From: Charlie Jenkins <charlie@rivosinc.com>
To: Vineet Gupta <vineetg@rivosinc.com>
Cc: "Shuah Khan" <shuah@kernel.org>,
"Paul Walmsley" <paul.walmsley@sifive.com>,
"Palmer Dabbelt" <palmer@dabbelt.com>,
"Albert Ou" <aou@eecs.berkeley.edu>,
"Andy Chiu" <andy.chiu@sifive.com>,
"Björn Töpel" <bjorn@kernel.org>,
linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
linux-riscv@lists.infradead.org
Subject: Re: [PATCH] riscv: selftests: Add signal handling vector tests
Date: Wed, 3 Apr 2024 17:43:50 -0700 [thread overview]
Message-ID: <Zg33xhwOe6Zwi23A@ghost> (raw)
In-Reply-To: <56863fe7-4d82-4f37-a518-2acf368c3ea6@rivosinc.com>
On Wed, Apr 03, 2024 at 05:27:23PM -0700, Vineet Gupta wrote:
> On 4/3/24 16:50, Charlie Jenkins wrote:
> > Add two tests to check vector save/restore when a signal is received
> > during a vector routine. One test ensures that a value is not clobbered
> > during signal handling. The other verifies that vector registers
> > modified in the signal handler are properly reflected when the signal
> > handling is complete.
>
> Hmm, isn't this testing two contradictory things ?
>
> We do want V regs to be not clobbered across a handled signal, o/w V
> enabled code would just not work at all.
> That implies that anything done by signal handler should just be
> discarded - no ?
>
> Am I missing something.
>
> -Vineet
This is referring to the user-implemented signal handler. The test in
question is testing the ability of the user-implemented signal handler
to be able to modify the vector registers. In the test:
static void vector_override(int sig_no, siginfo_t *info, void *vcontext)
{ ...
*(int *)v_ext_state->datap = SIGNAL_HANDLER_OVERRIDE;
... }
This line is changing the saved state of the user program's vector
registers.
When the signal handling concludes and the original user program resumes
execution, the content of the user-context registers that were modified
by the signal handler should be restored into the hardware registers.
Hence the test case checks that the v0 register contains the value
SIGNAL_HANDLER_OVERRIDE.
EXPECT_EQ(SIGNAL_HANDLER_OVERRIDE, result);
- Charlie
>
> >
> > Signed-off-by: Charlie Jenkins <charlie@rivosinc.com>
> > ---
> > These tests came about to highlight the bug fixed in
> > https://lore.kernel.org/lkml/20240403072638.567446-1-bjorn@kernel.org/
> > and will only pass with that fix applied.
> > ---
> > tools/testing/selftests/riscv/Makefile | 2 +-
> > tools/testing/selftests/riscv/sigreturn/.gitignore | 1 +
> > tools/testing/selftests/riscv/sigreturn/Makefile | 12 ++++
> > .../testing/selftests/riscv/sigreturn/sigreturn.c | 82 ++++++++++++++++++++++
> > 4 files changed, 96 insertions(+), 1 deletion(-)
> >
> > diff --git a/tools/testing/selftests/riscv/Makefile b/tools/testing/selftests/riscv/Makefile
> > index 4a9ff515a3a0..7ce03d832b64 100644
> > --- a/tools/testing/selftests/riscv/Makefile
> > +++ b/tools/testing/selftests/riscv/Makefile
> > @@ -5,7 +5,7 @@
> > ARCH ?= $(shell uname -m 2>/dev/null || echo not)
> >
> > ifneq (,$(filter $(ARCH),riscv))
> > -RISCV_SUBTARGETS ?= hwprobe vector mm
> > +RISCV_SUBTARGETS ?= hwprobe vector mm sigreturn
> > else
> > RISCV_SUBTARGETS :=
> > endif
> > diff --git a/tools/testing/selftests/riscv/sigreturn/.gitignore b/tools/testing/selftests/riscv/sigreturn/.gitignore
> > new file mode 100644
> > index 000000000000..35002b8ae780
> > --- /dev/null
> > +++ b/tools/testing/selftests/riscv/sigreturn/.gitignore
> > @@ -0,0 +1 @@
> > +sigreturn
> > diff --git a/tools/testing/selftests/riscv/sigreturn/Makefile b/tools/testing/selftests/riscv/sigreturn/Makefile
> > new file mode 100644
> > index 000000000000..eb8bac9279a8
> > --- /dev/null
> > +++ b/tools/testing/selftests/riscv/sigreturn/Makefile
> > @@ -0,0 +1,12 @@
> > +# SPDX-License-Identifier: GPL-2.0
> > +# Copyright (C) 2021 ARM Limited
