Greg Kurz <[email protected]> writes:
> On Fri, 11 Sep 2020 07:30:37 +0200
> Markus Armbruster <[email protected]> wrote:
[...]
>> No, the patch is okay as is.
>>
>> Dumbing it down to keep it simple would also be okay.
>>
>
> What about Ari's proposal to add ERRP_GUARD() and check errors
> with "if (*errp)" like we do for void functions ?
Up to the maintainer. I prefer this patch.
>> Regarding the proposed assertion: do we protect similar conversions from
>> over-wide negative errno int elsewhere?
>>
>
> We do protect int64_t->int conversions in a few places, eg.
> qcow2_write_snapshots() or qemu_spice_create_host_primary(),
> but I couldn't find one about a negarive errno.
Then I'd not protect it here, either.
>> >>> goto fail_log;
>> >>> }
>> >>>
>> >>> + s->cur_log_sector = cur_log_sector;
>> >>> s->nr_entries = le64_to_cpu(log_sb.nr_entries);
>> >>> }
>> >>> } else {
>>
