[PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Alexey Klimov]

This patch fixes the calculation of pcc_chan for non-zero id.
After the compiler ignores the (unsigned long) cast the
pcc_mbox_channels pointer is type-cast and then the type-cast
offset is added which results in address outside of the range
leading to the kernel crashing.

We might add braces and make it:

pcc_chan = (struct mbox_chan *)
		((unsigned long) pcc_mbox_channels +
		(id * sizeof(*pcc_chan)));

but let's go with array approach here and use id as index.

Tested on Juno board.

Acked-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Alexey Klimov <alexey.klimov@arm.com>
---
 drivers/mailbox/pcc.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/drivers/mailbox/pcc.c b/drivers/mailbox/pcc.c
index 45d85ae..8f779a1 100644
--- a/drivers/mailbox/pcc.c
+++ b/drivers/mailbox/pcc.c
@@ -81,16 +81,10 @@ static struct mbox_controller pcc_mbox_ctrl = {};
  */
 static struct mbox_chan *get_pcc_channel(int id)
 {
-	struct mbox_chan *pcc_chan;
-
 	if (id < 0 || id > pcc_mbox_ctrl.num_chans)
 		return ERR_PTR(-ENOENT);
 
-	pcc_chan = (struct mbox_chan *)
-		(unsigned long) pcc_mbox_channels +
-		(id * sizeof(*pcc_chan));
-
-	return pcc_chan;
+	return &pcc_mbox_channels[id];
 }
 
 /**
-- 
1.9.1

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Ashwin Chaugule]

On 10 December 2015 at 12:28, Alexey Klimov <alexey.klimov@arm.com> wrote:
> This patch fixes the calculation of pcc_chan for non-zero id.
> After the compiler ignores the (unsigned long) cast the
> pcc_mbox_channels pointer is type-cast and then the type-cast
> offset is added which results in address outside of the range
> leading to the kernel crashing.
>
> We might add braces and make it:
>
> pcc_chan = (struct mbox_chan *)
>                 ((unsigned long) pcc_mbox_channels +
>                 (id * sizeof(*pcc_chan)));
>
> but let's go with array approach here and use id as index.
>
> Tested on Juno board.
>
> Acked-by: Sudeep Holla <sudeep.holla@arm.com>
> Signed-off-by: Alexey Klimov <alexey.klimov@arm.com>
> ---
>  drivers/mailbox/pcc.c | 8 +-------
>  1 file changed, 1 insertion(+), 7 deletions(-)
>
> diff --git a/drivers/mailbox/pcc.c b/drivers/mailbox/pcc.c
> index 45d85ae..8f779a1 100644
> --- a/drivers/mailbox/pcc.c
> +++ b/drivers/mailbox/pcc.c
> @@ -81,16 +81,10 @@ static struct mbox_controller pcc_mbox_ctrl = {};
>   */
>  static struct mbox_chan *get_pcc_channel(int id)
>  {
> -       struct mbox_chan *pcc_chan;
> -
>         if (id < 0 || id > pcc_mbox_ctrl.num_chans)
>                 return ERR_PTR(-ENOENT);
>
> -       pcc_chan = (struct mbox_chan *)
> -               (unsigned long) pcc_mbox_channels +
> -               (id * sizeof(*pcc_chan));
> -
> -       return pcc_chan;
> +       return &pcc_mbox_channels[id];
>  }
>


Strange that we didn't catch this even with a non-zero id. But the
change makes sense so..

Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>

Thanks,
Ashwin.

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Ashwin Chaugule]

Jassi,

On 10 December 2015 at 13:19, Ashwin Chaugule
<ashwin.chaugule@linaro.org> wrote:
> On 10 December 2015 at 12:28, Alexey Klimov <alexey.klimov@arm.com> wrote:
>> This patch fixes the calculation of pcc_chan for non-zero id.
>> After the compiler ignores the (unsigned long) cast the
>> pcc_mbox_channels pointer is type-cast and then the type-cast
>> offset is added which results in address outside of the range
>> leading to the kernel crashing.
>>
>> We might add braces and make it:
>>
>> pcc_chan = (struct mbox_chan *)
>>                 ((unsigned long) pcc_mbox_channels +
>>                 (id * sizeof(*pcc_chan)));
>>
>> but let's go with array approach here and use id as index.
>>
>> Tested on Juno board.
>>
>> Acked-by: Sudeep Holla <sudeep.holla@arm.com>
>> Signed-off-by: Alexey Klimov <alexey.klimov@arm.com>
>> ---
>>  drivers/mailbox/pcc.c | 8 +-------
>>  1 file changed, 1 insertion(+), 7 deletions(-)
>>
>> diff --git a/drivers/mailbox/pcc.c b/drivers/mailbox/pcc.c
>> index 45d85ae..8f779a1 100644
>> --- a/drivers/mailbox/pcc.c
>> +++ b/drivers/mailbox/pcc.c
>> @@ -81,16 +81,10 @@ static struct mbox_controller pcc_mbox_ctrl = {};
>>   */
>>  static struct mbox_chan *get_pcc_channel(int id)
>>  {
>> -       struct mbox_chan *pcc_chan;
>> -
>>         if (id < 0 || id > pcc_mbox_ctrl.num_chans)
>>                 return ERR_PTR(-ENOENT);
>>
>> -       pcc_chan = (struct mbox_chan *)
>> -               (unsigned long) pcc_mbox_channels +
>> -               (id * sizeof(*pcc_chan));
>> -
>> -       return pcc_chan;
>> +       return &pcc_mbox_channels[id];
>>  }
>>
>
>
> Strange that we didn't catch this even with a non-zero id. But the
> change makes sense so..
>
> Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>

