Re: [PATCH v2] Fix %#0 misuses

[Dr. David Alan Gilbert]

* Philippe Mathieu-Daudé (f4bug@amsat.org) wrote:
> On 2/2/22 19:34, Dr. David Alan Gilbert (git) wrote:
> > From: "Dr. David Alan Gilbert" <dgilbert@redhat.com>
> > 
> > Joe Perches pointed out on lkml [1] that the format special %# (which
> > adds 0x) is included in the character count, i.e.:
> > 
> >    printf("0: %#08x\n0: %#08x\n", 0xabcdef01,1);
> > gives:
> > 0: 0xabcdef01
> > 0: 0x000001
> > 
> > rather than padding to the expected 8 data characters.
> > 
> > Replace all the '%#08' cases by '0x%08' and
> > also handle a '%#02'; there are some other cases mostly
> > in testing and a few that look like the authors have
> > thought about the size, that I've ignored for now.
> > 
> > (Note I've not managed to test most of these)
> > 
> > Signed-off-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
> > 
> > [1] 
> > https://lore.kernel.org/lkml/9499203f1e993872b384aabdec59ac223a8ab931.camel@perches.com/
> > 
> > Fixup %#
> > ---
> >   chardev/baum.c          | 2 +-
> >   disas/alpha.c           | 2 +-
> >   disas/sparc.c           | 2 +-
> >   hw/arm/omap1.c          | 2 +-
> >   hw/timer/a9gtimer.c     | 4 ++--
> >   include/hw/arm/omap.h   | 5 +++--
> >   softmmu/device_tree.c   | 2 +-
> >   target/openrisc/disas.c | 2 +-
> >   8 files changed, 11 insertions(+), 10 deletions(-)
> 
> Few more:
> 
> $ git grep -E '\%#[0-9]{1,2}[dDxX]'

Thanks, actually 1,3 is worth it :-)

> chardev/baum.c:369:            DPRINTF("Broken packet %#2x, tossing\n",
> req); \

Interestingly without the 0

> disas/alpha.c:1854:  (*info->fprintf_func) (info->stream, ".long %#08x",
> insn);
> hw/audio/ac97.c:244:    dolog ("bd %2d addr=%#x ctl=%#06x len=%#x(%d
> bytes)\n",
> hw/dma/i8257.c:265:        linfo ("write_cont: nport %#06x, ichan % 2d, val
> %#06x\n",
> hw/dma/i8257.c:290:    ldebug ("read_cont: nport %#06x, iport %#04x val
> %#x\n", nport, iport, val);

So some of these are a bit weird and might need thinking about; I
*think* they're 06 because the author realised the 0x was counted and
bumped it up from 04; but that actually means I'd need to understand the
device code before changing it.

> softmmu/device_tree.c:372:        error_report("%s: Couldn't set %s/%s =
> %#08x: %s", __func__,
> target/openrisc/disas.c:48:        output(".long", "%#08x", insn);
> tests/qemu-iotests/nbd-fault-injector.py:187: print('unrecognized command
> type %#02x' % req.type)
> tests/tcg/arm/fcvt.c:60:    printf("%02d   HALF: %#04x  (%#x => %s)\n",
> tests/tcg/multiarch/libs/float_helpers.c:72:    asprintf(&fmt, "f16(%#04x)",
> num);
> 
> In mood to add a regexp to checkpatch? :)

I think it already has one for new patches.

> Reviewed-by: Philippe Mathieu-Daudé <f4bug@amsat.org>

Thanks!

Dave

-- 
Dr. David Alan Gilbert / dgilbert@redhat.com / Manchester, UK