Re: [PATCH V2 3/3] rseq/selftests: Add support for riscv

From: Vincent Chen <vincent.chen@sifive.com>
To: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Palmer Dabbelt <palmer@dabbelt.com>,
	Paul Walmsley <paul.walmsley@sifive.com>,
	linux-kselftest <linux-kselftest@vger.kernel.org>,
	linux-riscv <linux-riscv@lists.infradead.org>
Subject: Re: [PATCH V2 3/3] rseq/selftests: Add support for riscv
Date: Fri, 27 Mar 2020 16:33:40 +0800	[thread overview]
Message-ID: <CABvJ_xiGzFY6mUAx7xZDUuk=bUWt3SDsDyzFFODmtuXA28NQzg@mail.gmail.com> (raw)
In-Reply-To: <490218026.13345.1585239426829.JavaMail.zimbra@efficios.com>

On Fri, Mar 27, 2020 at 12:17 AM Mathieu Desnoyers
<mathieu.desnoyers@efficios.com> wrote:
>
> ----- On Mar 26, 2020, at 11:49 AM, Palmer Dabbelt palmer@dabbelt.com wrote:
>
> > On Sun, 08 Mar 2020 22:59:52 PDT (-0700), vincent.chen@sifive.com wrote:
> >> Add support for risc-v in the rseq selftests, which covers both
> >> 64-bit and 32-bit ISA with little endian mode.
> >>
> >> Signed-off-by: Vincent Chen <vincent.chen@sifive.com>
> >> ---
> >>  tools/testing/selftests/rseq/param_test.c |  23 ++
> >>  tools/testing/selftests/rseq/rseq-riscv.h | 622 ++++++++++++++++++++++++++++++
> >>  tools/testing/selftests/rseq/rseq.h       |   2 +
> >>  3 files changed, 647 insertions(+)
> >>  create mode 100644 tools/testing/selftests/rseq/rseq-riscv.h
> >
> > There are a ton of checkpatch errors in here.
>
> Is it just my mail client or the main issue is:
>
> ERROR: DOS line endings
>
> ?
I am not sure, but I did not run into this error in my environment.

>
> As far as other issues are concerned, I know there are a few checkpatch
> false-positives that trigger for my rseq-{$ARCH}.h header, from which rseq-riscv.h
> is derived, because it has issues with extensive use of inline assembly.
>
> Thanks,
>
> Mathieu
>
Thank Mathieu for your explanation.

The errors reported by checkpatch.pl can be categorized into two
cases. The first one is "need consistent spacing around %". such as
ERROR: need consistent spacing around '%' (ctx:WxV)
#628: FILE: tools/testing/selftests/rseq/rseq-riscv.h:572:
+       RSEQ_ASM_DEFINE_EXIT_POINT(2f, %l[error2])
where RSEQ_ASM_DEFINE_EXIT_POINT is defined as below
#define RSEQ_ASM_DEFINE_EXIT_POINT(start_ip, exit_ip) \
    ".pushsection __rseq_exit_point_array, \"aw\"\n"        \
    ".quad " __rseq_str(start_ip) ", " __rseq_str(exit_ip) "\n"  \
    ".popsection\n"

These errors were mainly found in the rseq-riscv.h. As Mathieu
mentioned, the RSEQ_ASM_DEFINE_EXIT_POINT macro is used in the inline
assembly, which the second argument %l[error2] indicates the error2 is
a label and it locates outside of the inline assembly. To obey the
syntax, I cannot add a space after % to fix this bug.

The second kind of error is "Macros with complex values should be
enclosed in parentheses" such as
ERROR: Macros with complex values should be enclosed in parentheses
#27: FILE: tools/testing/selftests/rseq/param_test.c:210:
+#define RSEQ_INJECT_INPUT \
+       , [loop_cnt_1]"m"(loop_cnt[1]) \
+       , [loop_cnt_2]"m"(loop_cnt[2]) \
+       , [loop_cnt_3]"m"(loop_cnt[3]) \
+       , [loop_cnt_4]"m"(loop_cnt[4]) \
+       , [loop_cnt_5]"m"(loop_cnt[5]) \
+       , [loop_cnt_6]"m"(loop_cnt[6])
In this case, it was a input operand list of inline assembly, so I
could not add parentheses to enclose them. Except for these two kinds
of error, there are two erros could be solved by adding parentheses. I
also checked it and I think it would be safe. So, I mimic the
implementations of other architecture without using parenthese. If
needed, I think can add the parenthese to solve these two errors.
Thanks