Re: [GIT PULL] Core block bits for 2.6.40

[Jens Axboe]

On 2011-05-26 07:35, Eric Dumazet wrote:
> Le mercredi 25 mai 2011 Ã 15:03 +0200, Jens Axboe a Ãcrit :
>> Hi Linus,
>>
>> A real pull request this time. These are the core bits for the next
>> kernel. Nothing exciting in here in this round, with the churn in
>> the last release, this release will be mellow.
>>
>> - Optimizations for queue flushing on SATA devices, where we
>>   can't queue a FLUSH command. This fixes up a performance regression
>>   caused by a bug fix to prevent potential live lock on adding
>>   FLUSH commands at the head of the queue.
>>
>> - Make stats lockless for blk-cgroup.
>>
>> - Various bug fixes for blk-cgroup.
>>
>> - Various little fixes and improvements to CFQ.
>>
>> - Last round of media event changes.
>>
>> - Fixes for discard topology reporting and sysfs documentation update
>>   for the same.
>>
>> - Add support for batched discards. This greatly speeds up discarding
>>   an entire device.
>>
>> - Fix for bad return value in the discard ioctl.
>>
>> - Remove request plug state debug checks. Haven't triggered even during
>>   the 2.6.39-rc devel cycle, so get rid of them.
>>
>>
>> Please pull!
>>
>>
>>   git://git.kernel.dk/linux-2.6-block.git for-2.6.40/core
>>
> 
>> Tao Ma (2):
>>       block: Remove 'plug/unplug' comment in blk_execute_rq_nowait
>>       block: Remove extra discard_alignment from hd_struct.
> 
> At boot, I have a crash in part_discard_alignment_show+0x1b/0x50
> 
> CR2 : 000006ac
> 
> fault in : mov    0x2c(%rcx),%edx
> 
> I suspect commit 23ceb5b7719e9276d4 (block: Remove extra
> discard_alignment from hd_struct) being in fault

Hmm, that must be disk->queue == NULL. Do you have a trace of the crash,
with the surrounding boot messages? I wonder what is special about your
setup to trigger something like this.

diff --git a/fs/partitions/check.c b/fs/partitions/check.c
index 8ed4d34..f82e762 100644
--- a/fs/partitions/check.c
+++ b/fs/partitions/check.c
@@ -256,10 +256,12 @@ ssize_t part_discard_alignment_show(struct device *dev,
 {
 	struct hd_struct *p = dev_to_part(dev);
 	struct gendisk *disk = dev_to_disk(dev);
+	unsigned int alignment = 0;
 
-	return sprintf(buf, "%u\n",
-			queue_limit_discard_alignment(&disk->queue->limits,
-							p->start_sect));
+	if (disk->queue)
+		alignment = queue_limit_discard_alignment(&disk->queue->limits,
+								p->start_sect);
+	return sprintf(buf, "%u\n", alignment);
 }
 
 ssize_t part_stat_show(struct device *dev,


-- 
Jens Axboe

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/