> > +# Originally tools/testing/arm64/abi/Makefile
> > +
> > +CFLAGS += -I$(top_srcdir)/tools/include
> > +
> > +TEST_GEN_PROGS := sigreturn
> > +
> > +include ../../lib.mk
> > +
> > +$(OUTPUT)/sigreturn: sigreturn.c
> > + $(CC) -static -o$@ $(CFLAGS) $(LDFLAGS) $^
> > diff --git a/tools/testing/selftests/riscv/sigreturn/sigreturn.c b/tools/testing/selftests/riscv/sigreturn/sigreturn.c
> > new file mode 100644
> > index 000000000000..62397d5934f1
> > --- /dev/null
> > +++ b/tools/testing/selftests/riscv/sigreturn/sigreturn.c
> > @@ -0,0 +1,82 @@
> > +// SPDX-License-Identifier: GPL-2.0-only
> > +#include <signal.h>
> > +#include <stdio.h>
> > +#include <stdlib.h>
> > +#include <ucontext.h>
> > +#include <linux/ptrace.h>
> > +#include "../../kselftest_harness.h"
> > +
> > +#define RISCV_V_MAGIC 0x53465457
> > +#define DEFAULT_VALUE 2
> > +#define SIGNAL_HANDLER_OVERRIDE 3
> > +
> > +static void simple_handle(int sig_no, siginfo_t *info, void *vcontext)
> > +{
> > + ucontext_t *context = vcontext;
> > +
> > + context->uc_mcontext.__gregs[REG_PC] = context->uc_mcontext.__gregs[REG_PC] + 4;
> > +}
> > +
> > +static void vector_override(int sig_no, siginfo_t *info, void *vcontext)
> > +{
> > + ucontext_t *context = vcontext;
> > +
> > + // vector state
> > + struct __riscv_extra_ext_header *ext;
> > + struct __riscv_v_ext_state *v_ext_state;
> > +
> > + /* Find the vector context. */
> > + ext = (void *)(&context->uc_mcontext.__fpregs);
> > + if (ext->hdr.magic != RISCV_V_MAGIC) {
> > + fprintf(stderr, "bad vector magic: %x\n", ext->hdr.magic);
> > + abort();
> > + }
> > +
> > + v_ext_state = (void *)((char *)(ext) + sizeof(*ext));
> > +
> > + *(int *)v_ext_state->datap = SIGNAL_HANDLER_OVERRIDE;
> > +
> > + context->uc_mcontext.__gregs[REG_PC] = context->uc_mcontext.__gregs[REG_PC] + 4;
> > +}
> > +
> > +static int vector_sigreturn(int data, void (*handler)(int, siginfo_t *, void *))
> > +{
> > + int after_sigreturn;
> > + struct sigaction sig_action = {
> > + .sa_sigaction = handler,
> > + .sa_flags = SA_SIGINFO
> > + };
> > +
> > + sigaction(SIGSEGV, &sig_action, 0);
> > +
> > + asm(".option push \n\
> > + .option arch, +v \n\
> > + vsetivli x0, 1, e32, ta, ma \n\
> > + vmv.s.x v0, %1 \n\
> > + # Generate SIGSEGV \n\
> > + lw a0, 0(x0) \n\
> > + vmv.x.s %0, v0 \n\
> > + .option pop" : "=r" (after_sigreturn) : "r" (data));
> > +
> > + return after_sigreturn;
> > +}
> > +
> > +TEST(vector_restore)
> > +{
> > + int result;
> > +
> > + result = vector_sigreturn(DEFAULT_VALUE, &simple_handle);
> > +
> > + EXPECT_EQ(DEFAULT_VALUE, result);
> > +}
> > +
> > +TEST(vector_restore_signal_handler_override)
> > +{
> > + int result;
> > +
> > + result = vector_sigreturn(DEFAULT_VALUE, &vector_override);
> > +
> > + EXPECT_EQ(SIGNAL_HANDLER_OVERRIDE, result);
> > +}
> > +
> > +TEST_HARNESS_MAIN
> >
> > ---
> > base-commit: 4cece764965020c22cff7665b18a012006359095
> > change-id: 20240403-vector_sigreturn_tests-8118f0ac54fa
>
next prev parent reply other threads:[~2024-04-04 0:43 UTC|newest]
Thread overview: 7+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-04-03 23:50 [PATCH] riscv: selftests: Add signal handling vector tests Charlie Jenkins
2024-04-04 0:27 ` Vineet Gupta
2024-04-04 0:43 ` Charlie Jenkins [this message]
2024-04-04 5:46 ` Björn Töpel
2024-04-09 3:01 ` Andy Chiu
2024-05-22 16:40 ` Palmer Dabbelt
2024-05-22 23:51 ` patchwork-bot+linux-riscv
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Zg33xhwOe6Zwi23A@ghost \
--to=charlie@rivosinc.com \
--cc=andy.chiu@sifive.com \
--cc=aou@eecs.berkeley.edu \
--cc=bjorn@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=linux-riscv@lists.infradead.org \
--cc=palmer@dabbelt.com \
--cc=paul.walmsley@sifive.com \
--cc=shuah@kernel.org \
--cc=vineetg@rivosinc.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox