mbox series

[0/4] Explicitly deny IRQ0 in the libata drivers

Message ID 2e552c36-84a5-a642-a8f0-1395f77a1a6a@omprussia.ru
Headers show
Series Explicitly deny IRQ0 in the libata drivers | expand

Message

Sergey Shtylyov March 21, 2021, 6:50 p.m. UTC 
Here are 4 patches against the 'master' branch of the Jens Axboe's 'linux-blobk.git'
repo plus the 'pataep_93xx' driver patch re-posted yesterday.  The affected drivers
use platform_get_irq() which may return IRQ0 (considered invalid, according to Linus)
that means polling when passed to ata_host_activate() called at the end of the probe
methods. I think that the solution to this issue is to explicitly deny IRQ0 returned
by platform_get_irq()...

[1/4] pata_bk3710: deny IRQ0
[2/4] pata_ep93xx: deny IRQ0
[3/4] pata_ftide010: deny IRQ0
[4/4] pata_imx: deny IRQ0

Comments

Sergey Shtylyov May 10, 2021, 8:48 p.m. UTC | #1 
Hello!

   What about this series?
   I got no feedback whatsoever -- it seems to have been lost.

On 3/21/21 9:50 PM, Sergey Shtylyov wrote:

> Here are 4 patches against the 'master' branch of the Jens Axboe's 'linux-blobk.git'
> repo plus the 'pataep_93xx' driver patch re-posted yesterday.  The affected drivers
> use platform_get_irq() which may return IRQ0 (considered invalid, according to Linus)
> that means polling when passed to ata_host_activate() called at the end of the probe
> methods.

   I might not have been clear enough: 'irq == 0' means that the libata core would WARN about
the non-NULL 'handler' paramter which seems a to be a problem...

> I think that the solution to this issue is to explicitly deny IRQ0 returned
> by platform_get_irq()...
[...]

MBR, Sergei
Sergey Shtylyov Sept. 4, 2021, 6:57 p.m. UTC | #2 
On 5/10/21 11:48 PM, Sergey Shtylyov wrote:

> Hello!
> 
>    What about this series?
>    I got no feedback whatsoever -- it seems to have been lost.

   Almost 4 months have paseed from this reminder. Still no commnets whatsoever...

> On 3/21/21 9:50 PM, Sergey Shtylyov wrote:
> 
>> Here are 4 patches against the 'master' branch of the Jens Axboe's 'linux-blobk.git'
>> repo plus the 'pataep_93xx' driver patch re-posted yesterday.  The affected drivers
>> use platform_get_irq() which may return IRQ0 (considered invalid, according to Linus)
>> that means polling when passed to ata_host_activate() called at the end of the probe
>> methods.
> 
>    I might not have been clear enough: 'irq == 0' means that the libata core would WARN about
> the non-NULL 'handler' paramter which seems a to be a problem...
> 
>> I think that the solution to this issue is to explicitly deny IRQ0 returned
>> by platform_get_irq()...
[...]

MBR, Sergey
Jens Axboe Sept. 4, 2021, 9:25 p.m. UTC | #3 
On 9/4/21 12:57 PM, Sergey Shtylyov wrote:
> On 5/10/21 11:48 PM, Sergey Shtylyov wrote:
> 
>> Hello!
>>
>>    What about this series?
>>    I got no feedback whatsoever -- it seems to have been lost.
> 
>    Almost 4 months have paseed from this reminder. Still no commnets whatsoever...

You did get a review almost immediately, but it wasn't responded to.
Sergey Shtylyov Sept. 5, 2021, 9:47 a.m. UTC | #4 
On 05.09.2021 0:25, Jens Axboe wrote:

[...]
>>>     What about this series?
>>>     I got no feedback whatsoever -- it seems to have been lost.
>>
>>     Almost 4 months have paseed from this reminder. Still no commnets whatsoever...
> 
> You did get a review almost immediately, but it wasn't responded to.

     Oh, seeing it now (again)! Indeed, I forgot to reply back then, sorry!

MBR, Sergey