Re: [PATCH v1] media: staging: tegra-vde: Replace debug messages with trace points

[Dmitry Osipenko <digetx@gmail.com>]

On Wednesday, 25 July 2018 04:20:29 MSK Dmitry Osipenko wrote:
> On Wednesday, 25 July 2018 03:37:33 MSK Mauro Carvalho Chehab wrote:
> > Em Wed, 25 Jul 2018 01:38:37 +0300
> > 
> > Dmitry Osipenko <digetx@gmail.com> escreveu:
> > > On Wednesday, 25 July 2018 01:06:52 MSK Mauro Carvalho Chehab wrote:
> > > > Em Sat,  7 Jul 2018 19:20:49 +0300
> > > > 
> > > > Dmitry Osipenko <digetx@gmail.com> escreveu:
> > > > > Trace points are much more efficient than debug messages for
> > > > > extensive
> > > > > tracing and could be conveniently enabled / disabled dynamically,
> > > > > hence
> > > > > let's replace debug messages with the trace points.
> > > > 
> > > > This patch require some work:
> > > > 
> > > > $ make ARCH=i386  CF=-D__CHECK_ENDIAN__
> > > > CONFIG_DEBUG_SECTION_MISMATCH=y
> > > > C=1
> > > > W=1 CHECK='compile_checks' M=drivers/staging/media
> > > > 
> > > > ./include/linux/slab.h:631:13: error: undefined identifier
> > > > '__builtin_mul_overflow' ./include/linux/slab.h:631:13: warning: call
> > > > with
> > > > no type!
> > > > fixdep: error opening file: drivers/staging/media/tegra-vde/trace.h:
> > > > No
> > > > such file or directory
> > > > 
> > >   CHECK   drivers/staging/media/tegra-vde/tegra-vde.c
> > > 
> > > /bin/sh: compile_checks: command not found
> > > 
> > > Upstream kernel doesn't have "compile_checks" script and I can't find it
> > > anywhere else.
> > 
> > This is just a call for smatch/sparse:
> > 
> > #!/bin/bash
> > /devel/smatch/smatch -p=kernel $@
> > # This is too pedantic and produce lots of false-positives
> > #/devel/smatch/smatch --two-passes -- -p=kernel $@
> > /devel/sparse/sparse $@
> > 
> > However, the problem here is that you're doing a 64 bits division.
> > That causes compilation to break with 32 bits. you need to use
> > do_div & friends.
> 
> The tegra-vde driver code is fine, it is a known issue with the kernels
> checker [0]. Unfortunately the patch for the checker haven't been applied
> yet.
> 
> [0] https://www.spinics.net/lists/kernel/msg2824058.html

Mauro, are you going to un-reject patch [0] and apply it or there is some 
action needed from my side now?

[0] https://patchwork.linuxtv.org/patch/51002/