From 688cb30bdc3e398d97682a6a58f825821ee838c2 Mon Sep 17 00:00:00 2001 From: "David S. Miller" Date: Thu, 13 Oct 2005 22:15:24 -0700 Subject: [PATCH] [SPARC64]: Eliminate PCI IOMMU dma mapping size limit. The hairy fast allocator in the sparc64 PCI IOMMU code has a hard limit of 256 pages. Certain devices can exceed this when performing very large I/Os. So replace with a more simple allocator, based largely upon the arch/ppc64/kernel/iommu.c code. Signed-off-by: David S. Miller --- arch/sparc64/kernel/pci_iommu.c | 365 ++++++++++++-------------------- include/asm-sparc64/pbm.h | 28 +-- 2 files changed, 145 insertions(+), 248 deletions(-) diff --git a/arch/sparc64/kernel/pci_iommu.c b/arch/sparc64/kernel/pci_iommu.c index cdee7d9fed72..a11910be1013 100644 --- a/arch/sparc64/kernel/pci_iommu.c +++ b/arch/sparc64/kernel/pci_iommu.c @@ -49,12 +49,6 @@ static void __iommu_flushall(struct pci_iommu *iommu) /* Ensure completion of previous PIO writes. */ (void) pci_iommu_read(iommu->write_complete_reg); - - /* Now update everyone's flush point. */ - for (entry = 0; entry < PBM_NCLUSTERS; entry++) { - iommu->alloc_info[entry].flush = - iommu->alloc_info[entry].next; - } } #define IOPTE_CONSISTENT(CTX) \ @@ -80,9 +74,61 @@ static void inline iopte_make_dummy(struct pci_iommu *iommu, iopte_t *iopte) iopte_val(*iopte) = val; } +/* Based largely upon the ppc64 iommu allocator. */ +static long pci_arena_alloc(struct pci_iommu *iommu, unsigned long npages) +{ + struct pci_iommu_arena *arena = &iommu->arena; + unsigned long n, i, start, end, limit; + int pass; + + limit = arena->limit; + start = arena->hint; + pass = 0; + +again: + n = find_next_zero_bit(arena->map, limit, start); + end = n + npages; + if (unlikely(end >= limit)) { + if (likely(pass < 1)) { + limit = start; + start = 0; + __iommu_flushall(iommu); + pass++; + goto again; + } else { + /* Scanned the whole thing, give up. */ + return -1; + } + } + + for (i = n; i < end; i++) { + if (test_bit(i, arena->map)) { + start = i + 1; + goto again; + } + } + + for (i = n; i < end; i++) + __set_bit(i, arena->map); + + arena->hint = end; + + return n; +} + +static void pci_arena_free(struct pci_iommu_arena *arena, unsigned long base, unsigned long npages) +{ + unsigned long i; + + for (i = base; i < (base + npages); i++) + __clear_bit(i, arena->map); +} + void pci_iommu_table_init(struct pci_iommu *iommu, int tsbsize, u32 dma_offset, u32 dma_addr_mask) { - unsigned long i, tsbbase, order; + unsigned long i, tsbbase, order, sz, num_tsb_entries; + + num_tsb_entries = tsbsize / sizeof(iopte_t); /* Setup initial software IOMMU state. */ spin_lock_init(&iommu->lock); @@ -90,27 +136,16 @@ void pci_iommu_table_init(struct pci_iommu *iommu, int tsbsize, u32 dma_offset, iommu->page_table_map_base = dma_offset; iommu->dma_addr_mask = dma_addr_mask; - switch (tsbsize / (8 * 1024)) { - case 64: - iommu->page_table_sz_bits = 16; - break; - case 128: - iommu->page_table_sz_bits = 17; - break; - default: - prom_printf("PCI_IOMMU: Illegal TSB size %d\n", - tsbsize / (8 * 1024)); + /* Allocate and initialize the free area map. */ + sz = num_tsb_entries / 8; + sz = (sz + 7UL) & ~7UL; + iommu->arena.map = kmalloc(sz, GFP_KERNEL); + if (!iommu->arena.map) { + prom_printf("PCI_IOMMU: Error, kmalloc(arena.map) failed.