>-----Original Message-----
>From: Robin Getz [mailto:[email protected]]
>Sent: Tuesday, March 23, 2010 2:20 AM
>To: [email protected]
>Cc: [email protected]
>Subject: Re: [Linux-kernel-commits] [8528]
>trunk/drivers/serial/bfin_5xx.c: bfin-serial: use nosync mode
>for dma_disable_irq in timer function
>
>On Mon 22 Mar 2010 05:37, [email protected] pondered:
>> Revision
>>
>8528<http://blackfin.uclinux.org/gf/project/linux-kernel/scmsvn
>/?action=browse&path=/&view=rev&root=linux-kernel&revision=8528>
>> Author
>> bhsong<http://blackfin.uclinux.org/gf/user/bhsong/>
>> Date
>> 2010-03-22 05:37:11 -0400 (Mon, 22 Mar 2010)
>> Log Message
>>
>> bfin-serial: use nosync mode for dma_disable_irq in timer function
>
>Why? Does this fix a bug?
It doesn't fix any existing bug. Just call an atomic version in softirq
context to avoid possible deadlock.
>
>
>
>> Modified Paths
>>
>> * trunk/drivers/serial/bfin_5xx.c
>>
>> Diff
>> Modified: trunk/drivers/serial/bfin_5xx.c (8527 => 8528)
>>
>>
>> --- trunk/drivers/serial/bfin_5xx.c 2010-03-22 07:02:06
>UTC (rev 8527)
>> +++ trunk/drivers/serial/bfin_5xx.c 2010-03-22 09:37:11
>UTC (rev 8528)
>> @@ -486,7 +486,7 @@
>> {
>> int x_pos, pos;
>>
>> - dma_disable_irq(uart->rx_dma_channel);
>> + dma_disable_irq_nosync(uart->rx_dma_channel);
>> spin_lock_bh(&uart->rx_lock);
>>
>> /* 2D DMA RX buffer ring is used. Because curr_y_count and
>>
>>
>
_______________________________________________
Linux-kernel-commits mailing list
[email protected]
https://blackfin.uclinux.org/mailman/listinfo/linux-kernel-commits
