Re: [PATCH v4 0/5] siginfo: ABI fixes for TRAP_PERF

On Tue, 18 May 2021 at 05:47, Eric W. Biederman [off-list ref] wrote:

Marco Elver [off-list ref] writes:

quoted

On Mon, 17 May 2021 at 21:58, Eric W. Biederman [off-list ref] wrote:

quoted

During the merge window an issue with si_perf and the siginfo ABI came
up.  The alpha and sparc siginfo structure layout had changed with the
addition of SIGTRAP TRAP_PERF and the new field si_perf.

The reason only alpha and sparc were affected is that they are the
only architectures that use si_trapno.

Looking deeper it was discovered that si_trapno is used for only
a few select signals on alpha and sparc, and that none of the
other _sigfault fields past si_addr are used at all.  Which means
technically no regression on alpha and sparc.

While the alignment concerns might be dismissed the abuse of
si_errno by SIGTRAP TRAP_PERF does have the potential to cause
regressions in existing userspace.

While we still have time before userspace starts using and depending on
the new definition siginfo for SIGTRAP TRAP_PERF this set of changes
cleans up siginfo_t.

- The si_trapno field is demoted from magic alpha and sparc status and
  made an ordinary union member of the _sigfault member of siginfo_t.
  Without moving it of course.

- si_perf is replaced with si_perf_data and si_perf_type ending the
  abuse of si_errno.

- Unnecessary additions to signalfd_siginfo are removed.

v3: https://lkml.kernel.org/r/m1tuni8ano.fsf_-_@fess.ebiederm.org
v2: https://lkml.kernel.org/r/m14kfjh8et.fsf_-_@fess.ebiederm.org
v1: https://lkml.kernel.org/r/m1zgxfs7zq.fsf_-_@fess.ebiederm.org

This version drops the tests and fine grained handling of si_trapno
on alpha and sparc (replaced assuming si_trapno is valid for
all but the faults that defined different data).

And just to clarify, the rest of the series (including static-asserts)
for the next merge-window will be sent once this series is all sorted,
correct?

That is the plan.

I really wonder about alphas use of si_trapno, and alphas use send_sig
instead of force_sig.  It could be worth looking into those as it
has the potential to simplify the code.

quoted

quoted

Eric W. Biederman (5):
      siginfo: Move si_trapno inside the union inside _si_fault
      signal: Implement SIL_FAULT_TRAPNO
      signal: Factor force_sig_perf out of perf_sigtrap
      signal: Deliver all of the siginfo perf data in _perf
      signalfd: Remove SIL_PERF_EVENT fields from signalfd_siginfo

Looks good, thank you! I build-tested (defconfig -- x86_64, i386, arm,
arm64, m68k, sparc, alpha) this series together with a local patch to
pull in the static asserts from v3. Also re-ran perf_events kselftests
on x86_64 (native and 32bit compat).

Thanks,

Can I have your Tested-by?

Of course,

  Tested-by: Marco Elver [off-list ref]

Thanks,
-- Marco