2012-08-09 20:38:17

by Rafał Miłecki

[permalink] [raw]
Subject: [PATCH wireless?] bcma: keep info about parallel flash presentence

This allows drivers to access flash conditionally and stop crashing some
devices.

Signed-off-by: Rafał Miłecki <[email protected]>
Acked-by: Hauke Mehrtens <[email protected]>
---
John: I suggest taking it for wirelesss, AKA fix
---
drivers/bcma/driver_mips.c | 2 +-
include/linux/bcma/bcma_driver_chipcommon.h | 1 +
2 files changed, 2 insertions(+), 1 deletions(-)

diff --git a/drivers/bcma/driver_mips.c b/drivers/bcma/driver_mips.c
index cc65b45..55acba9 100644
--- a/drivers/bcma/driver_mips.c
+++ b/drivers/bcma/driver_mips.c
@@ -192,12 +192,12 @@ static void bcma_core_mips_flash_detect(struct bcma_drv_mips *mcore)
bcma_debug(bus, "Found parallel flash\n");
bus->drv_cc.pflash.window = 0x1c000000;
bus->drv_cc.pflash.window_size = 0x02000000;
-
if ((bcma_read32(bus->drv_cc.core, BCMA_CC_FLASH_CFG) &
BCMA_CC_FLASH_CFG_DS) == 0)
bus->drv_cc.pflash.buswidth = 1;
else
bus->drv_cc.pflash.buswidth = 2;
+ bus->drv_cc.pflash.present = true;
break;
default:
bcma_err(bus, "Flash type not supported\n");
diff --git a/include/linux/bcma/bcma_driver_chipcommon.h b/include/linux/bcma/bcma_driver_chipcommon.h
index 3fb8bba..16c266c 100644
--- a/include/linux/bcma/bcma_driver_chipcommon.h
+++ b/include/linux/bcma/bcma_driver_chipcommon.h
@@ -510,6 +510,7 @@ struct bcma_chipcommon_pmu {

#ifdef CONFIG_BCMA_DRIVER_MIPS
struct bcma_pflash {
+ bool present;
u8 buswidth;
u32 window;
u32 window_size;
--
1.7.7



2012-08-10 13:28:44

by John W. Linville

[permalink] [raw]
Subject: Re: [PATCH wireless?] bcma: keep info about parallel flash presentence

On Fri, Aug 10, 2012 at 02:25:48PM +0200, Rafał Miłecki wrote:
> 2012/8/10 Hauke Mehrtens <[email protected]>:
> > On 08/10/2012 11:32 AM, Rafał Miłecki wrote:
> >> 2012/8/10 Arend van Spriel <[email protected]>:
> >>> On 08/09/2012 10:38 PM, Rafał Miłecki wrote:
> >>>> This allows drivers to access flash conditionally and stop crashing some
> >>>> devices.
> >>>>
> >>>> Signed-off-by: Rafał Miłecki <[email protected]>
> >>>> Acked-by: Hauke Mehrtens <[email protected]>
> >>>> ---
> >>>> John: I suggest taking it for wirelesss, AKA fix
> >>>
> >>> Maybe you could mentioned devices and/or drivers crashes without this patch?
> >>
> >> Yes, sure! The problematic one for me is arch/mips/bcm47xx/nvram.c.
> >> Right now it assumes every BCMA board has pflash. I'll submit fix for
> >> mips tree when this patch hits it.
> >
> > Letting this patches get merged into the wireless tree and an other
> > patches which depends on this one into the mips tree will cause
> > problems, because the wireless tree only gets merged into the mips tree
> > indirectly round about the rc1 release in form of the rc1 release done
> > by Linus.
> >
> > The patch to arch/mips/bcm47xx/nvram.c will depend on this one and you
> > should submit them to the same tree.
>
> I'm planning to wait until rc2 (or rc3 if needed). I know the dependency :)

Which rc2? Because Ralf probably won't have it in his tree until
the 3.7-rc1 timeframe...

