Hi,

On Wed, Apr 19, 2023 at 3:57 PM Douglas Anderson [off-list ref] wrote:

This is an attempt to resurrect Sumit's old patch series [1] that
allowed us to use the arm64 pseudo-NMI to get backtraces of CPUs and
also to round up CPUs in kdb/kgdb. The last post from Sumit that I
could find was v7, so I called this series v8. I haven't copied all of
his old changelongs here, but you can find them from the link.

Since v7, I have:
* Addressed the small amount of feedback that was there for v7.
* Rebased.
* Added a new patch that prevents us from spamming the logs with idle
  tasks.
* Added an extra patch to gracefully fall back to regular IPIs if
  pseudo-NMIs aren't there.

Since there appear to be a few different patches series related to
being able to use NMIs to get stack traces of crashed systems, let me
try to organize them to the best of my understanding:

a) This series. On its own, a) will (among other things) enable stack
   traces of all running processes with the soft lockup detector if
   you've enabled the sysctl "kernel.softlockup_all_cpu_backtrace". On
   its own, a) doesn't give a hard lockup detector.

b) A different recently-posted series [2] that adds a hard lockup
   detector based on perf. On its own, b) gives a stack crawl of the
   locked up CPU but no stack crawls of other CPUs (even if they're
   locked too). Together with a) + b) we get everything (full lockup
   detect, full ability to get stack crawls).

c) The old Android "buddy" hard lockup detector [3] that I'm
   considering trying to upstream. If b) lands then I believe c) would
   be redundant (at least for arm64). c) on its own is really only
   useful on arm64 for platforms that can print CPU_DBGPCSR somehow
   (see [4]). a) + c) is roughly as good as a) + b).

[1] https://lore.kernel.org/linux-arm-kernel/1604317487-14543-1-git-send-email-sumit.garg@linaro.org/ (local)
[2] https://lore.kernel.org/linux-arm-kernel/20220903093415.15850-1-lecopzer.chen@mediatek.com/ (local)
[3] https://issuetracker.google.com/172213097
[4] https://issuetracker.google.com/172213129

Changes in v8:
- dynamic_ipi_setup() and dynamic_ipi_teardown() no longer take cpu param
- dynamic_ipi_setup() and dynamic_ipi_teardown() no longer take cpu param
- Add loongarch support, too
- Removed "#ifdef CONFIG_SMP" since arm64 is always SMP
- "Tag the arm64 idle functions as __cpuidle" new for v8
- "Provide a stub kgdb_nmicallback() if !CONFIG_KGDB" new for v8
- "Fallback to a regular IPI if NMI isn't enabled" new for v8

Douglas Anderson (3):
  arm64: idle: Tag the arm64 idle functions as __cpuidle
  kgdb: Provide a stub kgdb_nmicallback() if !CONFIG_KGDB
  arm64: ipi_nmi: Fallback to a regular IPI if NMI isn't enabled

Sumit Garg (7):
  arm64: Add framework to turn IPI as NMI
  irqchip/gic-v3: Enable support for SGIs to act as NMIs
  arm64: smp: Assign and setup an IPI as NMI
  nmi: backtrace: Allow runtime arch specific override
  arm64: ipi_nmi: Add support for NMI backtrace
  kgdb: Expose default CPUs roundup fallback mechanism
  arm64: kgdb: Roundup cpus using IPI as NMI

 arch/arm/include/asm/irq.h       |   2 +-
 arch/arm/kernel/smp.c            |   3 +-
 arch/arm64/include/asm/irq.h     |   4 ++
 arch/arm64/include/asm/nmi.h     |  17 +++++
 arch/arm64/kernel/Makefile       |   2 +-
 arch/arm64/kernel/idle.c         |   4 +-
 arch/arm64/kernel/ipi_nmi.c      | 103 +++++++++++++++++++++++++++++++
 arch/arm64/kernel/kgdb.c         |  18 ++++++
 arch/arm64/kernel/smp.c          |   8 +++
 arch/loongarch/include/asm/irq.h |   2 +-
 arch/loongarch/kernel/process.c  |   3 +-
 arch/mips/include/asm/irq.h      |   2 +-
 arch/mips/kernel/process.c       |   3 +-
 arch/powerpc/include/asm/nmi.h   |   2 +-
 arch/powerpc/kernel/stacktrace.c |   3 +-
 arch/sparc/include/asm/irq_64.h  |   2 +-
 arch/sparc/kernel/process_64.c   |   4 +-
 arch/x86/include/asm/irq.h       |   2 +-
 arch/x86/kernel/apic/hw_nmi.c    |   3 +-
 drivers/irqchip/irq-gic-v3.c     |  29 ++++++---
 include/linux/kgdb.h             |  13 ++++
 include/linux/nmi.h              |  12 ++--
 kernel/debug/debug_core.c        |   8 ++-
 23 files changed, 217 insertions(+), 32 deletions(-)

It's been 3 weeks and I haven't heard a peep on this series. That
means nobody has any objections and it's all good to land, right?
Right? :-P

Seriously, though, I really think we should figure out how to get this
landed. There's obviously interest from several different parties and
I'm chomping at the bit waiting to spin this series according to your
wishes. I also don't think there's anything super scary/ugly here. IMO
the ideal situation is that folks are OK with the idea presented in
the last patch in the series ("arm64: ipi_nmi: Fallback to a regular
IPI if NMI isn't enabled") and then I can re-spin this series to be
_much_ simpler. That being said, I'm also OK with dropping that patch
and starting the discussion for how to land the rest of the patches in
the series.

Please let me know!

-Doug