Re: [Qemu-devel] [PATCH for-2.10 1/3] block: Add errp to b{lk, drv}_truncate()

[Max Reitz <mreitz@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1685 bytes --]

On 07.03.2017 11:47, Kevin Wolf wrote:
> Am 06.03.2017 um 20:54 hat Max Reitz geschrieben:
>> For one thing, this allows us to drop the error message generation from
>> qemu-img.c and blockdev.c and instead have it unified in
>> bdrv_truncate().
>>
>> Signed-off-by: Max Reitz <mreitz@redhat.com>
> 
>>  block/commit.c                 |  5 +++--
>>  block/mirror.c                 |  2 +-
> 
> These still pass NULL after the series. Fixing it would require to add a
> way to complete jobs with an Error object, but maybe we should do this
> sooner or later anyway. Not necessarily part of this series, though.
> 
>>  block/vhdx.c                   |  6 +++---
>>  block/vpc.c                    |  2 +-
> 
> vpc can easily be fixed to pass the actual errp from vpc_create() to the
> blk_truncate() call. In vhdx.c, the blk_truncate() call is a bit more
> deeply nested, but it doesn't seem completely unreasonable there either.

I'll look into it, thanks.

>> --- a/qemu-io-cmds.c
>> +++ b/qemu-io-cmds.c
>> @@ -1569,7 +1569,7 @@ static int truncate_f(BlockBackend *blk, int argc, char **argv)
>>          return 0;
>>      }
>>  
>> -    ret = blk_truncate(blk, offset);
>> +    ret = blk_truncate(blk, offset, NULL);
>>      if (ret < 0) {
>>          printf("truncate: %s\n", strerror(-ret));
>>          return 0;
> 
> Come on, using NULL immediately before a printf()? Doing the right thing
> here is trivial.

Yeah, but it's also qemu-io, so I somehow felt both "Better not touch it
because it may break tests" and "Nobody really cares about a nice
message here anyway".

Will fix, since apparently at least you do care. ;-)

Max


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 512 bytes --]