diff options
author | Oliver Upton <oliver.upton@linux.dev> | 2024-03-16 03:24:57 +0300 |
---|---|---|
committer | Linus Torvalds <torvalds@linux-foundation.org> | 2024-03-16 03:50:22 +0300 |
commit | 82affc97affbdfe83c807ad17147a1220031fd90 (patch) | |
tree | e36a50ecc67cc15c16403d35f9843597e1dca3fd /arch/powerpc/kernel/pci_32.c | |
parent | 277100b3d5fefacba4f5ff18e2e52a9553eb6e3f (diff) | |
download | linux-82affc97affbdfe83c807ad17147a1220031fd90.tar.xz |
Revert "KVM: arm64: Snapshot all non-zero RES0/RES1 sysreg fields for later checking"
This reverts commits 99101dda29e3186b1356b0dc4dbb835c02c71ac9 and
b80b701d5a67d07f4df4a21e09cb31f6bc1feeca.
Linus reports that the sysreg reserved bit checks in KVM have led to
build failures, arising from commit fdd867fe9b32 ("arm64/sysreg: Add
register fields for ID_AA64DFR1_EL1") giving meaning to fields that were
previously RES0.
Of course, this is a genuine issue, since KVM's sysreg emulation depends
heavily on the definition of reserved fields. But at this point the
build breakage is far more offensive, and the right course of action is
to revert and retry later.
All of these build-time assertions were on by default before
commit 99101dda29e3 ("KVM: arm64: Make build-time check of RES0/RES1
bits optional"), so deliberately revert it all atomically to avoid
introducing further breakage of bisection.
Link: https://lore.kernel.org/all/CAHk-=whCvkhc8BbFOUf1ddOsgSGgEjwoKv77=HEY1UiVCydGqw@mail.gmail.com/
Acked-by: Marc Zyngier <maz@kernel.org>
Signed-off-by: Oliver Upton <oliver.upton@linux.dev>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Diffstat (limited to 'arch/powerpc/kernel/pci_32.c')
0 files changed, 0 insertions, 0 deletions