Hi,
These two patches fix some minor error path mistakes in the device
module.
Changes
-------
v1->v2
* Add fixes tag
* Add imperative statement in the commit description
v2->v3
* Add a goto exit label kunit_device_register_internal
v3->v4
* Remove some changes requested by Marcus Elfring, as I was alerted he
is a known troll.
Wander Lairson Costa (2):
kunit: unregister the device on error
kunit: avoid memory leak on device register error
lib/kunit/device.c | 13 +++++++------
1 file changed, 7 insertions(+), 6 deletions(-)
--
2.44.0
kunit_init_device() should unregister the device on bus register error,
but mistakenly it tries to unregister the bus.
Unregister the device instead of the bus.
Signed-off-by: Wander Lairson Costa <[email protected]>
Fixes: d03c720e03bd ("kunit: Add APIs for managing devices")
---
lib/kunit/device.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/lib/kunit/device.c b/lib/kunit/device.c
index abc603730b8e..25c81ed465fb 100644
--- a/lib/kunit/device.c
+++ b/lib/kunit/device.c
@@ -51,7 +51,7 @@ int kunit_bus_init(void)
error = bus_register(&kunit_bus_type);
if (error)
- bus_unregister(&kunit_bus_type);
+ root_device_unregister(kunit_bus_device);
return error;
}
--
2.44.0
If the device register fails, free the allocated memory before
returning.
Signed-off-by: Wander Lairson Costa <[email protected]>
Fixes: d03c720e03bd ("kunit: Add APIs for managing devices")
---
lib/kunit/device.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/lib/kunit/device.c b/lib/kunit/device.c
index 25c81ed465fb..d8c09dcb3e79 100644
--- a/lib/kunit/device.c
+++ b/lib/kunit/device.c
@@ -131,6 +131,7 @@ static struct kunit_device *kunit_device_register_internal(struct kunit *test,
err = device_register(&kunit_dev->dev);
if (err) {
put_device(&kunit_dev->dev);
+ kfree(kunit_dev);
return ERR_PTR(err);
}
--
2.44.0
On Fri, Apr 19, 2024 at 10:25:02AM -0300, Wander Lairson Costa wrote:
> If the device register fails, free the allocated memory before
> returning.
>
> Signed-off-by: Wander Lairson Costa <[email protected]>
> Fixes: d03c720e03bd ("kunit: Add APIs for managing devices")
> ---
> lib/kunit/device.c | 1 +
> 1 file changed, 1 insertion(+)
>
> diff --git a/lib/kunit/device.c b/lib/kunit/device.c
> index 25c81ed465fb..d8c09dcb3e79 100644
> --- a/lib/kunit/device.c
> +++ b/lib/kunit/device.c
> @@ -131,6 +131,7 @@ static struct kunit_device *kunit_device_register_internal(struct kunit *test,
> err = device_register(&kunit_dev->dev);
> if (err) {
> put_device(&kunit_dev->dev);
> + kfree(kunit_dev);
This still looks wrong, the release function for the device should free
the memory here, not this kfree, as the reference count in the embedded
'struct device' handles the memory logic for the whole structure (if
not, then something is REALLY wrong...)
You _do_ have a release function for the device, right? If not, you
should be getting loud messages in the kernel log when releasing a
device here.
thanks,
greg k-h
On Fri, Apr 19, 2024 at 10:25:01AM -0300, Wander Lairson Costa wrote:
> kunit_init_device() should unregister the device on bus register error,
> but mistakenly it tries to unregister the bus.
>
> Unregister the device instead of the bus.
>
> Signed-off-by: Wander Lairson Costa <[email protected]>
> Fixes: d03c720e03bd ("kunit: Add APIs for managing devices")
Reviewed-by: Greg Kroah-Hartman <[email protected]>
On Fri, Apr 19, 2024 at 11:03 AM Greg Kroah-Hartman
<[email protected]> wrote:
>
> On Fri, Apr 19, 2024 at 10:25:02AM -0300, Wander Lairson Costa wrote:
> > If the device register fails, free the allocated memory before
> > returning.
> >
> > Signed-off-by: Wander Lairson Costa <[email protected]>
> > Fixes: d03c720e03bd ("kunit: Add APIs for managing devices")
> > ---
> > lib/kunit/device.c | 1 +
> > 1 file changed, 1 insertion(+)
> >
> > diff --git a/lib/kunit/device.c b/lib/kunit/device.c
> > index 25c81ed465fb..d8c09dcb3e79 100644
> > --- a/lib/kunit/device.c
> > +++ b/lib/kunit/device.c
> > @@ -131,6 +131,7 @@ static struct kunit_device *kunit_device_register_internal(struct kunit *test,
> > err = device_register(&kunit_dev->dev);
> > if (err) {
> > put_device(&kunit_dev->dev);
> > + kfree(kunit_dev);
>
> This still looks wrong, the release function for the device should free
> the memory here, not this kfree, as the reference count in the embedded
> 'struct device' handles the memory logic for the whole structure (if
> not, then something is REALLY wrong...)
