NUL byte can make strncmp() report a match while
count_first is already greater than the actual length of the action
string.  For example, the input "bind\0 ..." makes the parser access
kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
string.

Use strlen() to verify that the input action length exactly matches the
known action string before accepting the match.  This avoids indexing
the action string with an out-of-bounds offset.

Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
---
lib/kobject_uevent.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
index ddbc4d7482d2..b03562301bfe 100644
--- a/lib/kobject_uevent.c
+++ b/lib/kobject_uevent.c
@@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
        if (strncmp(kobject_actions[action], buf, count_first) != 0)
                continue;
-             if (kobject_actions[action][count_first] != '\0')
+             if (strlen(kobject_actions[action]) != count_first)
                continue;
        if (args)
                *args = args_start;

Re: [PATCH] kobject: fix out-of-bounds read in action parser

[Vishal Moola (Fractile)]

I was looking at reviewing your patch. Unfortunately, it doesn't appear
to apply cleanly; I'm seeing:

Applying: kobject: fix out-of-bounds read in action parser
error: corrupt patch at .git/rebase-apply/patch:10
Patch failed at 0001 kobject: fix out-of-bounds read in action parser

It looks like sashiko failed to apply it as well[1]. Perhaps you should
reformat and resend?

On Tue, Aug 18, 2026 at 02:26:14PM +0800, Jiacheng Xu wrote:
> kobject_action_type() uses the position of the first space in the input
> as the length of the uevent action.  It then checks the byte at that
> position in the corresponding action string.
> 
> An embedded NUL byte can make strncmp() report a match while
> count_first is already greater than the actual length of the action
> string.  For example, the input "bind\0 ..." makes the parser access
> kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
> string.
> 
> Use strlen() to verify that the input action length exactly matches the
> known action string before accepting the match.  This avoids indexing
> the action string with an out-of-bounds offset.
> 
> Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
> Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
> ---
> lib/kobject_uevent.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
> index ddbc4d7482d2..b03562301bfe 100644
> --- a/lib/kobject_uevent.c
> +++ b/lib/kobject_uevent.c
> @@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
> for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
>         if (strncmp(kobject_actions[action], buf, count_first) != 0)
>                 continue;
> -             if (kobject_actions[action][count_first] != '\0')
> +             if (strlen(kobject_actions[action]) != count_first)
>                 continue;
>         if (args)
>                 *args = args_start;

[1] https://sashiko.dev/#/patchset/200d27f7.1471a.1a0138c92da.Coremail.stitch%40zju.edu.cn

Re: Re: [PATCH] kobject: fix out-of-bounds read in action parser

[Jiacheng Xu]

Thanks for pointing this out. I'll reformat and resend the patch as v2. Sorry for the noise.

> -----原始邮件-----
> 发件人: "Vishal Moola (Fractile)" <vishal.moola@gmail.com>
> 发送时间:2026-08-25 01:32:54 (星期二)
> 收件人: "Jiacheng Xu" <stitch@zju.edu.cn>
> 抄送: gregkh@linuxfoundation.org, rafael@kernel.org, dakr@kernel.org, akpm@linux-foundation.org, driver-core@lists.linux.dev
> 主题: Re: [PATCH] kobject: fix out-of-bounds read in action parser
> 
> I was looking at reviewing your patch. Unfortunately, it doesn't appear
> to apply cleanly; I'm seeing:
> 
> Applying: kobject: fix out-of-bounds read in action parser
> error: corrupt patch at .git/rebase-apply/patch:10
> Patch failed at 0001 kobject: fix out-of-bounds read in action parser
> 
> It looks like sashiko failed to apply it as well[1]. Perhaps you should
> reformat and resend?
> 
> On Tue, Aug 18, 2026 at 02:26:14PM +0800, Jiacheng Xu wrote:
> > kobject_action_type() uses the position of the first space in the input
> > as the length of the uevent action.  It then checks the byte at that
> > position in the corresponding action string.
> > 
> > An embedded NUL byte can make strncmp() report a match while
> > count_first is already greater than the actual length of the action
> > string.  For example, the input "bind\0 ..." makes the parser access
> > kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
> > string.
> > 
> > Use strlen() to verify that the input action length exactly matches the
> > known action string before accepting the match.  This avoids indexing
> > the action string with an out-of-bounds offset.
> > 
> > Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
> > Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
> > ---
> > lib/kobject_uevent.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
> > index ddbc4d7482d2..b03562301bfe 100644
> > --- a/lib/kobject_uevent.c
> > +++ b/lib/kobject_uevent.c
> > @@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
> > for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
> >         if (strncmp(kobject_actions[action], buf, count_first) != 0)
> >                 continue;
> > -             if (kobject_actions[action][count_first] != '\0')
> > +             if (strlen(kobject_actions[action]) != count_first)
> >                 continue;
> >         if (args)
> >                 *args = args_start;
> 
> [1] https://sashiko.dev/#/patchset/200d27f7.1471a.1a0138c92da.Coremail.stitch%40zju.edu.cn

