On 10/27/26 06:06, Richard Henderson wrote:
On 10/26/26 22:28, Brian Cain wrote:
The following changes since commit
d73149283205645625bfa9fe029cfb68e4c4a9ae:
Merge tag 'samuel-thibault' ofhttps://people.debian.org/~sthibault/
qemu into staging (2026-10-24 16:09:37 +0200)
are available in the Git repository at:
https://github.com/qualcomm/qemu tags/pull-hex-20261026
for you to fetch changes up to f6beebf1bb08b43d9bb8248ae49fbb9bbd4006d5:
hw/intc/hex-l2vic: find the next pending irq a word at a time
(2026-10-26 18:28:35 -0700)
v3: added "serialize config table as little endian" fix for test
failure in "ubuntu-24.04-s390x-all-system".
----------------------------------------------------------------
hexagon: v85, DSP machines, singlestep, bugfixes
- v85 scalar: atomics, CFI/_rb, vector stack
- new DSP machines: SM8975 NSP, SA8797P NSP0, SC8480XP NSP0, SA8775P
CDSP0,
QCS6490 CDSP.
- bugfixes:
- coverity, ubsan findings in semihosting, decoder
- handling framekey w/dealloc_return
- handle singlestep exception
- performance improvements: TLB flush, l2vic
- test coverage
- MTTCG fixes: enabling MTTCG exposed races in lock/interrupts and ciad.
Not sure what's going on, but there are new python failures in
functional testing:
https://gitlab.com/qemu-project/qemu/-/jobs/17499788773
https://gitlab.com/qemu-project/qemu/-/jobs/17499572582
As a warning, I'm now going to merge Pierrick's tcg helper patch set,
which conflicts with this patch set, so you'll need to fix that up too.
Ho hum. Those failures are not your fault, as they persisted across the
next job. They appear to be fixed now via flushing the runner caches.
You'll still have the merge conflict though, sorry.
r~
