[RFC] gpiolib: fix oops, if gpio name is NULL
diff mbox

Message ID 1447245413-32209-1-git-send-email-vz@mleia.com
State New
Headers show

Commit Message

Vladimir Zapolskiy Nov. 11, 2015, 12:36 p.m. UTC
Commit c0017ed71966 ("gpio: Introduce gpio descriptor 'name'") causes
OOPS on boot on LPC32xx boards:

    Unable to handle kernel NULL pointer dereference at virtual address 00000000
    CPU: 0 PID: 1 Comm: swapper Not tainted 4.3.0+ #707
    Hardware name: LPC32XX SoC (Flattened Device Tree)
    task: c381baa0 ti: c381e000 task.ti: c381e000
    PC is at strcmp+0x10/0x40
    LR is at gpiochip_add+0x3d0/0x4d4
    pc : [<>]    lr : [<>]    psr: a0000093
    sp : c381fd60  ip : c381fd70  fp : c381fd6c

    [snip]

    Backtrace:
    [<>] (strcmp) from [<>] (gpiochip_add+0x3d0/0x4d4)
    [<>] (gpiochip_add) from [<>] (lpc32xx_gpio_probe+0x44/0x60)
    [<>] (lpc32xx_gpio_probe) from [<>] (platform_drv_probe+0x40/0x8c)
    [<>] (platform_drv_probe) from [<>] (driver_probe_device+0x110/0x294)
    [<>] (driver_probe_device) from [<>] (__driver_attach+0x70/0x94)
    [<>] (__driver_attach) from [<>] (bus_for_each_dev+0x74/0x98)
    [<>] (bus_for_each_dev) from [<>] (driver_attach+0x20/0x28)
    [<>] (driver_attach) from [<>] (bus_add_driver+0xd4/0x1f0)
    [<>] (bus_add_driver) from [<>] (driver_register+0xa4/0xe8)
    [<>] (driver_register) from [<>] (__platform_driver_register+0x38/0x4c)
    [<>] (__platform_driver_register) from [<>] (lpc32xx_gpio_driver_init+0x18/0x20)
    [<>] (lpc32xx_gpio_driver_init) from [<>] (do_one_initcall+0x108/0x1c8)
    [<>] (do_one_initcall) from [<>] (kernel_init_freeable+0x10c/0x1d4)
    [<>] (kernel_init_freeable) from [<>] (kernel_init+0x10/0xec)
    [<>] (kernel_init) from [<>] (ret_from_fork+0x14/0x24)

This is caused by the fact that at the moment some GPIO names are set
to NULL, there is a hole in linear representation of one GPI bank, see
drivers/gpio/gpio-lpc32xx.c / gpi_p3_names[] for details.

The same problem most probably affects also gpio-cs5535.c, see
cs5535_gpio_names[].

Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
---
Linus, Markus,

I understand that LPC32xx GPIO driver is ugly (hopefully this
will be changed in future), but please account this problem.

I don't ask for inclusion of this change, because my own analysis is
not done completely.

Also please review, the same problem probably affects gpio-cs5535.c
as well, see cs5535_gpio_names[].

 drivers/gpio/gpiolib.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Comments

