Re: [PATCH] convert sbmac tx to spin_lock_irqsave to prevent early IRQ enable

Re: [PATCH] convert sbmac tx to spin_lock_irqsave to prevent early IRQ enable

[Ralf Baechle]

On Wed, Sep 17, 2008 at 07:40:51AM -0400, Jeff Garzik wrote:

> On Wed, Sep 17, 2008 at 10:25:37AM +0800, Weiwei Wang wrote:
> > Netpoll will call the interrupt handler with interrupts
> > disabled when using kgdboe, so spin_lock_irqsave() should
> > be used instead of spin_lock_irq() to prevent interrupts
> > from being incorrectly enabled.
> > 
> > Signed-off-by: Weiwei Wang <weiwei.wang@windriver.com>
> > ---
> >  drivers/net/sb1250-mac.c |   12 +++++++-----
> >  1 files changed, 7 insertions(+), 5 deletions(-)
> 
> Please send to jeff@garzik.org or jgarzik@pobox.com.

Jeff - I haven't looked at kgdboe but if he's right half of drivers/net
will need to be fixed ...

  Ralf

Re: [PATCH] convert sbmac tx to spin_lock_irqsave to prevent early IRQ enable

[Jeff Garzik]

Ralf Baechle wrote:
> On Wed, Sep 17, 2008 at 07:40:51AM -0400, Jeff Garzik wrote:
> 
>> On Wed, Sep 17, 2008 at 10:25:37AM +0800, Weiwei Wang wrote:
>>> Netpoll will call the interrupt handler with interrupts
>>> disabled when using kgdboe, so spin_lock_irqsave() should
>>> be used instead of spin_lock_irq() to prevent interrupts
>>> from being incorrectly enabled.
>>>
>>> Signed-off-by: Weiwei Wang <weiwei.wang@windriver.com>
>>> ---
>>>  drivers/net/sb1250-mac.c |   12 +++++++-----
>>>  1 files changed, 7 insertions(+), 5 deletions(-)
>> Please send to jeff@garzik.org or jgarzik@pobox.com.
> 
> Jeff - I haven't looked at kgdboe but if he's right half of drivers/net
> will need to be fixed ...

Oh indeed.  I mainly do it on a case-by-case basis where people care.  I 
overall think its a bogus change that deserves pushback, but that 
involves non-networking layers.  In the end, case-by-case application 
seemed to win.

	Jeff

Re: [PATCH] convert sbmac tx to spin_lock_irqsave to prevent early IRQ enable

[Ralf Baechle]

On Thu, Sep 25, 2008 at 03:33:40PM -0400, Jeff Garzik wrote:

>>> On Wed, Sep 17, 2008 at 10:25:37AM +0800, Weiwei Wang wrote:
>>>> Netpoll will call the interrupt handler with interrupts
>>>> disabled when using kgdboe, so spin_lock_irqsave() should
>>>> be used instead of spin_lock_irq() to prevent interrupts
>>>> from being incorrectly enabled.
>>>>
>>>> Signed-off-by: Weiwei Wang <weiwei.wang@windriver.com>
>>>> ---
>>>>  drivers/net/sb1250-mac.c |   12 +++++++-----
>>>>  1 files changed, 7 insertions(+), 5 deletions(-)
>>> Please send to jeff@garzik.org or jgarzik@pobox.com.
>>
>> Jeff - I haven't looked at kgdboe but if he's right half of drivers/net
>> will need to be fixed ...
>
> Oh indeed.  I mainly do it on a case-by-case basis where people care.  I  
> overall think its a bogus change that deserves pushback, but that  
> involves non-networking layers.  In the end, case-by-case application  
> seemed to win.

Turns out kgdboe isn't upstream yet - but netconsole apparently hast the
same issue.

  Ralf