Fixes and improvements around xnlock
Get series with b4
b4 am [email protected]Versions newest first
- v1 Latest4 patches 20 repliesIn review
Patches in v1 No applies recorded yet
| # | Subject and files | Size | Replies | Applied |
|---|---|---|---|---|
| 1/4 | Refactor generic system.hinclude/ | +204 | 5 replies | |
| 2/4 | Fix and optimize xnlock_putinclude/ | +19 | 6 replies | |
| 3/4 | Uninline heavy locking functionsinclude/ include/ | +26 | 5 replies | |
| 4/4 | Recursive FIFO ticket xnlockinclude/ ksrc/ | +112 | 4 replies |
Reviews trailers per patch
No Reviewed-by, Acked-by or Tested-by trailers for v1 yet.
Cover letter [Xenomai-core] [PATCH 0/4] Fixes and improvements around xnlock
Hi,
I'm happy to announce a patch series which actually started like "hey, let's check if we can have this nice FIFO ticket spinlock algorithm also for our recursive xnlock." The result is a critical bug fix and a massive text size reduction of nucleus and skins, both concerning SMP systems. Furthermore, there are now optional ticket locks available for larger SMP boxes. :)
This series consists of
1. Refactoring of asm-generic/system.h
(+ some optimization of xnlock_get)
2. SMP-critical memory barrier fix for xnlock_put (+ optimization for production code)
3. Uninlining of xnlock_get_irqsave/xnlock_put_irqrestore on SMP
(xeno_nucleus text, x86-64, non-debug: 112482 -> 79453 bytes!)
4. Optional FIFO ticket spinlock for strict deterministic spin times on xnlocks (still highly experimental!)
The second patch is stable stuff, so we should merge 1+2 into 2.4.x as well.
Jan