2014-06-09 22:21:49

by Alexander Holler

[permalink] [raw]
Subject: HID: hid-sensor-hub: fix sleeping function called from invalid context

Hello,

while rebasing the dozens (currently around 60) of refused, ignored or
similiar bugfixes and patches I need to use a Linux kernel, I've noticed
that Jiri Kosinas patch here

https://lkml.org/lkml/2013/8/9/214

doesn't apply anymore.

Looking why, I've seen that a similiar patch from Srinivas Pandruvada
named like the topic ended up in the kernel.

Looking at the differences, I wonder if not all spin_lock() calls in
hid-sensor-hub.c should be changed into spin_lock_irqsave() like the
patch from Jiri Kosina did.

Regards,

Alexander Holler

(sorry for the duplicate message, I've forgotten to add the
wanna-be-reviewer-school named lkml to the recipients)


2014-06-09 22:40:41

by Jiri Kosina

[permalink] [raw]
Subject: Re: HID: hid-sensor-hub: fix sleeping function called from invalid context

On Tue, 10 Jun 2014, Alexander Holler wrote:

> Hello,
>
> while rebasing the dozens (currently around 60) of refused, ignored or
> similiar bugfixes and patches I need to use a Linux kernel, I've noticed that
> Jiri Kosinas patch here
>
> https://lkml.org/lkml/2013/8/9/214
>
> doesn't apply anymore.
>
> Looking why, I've seen that a similiar patch from Srinivas Pandruvada named
> like the topic ended up in the kernel.

Hi Alexander,

thanks for drawing this one from the boulevard of forgotten patches.

I am actually not aware of any patch by Srinivas that would conflict with
the one I've written -- could you please elaborate where you believe the
conflict is?

> Looking at the differences, I wonder if not all spin_lock() calls in
> hid-sensor-hub.c should be changed into spin_lock_irqsave() like the
> patch from Jiri Kosina did.

Could you please give me your Tested-by:? I'll then be happily queuing the
fix. Thanks.

> (sorry for the duplicate message, I've forgotten to add the
> wanna-be-reviewer-school named lkml to the recipients)

You want to make it to LWN quotes of the week, don't you? :)

--
Jiri Kosina
SUSE Labs

2014-06-09 23:14:01

by srinivas pandruvada

[permalink] [raw]
Subject: Re: HID: hid-sensor-hub: fix sleeping function called from invalid context

On 06/09/2014 03:40 PM, Jiri Kosina wrote:
> On Tue, 10 Jun 2014, Alexander Holler wrote:
>
>> Hello,
>>
>> while rebasing the dozens (currently around 60) of refused, ignored or
>> similiar bugfixes and patches I need to use a Linux kernel, I've noticed that
>> Jiri Kosinas patch here
>>
>> https://lkml.org/lkml/2013/8/9/214
>>
>> doesn't apply anymore.
>>
>> Looking why, I've seen that a similiar patch from Srinivas Pandruvada named
>> like the topic ended up in the kernel.
>
> Hi Alexander,
>
> thanks for drawing this one from the boulevard of forgotten patches.
>
> I am actually not aware of any patch by Srinivas that would conflict with
> the one I've written -- could you please elaborate where you believe the
> conflict is?
>
>> Looking at the differences, I wonder if not all spin_lock() calls in
>> hid-sensor-hub.c should be changed into spin_lock_irqsave() like the
>> patch from Jiri Kosina did.
>

There is a patch submitted by "[email protected]"
[PATCH] HID: hid-sensor-hub: Fix lockdep warning for dynamic callback locks.

Already doing this, what you suggested?

+Reyad


Thanks,
Srinivas


> Could you please give me your Tested-by:? I'll then be happily queuing the
> fix. Thanks.
>
>> (sorry for the duplicate message, I've forgotten to add the
>> wanna-be-reviewer-school named lkml to the recipients)
>
> You want to make it to LWN quotes of the week, don't you? :)
>

2014-06-09 23:20:00

by Jiri Kosina

[permalink] [raw]
Subject: Re: HID: hid-sensor-hub: fix sleeping function called from invalid context

On Mon, 9 Jun 2014, Srinivas Pandruvada wrote:

> > > while rebasing the dozens (currently around 60) of refused, ignored
> > > or similiar bugfixes and patches I need to use a Linux kernel, I've
> > > noticed that Jiri Kosinas patch here
> > >
> > > https://lkml.org/lkml/2013/8/9/214
> > >
> > > doesn't apply anymore.
> > >
> > > Looking why, I've seen that a similiar patch from Srinivas Pandruvada
> > > named
> > > like the topic ended up in the kernel.
> >
> > Hi Alexander,
> >
> > thanks for drawing this one from the boulevard of forgotten patches.
> >
> > I am actually not aware of any patch by Srinivas that would conflict with
> > the one I've written -- could you please elaborate where you believe the
> > conflict is?
> >
> > > Looking at the differences, I wonder if not all spin_lock() calls in
> > > hid-sensor-hub.c should be changed into spin_lock_irqsave() like the
> > > patch from Jiri Kosina did.
> >
>
> There is a patch submitted by "[email protected]"
> [PATCH] HID: hid-sensor-hub: Fix lockdep warning for dynamic callback locks.
>
> Already doing this, what you suggested?

I don't see this in my mailbox at all .. so where was this submitted to,
and who merged it to which tree?

--
Jiri Kosina
SUSE Labs

2014-06-10 08:19:05

by Alexander Holler

[permalink] [raw]
Subject: Re: HID: hid-sensor-hub: fix sleeping function called from invalid context

Am 10.06.2014 00:40, schrieb Jiri Kosina:
> On Tue, 10 Jun 2014, Alexander Holler wrote:
>
>> Hello,
>>
>> while rebasing the dozens (currently around 60) of refused, ignored or
>> similiar bugfixes and patches I need to use a Linux kernel, I've noticed that
>> Jiri Kosinas patch here
>>
>> https://lkml.org/lkml/2013/8/9/214
>>
>> doesn't apply anymore.
>>
>> Looking why, I've seen that a similiar patch from Srinivas Pandruvada named
>> like the topic ended up in the kernel.
>
> Hi Alexander,
>
> thanks for drawing this one from the boulevard of forgotten patches.
>
> I am actually not aware of any patch by Srinivas that would conflict with
> the one I've written -- could you please elaborate where you believe the
> conflict is?

I actually don't rebase from main-version to main-version (that doesn't
really work), but I'm using format-patch and git am. It's a bit more
strict when that's ok. I haven't looked in deep which changes in 3.15 do
conflict with the patch based on 3.14, but it't can't be that much.

>> Looking at the differences, I wonder if not all spin_lock() calls in
>> hid-sensor-hub.c should be changed into spin_lock_irqsave() like the
>> patch from Jiri Kosina did.
>
> Could you please give me your Tested-by:? I'll then be happily queuing the
> fix. Thanks.

Sure, but I already did that in the reply to your patch.

>> (sorry for the duplicate message, I've forgotten to add the
>> wanna-be-reviewer-school named lkml to the recipients)
>
> You want to make it to LWN quotes of the week, don't you? :)

No, I'm just too often the object of their first steps and that's why I
won't post any patches on lkml anymore. It might be fun for them to do
reviews on patches from innocent people, but those victims usually don't
share their view. Maybe I will choose full disclosure or similiar I ever
will have an important patch poor Linux users should really have a need for.

Regards,

Alexander Holler

2014-06-10 09:08:22

by Jiri Kosina

[permalink] [raw]
Subject: Re: HID: hid-sensor-hub: fix sleeping function called from invalid context

On Tue, 10 Jun 2014, Alexander Holler wrote:

> >> Looking at the differences, I wonder if not all spin_lock() calls in
> >> hid-sensor-hub.c should be changed into spin_lock_irqsave() like the
> >> patch from Jiri Kosina did.
> >
> > Could you please give me your Tested-by:? I'll then be happily queuing the
> > fix. Thanks.
>
> Sure, but I already did that in the reply to your patch.

I probably missed in all the other noise in that thread. I've now pushed
out the staged patch and will be pushing it to Linus for 3.16 still.
Thanks,

--
Jiri Kosina
SUSE Labs