Markus Pargmann Nov. 17, 2015, 9:24 a.m. UTC | #1
Hi,

On Wednesday 11 November 2015 14:36:53 Vladimir Zapolskiy wrote:
> Commit c0017ed71966 ("gpio: Introduce gpio descriptor 'name'") causes
> OOPS on boot on LPC32xx boards:
> 
>     Unable to handle kernel NULL pointer dereference at virtual address 00000000
>     CPU: 0 PID: 1 Comm: swapper Not tainted 4.3.0+ #707
>     Hardware name: LPC32XX SoC (Flattened Device Tree)
>     task: c381baa0 ti: c381e000 task.ti: c381e000
>     PC is at strcmp+0x10/0x40
>     LR is at gpiochip_add+0x3d0/0x4d4
>     pc : [<>]    lr : [<>]    psr: a0000093
>     sp : c381fd60  ip : c381fd70  fp : c381fd6c
> 
>     [snip]
> 
>     Backtrace:
>     [<>] (strcmp) from [<>] (gpiochip_add+0x3d0/0x4d4)
>     [<>] (gpiochip_add) from [<>] (lpc32xx_gpio_probe+0x44/0x60)
>     [<>] (lpc32xx_gpio_probe) from [<>] (platform_drv_probe+0x40/0x8c)
>     [<>] (platform_drv_probe) from [<>] (driver_probe_device+0x110/0x294)
>     [<>] (driver_probe_device) from [<>] (__driver_attach+0x70/0x94)
>     [<>] (__driver_attach) from [<>] (bus_for_each_dev+0x74/0x98)
>     [<>] (bus_for_each_dev) from [<>] (driver_attach+0x20/0x28)
>     [<>] (driver_attach) from [<>] (bus_add_driver+0xd4/0x1f0)
>     [<>] (bus_add_driver) from [<>] (driver_register+0xa4/0xe8)
>     [<>] (driver_register) from [<>] (__platform_driver_register+0x38/0x4c)
>     [<>] (__platform_driver_register) from [<>] (lpc32xx_gpio_driver_init+0x18/0x20)
>     [<>] (lpc32xx_gpio_driver_init) from [<>] (do_one_initcall+0x108/0x1c8)
>     [<>] (do_one_initcall) from [<>] (kernel_init_freeable+0x10c/0x1d4)
>     [<>] (kernel_init_freeable) from [<>] (kernel_init+0x10/0xec)
>     [<>] (kernel_init) from [<>] (ret_from_fork+0x14/0x24)
> 
> This is caused by the fact that at the moment some GPIO names are set
> to NULL, there is a hole in linear representation of one GPI bank, see
> drivers/gpio/gpio-lpc32xx.c / gpi_p3_names[] for details.
> 
> The same problem most probably affects also gpio-cs5535.c, see
> cs5535_gpio_names[].
> 
> Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
> ---
> Linus, Markus,
> 
> I understand that LPC32xx GPIO driver is ugly (hopefully this
> will be changed in future), but please account this problem.
> 
> I don't ask for inclusion of this change, because my own analysis is
> not done completely.
> 
> Also please review, the same problem probably affects gpio-cs5535.c
> as well, see cs5535_gpio_names[].

Thanks for this hint. But I think these patches for gpio_name_to_desc won't get
mainline any time soon as this depends on the chardev interface and how it will
look like.

Best Regards,

Markus

> 
>  drivers/gpio/gpiolib.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> index a18f00f..2a91f32 100644
> --- a/drivers/gpio/gpiolib.c
> +++ b/drivers/gpio/gpiolib.c
> @@ -233,7 +233,7 @@ static struct gpio_desc *gpio_name_to_desc(const char * const name)
>  		for (i = 0; i != chip->ngpio; ++i) {
>  			struct gpio_desc *gpio = &chip->desc[i];
>  
> -			if (!gpio->name)
> +			if (!gpio->name || !name)
>  				continue;
>  
>  			if (!strcmp(gpio->name, name)) {
>
Vladimir Zapolskiy Nov. 17, 2015, 11:20 a.m. UTC | #2
Hi Markus,

