From 08be97962bf338161325d4901642f956ce8c1adb Mon Sep 17 00:00:00 2001 From: Thomas Gleixner Date: Wed, 14 Jul 2010 21:36:27 +0200 Subject: x86: Force HPET readback_cmp for all ATI chipsets commit 30a564be (x86, hpet: Restrict read back to affected ATI chipset) restricted the workaround for the HPET bug to SMX00 chipsets. This was reasonable as those were the only ones against which we ever got a bug report. Stephan Wolf reported now that this patch breaks his IXP400 based machine. Though it's confirmed to work on other IXP400 based systems. To error out on the safe side, we force the HPET readback workaround for all ATI SMbus class chipsets. Reported-by: Stephan Wolf LKML-Reference: Signed-off-by: Thomas Gleixner Tested-by: Stephan Wolf Acked-by: Borislav Petkov --- arch/x86/kernel/early-quirks.c | 18 ++++++++++++++++++ arch/x86/kernel/quirks.c | 5 ----- 2 files changed, 18 insertions(+), 5 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/early-quirks.c b/arch/x86/kernel/early-quirks.c index ebdb85cf2686..e5cc7e82e60d 100644 --- a/arch/x86/kernel/early-quirks.c +++ b/arch/x86/kernel/early-quirks.c @@ -18,6 +18,7 @@ #include #include #include +#include static void __init fix_hypertransport_config(int num, int slot, int func) { @@ -191,6 +192,21 @@ static void __init ati_bugs_contd(int num, int slot, int func) } #endif +/* + * Force the read back of the CMP register in hpet_next_event() + * to work around the problem that the CMP register write seems to be + * delayed. See hpet_next_event() for details. + * + * We do this on all SMBUS incarnations for now until we have more + * information about the affected chipsets. + */ +static void __init ati_hpet_bugs(int num, int slot, int func) +{ +#ifdef CONFIG_HPET_TIMER + hpet_readback_cmp = 1; +#endif +} + #define QFLAG_APPLY_ONCE 0x1 #define QFLAG_APPLIED 0x2 #define QFLAG_DONE (QFLAG_APPLY_ONCE|QFLAG_APPLIED) @@ -220,6 +236,8 @@ static struct chipset early_qrk[] __initdata = { PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs }, { PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS, PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_bugs_contd }, + { PCI_VENDOR_ID_ATI, PCI_ANY_ID, + PCI_CLASS_SERIAL_SMBUS, PCI_ANY_ID, 0, ati_hpet_bugs }, {} }; diff --git a/arch/x86/kernel/quirks.c b/arch/x86/kernel/quirks.c index e72d3fc6547d..939b9e98245f 100644 --- a/arch/x86/kernel/quirks.c +++ b/arch/x86/kernel/quirks.c @@ -498,15 +498,10 @@ void force_hpet_resume(void) * See erratum #27 (Misinterpreted MSI Requests May Result in * Corrupted LPC DMA Data) in AMD Publication #46837, * "SB700 Family Product Errata", Rev. 1.0, March 2010. - * - * Also force the read back of the CMP register in hpet_next_event() - * to work around the problem that the CMP register write seems to be - * delayed. See hpet_next_event() for details. */ static void force_disable_hpet_msi(struct pci_dev *unused) { hpet_msi_disable = 1; - hpet_readback_cmp = 1; } DECLARE_PCI_FIXUP_HEADER(PCI_VENDOR_ID_ATI, PCI_DEVICE_ID_ATI_SBX00_SMBUS, -- cgit v1.2.3 From fd19dce7ac07973f700b0f13fb7f94b951414a4c Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Thu, 15 Jul 2010 00:00:59 -0700 Subject: x86: Fix x2apic preenabled system with kexec Found one x2apic system kexec loop test failed when CONFIG_NMI_WATCHDOG=y (old) or CONFIG_LOCKUP_DETECTOR=y (current tip) first kernel can kexec second kernel, but second kernel can not kexec third one. it can be duplicated on another system with BIOS preenabled x2apic. First kernel can not kexec second kernel. It turns out, when kernel boot with pre-enabled x2apic, it will not execute disable_local_APIC on shutdown path. when init_apic_mappings() is