Russell, Santosh, the unneeded commit causing regression is still in place. Please try to compile an ARM kernel without CONFIG_ARM_PATCH_PHYS_VIRT and with CONFIG_ZONE_DMA and see for yourself, if you don't believe me.
Please be aware that this commit fixes nothing, its only function is causing the regression - so we don't lose anything by reverting it. If the attached wasn't clear, what the defective commit presently does is changing a perfectly valid code into a code referencing a variable which (without CONFIG_ARM_PATCH_PHYS_VIRT set) doesn't at all exist. With CONFIG_ARM_PATCH_PHYS_VIRT set, this commit does precisely nothing. grep __pv_phys_offset * -r gives: Broken by the commit in question: arch/arm/mm/init.c: arm_dma_limit = __pv_phys_offset + arm_dma_zone_size - 1; All of the following are only compiled/assembled with CONFIG_ARM_PATCH_PHYS_VIRT set: arch/arm/include/asm/memory.h:extern u64 __pv_phys_offset; arch/arm/include/asm/memory.h:#define PHYS_OFFSET __pv_phys_offset arch/arm/kernel/armksyms.c:EXPORT_SYMBOL(__pv_phys_offset); arch/arm/kernel/head.S: add r6, r6, r3 @ adjust __pv_phys_offset address arch/arm/kernel/head.S: str r8, [r6, #LOW_OFFSET] @ save computed PHYS_OFFSET to __pv_phys_offset arch/arm/kernel/head.S:2: .long __pv_phys_offset arch/arm/kernel/head.S: .globl __pv_phys_offset arch/arm/kernel/head.S: .type __pv_phys_offset, %object arch/arm/kernel/head.S:__pv_phys_offset: arch/arm/kernel/head.S: .size __pv_phys_offset, . -__pv_phys_offset In short, please revert 787b0d5c1ca7ff24feb6f92e4c7f4410ee7d81a8. Or, do you want me to send a patch which does that? HTH and TIA. > arch/arm/mm/init.c: In function 'setup_dma_zone': > arch/arm/mm/init.c:232:19: error: '__pv_phys_offset' undeclared (first use in > this function) > > Reverting the following commit fixes it: > > commit 787b0d5c1ca7ff24feb6f92e4c7f4410ee7d81a8 > Author: Santosh Shilimkar <santosh.shilim...@ti.com> > Date: Mon Dec 2 20:29:12 2013 +0100 > > ARM: 7908/1: mm: Fix the arm_dma_limit calculation > > Current code is using PHYS_OFFSET to calculate the arm_dma_limit which > will lead to wrong calculations in cases where PHYS_OFFSET is updated > runtime. > > So fix the code by using __pv_phys_offset instead of PHYS_OFFSET. > > It seems PHYS_OFFSET is equal to __pv_phys_offset if the latter is > available: > > arch/arm/include/asm/memory.h:#define PHYS_OFFSET __pv_phys_offset > > Otherwise (without CONFIG_ARM_PATCH_PHYS_VIRT) PHYS_OFFSET is a > hard-coded CONFIG value, or whatever the platform/CPU code needs: > > arch/arm/include/asm/memory.h:#define PHYS_OFFSET PLAT_PHYS_OFFSET > > Perhaps the patch in question was needed at some point but I think the > situation had changed before it was commited. -- Krzysztof Halasa Research Institute for Automation and Measurements PIAP Al. Jerozolimskie 202, 02-486 Warsaw, Poland -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/