kernel/git/stable/linux.git/lib/swiotlb.c, branch linux-2.6.23.y

Fix swiotlb_sync_single_range()

2007-07-22T00:49:14+00:00

If the swiotlb maps a multi-slab region, swiotlb_sync_single_range() can be invoked to sync a sub-region which does not include the first slab. Unfortunately io_tlb_orig_addr[] is only initialised for the first slab, and hence the call to sync_single() will read a garbage orig_addr in this case. This patch fixes the issue by initialising all mapped slabs in io_tlb_orig_addr[]. It also correctly adjusts the buffer pointer in sync_single() to handle the case that the given dma_addr is not aligned on a slab boundary. Signed-off-by: Keir Fraser Cc: "Luck, Tony" Acked-by: Andi Kleen Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

fix section mismatch warning in lib/swiotlb.c

2007-05-08T18:14:59+00:00

kbuild spits outs following warning on a defconfig x86_64 build: WARNING: swiotlb.o - Section mismatch: reference to .init.text:swiotlb_init from __ksymtab between '__ksymtab_swiotlb_init' (at offset 0xa0) and '__ksymtab_swiotlb_free_coherent' This warning happens because the function swiotlb_init is marked __init and EXPORT_SYMBOL(). A 'git grep swiotlb_init' showed no users in drivers/ so remove the EXPORT_SYMBOL. Signed-off-by: Sam Ravnborg Cc: Andi Kleen Cc: "Luck, Tony" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

Revert "[IA64] swiotlb abstraction (e.g. for Xen)"

2007-03-06T21:31:45+00:00

This reverts commit 51099005ab8e09d68a13fea8d55bc739c1040ca6.

[PATCH] swiotlb uninlinings

2007-02-12T17:48:29+00:00

Optimise swiotlb.c for size. text data bss dec hex filename 5009 89 64 5162 142a lib/swiotlb.o-before 4666 89 64 4819 12d3 lib/swiotlb.o-after For some reason my gcc (4.0.2) doesn't want to tailcall these things. swiotlb_sync_sg_for_device: pushq %rbp # movl $1, %r8d #, movq %rsp, %rbp #, call swiotlb_sync_sg # leave ret .size swiotlb_sync_sg_for_device, .-swiotlb_sync_sg_for_device .section .text.swiotlb_sync_sg_for_cpu,"ax",@progbits .globl swiotlb_sync_sg_for_cpu .type swiotlb_sync_sg_for_cpu, @function swiotlb_sync_sg_for_cpu: pushq %rbp # xorl %r8d, %r8d # movq %rsp, %rbp #, call swiotlb_sync_sg # leave ret Cc: Jan Beulich Cc: Andi Kleen Cc: "Luck, Tony" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds

[IA64] swiotlb abstraction (e.g. for Xen)

2007-02-06T02:53:04+00:00

Add abstraction so that the file can be used by environments other than IA64 and EM64T, namely for Xen. Signed-off-by: Jan Beulich Signed-off-by: Andrew Morton Signed-off-by: Tony Luck

[IA64] swiotlb cleanup

2007-02-06T02:51:25+00:00

- add proper __init decoration to swiotlb's init code (and the code calling it, where not already the case) - replace uses of 'unsigned long' with dma_addr_t where appropriate - do miscellaneous simplicfication and cleanup Signed-off-by: Jan Beulich Signed-off-by: Andrew Morton Signed-off-by: Tony Luck

[IA64] make swiotlb use bus_to_virt/virt_to_bus

2007-02-06T02:49:45+00:00

Convert all phys_to_virt/virt_to_phys uses to bus_to_virt/virt_to_bus, as is what is meant and what is needed in (at least) some virtualized environments like Xen. Signed-off-by: Jan Beulich Acked-by: Muli Ben-Yehuda Signed-off-by: Andrew Morton Signed-off-by: Tony Luck

[IA64] swiotlb bug fixes

2007-02-06T02:46:40+00:00

This patch fixes - marking I-cache clean of pages DMAed to now only done for IA64 - broken multiple inclusion in include/asm-x86_64/swiotlb.h - missing call to mark_clean in swiotlb_sync_sg() - a (perhaps only theoretical) issue in swiotlb_dma_supported() when io_tlb_end is exactly at the end of memory Signed-off-by: Jan Beulich Signed-off-by: Andrew Morton Signed-off-by: Tony Luck

BUG_ON() Conversion in lib/swiotlb.c

2006-03-24T17:47:11+00:00

this changes if() BUG(); constructs to BUG_ON() which is cleaner, contains unlikely() and can better optimized away. Signed-off-by: Eric Sesterhenn Signed-off-by: Adrian Bunk

[PATCH] x86_64: Use function pointers to call DMA mapping functions

2006-01-12T03:04:55+00:00

AK: I hacked Muli's original patch a lot and there were a lot of changes - all bugs are probably to blame on me now. There were also some changes in the fall back behaviour for swiotlb - in particular it doesn't try to use GFP_DMA now anymore. Also all DMA mapping operations use the same core dma_alloc_coherent code with proper fallbacks now. And various other changes and cleanups. Known problems: iommu=force swiotlb=force together breaks needs more testing. This patch cleans up x86_64's DMA mapping dispatching code. Right now we have three possible IOMMU types: AGP GART, swiotlb and nommu, and in the future we will also have Xen's x86_64 swiotlb and other HW IOMMUs for x86_64. In order to support all of them cleanly, this patch: - introduces a struct dma_mapping_ops with function pointers for each of the DMA mapping operations of gart (AMD HW IOMMU), swiotlb (software IOMMU) and nommu (no IOMMU). - gets rid of: if (swiotlb) return swiotlb_xxx(); - PCI_DMA_BUS_IS_PHYS is now checked against the dma_ops being set This makes swiotlb faster by avoiding double copying in some cases. Signed-Off-By: Muli Ben-Yehuda Signed-Off-By: Jon D. Mason Signed-off-by: Andi Kleen Signed-off-by: Linus Torvalds