On 17.11.2015 11:24, Markus Pargmann wrote:
> Hi,
> 
> On Wednesday 11 November 2015 14:36:53 Vladimir Zapolskiy wrote:
>> Commit c0017ed71966 ("gpio: Introduce gpio descriptor 'name'") causes
>> OOPS on boot on LPC32xx boards:
>>
>>     Unable to handle kernel NULL pointer dereference at virtual address 00000000
>>     CPU: 0 PID: 1 Comm: swapper Not tainted 4.3.0+ #707
>>     Hardware name: LPC32XX SoC (Flattened Device Tree)
>>     task: c381baa0 ti: c381e000 task.ti: c381e000
>>     PC is at strcmp+0x10/0x40
>>     LR is at gpiochip_add+0x3d0/0x4d4
>>     pc : [<>]    lr : [<>]    psr: a0000093
>>     sp : c381fd60  ip : c381fd70  fp : c381fd6c
>>
>>     [snip]
>>
>>     Backtrace:
>>     [<>] (strcmp) from [<>] (gpiochip_add+0x3d0/0x4d4)
>>     [<>] (gpiochip_add) from [<>] (lpc32xx_gpio_probe+0x44/0x60)
>>     [<>] (lpc32xx_gpio_probe) from [<>] (platform_drv_probe+0x40/0x8c)
>>     [<>] (platform_drv_probe) from [<>] (driver_probe_device+0x110/0x294)
>>     [<>] (driver_probe_device) from [<>] (__driver_attach+0x70/0x94)
>>     [<>] (__driver_attach) from [<>] (bus_for_each_dev+0x74/0x98)
>>     [<>] (bus_for_each_dev) from [<>] (driver_attach+0x20/0x28)
>>     [<>] (driver_attach) from [<>] (bus_add_driver+0xd4/0x1f0)
>>     [<>] (bus_add_driver) from [<>] (driver_register+0xa4/0xe8)
>>     [<>] (driver_register) from [<>] (__platform_driver_register+0x38/0x4c)
>>     [<>] (__platform_driver_register) from [<>] (lpc32xx_gpio_driver_init+0x18/0x20)
>>     [<>] (lpc32xx_gpio_driver_init) from [<>] (do_one_initcall+0x108/0x1c8)
>>     [<>] (do_one_initcall) from [<>] (kernel_init_freeable+0x10c/0x1d4)
>>     [<>] (kernel_init_freeable) from [<>] (kernel_init+0x10/0xec)
>>     [<>] (kernel_init) from [<>] (ret_from_fork+0x14/0x24)
>>
>> This is caused by the fact that at the moment some GPIO names are set
>> to NULL, there is a hole in linear representation of one GPI bank, see
>> drivers/gpio/gpio-lpc32xx.c / gpi_p3_names[] for details.
>>
>> The same problem most probably affects also gpio-cs5535.c, see
>> cs5535_gpio_names[].
>>
>> Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
>> ---
>> Linus, Markus,
>>
>> I understand that LPC32xx GPIO driver is ugly (hopefully this
>> will be changed in future), but please account this problem.
>>
>> I don't ask for inclusion of this change, because my own analysis is
>> not done completely.
>>
>> Also please review, the same problem probably affects gpio-cs5535.c
>> as well, see cs5535_gpio_names[].
> 
> Thanks for this hint. But I think these patches for gpio_name_to_desc won't get
> mainline any time soon as this depends on the chardev interface and how it will
> look like.
> 

sorry for misunderstanding, what do mean by "these patches"?

I see two options here, one is to fix or revert your c0017ed71966 due to
caused regressions, another one is to update affected GPIO drivers
removing discontinuity in GPIO enumeration inside a bank.

With best wishes,
Vladimir

>>
>>  drivers/gpio/gpiolib.c | 2 +-
>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
>> index a18f00f..2a91f32 100644
>> --- a/drivers/gpio/gpiolib.c
>> +++ b/drivers/gpio/gpiolib.c
>> @@ -233,7 +233,7 @@ static struct gpio_desc *gpio_name_to_desc(const char * const name)
>>  		for (i = 0; i != chip->ngpio; ++i) {
>>  			struct gpio_desc *gpio = &chip->desc[i];
>>  
>> -			if (!gpio->name)
>> +			if (!gpio->name || !name)
>>  				continue;
>>  
>>  			if (!strcmp(gpio->name, name)) {
>>
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Markus Pargmann Nov. 17, 2015, 11:50 a.m. UTC | #3
Hi Vladimir,

