From 8bee738bb1979c8bf7b42716b772522ab7d26b0c Mon Sep 17 00:00:00 2001 From: Masami Hiramatsu Date: Fri, 18 Dec 2009 10:40:13 -0500 Subject: [PATCH 1/8] x86: Fix objdump version check in chkobjdump.awk for different formats. Different version of objdump says its version in different way; GNU objdump 2.16.1 or GNU objdump version 2.19.51.0.14-1.fc11 20090722 This patch uses the first argument which starts with a number as version string. Changes in v2: - Remove unneeded increment. Signed-off-by: Masami Hiramatsu LKML-Reference: <20091218154012.16960.5113.stgit@dhcp-100-2-132.bos.redhat.com> Suggested-by: H. Peter Anvin Cc: Thomas Gleixner Cc: Ingo Molnar Cc: Andrew Morton Signed-off-by: H. Peter Anvin --- arch/x86/tools/chkobjdump.awk | 16 +++++++++++++--- 1 file changed, 13 insertions(+), 3 deletions(-) diff --git a/arch/x86/tools/chkobjdump.awk b/arch/x86/tools/chkobjdump.awk index 5bbb5a33f220..fd1ab80be0de 100644 --- a/arch/x86/tools/chkobjdump.awk +++ b/arch/x86/tools/chkobjdump.awk @@ -8,14 +8,24 @@ BEGIN { od_sver = 19; } -/^GNU/ { - split($3, ver, "."); +/^GNU objdump/ { + verstr = "" + for (i = 3; i <= NF; i++) + if (match($(i), "^[0-9]")) { + verstr = $(i); + break; + } + if (verstr == "") { + printf("Warning: Failed to find objdump version number.\n"); + exit 0; + } + split(verstr, ver, "."); if (ver[1] > od_ver || (ver[1] == od_ver && ver[2] >= od_sver)) { exit 1; } else { printf("Warning: objdump version %s is older than %d.%d\n", - $4, od_ver, od_sver); + verstr, od_ver, od_sver); print("Warning: Skipping posttest."); # Logic is inverted, because we just skip test without error. exit 0; From f797d9881b62c2ddb1d2e7bd80d87141949c84aa Mon Sep 17 00:00:00 2001 From: Shaun Ruffell Date: Thu, 17 Dec 2009 18:00:36 -0600 Subject: [PATCH 2/8] dma-debug: Do not add notifier when dma debugging is disabled. If CONFIG_HAVE_DMA_API_DEBUG is defined and "dma_debug=off" is specified on the kernel command line, when you detach a driver from a device you can cause the following NULL pointer dereference: BUG: unable to handle kernel NULL pointer dereference at (null) IP: [] dma_debug_device_change+0x5d/0x117 The problem is that the dma_debug_device_change notifier function is added to the bus notifier chain even though the dma_entry_hash array was never initialized. If dma debugging is disabled, this patch both prevents dma_debug_device_change notifiers from being added to the chain, and additionally ensures that the dma_debug_device_change notifier function is a no-op. Cc: stable@kernel.org Signed-off-by: Shaun Ruffell Signed-off-by: Joerg Roedel --- lib/dma-debug.c | 5 +++++ 1 file changed, 5 insertions(+) diff --git a/lib/dma-debug.c b/lib/dma-debug.c index d9b08e0f7f55..739974460c32 100644 --- a/lib/dma-debug.c +++ b/lib/dma-debug.c @@ -676,6 +676,8 @@ static int dma_debug_device_change(struct notifier_block *nb, struct device *dev = data; int count; + if (global_disable) + return; switch (action) { case BUS_NOTIFY_UNBOUND_DRIVER: @@ -697,6 +699,9 @@ void dma_debug_add_bus(struct bus_type *bus) { struct notifier_block *nb; + if (global_disable) + return; + nb = kzalloc(sizeof(struct notifier_block), GFP_KERNEL); if (nb == NULL) { pr_err("dma_debug_add_bus: