ARM: highbank: fix cache flush ordering for cpu hotplug
PATCH 1/1 ARM: highbank: fix cache flush ordering for cpu hotplug
From: Rob Herring <[email protected]>
The L1 data cache flush needs to be after highbank_set_cpu_jump call which pollutes the cache with the l2x0_lock. This causes other cores to deadlock waiting for the l2x0_lock. Moving the flush of the entire data cache after highbank_set_cpu_jump fixes the problem. Use flush_cache_louis instead of flush_cache_all are that is sufficient to flush only the L1 data cache. flush_cache_louis did not exist when highbank_cpu_die was originally written.
With PL310 errata 769419 enabled, a wmb is inserted into idle which takes the l2x0_lock. This makes the problem much more easily hit and causes reset to hang.
---
Olof, Arnd,
If you have any more 3.9 fixes, please send this in. Otherwise, it can wait for 3.10 and mark for stable.
Rob
arch/arm/mach-highbank/hotplug.c | 10 ++++------ 1 file changed, 4 insertions(+), 6 deletions(-)
arch/arm/mach-highbank/hotplug.c +4 -6
@@ -28,13 +28,11 @@ extern void secondary_startup(void);*/ void __ref highbank_cpu_die (unsigned int cpu) { -flush_cache_all (); -highbank_set_cpu_jump (cpu,phys_to_virt (0 )); -highbank_set_core_pwr (); -cpu_do_idle (); +flush_cache_louis (); +highbank_set_core_pwr (); -/* We should never return from idle */ -panic ("highbank: cpu %d unexpectedly exit from shutdown\n" , cpu); +while (1 ) +cpu_do_idle (); }
Signature · 1 line
1.7.10.4
Olof Johansson
AppliedPATCH 1/1 ARM: highbank: fix cache flush ordering for cpu hotplug
Quoted 14 lines · Rob Herring
From: Rob Herring <[email protected]>
The L1 data cache flush needs to be after
highbank_set_cpu_jumpcall which pollutes the cache with thel2x0_lock. This causes other cores to deadlock waiting for thel2x0_lock. Moving the flush of the entire data cache afterhighbank_set_cpu_jumpfixes the problem. Useflush_cache_louisinstead offlush_cache_allare that is sufficient to flush only the L1 data cache.flush_cache_louisdid not exist whenhighbank_cpu_diewas originally written.With PL310 errata 769419 enabled, a wmb is inserted into idle which takes the
l2x0_lock. This makes the problem much more easily hit and causes reset to hang.
Quoted 7 lines
Reported-by Paolo Pisati <[email protected]>Signed-off-by Rob Herring <[email protected]>---
Olof, Arnd,If you have any more 3.9 fixes, please send this in. Otherwise, it can wait for 3.10 and mark for stable.
Applied to fixes.
-Olof