On Tuesday 17 November 2015 13:20:40 Vladimir Zapolskiy wrote:
> Hi Markus,
> 
> On 17.11.2015 11:24, Markus Pargmann wrote:
> > Hi,
> > 
> > On Wednesday 11 November 2015 14:36:53 Vladimir Zapolskiy wrote:
> >> Commit c0017ed71966 ("gpio: Introduce gpio descriptor 'name'") causes
> >> OOPS on boot on LPC32xx boards:
> >>
> >>     Unable to handle kernel NULL pointer dereference at virtual address 00000000
> >>     CPU: 0 PID: 1 Comm: swapper Not tainted 4.3.0+ #707
> >>     Hardware name: LPC32XX SoC (Flattened Device Tree)
> >>     task: c381baa0 ti: c381e000 task.ti: c381e000
> >>     PC is at strcmp+0x10/0x40
> >>     LR is at gpiochip_add+0x3d0/0x4d4
> >>     pc : [<>]    lr : [<>]    psr: a0000093
> >>     sp : c381fd60  ip : c381fd70  fp : c381fd6c
> >>
> >>     [snip]
> >>
> >>     Backtrace:
> >>     [<>] (strcmp) from [<>] (gpiochip_add+0x3d0/0x4d4)
> >>     [<>] (gpiochip_add) from [<>] (lpc32xx_gpio_probe+0x44/0x60)
> >>     [<>] (lpc32xx_gpio_probe) from [<>] (platform_drv_probe+0x40/0x8c)
> >>     [<>] (platform_drv_probe) from [<>] (driver_probe_device+0x110/0x294)
> >>     [<>] (driver_probe_device) from [<>] (__driver_attach+0x70/0x94)
> >>     [<>] (__driver_attach) from [<>] (bus_for_each_dev+0x74/0x98)
> >>     [<>] (bus_for_each_dev) from [<>] (driver_attach+0x20/0x28)
> >>     [<>] (driver_attach) from [<>] (bus_add_driver+0xd4/0x1f0)
> >>     [<>] (bus_add_driver) from [<>] (driver_register+0xa4/0xe8)
> >>     [<>] (driver_register) from [<>] (__platform_driver_register+0x38/0x4c)
> >>     [<>] (__platform_driver_register) from [<>] (lpc32xx_gpio_driver_init+0x18/0x20)
> >>     [<>] (lpc32xx_gpio_driver_init) from [<>] (do_one_initcall+0x108/0x1c8)
> >>     [<>] (do_one_initcall) from [<>] (kernel_init_freeable+0x10c/0x1d4)
> >>     [<>] (kernel_init_freeable) from [<>] (kernel_init+0x10/0xec)
> >>     [<>] (kernel_init) from [<>] (ret_from_fork+0x14/0x24)
> >>
> >> This is caused by the fact that at the moment some GPIO names are set
> >> to NULL, there is a hole in linear representation of one GPI bank, see
> >> drivers/gpio/gpio-lpc32xx.c / gpi_p3_names[] for details.
> >>
> >> The same problem most probably affects also gpio-cs5535.c, see
> >> cs5535_gpio_names[].
> >>
> >> Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
> >> ---
> >> Linus, Markus,
> >>
> >> I understand that LPC32xx GPIO driver is ugly (hopefully this
> >> will be changed in future), but please account this problem.
> >>
> >> I don't ask for inclusion of this change, because my own analysis is
> >> not done completely.
> >>
> >> Also please review, the same problem probably affects gpio-cs5535.c
> >> as well, see cs5535_gpio_names[].
> > 
> > Thanks for this hint. But I think these patches for gpio_name_to_desc won't get
> > mainline any time soon as this depends on the chardev interface and how it will
> > look like.
> > 
> 
> sorry for misunderstanding, what do mean by "these patches"?
> 
> I see two options here, one is to fix or revert your c0017ed71966 due to
> caused regressions, another one is to update affected GPIO drivers
> removing discontinuity in GPIO enumeration inside a bank.

Sorry, seems I was on a wrong kernel version when checking this morning. Linus
removed lots of my patches from his tree after some correct comments so I
thought this patch was removed as well.

As it is mainline your patch looks good to me. Perhaps a comment on why we need
to check for !name at this point would be good.