out of memory\n"); From 0f764806438d5576ac58898332e5dcf30bb8a679 Mon Sep 17 00:00:00 2001 From: Joerg Roedel Date: Mon, 21 Dec 2009 15:51:23 +0100 Subject: [PATCH 3/8] x86/amd-iommu: Fix initialization failure panic The assumption that acpi_table_parse passes the return value of the hanlder function to the caller proved wrong recently. The return value of the handler function is totally ignored. This makes the initialization code for AMD IOMMU buggy in a way that could cause a kernel panic on initialization. This patch fixes the issue in the AMD IOMMU driver. Cc: stable@kernel.org Signed-off-by: Joerg Roedel --- arch/x86/kernel/amd_iommu_init.c | 10 ++++++++++ 1 file changed, 10 insertions(+) diff --git a/arch/x86/kernel/amd_iommu_init.c b/arch/x86/kernel/amd_iommu_init.c index 1dca9c34eaeb..fb490ce7dd55 100644 --- a/arch/x86/kernel/amd_iommu_init.c +++ b/arch/x86/kernel/amd_iommu_init.c @@ -137,6 +137,11 @@ int amd_iommus_present; /* IOMMUs have a non-present cache? */ bool amd_iommu_np_cache __read_mostly; +/* + * Set to true if ACPI table parsing and hardware intialization went properly + */ +static bool amd_iommu_initialized; + /* * List of protection domains - used during resume */ @@ -929,6 +934,8 @@ static int __init init_iommu_all(struct acpi_table_header *table) } WARN_ON(p != end); + amd_iommu_initialized = true; + return 0; } @@ -1263,6 +1270,9 @@ static int __init amd_iommu_init(void) if (acpi_table_parse("IVRS", init_iommu_all) != 0) goto free; + if (!amd_iommu_initialized) + goto free; + if (acpi_table_parse("IVRS", init_memory_definitions) != 0) goto free; From 17a2a9b57a9a7d2fd8f97df951b5e63e0bd56ef5 Mon Sep 17 00:00:00 2001 From: "H. Peter Anvin" Date: Fri, 25 Dec 2009 15:40:38 -0800 Subject: [PATCH 4/8] x86, compress: Force i386 instructions for the decompressor Recently, some distros have started shipping versions of gcc which default to -march=i686. This breaks building kernels for pre-i686 machines, even if they have been selected in Kconfig, due to the generation of CMOV instructions. There isn't enough benefit to try to preserve the generation of these instructions even when selected, so simply force -march=i386 for the decompressor when building a 32-bit kernel. Reported-and-tested-by: Chris Rankin Signed-off-by: H. Peter Anvin LKML-Reference: <219280.97558.qm@web52907.mail.re2.yahoo.com> --- arch/x86/boot/compressed/Makefile | 1 + 1 file changed, 1 insertion(+) diff --git a/arch/x86/boot/compressed/Makefile b/arch/x86/boot/compressed/Makefile index f8ed0658404c..f25bbd37765a 100644 --- a/arch/x86/boot/compressed/Makefile +++ b/arch/x86/boot/compressed/Makefile @@ -9,6 +9,7 @@ targets := vmlinux.lds vmlinux vmlinux.bin vmlinux.bin.gz vmlinux.bin.bz2 vmlinu KBUILD_CFLAGS := -m$(BITS) -D__KERNEL__ $(LINUX_INCLUDE) -O2 KBUILD_CFLAGS += -fno-strict-aliasing -fPIC KBUILD_CFLAGS += -DDISABLE_BRANCH_PROFILING +cflags-$(CONFIG_X86_32) := -march=i386 cflags-$(CONFIG_X86_64) := -mcmodel=small KBUILD_CFLAGS += $(cflags-y) KBUILD_CFLAGS += $(call cc-option,-ffreestanding) From d015a092989d673df44a5ad6866dc5d5006b7a2a Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Mon, 28 Dec 2009 10:26:59 +0200 Subject: [PATCH 5/8] x86: Use