ARM: mvebu: Fix bug in coherency fabric low level init function
Subsystems (3)
PATCH 1/1 ARM: mvebu: Fix bug in coherency fabric low level init function
Hello Jason and Andrew,
This a new fix for 3.10-rc coming from Marvell. I wasn't impacted (until now) by the bug fixed by this commit, but I trust Nadav with his patch, and I didn't see any flaw in his code. Furthermore, I applied and tested it and saw no regression.
So please applied it for 3.10-rc and we can also consider to applied it to 3.8 and 3.9.
Thanks
Nadav Haklai (1):
ARM: mvebu: Fix bug in coherency fabric low level init function
arch/arm/mach-mvebu/coherency_ll.S | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-)
Signature · 1 line
1.8.1.2
PATCH 1/1 ARM: mvebu: Fix bug in coherency fabric low level init function
From: Nadav Haklai <[email protected]>
When adding CPU to the SMP group and enabling the coherency on this CPU we must protect the register access.
The previous implementation claims to be atomic but doesn't provide any protection against parallel access to the coherency fabric control and configuration registers.
This patch fixes this by using the ldrex and strex mechanism. This method should be used in all accesses to those registers.
--- arch/arm/mach-mvebu/coherency_ll.S | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-)
arch/arm/mach-mvebu/coherency_ll.S +11 -5
@@ -32,15 +32,21 @@ ENTRY(ll_set_cpu_coherent) /* Add CPU to SMP group - Atomic */ add r3, r0, #ARMADA_XP_CFB_CTL_REG_OFFSET - ldr r2, [r3] +1: + ldrex r2, [r3] orr r2, r2, r1 - str r2, [r3] + strex r0, r2, [r3] + cmp r0, #0 + bne 1b /* Enable coherency on CPU - Atomic */ - add r3, r0, #ARMADA_XP_CFB_CFG_REG_OFFSET - ldr r2, [r3] + add r3, r3, #ARMADA_XP_CFB_CFG_REG_OFFSET +1: + ldrex r2, [r3] orr r2, r2, r1 - str r2, [r3] + strex r0, r2, [r3] + cmp r0, #0 + bne 1b dsb
Signature · 1 line
1.8.1.2
Jason Cooper
AppliedPATCH 1/1 ARM: mvebu: Fix bug in coherency fabric low level init function
Quoted 11 lines · Gregory CLEMENT
From: Nadav Haklai <[email protected]>
When adding CPU to the SMP group and enabling the coherency on this CPU we must protect the register access.
The previous implementation claims to be atomic but doesn't provide any protection against parallel access to the coherency fabric control and configuration registers.This patch fixes this by using the ldrex and strex mechanism. This method should be used in all accesses to those registers.
Quoted 6 lines
[gregory.clement at free-electrons.com: fixed the commit's topic]Signed-off-by Nadav Haklai <[email protected]>Signed-off-by Gregory CLEMENT <[email protected]>--- arch/arm/mach-mvebu/coherency_ll.S | 16 +++++++++++----- 1 file changed, 11 insertions(+), 5 deletions(-)
Applied to mvebu/fixes
thx,
Jason.
PATCH 1/1 ARM: mvebu: Fix bug in coherency fabric low level init function
In reply to Gregory CLEMENTHi Gregory,
Quoted 10 lines · Gregory CLEMENT
From: Nadav Haklai <[email protected]>
When adding CPU to the SMP group and enabling the coherency on this CPU we must protect the register access.
The previous implementation claims to be atomic but doesn't provide any protection against parallel access to the coherency fabric control and configuration registers.This patch fixes this by using the ldrex and strex mechanism. This method should be used in all accesses to those registers.
I don't know how to tell whether the fix works since it's unclear to me what issue it fixes, but at least I can say that it doesn't break my armada370 on 3.10-rc2 + latest fixes from your dev branch.
Best regards,
Willy