[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Qemu-block] [PATCH v2 9/9] tests: Add coverage for recent block geo
From: |
Eric Blake |
Subject: |
Re: [Qemu-block] [PATCH v2 9/9] tests: Add coverage for recent block geometry fixes |
Date: |
Mon, 21 Nov 2016 10:16:00 -0600 |
User-agent: |
Mozilla/5.0 (X11; Linux x86_64; rv:45.0) Gecko/20100101 Thunderbird/45.4.0 |
On 11/21/2016 05:38 AM, Kevin Wolf wrote:
>> And now with my last response to patch 5, I do know. So because the
>> "align" option isn't set, the request_alignment defaults to 512.
>> blkdebug will accept requests and pass them on even if they're not
>> aligned to pdiscard_alignment, so that's similar to how iscsi only drops
>> discard requests that are not aligned to request_alignment but passes
>> everything on regardless of whether it's aligned to pdiscard_alignment.
>
> If the 512 byte alignment is important, please make it explicit. I have
> a patch to make blkdebug work with byte alignment and then the default
> will be 1.
In fact, I just rebased my patches on top of your blkdebug cleanup
(since it was visible on your remove-aio-em branch), even though it
hasn't been posted to the list yet, and doing that let me fix the
regression on iotest 98 that was pointed out by Max on 7/9. So I'll
wait to post v3 until your byte-based blkdebug patch has hit the list.
--
Eric Blake eblake redhat com +1-919-301-3266
Libvirt virtualization library http://libvirt.org
signature.asc
Description: OpenPGP digital signature
- Re: [Qemu-block] [PATCH v2 4/9] block: Return -ENOTSUP rather than assert on unaligned discards, (continued)
[Qemu-block] [PATCH v2 5/9] block: Pass unaligned discard requests to drivers, Eric Blake, 2016/11/17
Re: [Qemu-block] [PATCH v2 5/9] block: Pass unaligned discard requests to drivers, Max Reitz, 2016/11/17
Re: [Qemu-block] [PATCH v2 5/9] block: Pass unaligned discard requests to drivers, Kevin Wolf, 2016/11/22