Can you please include this patch in your pull request to Linus?

Thanks,
Ashwin.

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Ashwin Chaugule]

+ Jassi (Linaro addr)

On 15 January 2016 at 13:20, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
> Jassi,
>
> On 10 December 2015 at 13:19, Ashwin Chaugule
> <ashwin.chaugule@linaro.org> wrote:
>> On 10 December 2015 at 12:28, Alexey Klimov <alexey.klimov@arm.com> wrote:
>>> This patch fixes the calculation of pcc_chan for non-zero id.
>>> After the compiler ignores the (unsigned long) cast the
>>> pcc_mbox_channels pointer is type-cast and then the type-cast
>>> offset is added which results in address outside of the range
>>> leading to the kernel crashing.
>>>
>>> We might add braces and make it:
>>>
>>> pcc_chan = (struct mbox_chan *)
>>>                 ((unsigned long) pcc_mbox_channels +
>>>                 (id * sizeof(*pcc_chan)));
>>>
>>> but let's go with array approach here and use id as index.
>>>
>>> Tested on Juno board.
>>>
>>> Acked-by: Sudeep Holla <sudeep.holla@arm.com>
>>> Signed-off-by: Alexey Klimov <alexey.klimov@arm.com>
>>> ---
>>>  drivers/mailbox/pcc.c | 8 +-------
>>>  1 file changed, 1 insertion(+), 7 deletions(-)
>>>
>>> diff --git a/drivers/mailbox/pcc.c b/drivers/mailbox/pcc.c
>>> index 45d85ae..8f779a1 100644
>>> --- a/drivers/mailbox/pcc.c
>>> +++ b/drivers/mailbox/pcc.c
>>> @@ -81,16 +81,10 @@ static struct mbox_controller pcc_mbox_ctrl = {};
>>>   */
>>>  static struct mbox_chan *get_pcc_channel(int id)
>>>  {
>>> -       struct mbox_chan *pcc_chan;
>>> -
>>>         if (id < 0 || id > pcc_mbox_ctrl.num_chans)
>>>                 return ERR_PTR(-ENOENT);
>>>
>>> -       pcc_chan = (struct mbox_chan *)
>>> -               (unsigned long) pcc_mbox_channels +
>>> -               (id * sizeof(*pcc_chan));
>>> -
>>> -       return pcc_chan;
>>> +       return &pcc_mbox_channels[id];
>>>  }
>>>
>>
>>
>> Strange that we didn't catch this even with a non-zero id. But the
>> change makes sense so..
>>
>> Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
>
> Can you please include this patch in your pull request to Linus?
>
> Thanks,
> Ashwin.

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Alexey Klimov]

(adding Rafael and linux-acpi)