--
John W. Linville Someday the world will need a hero, and you
[email protected] might be all we have. Be ready.

2012-08-10 12:23:42

by Hauke Mehrtens

[permalink] [raw]
Subject: Re: [PATCH wireless?] bcma: keep info about parallel flash presentence

On 08/10/2012 11:32 AM, Rafał Miłecki wrote:
> 2012/8/10 Arend van Spriel <[email protected]>:
>> On 08/09/2012 10:38 PM, Rafał Miłecki wrote:
>>> This allows drivers to access flash conditionally and stop crashing some
>>> devices.
>>>
>>> Signed-off-by: Rafał Miłecki <[email protected]>
>>> Acked-by: Hauke Mehrtens <[email protected]>
>>> ---
>>> John: I suggest taking it for wirelesss, AKA fix
>>
>> Maybe you could mentioned devices and/or drivers crashes without this patch?
>
> Yes, sure! The problematic one for me is arch/mips/bcm47xx/nvram.c.
> Right now it assumes every BCMA board has pflash. I'll submit fix for
> mips tree when this patch hits it.

Letting this patches get merged into the wireless tree and an other
patches which depends on this one into the mips tree will cause
problems, because the wireless tree only gets merged into the mips tree
indirectly round about the rc1 release in form of the rc1 release done
by Linus.

The patch to arch/mips/bcm47xx/nvram.c will depend on this one and you
should submit them to the same tree.

Hauke


2012-08-10 09:02:45

by Arend van Spriel

[permalink] [raw]
Subject: Re: [PATCH wireless?] bcma: keep info about parallel flash presentence

On 08/09/2012 10:38 PM, Rafał Miłecki wrote:
> This allows drivers to access flash conditionally and stop crashing some
> devices.
>
> Signed-off-by: Rafał Miłecki <[email protected]>
> Acked-by: Hauke Mehrtens <[email protected]>
> ---
> John: I suggest taking it for wirelesss, AKA fix

Maybe you could mentioned devices and/or drivers crashes without this patch?

Gr. AvS


2012-08-10 09:32:10

by Rafał Miłecki

[permalink] [raw]
Subject: Re: [PATCH wireless?] bcma: keep info about parallel flash presentence

2012/8/10 Arend van Spriel <[email protected]>:
> On 08/09/2012 10:38 PM, Rafał Miłecki wrote:
>> This allows drivers to access flash conditionally and stop crashing some
>> devices.
>>
>> Signed-off-by: Rafał Miłecki <[email protected]>
>> Acked-by: Hauke Mehrtens <[email protected]>
>> ---
>> John: I suggest taking it for wirelesss, AKA fix
>
> Maybe you could mentioned devices and/or drivers crashes without this patch?

Yes, sure! The problematic one for me is arch/mips/bcm47xx/nvram.c.
Right now it assumes every BCMA board has pflash. I'll submit fix for
mips tree when this patch hits it.

--
Rafał

2012-08-10 09:32:55

by Rafał Miłecki

[permalink] [raw]
Subject: Re: [PATCH wireless?] bcma: keep info about parallel flash presentence

2012/8/10 Rafał Miłecki <[email protected]>:
> 2012/8/10 Arend van Spriel <[email protected]>:
>> On 08/09/2012 10:38 PM, Rafał Miłecki wrote:
>>> This allows drivers to access flash conditionally and stop crashing some
>>> devices.
>>>
>>> Signed-off-by: Rafał Miłecki <[email protected]>
>>> Acked-by: Hauke Mehrtens <[email protected]>
>>> ---
>>> John: I suggest taking it for wirelesss, AKA fix
>>
>> Maybe you could mentioned devices and/or drivers crashes without this patch?
>
> Yes, sure! The problematic one for me is arch/mips/bcm47xx/nvram.c.
> Right now it assumes every BCMA board has pflash. I'll submit fix for
> mips tree when this patch hits it.