called in setup_arch, it will skip setting of apic_phys when x2apic_mode is set. ( x2apic_mode is much early check_x2apic()) Then later, disable_local_APIC() will bail out early because !apic_phys. So check !x2apic_mode in x2apic_mode in disable_local_APIC with !apic_phys. another solution could be updating init_apic_mappings() to set apic_phys even for preenabled x2apic system. Actually even for x2apic system, that lapic address is mapped already in early stage. BTW: is there any x2apic preenabled system with apicid of boot cpu > 255? Signed-off-by: Yinghai Lu LKML-Reference: <4C3EB22B.3000701@kernel.org> Acked-by: Suresh Siddha Cc: stable@kernel.org Signed-off-by: H. Peter Anvin --- arch/x86/kernel/apic/apic.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/apic/apic.c b/arch/x86/kernel/apic/apic.c index c02cc692985c..a96489ee6cab 100644 --- a/arch/x86/kernel/apic/apic.c +++ b/arch/x86/kernel/apic/apic.c @@ -921,7 +921,7 @@ void disable_local_APIC(void) unsigned int value; /* APIC hasn't been mapped yet */ - if (!apic_phys) + if (!x2apic_mode && !apic_phys) return; clear_local_APIC(); -- cgit v1.2.3 From a197479848a2f1a2a5c07cffa6c31ab5e8c82797 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Fri, 16 Jul 2010 18:17:12 -0700 Subject: x86: kprobes: fix swapped segment registers in kretprobe In commit f007ea26, the order of the %es and %ds segment registers got accidentally swapped, so synthesized 'struct pt_regs' frames have the two values inverted. It's almost sure that these values never matter, and that they also never differ. But wrong is wrong. Signed-off-by: Roland McGrath --- arch/x86/kernel/kprobes.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/kprobes.c b/arch/x86/kernel/kprobes.c index 345a4b1fe144..675879b65ce6 100644 --- a/arch/x86/kernel/kprobes.c +++ b/arch/x86/kernel/kprobes.c @@ -640,8 +640,8 @@ static int __kprobes kprobe_handler(struct pt_regs *regs) /* Skip cs, ip, orig_ax and gs. */ \ " subl $16, %esp\n" \ " pushl %fs\n" \ - " pushl %ds\n" \ " pushl %es\n" \ + " pushl %ds\n" \ " pushl %eax\n" \ " pushl %ebp\n" \ " pushl %edi\n" \ -- cgit v1.2.3 From 087b255a2b43f417af83cb44e0bb02507f36b7fe Mon Sep 17 00:00:00 2001 From: Adam Lackorzynski Date: Tue, 20 Jul 2010 15:18:19 -0700 Subject: x86, i8259: Only register sysdev if we have a real 8259 PIC My platform makes use of the null_legacy_pic choice and oopses when doing a shutdown as the shutdown code goes through all the registered sysdevs and calls their shutdown method which in my case poke on a non-existing i8259. Imho the i8259 specific sysdev should only be registered if the i8259 is actually there. Do not register the sysdev function when the null_legacy_pic is used so that the i8259 resume, suspend and shutdown functions are not called. Signed-off-by: Adam Lackorzynski LKML-Reference: <201007202218.o6KMIJ3m020955@imap1.linux-foundation.org> Cc: Jacob Pan Cc: 2.6.34 Signed-off-by: Andrew Morton Signed-off-by: H. Peter Anvin --- arch/x86/kernel/i8259.c | 25 +++++++++++++++---------- 1 file changed, 15 insertions(+), 10 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/i8259.c b/arch/x86/kernel/i8259.c index 7c9f02c130f3..cafa7c80ac95 100644 --- a/arch/x86/kernel/i8259.c +++ b/arch/x86/kernel/i8259.c @@ -276,16 +276,6 @@ static struct sys_device device_i8259A = { .cls = &i8259_sysdev_class, }; -static int __init i8259A_init_sysfs(void) -{ - int error = sysdev_class_register(&i8259_sysdev_class); - if (!error) - error = sysdev_register(&device_i8259A); - return error; -} - -device_initcall(i8259A_init_sysfs); - static void mask_8259A(void) { unsigned long flags; @@ -407,3 +397,18 @@ struct legacy_pic