Best Regards,

Markus

> 
> With best wishes,
> Vladimir
> 
> >>
> >>  drivers/gpio/gpiolib.c | 2 +-
> >>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> >> index a18f00f..2a91f32 100644
> >> --- a/drivers/gpio/gpiolib.c
> >> +++ b/drivers/gpio/gpiolib.c
> >> @@ -233,7 +233,7 @@ static struct gpio_desc *gpio_name_to_desc(const char * const name)
> >>  		for (i = 0; i != chip->ngpio; ++i) {
> >>  			struct gpio_desc *gpio = &chip->desc[i];
> >>  
> >> -			if (!gpio->name)
> >> +			if (!gpio->name || !name)
> >>  				continue;
> >>  
> >>  			if (!strcmp(gpio->name, name)) {
> >>
> > 
>
Vladimir Zapolskiy Nov. 29, 2015, 12:44 p.m. UTC | #4
Hi Linus,

On 17.11.2015 13:50, Markus Pargmann wrote:
> Hi Vladimir,
> 
> On Tuesday 17 November 2015 13:20:40 Vladimir Zapolskiy wrote:
>> Hi Markus,
>>
>> On 17.11.2015 11:24, Markus Pargmann wrote:
>>> Hi,
>>>
>>> On Wednesday 11 November 2015 14:36:53 Vladimir Zapolskiy wrote:
>>>> Commit c0017ed71966 ("gpio: Introduce gpio descriptor 'name'") causes
>>>> OOPS on boot on LPC32xx boards:
>>>>
>>>>     Unable to handle kernel NULL pointer dereference at virtual address 00000000
>>>>     CPU: 0 PID: 1 Comm: swapper Not tainted 4.3.0+ #707
>>>>     Hardware name: LPC32XX SoC (Flattened Device Tree)
>>>>     task: c381baa0 ti: c381e000 task.ti: c381e000
>>>>     PC is at strcmp+0x10/0x40
>>>>     LR is at gpiochip_add+0x3d0/0x4d4
>>>>     pc : [<>]    lr : [<>]    psr: a0000093
>>>>     sp : c381fd60  ip : c381fd70  fp : c381fd6c
>>>>
>>>>     [snip]
>>>>
>>>>     Backtrace:
>>>>     [<>] (strcmp) from [<>] (gpiochip_add+0x3d0/0x4d4)
>>>>     [<>] (gpiochip_add) from [<>] (lpc32xx_gpio_probe+0x44/0x60)
>>>>     [<>] (lpc32xx_gpio_probe) from [<>] (platform_drv_probe+0x40/0x8c)
>>>>     [<>] (platform_drv_probe) from [<>] (driver_probe_device+0x110/0x294)
>>>>     [<>] (driver_probe_device) from [<>] (__driver_attach+0x70/0x94)
>>>>     [<>] (__driver_attach) from [<>] (bus_for_each_dev+0x74/0x98)
>>>>     [<>] (bus_for_each_dev) from [<>] (driver_attach+0x20/0x28)
>>>>     [<>] (driver_attach) from [<>] (bus_add_driver+0xd4/0x1f0)
>>>>     [<>] (bus_add_driver) from [<>] (driver_register+0xa4/0xe8)
>>>>     [<>] (driver_register) from [<>] (__platform_driver_register+0x38/0x4c)
>>>>     [<>] (__platform_driver_register) from [<>] (lpc32xx_gpio_driver_init+0x18/0x20)
>>>>     [<>] (lpc32xx_gpio_driver_init) from [<>] (do_one_initcall+0x108/0x1c8)
>>>>     [<>] (do_one_initcall) from [<>] (kernel_init_freeable+0x10c/0x1d4)
>>>>     [<>] (kernel_init_freeable) from [<>] (kernel_init+0x10/0xec)
>>>>     [<>] (kernel_init) from [<>] (ret_from_fork+0x14/0x24)
>>>>
>>>> This is caused by the fact that at the moment some GPIO names are set
>>>> to NULL, there is a hole in linear representation of one GPI bank, see
>>>> drivers/gpio/gpio-lpc32xx.c / gpi_p3_names[] for details.
>>>>
>>>> The same problem most probably affects also gpio-cs5535.c, see
>>>> cs5535_gpio_names[].
>>>>
>>>> Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
>>>> ---
>>>> Linus, Markus,
>>>>
>>>> I understand that LPC32xx GPIO driver is ugly (hopefully this
>>>> will be changed in future), but please account this problem.
>>>>
>>>> I don't ask for inclusion of this change, because my own analysis is
>>>> not done completely.
>>>>
>>>> Also please review, the same problem probably affects gpio-cs5535.c
>>>> as well, see cs5535_gpio_names[].
>>>
>>> Thanks for this hint. But I think these patches for gpio_name_to_desc won't get
>>> mainline any time soon as this depends on the chardev interface and how it will
>>> look like.
>>>
>>
>> sorry for misunderstanding, what do mean by "these patches"?
>>
>> I see two options here, one is to fix or revert your c0017ed71966 due to
>> caused regressions, another one is to update affected GPIO drivers
>> removing discontinuity in GPIO enumeration inside a bank.
> 
> Sorry, seems I was on a wrong kernel version when checking this morning. Linus
> removed lots of my patches from his tree after some correct comments so I
> thought this patch was removed as well.

