Re: [PATCH] scsi: mpt3sas: fix hang on ata passthru commands

From: Martin K. Petersen
Date: Tue Jan 17 2017 - 15:59:22 EST


>>>>> "Christoph" == Christoph Hellwig <hch@xxxxxx> writes:

Christoph> On Tue, Jan 17, 2017 at 07:43:51PM +0530, Sreekanth Reddy wrote:
>> [Sreekanth] Just for readability purpose, can use use "if
>> (test_bit(0, &sas_device_priv_data->ata_command_pending)" instead of
>> "if (sas_device_priv_data->ata_command_pending)". Since while
>> setting & clearing the bit we are using atomic bit operations. I
>> don't see any issue functionality wise.

Christoph> I agree. Also while we're into nitpicking - it would be good
Christoph> to give bit 0 of the bitmap a name instead of hardcoding the
Christoph> 0.

I tweaked the test case. We can name the bit later if more flags are
needed (and in that case the ata_command_pending would need to get
renamed too).

In any case. This issue has taken waaay too long to get resolved so the
patch is now queued up in 4.10/scsi-fixes.

Thanks everyone!

--
Martin K. Petersen Oracle Linux Engineering