default_legacy_pic = { }; struct legacy_pic *legacy_pic = &default_legacy_pic; + +static int __init i8259A_init_sysfs(void) +{ + int error; + + if (legacy_pic != &default_legacy_pic) + return 0; + + error = sysdev_class_register(&i8259_sysdev_class); + if (!error) + error = sysdev_register(&device_i8259A); + return error; +} + +device_initcall(i8259A_init_sysfs); -- cgit v1.2.3 From 9aebbdb637a73a6092e1456ebb4a2df32cc1f611 Mon Sep 17 00:00:00 2001 From: Yinghai Lu Date: Tue, 20 Jul 2010 13:24:30 -0700 Subject: x86, numa: fix boot without RAM on node0 again Commit e534c7c5f8d6 ("numa: x86_64: use generic percpu var numa_node_id() implementation") broke numa systems that don't have ram on node0 when MEMORY_HOTPLUG is enabled, because cpu_up() will call cpu_to_node() before per_cpu(numa_node) is setup for APs. When Node0 doesn't have RAM, on x86, cpus already round it to nearest node with RAM in x86_cpu_to_node_map. and per_cpu(numa_node) is not set up until in c_init for APs. When later cpu_up() calling cpu_to_node() will get 0 again, and make it online even there is no RAM on node0. so later all APs can not booted up, and later will have panic. [ 1.611101] On node 0 totalpages: 0 ......... [ 2.608558] On node 0 totalpages: 0 [ 2.612065] Brought up 1 CPUs [ 2.615199] Total of 1 processors activated (3990.31 BogoMIPS). ... 93.225341] calling loop_init+0x0/0x1a4 @ 1 [ 93.229314] PERCPU: allocation failed, size=80 align=8, failed to populate [ 93.246539] Pid: 1, comm: swapper Tainted: G W 2.6.35-rc4-tip-yh-04371-gd64e6c4-dirty #354 [ 93.264621] Call Trace: [ 93.266533] [] pcpu_alloc+0x83a/0x8e7 [ 93.270710] [] __alloc_percpu+0x10/0x12 [ 93.285849] [] alloc_disk_node+0x94/0x16d [ 93.291811] [] alloc_disk+0x11/0x13 [ 93.306157] [] loop_alloc+0xa7/0x180 [ 93.310538] [] loop_init+0x9b/0x1a4 [ 93.324909] [] ? loop_init+0x0/0x1a4 [ 93.329650] [] do_one_initcall+0x57/0x136 [ 93.345197] [] kernel_init+0x184/0x20e [ 93.348146] [] kernel_thread_helper+0x4/0x10 [ 93.365194] [] ? restore_args+0x0/0x30 [ 93.369305] [] ? kernel_init+0x0/0x20e [ 93.386011] [] ? kernel_thread_helper+0x0/0x10 [ 93.392047] loop: out of memory ... Try to assign per_cpu(numa_node) early [akpm@linux-foundation.org: tidy up code comment] Signed-off-by: Yinghai Cc: Thomas Gleixner Cc: Ingo Molnar Cc: "H. Peter Anvin" Cc: Tejun Heo Cc: Denys Vlasenko Acked-by: Lee Schermerhorn Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- arch/x86/kernel/setup_percpu.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index de3b63ae3da2..690c2c09faf3 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -238,6 +238,15 @@ void __init setup_per_cpu_areas(void) #ifdef CONFIG_NUMA per_cpu(x86_cpu_to_node_map, cpu) = early_per_cpu_map(x86_cpu_to_node_map, cpu); + /* + * Ensure taht the boot cpu numa_node is correct when the boot + * cpu is on a node that doesn't have memory installed. + * Also cpu_up() will call cpu_to_node() for APs when + * MEMORY_HOTPLUG is defined, before per_cpu(numa_node) is set + * up later with c_init aka intel_init/amd_init. + * So set them all (boot cpu and all APs). + */ + set_cpu_numa_node(cpu, early_cpu_to_node(cpu)); #endif #endif /* @@ -257,14 +266,6 @@ void __init setup_per_cpu_areas(void) early_per_cpu_ptr(x86_cpu_to_node_map) = NULL; #endif -#if defined(CONFIG_X86_64) && defined(CONFIG_NUMA) - /* - * make sure boot cpu numa_node is right, when boot cpu is on the - * node that doesn't have mem installed - */ - set_cpu_numa_node(boot_cpu_id, early_cpu_to_node(boot_cpu_id)); -#endif - /* Setup node to cpumask map */ setup_node_to_cpumask_map(); -- cgit v1.2.3 From a4ce96ac356e7024a7724ade9d18ba1bdf3c5c06 Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Wed, 21 Jul 2010 09:25:42 -0700 Subject: Fix up trivial spelling errors ('taht' -> 'that') Pointed out by Lucas who found the new one in a comment in setup_percpu.c. And then I fixed the others that I grepped for. Reported-by: Lucas Signed-off-by: Linus Torvalds --- arch/x86/kernel/setup_percpu.c | 2 +- drivers/usb/gadget/f_fs.c | 2 +- drivers/video/aty/radeon_pm.c | 2 +- fs/jffs2/xattr.c | 2 +- 4 files changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/setup_percpu.c b/arch/x86/kernel/setup_percpu.c index 690c2c09faf3..a60df9ae6454 100644 --- a/arch/x86/kernel/setup_percpu.c +++ b/arch/x86/kernel/setup_percpu.c @@ -239,7 +239,7 @@ void __init setup_per_cpu_areas(void) per_cpu(x86_cpu_to_node_map, cpu) = early_per_cpu_map(x86_cpu_to_node_map, cpu); /* - * Ensure taht the boot cpu numa_node is correct when the boot + * Ensure that the boot cpu numa_node is correct when the boot * cpu is on a node that doesn't have memory installed. * Also cpu_up() will call cpu_to_node() for APs when * MEMORY_HOTPLUG is defined, before per_cpu(numa_node) is set diff --git a/drivers/usb/gadget/f_fs.c b/drivers/usb/gadget/f_fs.c index d69eccf5f197..2aaa0f75c6cf 100644 --- a/drivers/usb/gadget/f_fs.c +++ b/drivers/usb/gadget/f_fs.c @@ -136,7 +136,7 @@ struct ffs_data { * handling setup requests immidiatelly user space may be so * slow that another setup will be sent to the gadget but this * time not to us but another function and then there could be - * a race. Is taht the case? Or maybe we can use cdev->req + * a race. Is that the case? Or maybe we can use cdev->req * after all, maybe we just need some spinlock for that? */ struct usb_request *ep0req; /* P: mutex */ struct completion ep0req_completion; /* P: mutex */ diff --git a/drivers/video/aty/radeon_pm.c b/drivers/video/aty/radeon_pm.c index 515cf1978d19..c4e17642d9c5 100644 --- a/drivers/video/aty/radeon_pm.c +++ b/drivers/video/aty/radeon_pm.c @@ -2872,7 +2872,7 @@ void radeonfb_pm_init(struct radeonfb_info *rinfo, int dynclk, int ignore_devlis } #if 0 - /* Power down TV DAC, taht saves a significant amount of power, + /* Power down TV DAC, that saves a significant amount of power, * we'll have something better once we actually have some TVOut * support */ diff --git a/fs/jffs2/xattr.c b/fs/jffs2/xattr.c index a2d58c96f1b4..d258e261bdc7 100644 --- a/fs/jffs2/xattr.c +++ b/fs/jffs2/xattr.c @@ -626,7 +626,7 @@ void jffs2_xattr_free_inode(struct jffs2_sb_info *c, struct jffs2_inode_cache *i static int check_xattr_ref_inode(struct jffs2_sb_info *c, struct jffs2_inode_cache *ic) { - /* success of check_xattr_ref_inode() means taht inode (ic) dose not have + /* success of check_xattr_ref_inode() means that inode (ic) dose not have * duplicate name/value pairs. If duplicate name/value pair would be found, * one will be removed. */ -- cgit v1.2.3 From 0327559151c6886814d6d5b373b4bf6de63fb9f6 Mon Sep 17 00:00:00 2001 From: Roland McGrath Date: Wed, 21 Jul 2010 17:44:12 -0700 Subject: x86: auditsyscall: fix fastpath return value after reschedule In the CONFIG_AUDITSYSCALL fast-path for x86 64-bit system calls, we can pass a bad return value and/or error indication for the system call to audit_syscall_exit(). This happens when TIF_NEED_RESCHED was set as the system call returned, so we went out to schedule() and came back to the exit-audit fast-path. The fix is to reload the user return value register from the pt_regs before using it for audit_syscall_exit(). Both the 32-bit kernel's fast path and the 64-bit kernel's 32-bit system call fast paths work slightly differently, so that they always leave the fast path entirely to reschedule