Re: [Qemu-devel] [PATCH v3 07/15] vfio: Report warnings with warn_report(), not error_printf()

[Markus Armbruster]

Alex Williamson <address@hidden> writes:

> On Thu, 18 Apr 2019 08:18:56 +0200
> Markus Armbruster <address@hidden> wrote:
>
>> Alex Williamson <address@hidden> writes:
>> 
>> > On Wed, 17 Apr 2019 21:06:33 +0200
>> > Markus Armbruster <address@hidden> wrote:
>> >  
>> >> Cc: Alex Williamson <address@hidden>
>> >> Signed-off-by: Markus Armbruster <address@hidden>
>> >> ---
>> >>  hw/vfio/pci.c | 19 +++++++++++++------
>> >>  1 file changed, 13 insertions(+), 6 deletions(-)
>> >> 
>> >> diff --git a/hw/vfio/pci.c b/hw/vfio/pci.c
>> >> index 504019c458..0142819ea6 100644
>> >> --- a/hw/vfio/pci.c
>> >> +++ b/hw/vfio/pci.c
>> >> @@ -947,8 +947,10 @@ static void vfio_pci_size_rom(VFIOPCIDevice *vdev)
>> >>      if (vdev->pdev.romfile || !vdev->pdev.rom_bar) {
>> >>          /* Since pci handles romfile, just print a message and return */
>> >>          if (vfio_blacklist_opt_rom(vdev) && vdev->pdev.romfile) {
>> >> -            error_printf("Warning : Device at %s is known to cause 
>> >> system instability issues during option rom execution. Proceeding anyway 
>> >> since user specified romfile\n",
>> >> -                         vdev->vbasedev.name);
>> >> +            warn_report("Device at %s is known to cause system 
>> >> instability"
>> >> +                        " issues during option rom execution",
>> >> +                        vdev->vbasedev.name);
>> >> +            error_printf("Proceeding anyway since user specified 
>> >> romfile\n");  
>> >
>> > I'm confused, the original warning is "this device is know to have
>> > issues, proceeding because you asked me to".  Are we categorizing the
>> > first half as a warning and the latter as random uncategorized error
>> > spew?  Did an automated script chunk it this way because of the period
>> > and strict application of the "single phrase" specification of
>> > warn_report()?  If this is the recommended semantics, I'm not sure how
>> > I'd know to generate this myself for similar situations.  Should we
>> > instead try to express this in something acceptable as a single
>> > phrase?  Thanks,  
>> 
>> This is an instance of the following error reporting pattern:
>> 
>>     concise error / warning message (one line, single phrase)
>>     additional information (free format)
>> 
>> We use error_report() / warn_report() for the former (this adds
>> decorations to the message), and error_printf() for the latter.
>> 
>> "git-grep -w error_printf" will lead you to other instances.  Recommend
>> to look with this series applied, because it removes misuses of
>> error_printf().
>> 
>> Particularly relevant instances are error_report_err() and
>> warn_report_err(): these print the error proper with error_report() /
>> warn_report_err(), and the hint, if any, with error_printf().
>> 
>> Clearer now?
>
> I can't guarantee that I'd be able to reproduce these sorts of
> semantics without prompting, but yes, there does seem to be some method
> to the madness ;)  Thanks,

Hopefully, presence of good examples, absence of bad examples, and
review will do the trick often enough.

> Acked-by: Alex Williamson <address@hidden>

Thanks!