Re: [Qemu-devel] [PATCH] block/io.c: fix for the allocation failure

[Eric Blake]

On 8/19/19 2:46 PM, Denis V. Lunev wrote:
> On 8/17/19 5:56 PM, Eric Blake wrote:
>> On 8/17/19 9:49 AM, Eric Blake wrote:
>>
>>>> This change is a regression of sorts.  Now, you are unconditionally
>>>> attempting the fallback for ALL failures (such as EIO) and for all
>>>> drivers, even when that was not previously attempted and increases the
>>>> traffic.  I think we should revert this patch and instead fix the
>>>> fallocate() path to convert whatever ACTUAL errno you got from unaligned
>>>> fallocate failure into ENOTSUP (that is, just the file-posix.c location
>>>> that failed), while leaving all other errors as immediately fatal.
>> Or even better, fix the call site of fallocate() to skip attempting an
>> unaligned fallocate(), and just directly return ENOTSUP, rather than
>> trying to diagnose EINVAL after the fact.
>>
> No way. Single ENOTSUP will turn off fallocate() support on caller side
> while
> aligned (99.99% of calls) works normally.

I didn't mean skip fallocate() unconditionally, only when unaligned:

if (request not aligned enough)
   return -ENOTSUP;
fallocate() ...

so that the 99.99% requests that ARE aligned get to use fallocate()
normally.

-- 
Eric Blake, Principal Software Engineer
Red Hat, Inc.           +1-919-301-3226
Virtualization:  qemu.org | libvirt.org

signature.asc
Description: OpenPGP digital signature