\n"); prom_halt(); - break; - }; - - iommu->lowest_consistent_map = - 1 << (iommu->page_table_sz_bits - PBM_LOGCLUSTERS); - - for (i = 0; i < PBM_NCLUSTERS; i++) { - iommu->alloc_info[i].flush = 0; - iommu->alloc_info[i].next = 0; } + memset(iommu->arena.map, 0, sz); + iommu->arena.limit = num_tsb_entries; /* Allocate and initialize the dummy page which we * set inactive IO PTEs to point to. @@ -132,114 +167,24 @@ void pci_iommu_table_init(struct pci_iommu *iommu, int tsbsize, u32 dma_offset, } iommu->page_table = (iopte_t *)tsbbase; - for (i = 0; i < tsbsize / sizeof(iopte_t); i++) + for (i = 0; i < num_tsb_entries; i++) iopte_make_dummy(iommu, &iommu->page_table[i]); } -static iopte_t *alloc_streaming_cluster(struct pci_iommu *iommu, unsigned long npages) +static inline iopte_t *alloc_npages(struct pci_iommu *iommu, unsigned long npages) { - iopte_t *iopte, *limit, *first; - unsigned long cnum, ent, flush_point; - - cnum = 0; - while ((1UL << cnum) < npages) - cnum++; - iopte = (iommu->page_table + - (cnum << (iommu->page_table_sz_bits - PBM_LOGCLUSTERS))); - - if (cnum == 0) - limit = (iommu->page_table + - iommu->lowest_consistent_map); - else - limit = (iopte + - (1 << (iommu->page_table_sz_bits - PBM_LOGCLUSTERS))); + long entry; - iopte += ((ent = iommu->alloc_info[cnum].next) << cnum); - flush_point = iommu->alloc_info[cnum].flush; - - first = iopte; - for (;;) { - if (IOPTE_IS_DUMMY(iommu, iopte)) { - if ((iopte + (1 << cnum)) >= limit) - ent = 0; - else - ent = ent + 1; - iommu->alloc_info[cnum].next = ent; - if (ent == flush_point) - __iommu_flushall(iommu); - break; - } - iopte += (1 << cnum); - ent++; - if (iopte >= limit) { - iopte = (iommu->page_table + - (cnum << - (iommu->page_table_sz_bits - PBM_LOGCLUSTERS))); - ent = 0; - } - if (ent == flush_point) - __iommu_flushall(iommu); - if (iopte == first) - goto bad; - } - - /* I've got your streaming cluster right here buddy boy... */ - return iopte; + entry = pci_arena_alloc(iommu, npages); + if (unlikely(entry < 0)) + return NULL; -bad: - printk(KERN_EMERG "pci_iommu: alloc_streaming_cluster of npages(%ld) failed!\n", - npages); - return NULL; + return iommu->page_table + entry; } -static void free_streaming_cluster(struct pci_iommu *iommu, dma_addr_t base, - unsigned long npages, unsigned long ctx) +static inline void free_npages(struct pci_iommu *iommu, dma_addr_t base, unsigned long npages) { - unsigned long cnum, ent; - - cnum = 0; - while ((1UL << cnum) < npages) - cnum++; - - ent = (base << (32 - IO_PAGE_SHIFT + PBM_LOGCLUSTERS - iommu->page_table_sz_bits)) - >> (32 + PBM_LOGCLUSTERS + cnum - iommu->page_table_sz_bits); - - /* If the global flush might not have caught this entry, - * adjust the flush point such that we will