Linus, should I resend this fix for v4.4 or is it good enough for
application?

> As it is mainline your patch looks good to me. Perhaps a comment on why we need
> to check for !name at this point would be good.

There is such a description in the commit message, why this check is needed.

With best wishes,
Vladimir

> Best Regards,
> 
> Markus
> 
>>
>> With best wishes,
>> Vladimir
>>
>>>>
>>>>  drivers/gpio/gpiolib.c | 2 +-
>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>
>>>> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
>>>> index a18f00f..2a91f32 100644
>>>> --- a/drivers/gpio/gpiolib.c
>>>> +++ b/drivers/gpio/gpiolib.c
>>>> @@ -233,7 +233,7 @@ static struct gpio_desc *gpio_name_to_desc(const char * const name)
>>>>  		for (i = 0; i != chip->ngpio; ++i) {
>>>>  			struct gpio_desc *gpio = &chip->desc[i];
>>>>  
>>>> -			if (!gpio->name)
>>>> +			if (!gpio->name || !name)
>>>>  				continue;
>>>>  
>>>>  			if (!strcmp(gpio->name, name)) {
>>>>
>>>
>>
> 
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Linus Walleij Nov. 29, 2015, 9:18 p.m. UTC | #5
On Wed, Nov 11, 2015 at 1:36 PM, Vladimir Zapolskiy <vz@mleia.com> wrote:

> Commit c0017ed71966 ("gpio: Introduce gpio descriptor 'name'") causes
> OOPS on boot on LPC32xx boards:
>
>     Unable to handle kernel NULL pointer dereference at virtual address 00000000
>     CPU: 0 PID: 1 Comm: swapper Not tainted 4.3.0+ #707
>     Hardware name: LPC32XX SoC (Flattened Device Tree)
>     task: c381baa0 ti: c381e000 task.ti: c381e000
>     PC is at strcmp+0x10/0x40
>     LR is at gpiochip_add+0x3d0/0x4d4
>     pc : [<>]    lr : [<>]    psr: a0000093
>     sp : c381fd60  ip : c381fd70  fp : c381fd6c
>
>     [snip]
>
>     Backtrace:
>     [<>] (strcmp) from [<>] (gpiochip_add+0x3d0/0x4d4)
>     [<>] (gpiochip_add) from [<>] (lpc32xx_gpio_probe+0x44/0x60)
>     [<>] (lpc32xx_gpio_probe) from [<>] (platform_drv_probe+0x40/0x8c)
>     [<>] (platform_drv_probe) from [<>] (driver_probe_device+0x110/0x294)
>     [<>] (driver_probe_device) from [<>] (__driver_attach+0x70/0x94)
>     [<>] (__driver_attach) from [<>] (bus_for_each_dev+0x74/0x98)
>     [<>] (bus_for_each_dev) from [<>] (driver_attach+0x20/0x28)
>     [<>] (driver_attach) from [<>] (bus_add_driver+0xd4/0x1f0)
>     [<>] (bus_add_driver) from [<>] (driver_register+0xa4/0xe8)
>     [<>] (driver_register) from [<>] (__platform_driver_register+0x38/0x4c)
>     [<>] (__platform_driver_register) from [<>] (lpc32xx_gpio_driver_init+0x18/0x20)
>     [<>] (lpc32xx_gpio_driver_init) from [<>] (do_one_initcall+0x108/0x1c8)
>     [<>] (do_one_initcall) from [<>] (kernel_init_freeable+0x10c/0x1d4)
>     [<>] (kernel_init_freeable) from [<>] (kernel_init+0x10/0xec)
>     [<>] (kernel_init) from [<>] (ret_from_fork+0x14/0x24)
>
> This is caused by the fact that at the moment some GPIO names are set
> to NULL, there is a hole in linear representation of one GPI bank, see
> drivers/gpio/gpio-lpc32xx.c / gpi_p3_names[] for details.
>
> The same problem most probably affects also gpio-cs5535.c, see
> cs5535_gpio_names[].
>
> Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>

