diff options
| author | Arthur Hanson Rasmusson <arthur@vgpu.io> | 2026-10-05 00:12:56 -0400 |
|---|---|---|
| committer | Arthur Hanson Rasmusson <arthur@vgpu.io> | 2026-10-05 00:12:56 -0400 |
| commit | a9d776627f5057fd11c8d0a94bbfcca2358fe507 (patch) | |
| tree | 93cd901a4f16da27919a453427593f97aeccf6ae /kernel/gsp_sema_service.c | |
| parent | 0e073122a56cb2b8498aaacb79968bdf8309b309 (diff) | |
| download | open-iov-a9d776627f5057fd11c8d0a94bbfcca2358fe507.tar.gz open-iov-a9d776627f5057fd11c8d0a94bbfcca2358fe507.zip | |
vgsp: widen the CE pending-entry grace to 15 s
A consumed entry's work is never run; if its release lands via the
fallback before the entry's copies have executed, the client refills the
staging slot the copies then read (observed as the deterministic
one-chunk vadd input shift). Keep untranslatable CE entries pending
longer so their copies run before any release reaches the client.
8-cycle pilot (141-148): zero corruption; a red-fingerprint cycle
(held=1165, ring_auto=6) healed to ring_auto=26 with faked=0 and the
exact checksum, where the 8 s window recorded 1 corruption in 36.
Diffstat (limited to 'kernel/gsp_sema_service.c')
| -rw-r--r-- | kernel/gsp_sema_service.c | 2 |
1 files changed, 1 insertions, 1 deletions
diff --git a/kernel/gsp_sema_service.c b/kernel/gsp_sema_service.c index a62e3e4..be1d88a 100644 --- a/kernel/gsp_sema_service.c +++ b/kernel/gsp_sema_service.c @@ -6670,7 +6670,7 @@ u64 gsp_sema_copy_dst2; * write the next chunk's data over a finished one), so the hold is the only * chance the work has. */ -unsigned gsp_sema_ce_grace_ms = 8000; +unsigned gsp_sema_ce_grace_ms = 15000; /* * Bounded work-before-release hold (ms) for an unreadable CE entry. The |