>
> You _do_ have a release function for the device, right? If not, you
> should be getting loud messages in the kernel log when releasing a
> device here.
>
Ok, I got it. Yes, there is a release function. So this patch is
wrong, ignore it. Should I send a v5 with only the other patch?
> thanks,
>
> greg k-h
>
…
> * Remove some changes requested by Marcus Elfring,
I became curious how affected software components can evolve further.
> as I was alerted he is a known troll.
I would appreciate if this interpretation will be reconsidered somehow.
Regards,
Markus
> > kunit_init_device() should unregister the device on bus register error,
> > but mistakenly it tries to unregister the bus.
> >
> > Unregister the device instead of the bus.
…
> Reviewed-by: Greg Kroah-Hartman <[email protected]>
Would you ever like to distinguish hardware register errors from
item registration failures according to further improved commit messages?
Regards,
Markus
On Fri, Apr 19, 2024 at 06:32:05PM +0200, Markus Elfring wrote:
> > > kunit_init_device() should unregister the device on bus register error,
> > > but mistakenly it tries to unregister the bus.
> > >
> > > Unregister the device instead of the bus.
> …
> > Reviewed-by: Greg Kroah-Hartman <[email protected]>
>
> Would you ever like to distinguish hardware register errors from
> item registration failures according to further improved commit messages?
Hi,
This is the semi-friendly patch-bot of Greg Kroah-Hartman.
Markus, you seem to have sent a nonsensical or otherwise pointless
review comment to a patch submission on a Linux kernel developer mailing
list. I strongly suggest that you not do this anymore. Please do not
bother developers who are actively working to produce patches and
features with comments that, in the end, are a waste of time.
Patch submitter, please ignore Markus's suggestion; you do not need to
follow it at all. The person/bot/AI that sent it is being ignored by
almost all Linux kernel maintainers for having a persistent pattern of
behavior of producing distracting and pointless commentary, and
inability to adapt to feedback. Please feel free to also ignore emails
from them.
thanks,
greg k-h's patch email bot
On Fri, 19 Apr 2024 at 22:11, Wander Lairson Costa <[email protected]> wrote:
>
> On Fri, Apr 19, 2024 at 11:03 AM Greg Kroah-Hartman
> <[email protected]> wrote:
> >
> > On Fri, Apr 19, 2024 at 10:25:02AM -0300, Wander Lairson Costa wrote:
> > > If the device register fails, free the allocated memory before
> > > returning.
> > >
> > > Signed-off-by: Wander Lairson Costa <[email protected]>
> > > Fixes: d03c720e03bd ("kunit: Add APIs for managing devices")
> > > ---
> > > lib/kunit/device.c | 1 +
> > > 1 file changed, 1 insertion(+)
> > >
> > > diff --git a/lib/kunit/device.c b/lib/kunit/device.c
> > > index 25c81ed465fb..d8c09dcb3e79 100644
> > > --- a/lib/kunit/device.c
> > > +++ b/lib/kunit/device.c
> > > @@ -131,6 +131,7 @@ static struct kunit_device *kunit_device_register_internal(struct kunit *test,
> > > err = device_register(&kunit_dev->dev);
> > > if (err) {
> > > put_device(&kunit_dev->dev);
> > > + kfree(kunit_dev);
> >
> > This still looks wrong, the release function for the device should free
> > the memory here, not this kfree, as the reference count in the embedded
> > 'struct device' handles the memory logic for the whole structure (if
> > not, then something is REALLY wrong...)
> >
> > You _do_ have a release function for the device, right? If not, you
> > should be getting loud messages in the kernel log when releasing a
> > device here.
> >
>
> Ok, I got it. Yes, there is a release function. So this patch is
> wrong, ignore it. Should I send a v5 with only the other patch?
>
Thanks. Don't worry about sending a v5: just patch 1 of v4 is now in
the kunit branch:
https://git.kernel.org/pub/scm/linux/kernel/git/shuah/linux-kselftest.git/commit/?h=kunit&id=63761ec5971ea47c1f2d7698f03e1c6ffb9fb22a
Cheers,
-- David