Patch applied for fixes. Sorry for the delay.

Yours,
Linus Walleij
--
To unsubscribe from this list: send the line "unsubscribe linux-gpio" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Markus Pargmann Nov. 30, 2015, 9:51 a.m. UTC | #6
Hi,

On Sunday 29 November 2015 14:44:30 Vladimir Zapolskiy wrote:
> Hi Linus,
> 
> On 17.11.2015 13:50, Markus Pargmann wrote:
> > Hi Vladimir,
> > 
> > On Tuesday 17 November 2015 13:20:40 Vladimir Zapolskiy wrote:
> >> Hi Markus,
> >>
> >> On 17.11.2015 11:24, Markus Pargmann wrote:
> >>> Hi,
> >>>
> >>> On Wednesday 11 November 2015 14:36:53 Vladimir Zapolskiy wrote:
> >>>> Commit c0017ed71966 ("gpio: Introduce gpio descriptor 'name'") causes
> >>>> OOPS on boot on LPC32xx boards:
> >>>>
> >>>>     Unable to handle kernel NULL pointer dereference at virtual address 00000000
> >>>>     CPU: 0 PID: 1 Comm: swapper Not tainted 4.3.0+ #707
> >>>>     Hardware name: LPC32XX SoC (Flattened Device Tree)
> >>>>     task: c381baa0 ti: c381e000 task.ti: c381e000
> >>>>     PC is at strcmp+0x10/0x40
> >>>>     LR is at gpiochip_add+0x3d0/0x4d4
> >>>>     pc : [<>]    lr : [<>]    psr: a0000093
> >>>>     sp : c381fd60  ip : c381fd70  fp : c381fd6c
> >>>>
> >>>>     [snip]
> >>>>
> >>>>     Backtrace:
> >>>>     [<>] (strcmp) from [<>] (gpiochip_add+0x3d0/0x4d4)
> >>>>     [<>] (gpiochip_add) from [<>] (lpc32xx_gpio_probe+0x44/0x60)
> >>>>     [<>] (lpc32xx_gpio_probe) from [<>] (platform_drv_probe+0x40/0x8c)
> >>>>     [<>] (platform_drv_probe) from [<>] (driver_probe_device+0x110/0x294)
> >>>>     [<>] (driver_probe_device) from [<>] (__driver_attach+0x70/0x94)
> >>>>     [<>] (__driver_attach) from [<>] (bus_for_each_dev+0x74/0x98)
> >>>>     [<>] (bus_for_each_dev) from [<>] (driver_attach+0x20/0x28)
> >>>>     [<>] (driver_attach) from [<>] (bus_add_driver+0xd4/0x1f0)
> >>>>     [<>] (bus_add_driver) from [<>] (driver_register+0xa4/0xe8)
> >>>>     [<>] (driver_register) from [<>] (__platform_driver_register+0x38/0x4c)
> >>>>     [<>] (__platform_driver_register) from [<>] (lpc32xx_gpio_driver_init+0x18/0x20)
> >>>>     [<>] (lpc32xx_gpio_driver_init) from [<>] (do_one_initcall+0x108/0x1c8)
> >>>>     [<>] (do_one_initcall) from [<>] (kernel_init_freeable+0x10c/0x1d4)
> >>>>     [<>] (kernel_init_freeable) from [<>] (kernel_init+0x10/0xec)
> >>>>     [<>] (kernel_init) from [<>] (ret_from_fork+0x14/0x24)
> >>>>
> >>>> This is caused by the fact that at the moment some GPIO names are set
> >>>> to NULL, there is a hole in linear representation of one GPI bank, see
> >>>> drivers/gpio/gpio-lpc32xx.c / gpi_p3_names[] for details.
> >>>>
> >>>> The same problem most probably affects also gpio-cs5535.c, see
> >>>> cs5535_gpio_names[].
> >>>>
> >>>> Signed-off-by: Vladimir Zapolskiy <vz@mleia.com>
> >>>> ---
> >>>> Linus, Markus,
> >>>>
> >>>> I understand that LPC32xx GPIO driver is ugly (hopefully this
> >>>> will be changed in future), but please account this problem.
> >>>>
> >>>> I don't ask for inclusion of this change, because my own analysis is
> >>>> not done completely.
> >>>>
> >>>> Also please review, the same problem probably affects gpio-cs5535.c
> >>>> as well, see cs5535_gpio_names[].
> >>>
> >>> Thanks for this hint. But I think these patches for gpio_name_to_desc won't get
> >>> mainline any time soon as this depends on the chardev interface and how it will
> >>> look like.
> >>>
> >>
> >> sorry for misunderstanding, what do mean by "these patches"?
> >>
> >> I see two options here, one is to fix or revert your c0017ed71966 due to
> >> caused regressions, another one is to update affected GPIO drivers
> >> removing discontinuity in GPIO enumeration inside a bank.
> > 
> > Sorry, seems I was on a wrong kernel version when checking this morning. Linus
> > removed lots of my patches from his tree after some correct comments so I
> > thought this patch was removed as well.
> 
> Linus, should I resend this fix for v4.4 or is it good enough for
> application?
> 
> > As it is mainline your patch looks good to me. Perhaps a comment on why we need
> > to check for !name at this point would be good.
> 
> There is such a description in the commit message, why this check is needed.