KERN_DEFAULT log-level in __show_regs() Andrew Morton reported a strange looking kmemcheck warning: WARNING: kmemcheck: Caught 32-bit read from uninitialized memory (ffff88004fba6c20) 0000000000000000310000000000000000000000000000002413000000c9ffff u u u u u u u u u u u u u u u u i i i i i i i i u u u u u u u u [] kmemleak_scan+0x25a/0x540 [] kmemleak_scan_thread+0x5b/0xe0 [] kthread+0x9e/0xb0 [] kernel_thread_helper+0x4/0x10 [] 0xffffffffffffffff The above printout is missing register dump completely. The problem here is that the output comes from syslog which doesn't show KERN_INFO log-level messages. We didn't see this before because both of us were testing on 32-bit kernels which use the _default_ log-level. Fix that up by explicitly using KERN_DEFAULT log-level for __show_regs() printks. Signed-off-by: Pekka Enberg Cc: Vegard Nossum Cc: Andrew Morton Cc: Arjan van de Ven Cc: Linus Torvalds LKML-Reference: <1261988819.4641.2.camel@penberg-laptop> Signed-off-by: Ingo Molnar --- arch/x86/kernel/process.c | 4 ++-- arch/x86/kernel/process_32.c | 14 +++++++------- arch/x86/kernel/process_64.c | 24 ++++++++++++------------ 3 files changed, 21 insertions(+), 21 deletions(-) diff --git a/arch/x86/kernel/process.c b/arch/x86/kernel/process.c index 98c2cdeb599e..c6ee241c8a98 100644 --- a/arch/x86/kernel/process.c +++ b/arch/x86/kernel/process.c @@ -103,8 +103,8 @@ void show_regs_common(void) if (!product) product = ""; - printk("\n"); - printk(KERN_INFO "Pid: %d, comm: %.20s %s %s %.*s %s/%s\n", + printk(KERN_CONT "\n"); + printk(KERN_DEFAULT "Pid: %d, comm: %.20s %s %s %.*s %s/%s\n", current->pid, current->comm, print_tainted(), init_utsname()->release, (int)strcspn(init_utsname()->version, " "), diff --git a/arch/x86/kernel/process_32.c b/arch/x86/kernel/process_32.c index 9c517b5858f0..37ad1e046aae 100644 --- a/arch/x86/kernel/process_32.c +++ b/arch/x86/kernel/process_32.c @@ -139,16 +139,16 @@ void __show_regs(struct pt_regs *regs, int all) show_regs_common(); - printk("EIP: %04x:[<%08lx>] EFLAGS: %08lx CPU: %d\n", + printk(KERN_DEFAULT "EIP: %04x:[<%08lx>] EFLAGS: %08lx CPU: %d\n", (u16)regs->cs, regs->ip, regs->flags, smp_processor_id()); print_symbol("EIP is at %s\n", regs->ip); - printk("EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n", + printk(KERN_DEFAULT "EAX: %08lx EBX: %08lx ECX: %08lx EDX: %08lx\n", regs->ax, regs->bx, regs->cx, regs->dx); - printk("ESI: %08lx EDI: %08lx EBP: %08lx ESP: %08lx\n", + printk(KERN_DEFAULT "ESI: %08lx EDI: %08lx EBP: %08lx ESP: %08lx\n", regs->si, regs->di, regs->bp, sp); - printk(" DS: %04x ES: %04x FS: %04x GS: %04x SS: %04x\n", + printk(KERN_DEFAULT " DS: %04x ES: %04x FS: %04x GS: %04x SS: %04x\n", (u16)regs->ds, (u16)regs->es, (u16)regs->fs, gs, ss); if (!all) @@ -158,19 +158,19 @@ void __show_regs(struct pt_regs *regs, int all) cr2 = read_cr2(); cr3 = read_cr3(); cr4 = read_cr4_safe(); - printk("CR0: %08lx CR2: %08lx CR3: %08lx CR4: %08lx\n", + printk(KERN_DEFAULT "CR0: %08lx CR2: %08lx CR3: %08lx CR4: %08lx\n", cr0, cr2, cr3, cr4); get_debugreg(d0, 0); get_debugreg(d1, 1); get_debugreg(d2, 2); get_debugreg(d3, 3); - printk("DR0: %08lx DR1: %08lx DR2: %08lx DR3: %08lx\n", + printk(KERN_DEFAULT "DR0: %08lx DR1: %08lx DR2: %08lx DR3: %08lx\n", d0, d1, d2, d3); get_debugreg(d6, 6); get_debugreg(d7, 7); - printk("DR6: %08lx DR7: %08lx\n", + printk(KERN_DEFAULT "DR6: %08lx DR7: %08lx\n", d6, d7); } diff --git a/arch/x86/kernel/process_64.c b/arch/x86/kernel/process_64.c index 52fbd0c60198..f9e033150cdf 100644 --- a/arch/x86/kernel/process_64.c +++ b/arch/x86/kernel/process_64.c @@ -161,19 +161,19 @@ void __show_regs(struct pt_regs *regs, int all) unsigned int ds, cs, es; show_regs_common(); - printk(KERN_INFO "RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->ip); + printk(KERN_DEFAULT "RIP: %04lx:[<%016lx>] ", regs->cs & 0xffff, regs->ip); printk_address(regs->ip, 1); - printk(KERN_INFO "RSP: %04lx:%016lx EFLAGS: %08lx\n", regs->ss, + printk(KERN_DEFAULT "RSP: %04lx:%016lx EFLAGS: %08lx\n", regs->ss, regs->sp, regs->flags); - printk(KERN_INFO "RAX: %016lx RBX: %016lx RCX: %016lx\n", + printk(KERN_DEFAULT "RAX: %016lx RBX: %016lx RCX: %016lx\n", regs->ax, regs->bx, regs->cx); - printk(KERN_INFO "RDX: %016lx RSI: %016lx RDI: %016lx\n", + printk(KERN_DEFAULT "RDX: %016lx RSI: %016lx RDI: %016lx\n", regs->dx, regs->si, regs->di); - printk(KERN_INFO "RBP: %016lx R08: %016lx R09: %016lx\n", + printk(KERN_DEFAULT "RBP: %016lx R08: %016lx R09: %016lx\n", regs->bp, regs->r8, regs->r9); - printk(KERN_INFO "R10: %016lx R11: %016lx R12: %016lx\n", + printk(KERN_DEFAULT "R10: %016lx R11: %016lx R12: %016lx\n", regs->r10, regs->r11, regs->r12); - printk(KERN_INFO "R13: %016lx R14: %016lx R15: %016lx\n", + printk(KERN_DEFAULT "R13: %016lx R14: %016lx R15: %016lx\n", regs->r13, regs->r14, regs->r15); asm("movl %%ds,%0" : "=r" (ds)); @@ -194,21 +194,21 @@ void __show_regs(struct pt_regs *regs, int all) cr3 = read_cr3(); cr4 = read_cr4(); - printk(KERN_INFO "FS: %016lx(%04x) GS:%016lx(%04x) knlGS:%016lx\n", + printk(KERN_DEFAULT "FS: %016lx(%04x) GS:%016lx(%04x) knlGS:%016lx\n", fs, fsindex, gs, gsindex, shadowgs); - printk(KERN_INFO "CS: %04x DS: %04x ES: %04x CR0: %016lx\n", cs, ds, + printk(KERN_DEFAULT "CS: %04x DS: %04x ES: %04x CR0: %016lx\n", cs, ds, es, cr0); - printk(KERN_INFO "CR2: %016lx CR3: %016lx CR4: %016lx\n", cr2, cr3, + printk(KERN_DEFAULT "CR2: %016lx CR3: %016lx CR4: %016lx\n", cr2, cr3, cr4); get_debugreg(d0, 0); get_debugreg(d1, 1); get_debugreg(d2, 2); - printk(KERN_INFO "DR0: %016lx DR1: %016lx DR2: %016lx\n", d0, d1, d2); + printk(KERN_DEFAULT "DR0: %016lx DR1: %016lx DR2: %016lx\n", d0, d1, d2); get_debugreg(d3, 3); get_debugreg(d6, 6); get_debugreg(d7, 7); - printk(KERN_INFO "DR3: %016lx DR6: %016lx DR7: %016lx\n", d3, d6, d7); + printk(KERN_DEFAULT "DR3: %016lx DR6: %016lx DR7: %016lx\n", d3, d6, d7); } void show_regs(struct pt_regs *regs) From c0ca9da442df82b67095f230f24762042f9f3b7d Mon Sep 17 00:00:00 2001 From: Pekka Enberg Date: Mon, 28 Dec 2009 11:02:15 +0200 Subject: [PATCH 6/8] x86, kmemcheck: Use KERN_WARNING for error reporting As suggested by Vegard Nossum, use KERN_WARNING for error reporting to make sure kmemcheck reports end up in syslog. Suggested-by: Vegard Nossum Signed-off-by: Pekka Enberg Cc: Andrew Morton LKML-Reference: <1261990935.4641.7.camel@penberg-laptop> Signed-off-by: Ingo Molnar --- arch/x86/mm/kmemcheck/error.c | 19 +++++++++---------- 1 file changed, 9 insertions(+), 10 deletions(-) diff --git a/arch/x86/mm/kmemcheck/error.c b/arch/x86/mm/kmemcheck/error.c index 4901d0dafda6..af3b6c8a436f 100644 --- a/arch/x86/mm/kmemcheck/error.c +++ b/arch/x86/mm/kmemcheck/error.c @@ -106,26 +106,25 @@ void kmemcheck_error_recall(void) switch (e->type) { case KMEMCHECK_ERROR_INVALID_ACCESS: - printk(KERN_ERR "WARNING: kmemcheck: Caught %d-bit read " - "from %s memory (%p)\n", + printk(KERN_WARNING "WARNING: kmemcheck: Caught %d-bit read from %s memory (%p)\n", 8 * e->size, e->state < ARRAY_SIZE(desc) ? desc[e->state] : "(invalid shadow state)", (void *) e->address); - printk(KERN_INFO); + printk(KERN_WARNING); for (i = 0; i < SHADOW_COPY_SIZE; ++i) - printk("%02x", e->memory_copy[i]); - printk("\n"); + printk(KERN_CONT "%02x", e->memory_copy[i]); + printk(KERN_CONT "\n"); - printk(KERN_INFO); + printk(KERN_WARNING); for (i = 0; i < SHADOW_COPY_SIZE; ++i) { if (e->shadow_copy[i] < ARRAY_SIZE(short_desc)) - printk(" %c", short_desc[e->shadow_copy[i]]); + printk(KERN_CONT " %c", short_desc[e->shadow_copy[i]]); else - printk(" ?"); + printk(KERN_CONT " ?"); } - printk("\n"); - printk(KERN_INFO "%*c\n", 2 + 2 + printk(KERN_CONT "\n"); + printk(KERN_WARNING "%*c\n", 2 + 2 * (int) (e->address & (SHADOW_COPY_SIZE - 1)), '^'); break; case KMEMCHECK_ERROR_BUG: From 39d30770992895d55789de64bad2349510af68d0 Mon Sep 17 00:00:00 2001 From: Mike Travis Date: Mon, 28 Dec 2009 13:28:25 -0800 Subject: [PATCH 7/8] x86: SGI UV: Fix writes to led registers on remote uv hubs The wrong address was being used to write the SCIR led regs on remote hubs. Also, there was an inconsistency between how BIOS and the kernel indexed these regs. Standardize on using the lower 6 bits of the APIC ID as the index. This patch fixes the problem of writing to an errant address to a cpu # >= 64. Signed-off-by: Mike Travis Reviewed-by: Jack Steiner Cc: Robin Holt Cc: Linus Torvalds Cc: stable@kernel.org LKML-Reference: <4B3922F9.3060905@sgi.com> [ v2: fix a number of annoying checkpatch artifacts and whitespace noise ] Signed-off-by: Ingo Molnar --- arch/x86/include/asm/uv/uv_hub.h | 86 +++++++++++++++++------------- arch/x86/kernel/apic/x2apic_uv_x.c | 12 ++--- 2 files changed, 54 insertions(+), 44 deletions(-) diff --git a/arch/x86/include/asm/uv/uv_hub.h b/arch/x86/include/asm/uv/uv_hub.h index 811bfabc80b7..bc54fa965af3 100644 --- a/arch/x86/include/asm/uv/uv_hub.h +++ b/arch/x86/include/asm/uv/uv_hub.h @@ -31,20 +31,20 @@ * contiguous (although various IO spaces may punch holes in * it).. * - * N - Number of bits in the node portion of a socket physical - * address. + * N - Number of bits in the node portion of a socket physical + * address. * - * NASID - network ID of a router, Mbrick or Cbrick. Nasid values of - * routers always have low bit of 1, C/MBricks have low bit - * equal to 0. Most addressing macros that target UV hub chips - * right shift the NASID by 1 to exclude the always-zero bit. - * NASIDs contain up to 15 bits. + * NASID - network ID of a router, Mbrick or Cbrick. Nasid values of + * routers always have low bit of 1, C/MBricks have low bit + * equal to 0. Most addressing macros