flush before - * ever trying to reuse it. - */ -#define between(X,Y,Z) (((Z) - (Y)) >= ((X) - (Y))) - if (between(ent, iommu->alloc_info[cnum].next, iommu->alloc_info[cnum].flush)) - iommu->alloc_info[cnum].flush = ent; -#undef between -} - -/* We allocate consistent mappings from the end of cluster zero. */ -static iopte_t *alloc_consistent_cluster(struct pci_iommu *iommu, unsigned long npages) -{ - iopte_t *iopte; - - iopte = iommu->page_table + (1 << (iommu->page_table_sz_bits - PBM_LOGCLUSTERS)); - while (iopte > iommu->page_table) { - iopte--; - if (IOPTE_IS_DUMMY(iommu, iopte)) { - unsigned long tmp = npages; - - while (--tmp) { - iopte--; - if (!IOPTE_IS_DUMMY(iommu, iopte)) - break; - } - if (tmp == 0) { - u32 entry = (iopte - iommu->page_table); - - if (entry < iommu->lowest_consistent_map) - iommu->lowest_consistent_map = entry; - return iopte; - } - } - } - return NULL; + pci_arena_free(&iommu->arena, base >> IO_PAGE_SHIFT, npages); } static int iommu_alloc_ctx(struct pci_iommu *iommu) @@ -279,7 +224,7 @@ void *pci_alloc_consistent(struct pci_dev *pdev, size_t size, dma_addr_t *dma_ad struct pcidev_cookie *pcp; struct pci_iommu *iommu; iopte_t *iopte; - unsigned long flags, order, first_page, ctx; + unsigned long flags, order, first_page; void *ret; int npages; @@ -297,9 +242,10 @@ void *pci_alloc_consistent(struct pci_dev *pdev, size_t size, dma_addr_t *dma_ad iommu = pcp->pbm->iommu; spin_lock_irqsave(&iommu->lock, flags); - iopte = alloc_consistent_cluster(iommu, size >> IO_PAGE_SHIFT); - if (iopte == NULL) { - spin_unlock_irqrestore(&iommu->lock, flags); + iopte = alloc_npages(iommu, size >> IO_PAGE_SHIFT); + spin_unlock_irqrestore(&iommu->lock, flags); + + if (unlikely(iopte == NULL)) { free_pages(first_page, order); return NULL; } @@ -308,31 +254,15 @@ void *pci_alloc_consistent(struct pci_dev *pdev, size_t size, dma_addr_t *dma_ad ((iopte - iommu->page_table) << IO_PAGE_SHIFT)); ret = (void *) first_page; npages = size >> IO_PAGE_SHIFT; - ctx = 0; - if (iommu->iommu_ctxflush) - ctx = iommu_alloc_ctx(iommu); first_page = __pa(first_page); while (npages--) { - iopte_val(*iopte) = (IOPTE_CONSISTENT(ctx) | + iopte_val(*iopte) = (IOPTE_CONSISTENT(0UL) | IOPTE_WRITE | (first_page & IOPTE_PAGE)); iopte++; first_page += IO_PAGE_SIZE; } - { - int i; - u32 daddr = *dma_addrp; - - npages = size >> IO_PAGE_SHIFT; - for (i = 0; i < npages; i++) { - pci_iommu_write(iommu->iommu_flush, daddr); - daddr += IO_PAGE_SIZE; - } - } - - spin_unlock_irqrestore(&iommu->lock, flags); - return ret; } @@ -342,7 +272,7 @@ void pci_free_consistent(struct pci_dev *pdev, size_t size, void *cpu, dma_addr_ struct pcidev_cookie *pcp; struct pci_iommu *iommu; iopte_t *iopte; - unsigned long flags, order, npages, i, ctx; + unsigned long flags, order, npages; npages = IO_PAGE_ALIGN(size) >> IO_PAGE_SHIFT; pcp = pdev->sysdata; @@ -352,46 +282,7 @@ void pci_free_consistent(struct pci_dev *pdev, size_t size, void *cpu, dma_addr_ spin_lock_irqsave(&iommu->lock, flags); - if ((iopte - iommu->page_table) == - iommu->lowest_consistent_map) { - iopte_t *walk = iopte + npages; - iopte_t *limit; - - limit = (iommu->page_table + - (1 << (iommu->page_table_sz_bits - PBM_LOGCLUSTERS))); - while (walk < limit) { - if (!IOPTE_IS_DUMMY(iommu, walk)) - break; - walk++; - } - iommu->lowest_consistent_map = - (walk - iommu->page_table); - } - - /* Data for consistent mappings cannot enter the streaming - * buffers, so we only need to update the TSB. We flush - * the IOMMU here as well to prevent conflicts with the - * streaming mapping deferred tlb flush scheme. - */ - - ctx = 0; - if (iommu->iommu_ctxflush) - ctx = (iopte_val(*iopte) & IOPTE_CONTEXT) >> 47UL; - - for (i = 0; i < npages; i++, iopte++) - iopte_make_dummy(iommu, iopte); - - if (iommu->iommu_ctxflush) { - pci_iommu_write(iommu->iommu_ctxflush, ctx); - } else { - for (i = 0; i < npages; i++) { - u32 daddr = dvma + (i << IO_PAGE_SHIFT); - - pci_iommu_write(iommu->iommu_flush, daddr); - } - } - - iommu_free_ctx(iommu, ctx); + free_npages(iommu, dvma, npages); spin_unlock_irqrestore(&iommu->lock, flags); @@ -418,25 +309,27 @@ dma_addr_t pci_map_single(struct pci_dev *pdev, void *ptr, size_t sz, int direct iommu = pcp->pbm->iommu; strbuf = &pcp->pbm->stc; - if (direction == PCI_DMA_NONE) - BUG(); + if (unlikely(direction == PCI_DMA_NONE)) + goto bad_no_ctx; oaddr = (unsigned long)ptr; npages = IO_PAGE_ALIGN(oaddr + sz) - (oaddr & IO_PAGE_MASK); npages >>= IO_PAGE_SHIFT; spin_lock_irqsave(&iommu->lock, flags); + base = alloc_npages(iommu, npages); + ctx = 0; + if (iommu->iommu_ctxflush) + ctx = iommu_alloc_ctx(iommu); + spin_unlock_irqrestore(&iommu->lock, flags); - base = alloc_streaming_cluster(iommu, npages); - if (base == NULL) + if (unlikely(!base)) goto bad; + bus_addr = (iommu->page_table_map_base + ((base - iommu->page_table) << IO_PAGE_SHIFT)); ret = bus_addr | (oaddr & ~IO_PAGE_MASK); base_paddr = __pa(oaddr & IO_PAGE_MASK); - ctx = 0; - if (iommu->iommu_ctxflush) - ctx = iommu_alloc_ctx(iommu); if (strbuf->strbuf_enabled) iopte_protection = IOPTE_STREAMING(ctx); else @@ -447,12 +340,13 @@ dma_addr_t pci_map_single(struct pci_dev *pdev, void *ptr, size_t sz, int direct for (i = 0; i < npages; i++, base++, base_paddr += IO_PAGE_SIZE) iopte_val(*base) = iopte_protection | base_paddr; - spin_unlock_irqrestore(&iommu->lock, flags); - return ret; bad: - spin_unlock_irqrestore(&iommu->lock, flags); + iommu_free_ctx(iommu, ctx); +bad_no_ctx: + if (printk_ratelimit()) + WARN_ON(1); return PCI_DMA_ERROR_CODE; } @@ -527,10 +421,13 @@ void pci_unmap_single(struct pci_dev *pdev, dma_addr_t bus_addr, size_t sz, int struct pci_iommu *iommu; struct pci_strbuf *strbuf; iopte_t *base; - unsigned long flags, npages, ctx; + unsigned long flags, npages, ctx, i; - if (direction == PCI_DMA_NONE) - BUG(); + if (unlikely(direction == PCI_DMA_NONE)) { + if (printk_ratelimit()) + WARN_ON(1); + return; + } pcp = pdev->sysdata; iommu = pcp->pbm->iommu; @@ -556,13 +453,14 @@ void pci_unmap_single(struct pci_dev *pdev, dma_addr_t