Re: [RFC PATCH] checkpatch: check for function calls with struct or union on stack
From: Andrew Morton
Date: Thu Jul 26 2018 - 16:38:36 EST
On Thu, 26 Jul 2018 13:05:29 -0700 Joe Perches <joe@xxxxxxxxxxx> wrote:
> On Thu, 2018-07-26 at 12:28 -0700, Andrew Morton wrote:
> > On Thu, 26 Jul 2018 12:25:33 -0700 Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx> wrote:
> > > I'll give it a spin, see how noisy it is.
> > Actually, I would prefer if the message, changelog and title
> > used the term "passed by value". It's a more familiar term
> > and it is possible for a passed-by-value aggregate to in fact
> > be passed in registers.
> RFC, No worries, I'll change it if it's OK.
> I'm testing it right now against the last 5000 commits
> (which takes awhile here) via
> $ git log --no-merges --format=oneline -5000 | \
> cut -f1 -d" " | \
> while read commit ; do \
> echo $commit; \
> ./scripts/checkpatch.pl --git $commit --types=aggregate_on_stack --quiet --no-summary ; \
> It doesn't seem noisy at all, but maybe there are a few
> known structs like "struct timespec64" that could be
> The only real hits so far are:
> commit f2fb56afba11426ee5c9603b28a9827c530909c0
> WARNING: Unusual use of 'struct msm_display_topology' on stack
> #28374: FILE: drivers/gpu/drm/msm/disp/dpu1/dpu_rm.c:149:
> +enum dpu_rm_topology_name
> +dpu_rm_get_topology_name(struct msm_display_topology topology)
hm. 12 bytes. I don't know if this would be more efficient than using
const struct msm_display_topology*.
> WARNING: Unusual use of 'struct cppc_perf_fb_ctrs' on stack
> #45: FILE: drivers/cpufreq/cppc_cpufreq.c:307:
> +static int cppc_get_rate_from_fbctrs(struct cppc_cpudata *cpu,
> + struct cppc_perf_fb_ctrs fb_ctrs_t0,
> + struct cppc_perf_fb_ctrs fb_ctrs_t1)
Two 32-byte structures? That seems excessive.
Yes, a warning which sends developers back for a bit more thinnking