that target UV hub chips + * right shift the NASID by 1 to exclude the always-zero bit. + * NASIDs contain up to 15 bits. * * GNODE - NASID right shifted by 1 bit. Most mmrs contain gnodes instead * of nasids. * - * PNODE - the low N bits of the GNODE. The PNODE is the most useful variant - * of the nasid for socket usage. + * PNODE - the low N bits of the GNODE. The PNODE is the most useful variant + * of the nasid for socket usage. * * * NumaLink Global Physical Address Format: @@ -71,12 +71,12 @@ * * * APICID format - * NOTE!!!!!! This is the current format of the APICID. However, code - * should assume that this will change in the future. Use functions - * in this file for all APICID bit manipulations and conversion. + * NOTE!!!!!! This is the current format of the APICID. However, code + * should assume that this will change in the future. Use functions + * in this file for all APICID bit manipulations and conversion. * - * 1111110000000000 - * 5432109876543210 + * 1111110000000000 + * 5432109876543210 * pppppppppplc0cch * sssssssssss * @@ -89,9 +89,9 @@ * Note: Processor only supports 12 bits in the APICID register. The ACPI * tables hold all 16 bits. Software needs to be aware of this. * - * Unless otherwise specified, all references to APICID refer to - * the FULL value contained in ACPI tables, not the subset in the - * processor APICID register. + * Unless otherwise specified, all references to APICID refer to + * the FULL value contained in ACPI tables, not the subset in the + * processor APICID register. */ @@ -151,16 +151,16 @@ struct uv_hub_info_s { }; DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); -#define uv_hub_info (&__get_cpu_var(__uv_hub_info)) +#define uv_hub_info (&__get_cpu_var(__uv_hub_info)) #define uv_cpu_hub_info(cpu) (&per_cpu(__uv_hub_info, cpu)) /* * Local & Global MMR space macros. - * Note: macros are intended to be used ONLY by inline functions - * in this file - not by other kernel code. - * n - NASID (full 15-bit global nasid) - * g - GNODE (full 15-bit global nasid, right shifted 1) - * p - PNODE (local part of nsids, right shifted 1) + * Note: macros are intended to be used ONLY by inline functions + * in this file - not by other kernel code. + * n - NASID (full 15-bit global nasid) + * g - GNODE (full 15-bit global nasid, right shifted 1) + * p - PNODE (local part of nsids, right shifted 1) */ #define UV_NASID_TO_PNODE(n) (((n) >> 1) & uv_hub_info->pnode_mask) #define UV_PNODE_TO_GNODE(p) ((p) |uv_hub_info->gnode_extra) @@ -215,8 +215,8 @@ DECLARE_PER_CPU(struct uv_hub_info_s, __uv_hub_info); /* * Macros for converting between kernel virtual addresses, socket local physical * addresses, and UV global physical addresses. - * Note: use the standard __pa() & __va() macros for converting - * between socket virtual and socket physical addresses. + * Note: use the standard __pa() & __va() macros for converting + * between socket virtual and socket physical addresses. */ /* socket phys RAM --> UV global physical address */ @@ -287,21 +287,18 @@ static inline int uv_apicid_to_pnode(int apicid) * Access global MMRs using the low memory MMR32 space. This region supports * faster MMR access but not all MMRs are