bus_addr, size_t sz, int /* Step 1: Kick data out of streaming buffers if necessary. */ if (strbuf->strbuf_enabled) - pci_strbuf_flush(strbuf, iommu, bus_addr, ctx, npages, direction); + pci_strbuf_flush(strbuf, iommu, bus_addr, ctx, + npages, direction); - /* Step 2: Clear out first TSB entry. */ - iopte_make_dummy(iommu, base); + /* Step 2: Clear out TSB entries. */ + for (i = 0; i < npages; i++) + iopte_make_dummy(iommu, base + i); - free_streaming_cluster(iommu, bus_addr - iommu->page_table_map_base, - npages, ctx); + free_npages(iommu, bus_addr - iommu->page_table_map_base, npages); iommu_free_ctx(iommu, ctx); @@ -667,6 +565,8 @@ int pci_map_sg(struct pci_dev *pdev, struct scatterlist *sglist, int nelems, int pci_map_single(pdev, (page_address(sglist->page) + sglist->offset), sglist->length, direction); + if (unlikely(sglist->dma_address == PCI_DMA_ERROR_CODE)) + return 0; sglist->dma_length = sglist->length; return 1; } @@ -675,21 +575,29 @@ int pci_map_sg(struct pci_dev *pdev, struct scatterlist *sglist, int nelems, int iommu = pcp->pbm->iommu; strbuf = &pcp->pbm->stc; - if (direction == PCI_DMA_NONE) - BUG(); + if (unlikely(direction == PCI_DMA_NONE)) + goto bad_no_ctx; /* Step 1: Prepare scatter list. */ npages = prepare_sg(sglist, nelems); - /* Step 2: Allocate a cluster. */ + /* Step 2: Allocate a cluster and context, if necessary. */ spin_lock_irqsave(&iommu->lock, flags); - base = alloc_streaming_cluster(iommu, npages); + base = alloc_npages(iommu, npages); + ctx = 0; + if (iommu->iommu_ctxflush) + ctx = iommu_alloc_ctx(iommu); + + spin_unlock_irqrestore(&iommu->lock, flags); + if (base == NULL) goto bad; - dma_base = iommu->page_table_map_base + ((base - iommu->page_table) << IO_PAGE_SHIFT); + + dma_base = iommu->page_table_map_base + + ((base - iommu->page_table) << IO_PAGE_SHIFT); /* Step 3: Normalize DMA addresses. */ used = nelems; @@ -702,30 +610,28 @@ int pci_map_sg(struct pci_dev *pdev, struct scatterlist *sglist, int nelems, int } used = nelems - used; - /* Step 4: Choose a context if necessary. */ - ctx = 0; - if (iommu->iommu_ctxflush) - ctx = iommu_alloc_ctx(iommu); - - /* Step 5: Create the mappings. */ + /* Step 4: Create the mappings. */ if (strbuf->strbuf_enabled) iopte_protection = IOPTE_STREAMING(ctx); else iopte_protection = IOPTE_CONSISTENT(ctx); if (direction != PCI_DMA_TODEVICE) iopte_protection |= IOPTE_WRITE; - fill_sg (base, sglist, used, nelems, iopte_protection); + + fill_sg(base, sglist, used, nelems, iopte_protection); + #ifdef VERIFY_SG verify_sglist(sglist, nelems, base, npages); #endif - spin_unlock_irqrestore(&iommu->lock, flags); - return used; bad: - spin_unlock_irqrestore(&iommu->lock, flags); - return PCI_DMA_ERROR_CODE; + iommu_free_ctx(iommu, ctx); +bad_no_ctx: + if (printk_ratelimit()) + WARN_ON(1); + return 0; } /* Unmap a set of streaming mode DMA translations. */ @@ -738,8 +644,10 @@ void pci_unmap_sg(struct pci_dev *pdev, struct scatterlist *sglist, int nelems, unsigned long flags, ctx, i, npages; u32 