and don't return there, so they don't have the analogous bugs. Reported-by: Alexander Viro Signed-off-by: Roland McGrath --- arch/x86/kernel/entry_64.S | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/entry_64.S b/arch/x86/kernel/entry_64.S index 0697ff139837..4db7c4d12ffa 100644 --- a/arch/x86/kernel/entry_64.S +++ b/arch/x86/kernel/entry_64.S @@ -571,8 +571,8 @@ auditsys: * masked off. */ sysret_audit: - movq %rax,%rsi /* second arg, syscall return value */ - cmpq $0,%rax /* is it < 0? */ + movq RAX-ARGOFFSET(%rsp),%rsi /* second arg, syscall return value */ + cmpq $0,%rsi /* is it < 0? */ setl %al /* 1 if so, 0 if not */ movzbl %al,%edi /* zero-extend that into %edi */ inc %edi /* first arg, 0->1(AUDITSC_SUCCESS), 1->2(AUDITSC_FAILURE) */ -- cgit v1.2.3 From 718be4aaf3613cf7c2d097f925abc3d3553c0605 Mon Sep 17 00:00:00 2001 From: Len Brown Date: Thu, 22 Jul 2010 16:54:27 -0400 Subject: ACPI: skip checking BM_STS if the BIOS doesn't ask for it It turns out that there is a bit in the _CST for Intel FFH C3 that tells the OS if we should be checking BM_STS or not. Linux has been unconditionally checking BM_STS. If the chip-set is configured to enable BM_STS, it can retard or completely prevent entry into deep C-states -- as illustrated by turbostat: http://userweb.kernel.org/~lenb/acpi/utils/pmtools/turbostat/ ref: Intel Processor Vendor-Specific ACPI Interface Specification table 4 "_CST FFH GAS Field Encoding" Bit 1: Set to 1 if OSPM should use Bus Master avoidance for this C-state https://bugzilla.kernel.org/show_bug.cgi?id=15886 Signed-off-by: Len Brown --- arch/x86/kernel/acpi/cstate.c | 9 +++++++++ drivers/acpi/processor_idle.c | 2 +- include/acpi/processor.h | 3 ++- 3 files changed, 12 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/acpi/cstate.c b/arch/x86/kernel/acpi/cstate.c index 2e837f5080fe..fb7a5f052e2b 100644 --- a/arch/x86/kernel/acpi/cstate.c +++ b/arch/x86/kernel/acpi/cstate.c @@ -145,6 +145,15 @@ int acpi_processor_ffh_cstate_probe(unsigned int cpu, percpu_entry->states[cx->index].eax = cx->address; percpu_entry->states[cx->index].ecx = MWAIT_ECX_INTERRUPT_BREAK; } + + /* + * For _CST FFH on Intel, if GAS.access_size bit 1 is cleared, + * then we should skip checking BM_STS for this C-state. + * ref: "Intel Processor Vendor-Specific ACPI Interface Specification" + */ + if ((c->x86_vendor == X86_VENDOR_INTEL) && !(reg->access_size & 0x2)) + cx->bm_sts_skip = 1; + return retval; } EXPORT_SYMBOL_GPL(acpi_processor_ffh_cstate_probe); diff --git a/drivers/acpi/processor_idle.c b/drivers/acpi/processor_idle.c index b1b385692f46..b351342f1faf 100644 --- a/drivers/acpi/processor_idle.c +++ b/drivers/acpi/processor_idle.c @@ -947,7 +947,7 @@ static int acpi_idle_enter_bm(struct cpuidle_device *dev, if (acpi_idle_suspend) return(acpi_idle_enter_c1(dev, state)); - if (acpi_idle_bm_check()) { + if (!cx->bm_sts_skip && acpi_idle_bm_check()) { if (dev->safe_state) { dev->last_state = dev->safe_state; return dev->safe_state->enter(dev, dev->safe_state); diff --git a/include/acpi/processor.h b/include/acpi/processor.h index da565a48240e..a68ca8a11a53 100644 --- a/include/acpi/processor.h +++ b/include/acpi/processor.h @@ -48,7 +48,7 @@ struct acpi_power_register { u8 space_id; u8 bit_width; u8 bit_offset; - u8 reserved; + u8 access_size; u64 address; } __attribute__ ((packed)); @@ -63,6 +63,7 @@ struct acpi_processor_cx { u32 power; u32 usage; u64 time; + u8 bm_sts_skip; char desc[ACPI_CX_DESC_LEN]; }; -- cgit v1.2.3 From ff4878089e1eaeac79d57878ad4ea32910fb4037 Mon Sep 17 00:00:00 2001 From: Stefano Stabellini Date: Wed, 21 Jul 2010 18:32:37 +0100 Subject: x86: Do not try to disable hpet if it hasn't been initialized before hpet_disable is called unconditionally on machine reboot if hpet support is compiled in the kernel. hpet_disable only checks if the machine is hpet capable but doesn't make sure that hpet has been initialized. [ tglx: Made it a one liner and removed the redundant hpet_address check ] Signed-off-by: Stefano Stabellini Acked-by: Venkatesh Pallipadi LKML-Reference: Cc: stable@kernel.org Signed-off-by: Thomas Gleixner --- arch/x86/kernel/hpet.