Re: [Qemu-devel] [PATCH 1/2] config: qemu_config_parse() return number of config groups

[Markus Armbruster]

Eduardo Habkost <address@hidden> writes:

> On Mon, Feb 27, 2017 at 06:59:51PM +0100, Markus Armbruster wrote:
>> Uh, I totally forgot about this series.  My apologies...
>> 
>> Eduardo Habkost <address@hidden> writes:
>> 
>> > Change qemu_config_parse() to return the number of config groups
>> > in success and -EINVAL on error. This will allow callers of
>> > qemu_config_parse() to check if something was really loaded from
>> > the config file.
>> >
>> > All existing callers of qemu_config_parse() and
>> > qemu_read_config_file() only check if the return value was
>> > negative, so the change shouldn't affect them.
>> 
>> Two of them:
>> 
>> * read_config() maps negative value to -EINVAL.  Callers:
>> 
>>   - blkdebug_open() passes it on.  As a .bdrv_file_open() method, it's
>>     supposed to return -errno on failure.  Good.
>> 
>> * qemu_read_config_file() maps non-zero value to -EINVAL.  Callers:
>> 
>>   - qemu_read_default_config_file() maps -EINVAL to zero.  WTF?
>
> qemu_read_config_file() maps -ENOENT to zero, not -EINVAL. It
> will ignore the error only if the default config file doesn't
> exist (if fopen() sets errno to ENOENT).

Crosseyed reviewer syndrome...

>>   - main() passes sterror(EINVAL) to error_report().  Good.
>> 
>> Also: qemu_config_parse() reports errors with error_report().  Let's
>> have another look at its callers:
>> 
>> * read_config() has an Error ** parameter.  Bad.  Care to convert the
>>   sucker to Error?
>
> I can do it in a separate patch, unless you believe it must be
> done by this patch patch in a single step.

It *should* be a separate patch.

>> 
>> * qemu_read_config_file() doesn't report errors.  Callers:
>> 
>>   - qemu_read_default_config_file() doesn't report errors.  Its called
>>     by main(), and ...
>
> This means fopen() errors won't get reported. Good catch. I will
> send a separate fix.
>
>> 
>>   - main() reports with error_report().  Good.
>> 
> [...]
>> 
>> I think this mapping to -EINVAL is also superfluous now:
>> 
>> diff --git a/block/blkdebug.c b/block/blkdebug.c
>> index 6117ce5..fbefa9e 100644
>> --- a/block/blkdebug.c
>> +++ b/block/blkdebug.c
>> @@ -252,7 +252,6 @@ static int read_config(BDRVBlkdebugState *s, const char 
>> *filename,
>>          ret = qemu_config_parse(f, config_groups, filename);
>>          if (ret < 0) {
>>              error_setg(errp, "Could not parse blkdebug config file");
>> -            ret = -EINVAL;
>>              goto fail;
>
> Good catch. I will squash it into the patch.

Looking forward to v2 :)