Skip to contentSkip to conversation

ARM: head: fix virt/phys patching for big-endian systems

PATCH New Will Deacon 1 participant 2 messages Apr 23, 2014 Arm (32-bit)

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 ARM: head: fix virt/phys patching for big-endian systems

Commit e26a9e00afc4 ("ARM: Better virt_to_page() handling") reworked the dynamic virt/phys patching to store the physical offset of the start of memory as a PFN (32-bit quantity) instead of a physical address (potentially a 64-bit quantity). Unfortunately, some of the old code was left behind, causing big-endian systems to store the calculated PFN to the next word *after* __pv_phys_pfn_offset and causing the kernel to treat PHYS_OFFSET as 0x0.

This patch fixes the patching code, so that we store the 32-bit PFN offset to the correct address, regardless of endianness.

---
 arch/arm/kernel/head.S | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/kernel/head.S +1 -1
index f8c08839edf3..591d6e4a6492 100644
@@ -587,7 +587,7 @@ __fixup_pv_table:
 	add	r6, r6, r3	@ adjust __pv_phys_pfn_offset address
 	add	r7, r7, r3	@ adjust __pv_offset address
 	mov	r0, r8, lsr #12	@ convert to PFN
-	str	r0, [r6, #LOW_OFFSET]	@ save computed PHYS_OFFSET to __pv_phys_pfn_offset
+	str	r0, [r6]	@ save computed PHYS_OFFSET to __pv_phys_pfn_offset
 	strcc	ip, [r7, #HIGH_OFFSET]	@ save to __pv_offset high bits
 	mov	r6, r3, lsr #24	@ constant for add/sub instructions
 	teq	r3, r6, lsl #24 @ must be 16MiB aligned
Signature · 1 line
1.9.2
#1 lore raw

PATCH 1/1 ARM: head: fix virt/phys patching for big-endian systems

Quoted 10 lines · Will Deacon

Commit e26a9e00afc4 ("ARM: Better virt_to_page() handling") reworked the dynamic virt/phys patching to store the physical offset of the start of memory as a PFN (32-bit quantity) instead of a physical address (potentially a 64-bit quantity). Unfortunately, some of the old code was left behind, causing big-endian systems to store the calculated PFN to the next word *after* __pv_phys_pfn_offset and causing the kernel to treat PHYS_OFFSET as 0x0.

This patch fixes the patching code, so that we store the 32-bit PFN offset to the correct address, regardless of endianness.

Aha, this was already fixed in 8033/1 which Russell has pushed out on his fixes branch.

Sorry for the noise,

Will

Quoted 21 lines
---
 arch/arm/kernel/head.S | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/kernel/head.S +1 -1
index f8c08839edf3..591d6e4a6492 100644
@@ -587,7 +587,7 @@ __fixup_pv_table:
 	add	r6, r6, r3	@ adjust __pv_phys_pfn_offset address
 	add	r7, r7, r3	@ adjust __pv_offset address
 	mov	r0, r8, lsr #12	@ convert to PFN
-	str	r0, [r6, #LOW_OFFSET]	@ save computed PHYS_OFFSET to __pv_phys_pfn_offset
+	str	r0, [r6]	@ save computed PHYS_OFFSET to __pv_phys_pfn_offset
 	strcc	ip, [r7, #HIGH_OFFSET]	@ save to __pv_offset high bits
 	mov	r6, r3, lsr #24	@ constant for add/sub instructions
 	teq	r3, r6, lsl #24 @ must be 16MiB aligned
Signature · 1 line
1.9.2