bus_addr; - if (direction == PCI_DMA_NONE) - BUG(); + if (unlikely(direction == PCI_DMA_NONE)) { + if (printk_ratelimit()) + WARN_ON(1); + } pcp = pdev->sysdata; iommu = pcp->pbm->iommu; @@ -751,7 +659,8 @@ void pci_unmap_sg(struct pci_dev *pdev, struct scatterlist *sglist, int nelems, if (sglist[i].dma_length == 0) break; i--; - npages = (IO_PAGE_ALIGN(sglist[i].dma_address + sglist[i].dma_length) - bus_addr) >> IO_PAGE_SHIFT; + npages = (IO_PAGE_ALIGN(sglist[i].dma_address + sglist[i].dma_length) - + bus_addr) >> IO_PAGE_SHIFT; base = iommu->page_table + ((bus_addr - iommu->page_table_map_base) >> IO_PAGE_SHIFT); @@ -772,11 +681,11 @@ void pci_unmap_sg(struct pci_dev *pdev, struct scatterlist *sglist, int nelems, if (strbuf->strbuf_enabled) pci_strbuf_flush(strbuf, iommu, bus_addr, ctx, npages, direction); - /* Step 2: Clear out first TSB entry. */ - iopte_make_dummy(iommu, base); + /* Step 2: Clear out the TSB entries. */ + for (i = 0; i < npages; i++) + iopte_make_dummy(iommu, base + i); - free_streaming_cluster(iommu, bus_addr - iommu->page_table_map_base, - npages, ctx); + free_npages(iommu, bus_addr - iommu->page_table_map_base, npages); iommu_free_ctx(iommu, ctx); diff --git a/include/asm-sparc64/pbm.h b/include/asm-sparc64/pbm.h index c067407de0b0..dd35a2c7798a 100644 --- a/include/asm-sparc64/pbm.h +++ b/include/asm-sparc64/pbm.h @@ -27,23 +27,27 @@ * PCI bus. */ -#define PBM_LOGCLUSTERS 3 -#define PBM_NCLUSTERS (1 << PBM_LOGCLUSTERS) - struct pci_controller_info; /* This contains the software state necessary to drive a PCI * controller's IOMMU. */ +struct pci_iommu_arena { + unsigned long *map; + unsigned int hint; + unsigned int limit; +}; + struct pci_iommu { /* This protects the controller's IOMMU and all * streaming buffers underneath. */ spinlock_t lock; + struct pci_iommu_arena arena; + /* IOMMU page table, a linear array of ioptes. */ iopte_t *page_table; /* The page table itself. */ - int page_table_sz_bits; /* log2 of ow many pages does it map? */ /* Base PCI memory space address where IOMMU mappings * begin. @@ -62,12 +66,6 @@ struct pci_iommu { */ unsigned long write_complete_reg; - /* The lowest used consistent mapping entry. Since - * we allocate consistent maps out of cluster 0 this - * is relative to the beginning of closter 0. - */ - u32 lowest_consistent_map; - /* In order to deal with some buggy third-party PCI bridges that * do wrong prefetching, we never mark valid mappings as invalid. * Instead we point them at this dummy page. @@ -75,16 +73,6 @@ struct pci_iommu { unsigned long dummy_page; unsigned long dummy_page_pa; - /* If PBM_NCLUSTERS is ever decreased to 4 or lower, - * or if largest supported page_table_sz * 8K goes above - * 2GB, you must increase the size of the type of - * these counters. You have been duly warned. -DaveM - */ - struct { - u16 next; - u16 flush; - } alloc_info[PBM_NCLUSTERS]; - /* CTX allocation. */ unsigned long ctx_lowest_free; unsigned long ctx_bitmap[IOMMU_NUM_CTXS / (sizeof(unsigned long) * 8)]; -- 2.39.2