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c index a198b7c87a12..ba390d731175 100644 --- a/arch/x86/kernel/hpet.c +++ b/arch/x86/kernel/hpet.c @@ -964,7 +964,7 @@ fs_initcall(hpet_late_init); void hpet_disable(void) { - if (is_hpet_capable()) { + if (is_hpet_capable() && hpet_virt_address) { unsigned int cfg = hpet_readl(HPET_CFG); if (hpet_legacy_int_enabled) { -- cgit v1.2.3 From 72ad5d77fb981963edae15eee8196c80238f5ed0 Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Fri, 23 Jul 2010 22:59:09 +0200 Subject: ACPI / Sleep: Allow the NVS saving to be skipped during suspend to RAM Commit 2a6b69765ad794389f2fc3e14a0afa1a995221c2 (ACPI: Store NVS state even when entering suspend to RAM) caused the ACPI suspend code save the NVS area during suspend and restore it during resume unconditionally, although it is known that some systems need to use acpi_sleep=s4_nonvs for hibernation to work. To allow the affected systems to avoid saving and restoring the NVS area during suspend to RAM and resume, introduce kernel command line option acpi_sleep=nonvs and make acpi_sleep=s4_nonvs work as its alias temporarily (add acpi_sleep=s4_nonvs to the feature removal file). Addresses https://bugzilla.kernel.org/show_bug.cgi?id=16396 . Signed-off-by: Rafael J. Wysocki Reported-and-tested-by: tomas m Signed-off-by: Len Brown --- Documentation/feature-removal-schedule.txt | 7 ++++++ Documentation/kernel-parameters.txt | 4 ++-- arch/x86/kernel/acpi/sleep.c | 9 ++++++-- drivers/acpi/sleep.c | 35 +++++++++++++++--------------- include/linux/acpi.h | 2 +- 5 files changed, 34 insertions(+), 23 deletions(-) (limited to 'arch/x86/kernel') diff --git a/Documentation/feature-removal-schedule.txt b/Documentation/feature-removal-schedule.txt index c268783bc4e7..1571c0c83dba 100644 --- a/Documentation/feature-removal-schedule.txt +++ b/Documentation/feature-removal-schedule.txt @@ -647,3 +647,10 @@ Who: Stefan Richter ---------------------------- +What: The acpi_sleep=s4_nonvs command line option +When: 2.6.37 +Files: arch/x86/kernel/acpi/sleep.c +Why: superseded by acpi_sleep=nonvs +Who: Rafael J. Wysocki + +---------------------------- diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index 4ddb58df081e..2b2407d9a6d0 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -254,8 +254,8 @@ and is between 256 and 4096 characters. It is defined in the file control method, with respect to putting devices into low power states, to be enforced (the ACPI 2.0 ordering of _PTS is used by default). - s4_nonvs prevents the kernel from saving/restoring the - ACPI NVS memory during hibernation. + nonvs prevents the kernel from saving/restoring the + ACPI NVS memory during suspend/hibernation and resume. sci_force_enable causes the kernel to set SCI_EN directly on resume from S1/S3 (which is against the ACPI spec, but some broken systems don't work without it). diff --git a/arch/x86/kernel/acpi/sleep.c b/arch/x86/kernel/acpi/sleep.c index 82e508677b91..fcc3c61fdecc 100644 --- a/arch/x86/kernel/acpi/sleep.c +++ b/arch/x86/kernel/acpi/sleep.c @@ -157,9 +157,14 @@ static int __init acpi_sleep_setup(char *str) #ifdef CONFIG_HIBERNATION if (strncmp(str, "s4_nohwsig", 10) == 0) acpi_no_s4_hw_signature(); - if (strncmp(str, "s4_nonvs", 8) == 0) - acpi_s4_no_nvs(); + if (strncmp(str, "s4_nonvs", 8) == 0) { + pr_warning("ACPI: acpi_sleep=s4_nonvs is deprecated, " + "please use acpi_sleep=nonvs instead"); + acpi_nvs_nosave(); + } #endif + if (strncmp(str, "nonvs", 5) == 0) + acpi_nvs_nosave(); if (strncmp(str, "old_ordering", 12) == 0) acpi_old_suspend_ordering(); str = strchr(str, ','); diff --git a/drivers/acpi/sleep.c b/drivers/acpi/sleep.c index 5b7c52e4a00f..2862c781b372 100644 --- a/drivers/acpi/sleep.c +++ b/drivers/acpi/sleep.c @@ -81,6 +81,20 @@ static int acpi_sleep_prepare(u32 acpi_state) #ifdef CONFIG_ACPI_SLEEP static u32 acpi_target_sleep_state = ACPI_STATE_S0; +/* + * The ACPI specification wants us to save NVS memory regions during hibernation + * and to restore them during the subsequent resume. Windows does that also for + * suspend to RAM. However, it is known that this mechanism does not work on + * all machines, so we allow the user to disable it with the help of the + * 'acpi_sleep=nonvs' kernel command line option. + */ +static bool nvs_nosave; + +void __init acpi_nvs_nosave(void) +{ + nvs_nosave = true; +} + /* * ACPI 1.0 wants us to execute _PTS before suspending devices, so we allow the * user to request that behavior by using the 'acpi_old_suspend_ordering' @@ -197,8 +211,7 @@ static int acpi_suspend_begin(suspend_state_t pm_state) u32 acpi_state = acpi_suspend_states[pm_state]; int error = 0; - error = suspend_nvs_alloc(); - + error = nvs_nosave ? 0 : suspend_nvs_alloc(); if (error) return error; @@ -388,20 +401,6 @@ static struct dmi_system_id __initdata acpisleep_dmi_table[] = { #endif /* CONFIG_SUSPEND */ #ifdef CONFIG_HIBERNATION -/* - * The ACPI specification wants us to save NVS memory regions during hibernation - * and to restore them during the subsequent resume. However, it is not certain - * if this mechanism is going to work on all machines, so we allow the user to - * disable this mechanism using the 'acpi_sleep=s4_nonvs' kernel command line - * option. - */ -static bool s4_no_nvs; - -void __init acpi_s4_no_nvs(void) -{ - s4_no_nvs = true; -} - static unsigned long s4_hardware_signature; static struct acpi_table_facs *facs; static bool nosigcheck; @@ -415,7 +414,7 @@ static int acpi_hibernation_begin(void) { int error; - error = s4_no_nvs ? 0 : suspend_nvs_alloc(); + error = nvs_nosave ? 0 : suspend_nvs_alloc(); if (!error) { acpi_target_sleep_state = ACPI_STATE_S4; acpi_sleep_tts_switch(acpi_target_sleep_state); @@ -510,7 +509,7 @@ static int acpi_hibernation_begin_old(void) error = acpi_sleep_prepare(ACPI_STATE_S4); if (!error) { - if (!s4_no_nvs) + if (!nvs_nosave) error = suspend_nvs_alloc(); if (!error) acpi_target_sleep_state = ACPI_STATE_S4; diff --git a/include/linux/acpi.h b/include/linux/acpi.h index 224a38c960d4..ccf94dc5acdf 100644 --- a/include/linux/acpi.h +++ b/include/linux/acpi.h @@ -253,7 +253,7 @@ int acpi_resources_are_enforced(void); #ifdef CONFIG_PM_SLEEP void __init acpi_no_s4_hw_signature(void); void __init acpi_old_suspend_ordering(void); -void __init acpi_s4_no_nvs(void); +void __init acpi_nvs_nosave(void); #endif /* CONFIG_PM_SLEEP */ struct acpi_osc_context { -- cgit v1.2.3 From 47f8bcf362410b631a4d99ff5c79ec6b9dd3ace6 Mon Sep 17 00:00:00 2001 From: Matthew Garrett Date: Tue, 20 Jul 2010 13:52:00 -0400 Subject: [CPUFREQ] pcc driver should check for pcch method before calling _OSC The pcc specification documents an _OSC method that's incompatible with the one defined as part of the ACPI spec. This shouldn't be a problem as both are supposed to be guarded with a UUID. Unfortunately