Ah, and device. My Netgear WNDR4500 has serial flash and NAND flash.
It doesn't have parallel one, so this code crashes my device.

--
Rafał

2012-08-10 13:36:28

by Rafał Miłecki

[permalink] [raw]
Subject: Re: [PATCH wireless?] bcma: keep info about parallel flash presentence

2012/8/10 John W. Linville <[email protected]>:
> On Fri, Aug 10, 2012 at 02:25:48PM +0200, Rafał Miłecki wrote:
>> 2012/8/10 Hauke Mehrtens <[email protected]>:
>> > On 08/10/2012 11:32 AM, Rafał Miłecki wrote:
>> >> 2012/8/10 Arend van Spriel <[email protected]>:
>> >>> On 08/09/2012 10:38 PM, Rafał Miłecki wrote:
>> >>>> This allows drivers to access flash conditionally and stop crashing some
>> >>>> devices.
>> >>>>
>> >>>> Signed-off-by: Rafał Miłecki <[email protected]>
>> >>>> Acked-by: Hauke Mehrtens <[email protected]>
>> >>>> ---
>> >>>> John: I suggest taking it for wirelesss, AKA fix
>> >>>
>> >>> Maybe you could mentioned devices and/or drivers crashes without this patch?
>> >>
>> >> Yes, sure! The problematic one for me is arch/mips/bcm47xx/nvram.c.
>> >> Right now it assumes every BCMA board has pflash. I'll submit fix for
>> >> mips tree when this patch hits it.
>> >
>> > Letting this patches get merged into the wireless tree and an other
>> > patches which depends on this one into the mips tree will cause
>> > problems, because the wireless tree only gets merged into the mips tree
>> > indirectly round about the rc1 release in form of the rc1 release done
>> > by Linus.
>> >
>> > The patch to arch/mips/bcm47xx/nvram.c will depend on this one and you
>> > should submit them to the same tree.
>>
>> I'm planning to wait until rc2 (or rc3 if needed). I know the dependency :)
>
> Which rc2? Because Ralf probably won't have it in his tree until
> the 3.7-rc1 timeframe...

Ouch, I hoped bcma fix will go into 3.6-rc2 or 3.6-rc3 and Ralf will
rebase his tree against Linus at some point...

OK, I'll try to send additional bcm47xx fix through yours tree (after
Ralf accepts it).

--
Rafał

2012-08-10 12:25:49

by Rafał Miłecki

[permalink] [raw]
Subject: Re: [PATCH wireless?] bcma: keep info about parallel flash presentence

2012/8/10 Hauke Mehrtens <[email protected]>:
> On 08/10/2012 11:32 AM, Rafał Miłecki wrote:
>> 2012/8/10 Arend van Spriel <[email protected]>:
>>> On 08/09/2012 10:38 PM, Rafał Miłecki wrote:
>>>> This allows drivers to access flash conditionally and stop crashing some
>>>> devices.
>>>>
>>>> Signed-off-by: Rafał Miłecki <[email protected]>
>>>> Acked-by: Hauke Mehrtens <[email protected]>
>>>> ---
>>>> John: I suggest taking it for wirelesss, AKA fix
>>>
>>> Maybe you could mentioned devices and/or drivers crashes without this patch?
>>
>> Yes, sure! The problematic one for me is arch/mips/bcm47xx/nvram.c.
>> Right now it assumes every BCMA board has pflash. I'll submit fix for
>> mips tree when this patch hits it.
>
> Letting this patches get merged into the wireless tree and an other
> patches which depends on this one into the mips tree will cause
> problems, because the wireless tree only gets merged into the mips tree
> indirectly round about the rc1 release in form of the rc1 release done
> by Linus.
>
> The patch to arch/mips/bcm47xx/nvram.c will depend on this one and you
> should submit them to the same tree.

I'm planning to wait until rc2 (or rc3 if needed). I know the dependency :)

--
Rafał