RFC 2nd attempt at PR KVM + SCV and syscall
Get series with b4
b4 am [email protected]Versions newest first
- v1 RFC Latest2 patches 1 replyIn review
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/2 | powerpc/64: remove system call instruction emulationarch/ | +0 | 1 reply | |
| 2/2 | KVM: PPC: Book3S PR: Disable SCV when running AIL is disabledarch/ arch/ arch/ | +33 | 0 replies |
Reviews trailers per patch
1 of 2 patches has a Reviewed-by or Acked-by.
| Patch | Naveen N. Rao |
|---|---|
| 1/2 | A-b |
| 2/2 |
Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.
Cover letter [RFC PATCH 0/2] 2nd attempt at PR KVM + SCV and syscall
The more I looked into system call emulation, the harder it seems to get.
Second proposal is remove it entirely because it's already broken, and just boot-time restrict SCV support if we run with PR possible, in hash mode, on pseries. Unfortunately that catches Power9 PowerVM, but at least OpenPOWER and Power10 by default is okay.
We probably have to do this as a minimal backport at first even if we did later decide we need to fix things in a better way because it looks like it would take a lot of work.
Any comments? mpe helped with looking at the tracing logic and history of the code (thanks) but any mistakes are mine, Naveen are we on the right track here?
Thanks,
Nick
Nicholas Piggin (2):
powerpc/64: remove system call instruction emulation
KVM: PPC: Book3S PR: Disable SCV when running AIL is disabled
arch/powerpc/kernel/exceptions-64s.S | 4 ++++ arch/powerpc/kernel/setup_64.c | 15 ++++++++++++ arch/powerpc/kvm/book3s_pr.c | 20 +++++++++++----- arch/powerpc/lib/sstep.c | 36 ---------------------------- 4 files changed, 33 insertions(+), 42 deletions(-)
Signature · 1 line
2.23.0