[PATCH v2] kobject: fix out-of-bounds read in action parser

[Jiacheng Xu]

kobject_action_type() uses the position of the first space in the input
as the length of the uevent action. It then checks the byte at that
position in the corresponding action string.

An embedded NUL byte can make strncmp() report a match while
count_first is already greater than the actual length of the action
string. For example, the input "bind\0 ..." makes the parser access
kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
string.

Use strlen() to verify that the input action length exactly matches the
known action string before accepting the match. This avoids indexing
the action string with an out-of-bounds offset.

Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
---
 lib/kobject_uevent.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
index ddbc4d7482d2..b03562301bfe 100644
--- a/lib/kobject_uevent.c
+++ b/lib/kobject_uevent.c
@@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
        for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
                if (strncmp(kobject_actions[action], buf, count_first) != 0)
                        continue;
-               if (kobject_actions[action][count_first] != '\0')
+               if (strlen(kobject_actions[action]) != count_first)
                        continue;
                if (args)
                        *args = args_start;
-- 
2.51.0


> -----原始邮件-----
> 发件人: "Vishal Moola (Fractile)" <vishal.moola@gmail.com>
> 发送时间:2026-08-25 01:32:54 (星期二)
> 收件人: "Jiacheng Xu" <stitch@zju.edu.cn>
> 抄送: gregkh@linuxfoundation.org, rafael@kernel.org, dakr@kernel.org, akpm@linux-foundation.org, driver-core@lists.linux.dev
> 主题: Re: [PATCH] kobject: fix out-of-bounds read in action parser
> 
> I was looking at reviewing your patch. Unfortunately, it doesn't appear
> to apply cleanly; I'm seeing:
> 
> Applying: kobject: fix out-of-bounds read in action parser
> error: corrupt patch at .git/rebase-apply/patch:10
> Patch failed at 0001 kobject: fix out-of-bounds read in action parser
> 
> It looks like sashiko failed to apply it as well[1]. Perhaps you should
> reformat and resend?
> 
> On Tue, Aug 18, 2026 at 02:26:14PM +0800, Jiacheng Xu wrote:
> > kobject_action_type() uses the position of the first space in the input
> > as the length of the uevent action.  It then checks the byte at that
> > position in the corresponding action string.
> > 
> > An embedded NUL byte can make strncmp() report a match while
> > count_first is already greater than the actual length of the action
> > string.  For example, the input "bind\0 ..." makes the parser access
> > kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
> > string.
> > 
> > Use strlen() to verify that the input action length exactly matches the
> > known action string before accepting the match.  This avoids indexing
> > the action string with an out-of-bounds offset.
> > 
> > Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
> > Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
> > ---
> > lib/kobject_uevent.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> > 
> > diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
> > index ddbc4d7482d2..b03562301bfe 100644
> > --- a/lib/kobject_uevent.c
> > +++ b/lib/kobject_uevent.c
> > @@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
> > for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
> >         if (strncmp(kobject_actions[action], buf, count_first) != 0)
> >                 continue;
> > -             if (kobject_actions[action][count_first] != '\0')
> > +             if (strlen(kobject_actions[action]) != count_first)
> >                 continue;
> >         if (args)
> >                 *args = args_start;
> 
> [1] https://sashiko.dev/#/patchset/200d27f7.1471a.1a0138c92da.Coremail.stitch%40zju.edu.cn

Re: [PATCH v2] kobject: fix out-of-bounds read in action parser

[Greg KH]

On Tue, Aug 25, 2026 at 11:57:39AM +0800, Jiacheng Xu wrote:
> kobject_action_type() uses the position of the first space in the input
> as the length of the uevent action. It then checks the byte at that
> position in the corresponding action string.