accessible in this space. */ -static inline unsigned long *uv_global_mmr32_address(int pnode, - unsigned long offset) +static inline unsigned long *uv_global_mmr32_address(int pnode, unsigned long offset) { return __va(UV_GLOBAL_MMR32_BASE | UV_GLOBAL_MMR32_PNODE_BITS(pnode) | offset); } -static inline void uv_write_global_mmr32(int pnode, unsigned long offset, - unsigned long val) +static inline void uv_write_global_mmr32(int pnode, unsigned long offset, unsigned long val) { writeq(val, uv_global_mmr32_address(pnode, offset)); } -static inline unsigned long uv_read_global_mmr32(int pnode, - unsigned long offset) +static inline unsigned long uv_read_global_mmr32(int pnode, unsigned long offset) { return readq(uv_global_mmr32_address(pnode, offset)); } @@ -310,21 +307,18 @@ static inline unsigned long uv_read_global_mmr32(int pnode, * Access Global MMR space using the MMR space located at the top of physical * memory. */ -static inline unsigned long *uv_global_mmr64_address(int pnode, - unsigned long offset) +static inline unsigned long *uv_global_mmr64_address(int pnode, unsigned long offset) { return __va(UV_GLOBAL_MMR64_BASE | UV_GLOBAL_MMR64_PNODE_BITS(pnode) | offset); } -static inline void uv_write_global_mmr64(int pnode, unsigned long offset, - unsigned long val) +static inline void uv_write_global_mmr64(int pnode, unsigned long offset, unsigned long val) { writeq(val, uv_global_mmr64_address(pnode, offset)); } -static inline unsigned long uv_read_global_mmr64(int pnode, - unsigned long offset) +static inline unsigned long uv_read_global_mmr64(int pnode, unsigned long offset) { return readq(uv_global_mmr64_address(pnode, offset)); } @@ -338,6 +332,16 @@ static inline unsigned long uv_global_gru_mmr_address(int pnode, unsigned long o return UV_GLOBAL_GRU_MMR_BASE | offset | (pnode << uv_hub_info->m_val); } +static inline void uv_write_global_mmr8(int pnode, unsigned long offset, unsigned char val) +{ + writeb(val, uv_global_mmr64_address(pnode, offset)); +} + +static inline unsigned char uv_read_global_mmr8(int pnode, unsigned long offset) +{ + return readb(uv_global_mmr64_address(pnode, offset)); +} + /* * Access hub local MMRs. Faster than using global space but only local MMRs * are accessible. @@ -457,11 +461,17 @@ static inline void uv_set_scir_bits(unsigned char value) } } +static inline unsigned long uv_scir_offset(int apicid) +{ + return SCIR_LOCAL_MMR_BASE | (apicid & 0x3f); +} + static inline void uv_set_cpu_scir_bits(int cpu, unsigned char value) { if (uv_cpu_hub_info(cpu)->scir.state != value) { + uv_write_global_mmr8(uv_cpu_to_pnode(cpu), + uv_cpu_hub_info(cpu)->scir.offset, value); uv_cpu_hub_info(cpu)->scir.state = value; - uv_write_local_mmr8(uv_cpu_hub_info(cpu)->scir.offset, value); } } diff --git a/arch/x86/kernel/apic/x2apic_uv_x.c b/arch/x86/kernel/apic/x2apic_uv_x.c index d56b0efb2057..5f92494dab61 100644 --- a/arch/x86/kernel/apic/x2apic_uv_x.c +++ b/arch/x86/kernel/apic/x2apic_uv_x.c @@ -629,8 +629,10 @@ void __init uv_system_init(void) uv_rtc_init(); for_each_present_cpu(cpu) { + int apicid = per_cpu(x86_cpu_to_apicid, cpu); + nid = cpu_to_node(cpu); - pnode = uv_apicid_to_pnode(per_cpu(x86_cpu_to_apicid, cpu)); + pnode = uv_apicid_to_pnode(apicid); blade = boot_pnode_to_blade(pnode); lcpu = uv_blade_info[blade].nr_possible_cpus; uv_blade_info[blade].nr_possible_cpus++; @@ -651,15 +653,13 @@ void __init uv_system_init(void) uv_cpu_hub_info(cpu)->gnode_extra = gnode_extra; uv_cpu_hub_info(cpu)->global_mmr_base = mmr_base; uv_cpu_hub_info(cpu)->coherency_domain_number = sn_coherency_id; - uv_cpu_hub_info(cpu)->scir.offset = SCIR_LOCAL_MMR_BASE + lcpu; + uv_cpu_hub_info(cpu)->scir.offset = uv_scir_offset(apicid); uv_node_to_blade[nid] = blade; uv_cpu_to_blade[cpu] = blade; max_pnode = max(pnode, max_pnode); - printk(KERN_DEBUG "UV: cpu %d, apicid 0x%x, pnode %d, nid %d, " - "lcpu %d, blade %d\n", - cpu, per_cpu(x86_cpu_to_apicid, cpu), pnode, nid, - lcpu, blade); + printk(KERN_DEBUG "UV: cpu %d, apicid 0x%x, pnode %d, nid %d, lcpu %d, blade %d\n", + cpu, apicid, pnode, nid, lcpu, blade); } /* Add blade/pnode info for nodes without cpus */ From f405d2c02395a74d3883bd03ded36457aa3697ad Mon Sep 17 00:00:00 2001 From: FUJITA Tomonori Date: Mon, 28 Dec 2009 18:11:56 +0900 Subject: [PATCH 8/8] x86/agp: Fix agp_amd64_init() initialization with CONFIG_GART_IOMMU enabled with CONFIG_GART_IOMMU enabled drivers/char/agp/amd64-agp.c has: #ifndef CONFIG_GART_IOMMU module_init(agp_amd64_init); module_exit(agp_amd64_cleanup); #endif agp_amd64_init() was called via gart_iommu_init with CONFIG_GART_IOMMU=y agp_amd64_init() was called via module_init with CONFIG_GART_IOMMU=n The commit 75f1cdf1dda92cae037ec848ae63690d91913eac changes the x86 dma initialization routine: gart_iommu_init() is called only when GART IOMMU is detected. So when GART IOMMU isn't detected, agp_amd64_init isn't called. Marin Mitov reported this issue: http://marc.info/?l=linux-kernel&m=126192729110083&w=2 With this patch, agp_amd64_init() is always called via module_init (the above ifndef is removed). If agp_amd64_init() is called via gart_iommu_init() earlier, agp_amd64_init() finishes without doing anything (when it is called via module_init). Reported-by: Marin Mitov Tested-by: Marin Mitov Signed-off-by: FUJITA Tomonori Cc: davej@redhat.com LKML-Reference: <20091228181118C.fujita.tomonori@lab.ntt.co.jp> Signed-off-by: Ingo Molnar --- drivers/char/agp/amd64-agp.c | 9 +++++---- 1 file changed, 5 insertions(+), 4 deletions(-) diff --git a/drivers/char/agp/amd64-agp.c b/drivers/char/agp/amd64-agp.c index 2fb2e6cc322a..5aa7a586a7ff 100644 --- a/drivers/char/agp/amd64-agp.c +++ b/drivers/char/agp/amd64-agp.c @@ -725,9 +725,14 @@ static struct pci_driver agp_amd64_pci_driver = { int __init agp_amd64_init(void) { int err = 0; + static int done = 0; if (agp_off) return -EINVAL; + + if (done++) + return agp_bridges_found ? 0 : -ENODEV; + err = pci_register_driver(&agp_amd64_pci_driver); if (err < 0) return err; @@ -771,12 +776,8 @@ static void __exit agp_amd64_cleanup(void) pci_unregister_driver(&agp_amd64_pci_driver); } -/* On AMD64 the PCI driver needs to initialize this driver early - for the IOMMU, so it has to be called via a backdoor. */ -#ifndef CONFIG_GART_IOMMU module_init(agp_amd64_init); module_exit(agp_amd64_cleanup); -#endif MODULE_AUTHOR("Dave Jones , Andi Kleen"); module_param(agp_try_unsupported, bool, 0);