Re: [Qemu-block] [Qemu-devel] [PATCH for-2.11] iotests: Fix 176 on 32-bit host

[Eric Blake]

On 11/17/2017 04:59 PM, John Snow wrote:
>>> So for this test, the easiest solution is to filter out the
>>> actual hash value.  Broken in commit 4096974e.
>>
>> Of course, if Kevin sends a v2 pull, it's probably better to just squash
>> this in to my original testsuite change (since a v2 would probably
>> invalidate that commit id).
>>
> 
> Whichever way you go,

Commit 4096974e has already landed in master, so this commit message is
correct as-is and belongs in the next pull request (whether for -rc2 or
-rc3 depends on timing).

> 
> Reviewed-by: John Snow 
> 
> I don't know how important it is to nail this down, but the purpose of
> this command is primarily for sanity testing and not necessarily between
> architectures, so this might just be a limitation to document.
> 
> Also, don't use this for anything except testing.

Indeed, the name of x-debug-block-dirty-bitmap-sha256 already implies
its limited portability; but it also gives us free reign to change it's
behavior if we find something we like better for the same ease of bitmap
testing/debugging.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.   +1-919-301-3266
Virtualization:  qemu.org | libvirt.org



signature.asc
Description: OpenPGP digital signature

Re: [Qemu-block] [Qemu-devel] [PATCH for-2.11] iotests: Fix 176 on 32-bit host

[John Snow]

On 11/17/2017 02:10 PM, Eric Blake wrote:
> On 11/17/2017 01:04 PM, Eric Blake wrote:
>> The contents of a qcow2 bitmap are rounded up to a size that
>> matches the number of bits available for the granularity, but
>> that granularity differs for 32-bit hosts (our default 64k
>> cluster allows for 2M bitmap coverage per 'long') and 64-bit
>> hosts (4M bitmap per 'long').  If the image is a multiple of
>> 2M but not 4M, then the number of bytes occupied by the array
>> of longs in memory differs between architecture, thus
>> resulting in different SHA256 hashes.
>>
>> Furthermore (but untested by me), if our computation of the
>> SHA256 hash is at all endian-dependent because of how we store
>> data in memory, that's another variable we'd have to account
>> for (ideally, we specified the bitmap stored in qcow2 as
>> fixed-endian on disk, because the same qcow2 file must be
>> usable across any architecture; but that says nothing about
>> how we represent things in memory).  But we already have test
>> 165 to validate that bitmaps are stored correctly on disk,
>> while this test is merely testing that the bitmap exists.
>>
>> So for this test, the easiest solution is to filter out the
>> actual hash value.  Broken in commit 4096974e.
> 
> Of course, if Kevin sends a v2 pull, it's probably better to just squash
> this in to my original testsuite change (since a v2 would probably
> invalidate that commit id).
> 

Whichever way you go,

Reviewed-by: John Snow 

I don't know how important it is to nail this down, but the purpose of
this command is primarily for sanity testing and not necessarily between
architectures, so this might just be a limitation to document.

Also, don't use this for anything except testing.

:(

Re: [Qemu-block] [Qemu-devel] [PATCH for-2.11] iotests: Fix 176 on 32-bit host

[Eric Blake]

On 11/17/2017 01:04 PM, Eric Blake wrote:
> The contents of a qcow2 bitmap are rounded up to a size that
> matches the number of bits available for the granularity, but
> that granularity differs for 32-bit hosts (our default 64k
> cluster allows for 2M bitmap coverage per 'long') and 64-bit
> hosts (4M bitmap per 'long').  If the image is a multiple of
> 2M but not 4M, then the number of bytes occupied by the array
> of longs in memory differs between architecture, thus
> resulting in different SHA256 hashes.
> 
> Furthermore (but untested by me), if our computation of the
> SHA256 hash is at all endian-dependent because of how we store
> data in memory, that's another variable we'd have to account
> for (ideally, we specified the bitmap stored in qcow2 as
> fixed-endian on disk, because the same qcow2 file must be
> usable across any architecture; but that says nothing about
> how we represent things in memory).  But we already have test
> 165 to validate that bitmaps are stored correctly on disk,
> while this test is merely testing that the bitmap exists.
> 
> So for this test, the easiest solution is to filter out the
> actual hash value.  Broken in commit 4096974e.

Of course, if Kevin sends a v2 pull, it's probably better to just squash
this in to my original testsuite change (since a v2 would probably
invalidate that commit id).

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.   +1-919-301-3266
Virtualization:  qemu.org | libvirt.org



signature.asc
Description: OpenPGP digital signature