<snip>

This is being sent in base64, and the tools can not apply it :(

Please fix your email client to not compress/encode the message.

thanks,

greg k-h

[PATCH] kobject: fix out-of-bounds read in action parser

[Jiacheng Xu]

kobject_action_type() uses the position of the first space in the input
as the length of the uevent action. It then checks the byte at that
position in the corresponding action string.

An embedded NUL byte can make strncmp() report a match while
count_first is already greater than the actual length of the action
string. For example, the input "bind\0 ..." makes the parser access
kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
string.

Use strlen() to verify that the input action length exactly matches the
known action string before accepting the match. This avoids indexing
the action string with an out-of-bounds offset.

Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
---
 lib/kobject_uevent.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
index ddbc4d7482d2..b03562301bfe 100644
--- a/lib/kobject_uevent.c
+++ b/lib/kobject_uevent.c
@@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
 	for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
 		if (strncmp(kobject_actions[action], buf, count_first) != 0)
 			continue;
-		if (kobject_actions[action][count_first] != '\0')
+		if (strlen(kobject_actions[action]) != count_first)
 			continue;
 		if (args)
 			*args = args_start;
-- 
2.51.0

Re: [PATCH] kobject: fix out-of-bounds read in action parser

[Greg KH]

On Tue, Aug 25, 2026 at 02:26:20PM +0800, Jiacheng Xu wrote:
> kobject_action_type() uses the position of the first space in the input
> as the length of the uevent action. It then checks the byte at that
> position in the corresponding action string.
> 
> An embedded NUL byte can make strncmp() report a match while
> count_first is already greater than the actual length of the action
> string. For example, the input "bind\0 ..." makes the parser access
> kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
> string.
> 
> Use strlen() to verify that the input action length exactly matches the
> known action string before accepting the match. This avoids indexing
> the action string with an out-of-bounds offset.
> 
> Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
> Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
> ---
>  lib/kobject_uevent.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
> index ddbc4d7482d2..b03562301bfe 100644
> --- a/lib/kobject_uevent.c
> +++ b/lib/kobject_uevent.c
> @@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
>  	for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
>  		if (strncmp(kobject_actions[action], buf, count_first) != 0)
>  			continue;
> -		if (kobject_actions[action][count_first] != '\0')
> +		if (strlen(kobject_actions[action]) != count_first)
>  			continue;
>  		if (args)
>  			*args = args_start;
> -- 
> 2.51.0
> 
> 

Hi,

This is the friendly patch-bot of Greg Kroah-Hartman.  You have sent him
a patch that has triggered this response.  He used to manually respond
to these common problems, but in order to save his sanity (he kept
writing the same thing over and over, yet to different people), I was
created.  Hopefully you will not take offence and will fix the problem
in your patch and resubmit it so that it can be accepted into the Linux
kernel tree.

You are receiving this message because of the following common error(s)
as indicated below:

- This looks like a new version of a previously submitted patch, but you
  did not list below the --- line any changes from the previous version.
  Please read the section entitled "The canonical patch format" in the
  kernel file, Documentation/process/submitting-patches.rst for what
  needs to be done here to properly describe this.

If you wish to discuss this problem further, or you have questions about
how to resolve this issue, please feel free to respond to this email and
Greg will reply once he has dug out from the pending patches received
from other developers.

thanks,

greg k-h's patch email bot

[PATCH v2 RESEND] kobject: fix out-of-bounds access in kobject_action_type()

[Jiacheng Xu]

kobject_action_type() uses the position of the first space in the input
as the length of the uevent action. It then checks the byte at that
position in the corresponding action string.

An embedded NUL byte can make strncmp() report a match while
count_first is already greater than the actual length of the action
string. For example, the input "bind\0 ..." makes the parser access
kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
string.

Use strlen() to verify that the input action length exactly matches the
known action string before accepting the match. This avoids indexing
the action string with an out-of-bounds offset.

Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
---
 lib/kobject_uevent.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
index ddbc4d7482d2..b03562301bfe 100644
--- a/lib/kobject_uevent.c
+++ b/lib/kobject_uevent.c
@@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
 	for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
 		if (strncmp(kobject_actions[action], buf, count_first) != 0)
 			continue;
-		if (kobject_actions[action][count_first] != '\0')
+		if (strlen(kobject_actions[action]) != count_first)
 			continue;
 		if (args)
 			*args = args_start;
-- 
2.51.0

Re: [PATCH v2 RESEND] kobject: fix out-of-bounds access in kobject_action_type()

[Greg KH]

On Tue, Aug 25, 2026 at 03:05:25PM +0800, Jiacheng Xu wrote:
> kobject_action_type() uses the position of the first space in the input
> as the length of the uevent action. It then checks the byte at that
> position in the corresponding action string.
> 
> An embedded NUL byte can make strncmp() report a match while
> count_first is already greater than the actual length of the action
> string. For example, the input "bind\0 ..." makes the parser access
> kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
> string.
> 
> Use strlen() to verify that the input action length exactly matches the
> known action string before accepting the match. This avoids indexing
> the action string with an out-of-bounds offset.
> 
> Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")
> Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>
> ---
>  lib/kobject_uevent.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
> index ddbc4d7482d2..b03562301bfe 100644
> --- a/lib/kobject_uevent.c
> +++ b/lib/kobject_uevent.c
> @@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
>  	for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
>  		if (strncmp(kobject_actions[action], buf, count_first) != 0)
>  			continue;
> -		if (kobject_actions[action][count_first] != '\0')
> +		if (strlen(kobject_actions[action]) != count_first)
>  			continue;
>  		if (args)
>  			*args = args_start;
> -- 
> 2.51.0
> 

Hi,

This is the friendly patch-bot of Greg Kroah-Hartman.  You have sent him
a patch that has triggered this response.  He used to manually respond
to these common problems, but in order to save his sanity (he kept
writing the same thing over and over, yet to different people), I was
created.  Hopefully you will not take offence and will fix the problem
in your patch and resubmit it so that it can be accepted into the Linux
kernel tree.

You are receiving this message because of the following common error(s)
as indicated below:

- This looks like a new version of a previously submitted patch, but you
  did not list below the --- line any changes from the previous version.
  Please read the section entitled "The canonical patch format" in the
  kernel file, Documentation/process/submitting-patches.rst for what
  needs to be done here to properly describe this.

If you wish to discuss this problem further, or you have questions about
how to resolve this issue, please feel free to respond to this email and
Greg will reply once he has dug out from the pending patches received
from other developers.

thanks,

greg k-h's patch email bot

Re: [PATCH v2 RESEND] kobject: fix out-of-bounds access in kobject_action_type()

[Vishal Moola (Fractile)]

On Tue, Aug 25, 2026 at 03:05:25PM +0800, Jiacheng Xu wrote:
> kobject_action_type() uses the position of the first space in the input
> as the length of the uevent action. It then checks the byte at that
> position in the corresponding action string.
> 
> An embedded NUL byte can make strncmp() report a match while
> count_first is already greater than the actual length of the action
> string. For example, the input "bind\0 ..." makes the parser access
> kobject_actions[KOBJ_BIND][5], which is beyond the end of the "bind"
> string.
> 
> Use strlen() to verify that the input action length exactly matches the
> known action string before accepting the match. This avoids indexing
> the action string with an out-of-bounds offset.

Sounds sensible to me. This doesn't seem like something anyone should
intentionally be doing, and I assume users won't care about the additional
minor overhead of strlen().

From what I can tell, this does change the behavior for actions with Nul
embedded strings from 'undefined' to 'expected to fail.' I think that's worth
mentioning in the commit log.

> Fixes: f36776fafbaa ("kobject: support passing in variables for synthetic uevents")

Also, it appears the pattern predates this commit, so definitely double
check that tag :)

> Signed-off-by: Jiacheng Xu <stitch@zju.edu.cn>

With those mentioned, feel free to add my tag:

Reviewed-by: Vishal Moola (Fractile) <vishal.moola@gmail.com>

> ---
>  lib/kobject_uevent.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/lib/kobject_uevent.c b/lib/kobject_uevent.c
> index ddbc4d7482d2..b03562301bfe 100644
> --- a/lib/kobject_uevent.c
> +++ b/lib/kobject_uevent.c
> @@ -83,7 +83,7 @@ static int kobject_action_type(const char *buf, size_t count,
>  	for (action = 0; action < ARRAY_SIZE(kobject_actions); action++) {
>  		if (strncmp(kobject_actions[action], buf, count_first) != 0)
>  			continue;
> -		if (kobject_actions[action][count_first] != '\0')
> +		if (strlen(kobject_actions[action]) != count_first)
>  			continue;
>  		if (args)
>  			*args = args_start;
> -- 
> 2.51.0
>