I meant a code comment. But that's not too important.

Best Regards,

Markus

> 
> With best wishes,
> Vladimir
> 
> > Best Regards,
> > 
> > Markus
> > 
> >>
> >> With best wishes,
> >> Vladimir
> >>
> >>>>
> >>>>  drivers/gpio/gpiolib.c | 2 +-
> >>>>  1 file changed, 1 insertion(+), 1 deletion(-)
> >>>>
> >>>> diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
> >>>> index a18f00f..2a91f32 100644
> >>>> --- a/drivers/gpio/gpiolib.c
> >>>> +++ b/drivers/gpio/gpiolib.c
> >>>> @@ -233,7 +233,7 @@ static struct gpio_desc *gpio_name_to_desc(const char * const name)
> >>>>  		for (i = 0; i != chip->ngpio; ++i) {
> >>>>  			struct gpio_desc *gpio = &chip->desc[i];
> >>>>  
> >>>> -			if (!gpio->name)
> >>>> +			if (!gpio->name || !name)
> >>>>  				continue;
> >>>>  
> >>>>  			if (!strcmp(gpio->name, name)) {
> >>>>
> >>>
> >>
> > 
>

Patch
diff mbox

diff --git a/drivers/gpio/gpiolib.c b/drivers/gpio/gpiolib.c
index a18f00f..2a91f32 100644
--- a/drivers/gpio/gpiolib.c
+++ b/drivers/gpio/gpiolib.c
@@ -233,7 +233,7 @@  static struct gpio_desc *gpio_name_to_desc(const char * const name)
 		for (i = 0; i != chip->ngpio; ++i) {
 			struct gpio_desc *gpio = &chip->desc[i];
 
-			if (!gpio->name)
+			if (!gpio->name || !name)
 				continue;
 
 			if (!strcmp(gpio->name, name)) {