Message ID | 1626902375-7002-1-git-send-email-tsimpson@quicinc.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
On Wed, 21 Jul 2021 at 22:19, Taylor Simpson <tsimpson@quicinc.com> wrote: > > The following changes since commit 7457b407edd6e8555e4b46488aab2f13959fccf8: > > Merge remote-tracking branch 'remotes/thuth-gitlab/tags/pull-request-2021-07-19' into staging (2021-07-19 11:34:08 +0100) > > are available in the git repository at: > > https://github.com/quic/qemu tags/pull-hex-20210721 > > for you to fetch changes up to 953ea3e4b426ee0c8349343c53e3358cfec720f2: > > linux-test (tests/tcg/multiarch/linux-test.c) add check (2021-07-21 15:54:28 -0500) > > ---------------------------------------------------------------- > The Hexagon target was silently failing the SIGSEGV test because > the signal handler was not called. > > Patch 1/3 fixes the Hexagon target > Patch 2/3 drops include qemu.h from target/hexagon/op_helper.c > Patch 3/3 adds a check that the signal handler is called > > ---------------------------------------------------------------- Hi; the check added in patch 2 seems to fire about 50% of the time for qemu-riscv64, causing 'make check-tcg' to fail. $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test /mnt/nvmedisk/linaro/qemu-for-merges/tests/tcg/multiarch/linux-test.c:500: SIGSEGV handler not called $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test /mnt/nvmedisk/linaro/qemu-for-merges/tests/tcg/multiarch/linux-test.c:500: SIGSEGV handler not called $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test /mnt/nvmedisk/linaro/qemu-for-merges/tests/tcg/multiarch/linux-test.c:500: SIGSEGV handler not called thanks -- PMM
I've added the riscv maintainers and Laurent and Alex to the CC list. Please advise on how to proceed. Is this a known issue with riscv? Should I try to debug the riscv target or remove the change to linux-test.c from the pull request? Thanks, Taylor > -----Original Message----- > From: Peter Maydell <peter.maydell@linaro.org> > Sent: Thursday, July 22, 2021 12:31 PM > To: Taylor Simpson <tsimpson@quicinc.com> > Cc: QEMU Developers <qemu-devel@nongnu.org>; Richard Henderson > <richard.henderson@linaro.org>; Philippe Mathieu-Daudé > <philmd@redhat.com>; Alessandro Di Federico <ale@rev.ng>; Brian Cain > <bcain@quicinc.com> > Subject: Re: [PULL 0/3] SIGSEGV fixes > > On Wed, 21 Jul 2021 at 22:19, Taylor Simpson <tsimpson@quicinc.com> > wrote: > > > > The following changes since commit > 7457b407edd6e8555e4b46488aab2f13959fccf8: > > > > Merge remote-tracking branch > > 'remotes/thuth-gitlab/tags/pull-request-2021-07-19' into staging > > (2021-07-19 11:34:08 +0100) > > > > are available in the git repository at: > > > > https://github.com/quic/qemu tags/pull-hex-20210721 > > > > for you to fetch changes up to > 953ea3e4b426ee0c8349343c53e3358cfec720f2: > > > > linux-test (tests/tcg/multiarch/linux-test.c) add check (2021-07-21 > > 15:54:28 -0500) > > > > ---------------------------------------------------------------- > > The Hexagon target was silently failing the SIGSEGV test because the > > signal handler was not called. > > > > Patch 1/3 fixes the Hexagon target > > Patch 2/3 drops include qemu.h from target/hexagon/op_helper.c Patch > > 3/3 adds a check that the signal handler is called > > > > ---------------------------------------------------------------- > > Hi; the check added in patch 2 seems to fire about 50% of the time for qemu- > riscv64, causing 'make check-tcg' to fail. > > $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test > $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test > $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test > $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test > /mnt/nvmedisk/linaro/qemu-for-merges/tests/tcg/multiarch/linux- > test.c:500: > SIGSEGV handler not called > $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test > $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test > /mnt/nvmedisk/linaro/qemu-for-merges/tests/tcg/multiarch/linux- > test.c:500: > SIGSEGV handler not called > $ ./qemu-riscv64 ./tests/tcg/riscv64-linux-user/linux-test > /mnt/nvmedisk/linaro/qemu-for-merges/tests/tcg/multiarch/linux- > test.c:500: > SIGSEGV handler not called
On 7/23/21 8:48 AM, Taylor Simpson wrote: > I've added the riscv maintainers and Laurent and Alex to the CC list. > > Please advise on how to proceed. Is this a known issue with riscv? Should I try to debug the riscv target or remove the change to linux-test.c from the pull request? Remove the linux-test.c from the pull request. Someone will need to debug riscv, but I don't think you want that to block the bug fix for hexagon. r~ PS: Please use some subject-line indicator for what component you're patching. "SIGSEGV fixes" is impossibly generic.
On 7/23/21 9:10 PM, Richard Henderson wrote: > On 7/23/21 8:48 AM, Taylor Simpson wrote: >> I've added the riscv maintainers and Laurent and Alex to the CC list. >> >> Please advise on how to proceed. Is this a known issue with riscv? >> Should I try to debug the riscv target or remove the change to >> linux-test.c from the pull request? > > Remove the linux-test.c from the pull request. > > Someone will need to debug riscv, but I don't think you want that to > block the bug fix for hexagon. > > > r~ > > > PS: Please use some subject-line indicator for what component you're > patching. "SIGSEGV fixes" is impossibly generic. Although more generic is possible: "fixes"