On Fri, Jan 15, 2016 at 6:22 PM, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
> + Jassi (Linaro addr)
>
> On 15 January 2016 at 13:20, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
>> Jassi,
>>
>> On 10 December 2015 at 13:19, Ashwin Chaugule
>> <ashwin.chaugule@linaro.org> wrote:
>>> On 10 December 2015 at 12:28, Alexey Klimov <alexey.klimov@arm.com> wrote:
>>>> This patch fixes the calculation of pcc_chan for non-zero id.
>>>> After the compiler ignores the (unsigned long) cast the
>>>> pcc_mbox_channels pointer is type-cast and then the type-cast
>>>> offset is added which results in address outside of the range
>>>> leading to the kernel crashing.
>>>>
>>>> We might add braces and make it:
>>>>
>>>> pcc_chan = (struct mbox_chan *)
>>>>                 ((unsigned long) pcc_mbox_channels +
>>>>                 (id * sizeof(*pcc_chan)));
>>>>
>>>> but let's go with array approach here and use id as index.
>>>>
>>>> Tested on Juno board.
>>>>
>>>> Acked-by: Sudeep Holla <sudeep.holla@arm.com>
>>>> Signed-off-by: Alexey Klimov <alexey.klimov@arm.com>
>>>> ---
>>>>  drivers/mailbox/pcc.c | 8 +-------
>>>>  1 file changed, 1 insertion(+), 7 deletions(-)
>>>>
>>>> diff --git a/drivers/mailbox/pcc.c b/drivers/mailbox/pcc.c
>>>> index 45d85ae..8f779a1 100644
>>>> --- a/drivers/mailbox/pcc.c
>>>> +++ b/drivers/mailbox/pcc.c
>>>> @@ -81,16 +81,10 @@ static struct mbox_controller pcc_mbox_ctrl = {};
>>>>   */
>>>>  static struct mbox_chan *get_pcc_channel(int id)
>>>>  {
>>>> -       struct mbox_chan *pcc_chan;
>>>> -
>>>>         if (id < 0 || id > pcc_mbox_ctrl.num_chans)
>>>>                 return ERR_PTR(-ENOENT);
>>>>
>>>> -       pcc_chan = (struct mbox_chan *)
>>>> -               (unsigned long) pcc_mbox_channels +
>>>> -               (id * sizeof(*pcc_chan));
>>>> -
>>>> -       return pcc_chan;
>>>> +       return &pcc_mbox_channels[id];
>>>>  }
>>>>
>>>
>>>
>>> Strange that we didn't catch this even with a non-zero id. But the
>>> change makes sense so..
>>>
>>> Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
>>
>> Can you please include this patch in your pull request to Linus?

Hi Rafael,

any chance you can take this fix via your tree?
I can resend patch if you want.
Looks like Jassi doesn't have time or doesn't care.

Best regards,
Alexey

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Jassi Brar]

On Mon, Feb 1, 2016 at 8:49 PM, Alexey Klimov <alexey.klimov@arm.com> wrote:
> (adding Rafael and linux-acpi)
>
> On Fri, Jan 15, 2016 at 6:22 PM, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
>> + Jassi (Linaro addr)
>>
>> On 15 January 2016 at 13:20, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
>>> Jassi,
>>>
>>> On 10 December 2015 at 13:19, Ashwin Chaugule
>>> <ashwin.chaugule@linaro.org> wrote:
>>>> On 10 December 2015 at 12:28, Alexey Klimov <alexey.klimov@arm.com> wrote:
>>>>> This patch fixes the calculation of pcc_chan for non-zero id.
>>>>> After the compiler ignores the (unsigned long) cast the
>>>>> pcc_mbox_channels pointer is type-cast and then the type-cast
>>>>> offset is added which results in address outside of the range
>>>>> leading to the kernel crashing.
>>>>>
>>>>> We might add braces and make it:
>>>>>
>>>>> pcc_chan = (struct mbox_chan *)
>>>>>                 ((unsigned long) pcc_mbox_channels +
>>>>>                 (id * sizeof(*pcc_chan)));
>>>>>
>>>>> but let's go with array approach here and use id as index.
>>>>>
>>>>> Tested on Juno board.
>>>>>
>>>>> Acked-by: Sudeep Holla <sudeep.holla@arm.com>
>>>>> Signed-off-by: Alexey Klimov <alexey.klimov@arm.com>
>>>>> ---
>>>>>  drivers/mailbox/pcc.c | 8 +-------
>>>>>  1 file changed, 1 insertion(+), 7 deletions(-)
>>>>>
>>>>> diff --git a/drivers/mailbox/pcc.c b/drivers/mailbox/pcc.c
>>>>> index 45d85ae..8f779a1 100644
>>>>> --- a/drivers/mailbox/pcc.c
>>>>> +++ b/drivers/mailbox/pcc.c
>>>>> @@ -81,16 +81,10 @@ static struct mbox_controller pcc_mbox_ctrl = {};
>>>>>   */
>>>>>  static struct mbox_chan *get_pcc_channel(int id)
>>>>>  {
>>>>> -       struct mbox_chan *pcc_chan;
>>>>> -
>>>>>         if (id < 0 || id > pcc_mbox_ctrl.num_chans)
>>>>>                 return ERR_PTR(-ENOENT);
>>>>>
>>>>> -       pcc_chan = (struct mbox_chan *)
>>>>> -               (unsigned long) pcc_mbox_channels +
>>>>> -               (id * sizeof(*pcc_chan));
>>>>> -
>>>>> -       return pcc_chan;
>>>>> +       return &pcc_mbox_channels[id];
>>>>>  }
>>>>>
>>>>
>>>>
>>>> Strange that we didn't catch this even with a non-zero id. But the
>>>> change makes sense so..
>>>>
>>>> Acked-by: Ashwin Chaugule <ashwin.chaugule@linaro.org>
>>>
>>> Can you please include this patch in your pull request to Linus?
>
> Hi Rafael,
>
> any chance you can take this fix via your tree?
> I can resend patch if you want.
> Looks like Jassi doesn't have time or doesn't care.
>
Ouch... sorry. Ashwin's reminder came after I had sent the pull
request. And then it just got overlooked. If Rafael doesn't pick it as
my punishment, I'll tomorow :)

Cheers!

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Rafael J. Wysocki]

On Mon, Feb 1, 2016 at 4:43 PM, Jassi Brar <jassisinghbrar@gmail.com> wrote:
> On Mon, Feb 1, 2016 at 8:49 PM, Alexey Klimov <alexey.klimov@arm.com> wrote:
>> (adding Rafael and linux-acpi)
>>
>> On Fri, Jan 15, 2016 at 6:22 PM, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
>>> + Jassi (Linaro addr)
>>>
>>> On 15 January 2016 at 13:20, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:

[cut]

>>>> Can you please include this patch in your pull request to Linus?
>>
>> Hi Rafael,
>>
>> any chance you can take this fix via your tree?
>> I can resend patch if you want.
>> Looks like Jassi doesn't have time or doesn't care.
>>
> Ouch... sorry. Ashwin's reminder came after I had sent the pull
> request. And then it just got overlooked. If Rafael doesn't pick it as
> my punishment, I'll tomorow :)

I can apply it, but can anyone please point me to the original patch?

Or, maybe better, resend it with a CC to linux-acpi and all of the
tags collected so far?

Thanks,
Rafael

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Jassi Brar]

On Tue, Feb 2, 2016 at 2:21 AM, Rafael J. Wysocki <rafael@kernel.org> wrote:
> On Mon, Feb 1, 2016 at 4:43 PM, Jassi Brar <jassisinghbrar@gmail.com> wrote:
>> On Mon, Feb 1, 2016 at 8:49 PM, Alexey Klimov <alexey.klimov@arm.com> wrote:
>>> (adding Rafael and linux-acpi)
>>>
>>> On Fri, Jan 15, 2016 at 6:22 PM, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
>>>> + Jassi (Linaro addr)
>>>>
>>>> On 15 January 2016 at 13:20, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
>
> [cut]
>
>>>>> Can you please include this patch in your pull request to Linus?
>>>
>>> Hi Rafael,
>>>
>>> any chance you can take this fix via your tree?
>>> I can resend patch if you want.
>>> Looks like Jassi doesn't have time or doesn't care.
>>>
>> Ouch... sorry. Ashwin's reminder came after I had sent the pull
>> request. And then it just got overlooked. If Rafael doesn't pick it as
>> my punishment, I'll tomorow :)
>
> I can apply it, but can anyone please point me to the original patch?
>
> Or, maybe better, resend it with a CC to linux-acpi and all of the
> tags collected so far?
>
or even simpler ... I queued it already
https://git.linaro.org/landing-teams/working/fujitsu/integration.git/log/refs/heads/mailbox-devel
 ?

Thanks.

Re: [PATCH] mailbox: pcc: fix channel calculation in get_pcc_channel()

[Rafael J. Wysocki]

On Tuesday, February 02, 2016 04:42:52 PM Jassi Brar wrote:
> On Tue, Feb 2, 2016 at 2:21 AM, Rafael J. Wysocki <rafael@kernel.org> wrote:
> > On Mon, Feb 1, 2016 at 4:43 PM, Jassi Brar <jassisinghbrar@gmail.com> wrote:
> >> On Mon, Feb 1, 2016 at 8:49 PM, Alexey Klimov <alexey.klimov@arm.com> wrote:
> >>> (adding Rafael and linux-acpi)
> >>>
> >>> On Fri, Jan 15, 2016 at 6:22 PM, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
> >>>> + Jassi (Linaro addr)
> >>>>
> >>>> On 15 January 2016 at 13:20, Ashwin Chaugule <ashwin.chaugule@linaro.org> wrote:
> >
> > [cut]
> >
> >>>>> Can you please include this patch in your pull request to Linus?
> >>>
> >>> Hi Rafael,
> >>>
> >>> any chance you can take this fix via your tree?
> >>> I can resend patch if you want.
> >>> Looks like Jassi doesn't have time or doesn't care.
> >>>
> >> Ouch... sorry. Ashwin's reminder came after I had sent the pull
> >> request. And then it just got overlooked. If Rafael doesn't pick it as
> >> my punishment, I'll tomorow :)
> >
> > I can apply it, but can anyone please point me to the original patch?
> >
> > Or, maybe better, resend it with a CC to linux-acpi and all of the
> > tags collected so far?
> >
> or even simpler ... I queued it already
> https://git.linaro.org/landing-teams/working/fujitsu/integration.git/log/refs/heads/mailbox-devel
>  ?

That works for me too. :-)