Re: [PATCH v2 1/2] block: report errno when flock fcntl fails

[David Edmondson]

On Wednesday, 2021-01-06 at 10:48:09 +01, Max Reitz wrote:

> On 21.12.20 14:49, David Edmondson wrote:
>> When a call to fcntl(2) for the purpose of manipulating file locks
>> fails, report the error returned by fcntl.
>> Signed-off-by: David Edmondson <david.edmondson@oracle.com>
>> ---
>>   block/file-posix.c         | 20 +++++-----
>>   tests/qemu-iotests/153.out | 76 +++++++++++++++++++-------------------
>>   tests/qemu-iotests/182.out |  2 +-
>>   3 files changed, 49 insertions(+), 49 deletions(-)
>
> tests/qemu-iotests/296.out also needs to be adjusted (found by
> grepping for 'Failed to get').

I will check and add it.

> And now I might as well add this idea: EAGAIN is the most common errno
> when flock fcntl fails, so would it make sense to generate a custom
> error message then?  I’d like to think we could do better than
> “Resource temporarily unavailable”, e.g. perhaps “Lock is already
> taken”.

Kevin previously suggested that we should elide the detail in this
common case (please correct me if that's a misrepresentation), but there
didn't seem to be any consensus for that.

> OTOH, “Resource temporarily unavailable” isn’t *that* bad, so if you
> don’t want to, I won’t push for it.

Other than that it's just more code (which in general is not good), I
don't have a strong opinion.

dme.
-- 
Facts don't do what I want them to.