aboutsummaryrefslogtreecommitdiffstats
path: root/include/asm-powerpc
diff options
context:
space:
mode:
authorLinas Vepstas <linas@austin.ibm.com>2006-10-30 00:15:59 -0500
committerPaul Mackerras <paulus@samba.org>2006-10-31 22:52:48 -0500
commit5d2efba64b231a1733c4048d1708d77e07f26426 (patch)
tree2893dd45b9c26cef6cddb5fef0c6f820c5eb534e /include/asm-powerpc
parentdd6c89f686bdb2a5de72fab636fc839e5a0add6d (diff)
[POWERPC] Use 4kB iommu pages even on 64kB-page systems
The 10Gigabit ethernet device drivers appear to be able to chew up all 256MB of TCE mappings on pSeries systems, as evidenced by numerous error messages: iommu_alloc failed, tbl c0000000010d5c48 vaddr c0000000d875eff0 npages 1 Some experimentation indicates that this is essentially because one 1500 byte ethernet MTU gets mapped as a 64K DMA region when the large 64K pages are enabled. Thus, it doesn't take much to exhaust all of the available DMA mappings for a high-speed card. This patch changes the iommu allocator to work with its own unique, distinct page size. Although the patch is long, its actually quite simple: it just #defines a distinct IOMMU_PAGE_SIZE and then uses this in all the places that matter. As a side effect, it also dramatically improves network performance on platforms with H-calls on iommu translation inserts/removes (since we no longer call it 16 times for a 1500 bytes packet when the iommu HW is still 4k). In the future, we might want to make the IOMMU_PAGE_SIZE a variable in the iommu_table instance, thus allowing support for different HW page sizes in the iommu itself. Signed-off-by: Linas Vepstas <linas@austin.ibm.com> Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org> Acked-by: Olof Johansson <olof@lixom.net> Acked-by: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Paul Mackerras <paulus@samba.org>
Diffstat (limited to 'include/asm-powerpc')
-rw-r--r--include/asm-powerpc/iommu.h22
-rw-r--r--include/asm-powerpc/tce.h3
2 files changed, 22 insertions, 3 deletions
diff --git a/include/asm-powerpc/iommu.h b/include/asm-powerpc/iommu.h
index a5e98641a2ae..39fad685ffab 100644
--- a/include/asm-powerpc/iommu.h
+++ b/include/asm-powerpc/iommu.h
@@ -22,17 +22,35 @@
22#define _ASM_IOMMU_H 22#define _ASM_IOMMU_H
23#ifdef __KERNEL__ 23#ifdef __KERNEL__
24 24
25#include <asm/types.h> 25#include <linux/compiler.h>
26#include <linux/spinlock.h> 26#include <linux/spinlock.h>
27#include <linux/device.h> 27#include <linux/device.h>
28#include <linux/dma-mapping.h> 28#include <linux/dma-mapping.h>
29#include <asm/types.h>
30#include <asm/bitops.h>
31
32#define IOMMU_PAGE_SHIFT 12
33#define IOMMU_PAGE_SIZE (ASM_CONST(1) << IOMMU_PAGE_SHIFT)
34#define IOMMU_PAGE_MASK (~((1 << IOMMU_PAGE_SHIFT) - 1))
35#define IOMMU_PAGE_ALIGN(addr) _ALIGN_UP(addr, IOMMU_PAGE_SIZE)
36
37#ifndef __ASSEMBLY__
38
39/* Pure 2^n version of get_order */
40static __inline__ __attribute_const__ int get_iommu_order(unsigned long size)
41{
42 return __ilog2((size - 1) >> IOMMU_PAGE_SHIFT) + 1;
43}
44
45#endif /* __ASSEMBLY__ */
46
29 47
30/* 48/*
31 * IOMAP_MAX_ORDER defines the largest contiguous block 49 * IOMAP_MAX_ORDER defines the largest contiguous block
32 * of dma space we can get. IOMAP_MAX_ORDER = 13 50 * of dma space we can get. IOMAP_MAX_ORDER = 13
33 * allows up to 2**12 pages (4096 * 4096) = 16 MB 51 * allows up to 2**12 pages (4096 * 4096) = 16 MB
34 */ 52 */
35#define IOMAP_MAX_ORDER 13 53#define IOMAP_MAX_ORDER 13
36 54
37struct iommu_table { 55struct iommu_table {
38 unsigned long it_busno; /* Bus number this table belongs to */ 56 unsigned long it_busno; /* Bus number this table belongs to */
diff --git a/include/asm-powerpc/tce.h b/include/asm-powerpc/tce.h
index c9483adbf599..f663634cccc9 100644
--- a/include/asm-powerpc/tce.h
+++ b/include/asm-powerpc/tce.h
@@ -22,6 +22,8 @@
22#define _ASM_POWERPC_TCE_H 22#define _ASM_POWERPC_TCE_H
23#ifdef __KERNEL__ 23#ifdef __KERNEL__
24 24
25#include <asm/iommu.h>
26
25/* 27/*
26 * Tces come in two formats, one for the virtual bus and a different 28 * Tces come in two formats, one for the virtual bus and a different
27 * format for PCI 29 * format for PCI
@@ -33,7 +35,6 @@
33 35
34#define TCE_SHIFT 12 36#define TCE_SHIFT 12
35#define TCE_PAGE_SIZE (1 << TCE_SHIFT) 37#define TCE_PAGE_SIZE (1 << TCE_SHIFT)
36#define TCE_PAGE_FACTOR (PAGE_SHIFT - TCE_SHIFT)
37 38
38#define TCE_ENTRY_SIZE 8 /* each TCE is 64 bits */ 39#define TCE_ENTRY_SIZE 8 /* each TCE is 64 bits */
39 40