approximately nobody (including HP, who wrote this spec) properly check the UUID on entry to the _OSC call. Right now this could result in surprising behaviour if the pcc driver performs an _OSC call on a machine that doesn't implement the pcc specification. Check whether the PCCH method exists first in order to reduce this probability. Signed-off-by: Matthew Garrett Cc: Naga Chumbalkar Signed-off-by: Dave Jones --- arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c | 6 +++++- 1 file changed, 5 insertions(+), 1 deletion(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c index ce7cde713e71..01bd25c3c7ca 100644 --- a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c @@ -397,13 +397,17 @@ static int __init pcc_cpufreq_probe(void) struct pcc_memory_resource *mem_resource; struct pcc_register_resource *reg_resource; union acpi_object *out_obj, *member; - acpi_handle handle, osc_handle; + acpi_handle handle, osc_handle, pcch_handle; int ret = 0; status = acpi_get_handle(NULL, "\\_SB", &handle); if (ACPI_FAILURE(status)) return -ENODEV; + status = acpi_get_handle(handle, "PCCH", &pcch_handle); + if (ACPI_FAILURE(status)) + return -ENODEV; + status = acpi_get_handle(handle, "_OSC", &osc_handle); if (ACPI_SUCCESS(status)) { ret = pcc_cpufreq_do_osc(&osc_handle); -- cgit v1.2.3 From 3847d223f2e4da5ceb47ea8996618010192f3197 Mon Sep 17 00:00:00 2001 From: Daniel J Blueman Date: Fri, 23 Jul 2010 23:06:52 +0100 Subject: [CPUFREQ] fix double freeing in error path of pcc-cpufreq Prevent double freeing on error path. Signed-off-by: Daniel J Blueman Signed-off-by: Dave Jones --- arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c | 18 ++++++------------ 1 file changed, 6 insertions(+), 12 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c index 01bd25c3c7ca..900702888bfb 100644 --- a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c @@ -368,22 +368,16 @@ static int __init pcc_cpufreq_do_osc(acpi_handle *handle) return -ENODEV; out_obj = output.pointer; - if (out_obj->type != ACPI_TYPE_BUFFER) { - ret = -ENODEV; - goto out_free; - } + if (out_obj->type != ACPI_TYPE_BUFFER) + return -ENODEV; errors = *((u32 *)out_obj->buffer.pointer) & ~(1 << 0); - if (errors) { - ret = -ENODEV; - goto out_free; - } + if (errors) + return -ENODEV; supported = *((u32 *)(out_obj->buffer.pointer + 4)); - if (!(supported & 0x1)) { - ret = -ENODEV; - goto out_free; - } + if (!(supported & 0x1)) + return -ENODEV; out_free: kfree(output.pointer); -- cgit v1.2.3 From 179ee43465343d1f8f2a4af25ead4ae15e43fa6e Mon Sep 17 00:00:00 2001 From: Matthew Garrett Date: Thu, 15 Jul 2010 11:44:00 -0400 Subject: [CPUFREQ] Fix PCC driver error path The PCC cpufreq driver unmaps the mailbox address range if any CPUs fail to initialise, but doesn't do anything to remove the registered CPUs from the cpufreq core resulting in failures further down the line. We're better off simply returning a failure - the cpufreq core will unregister us cleanly if we end up with no successfully registered CPUs. Tidy up the failure path and also add a sanity check to ensure that the firmware gives us a realistic frequency - the core deals badly with that being set to 0. Signed-off-by: Matthew Garrett Cc: Naga Chumbalkar Signed-off-by: Dave Jones --- arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c | 17 +++++++++-------- 1 file changed, 9 insertions(+), 8 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c index 900702888bfb..a36de5bbb622 100644 --- a/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c +++ b/arch/x86/kernel/cpu/cpufreq/pcc-cpufreq.c @@ -541,13 +541,13 @@ static int pcc_cpufreq_cpu_init(struct cpufreq_policy *policy) if (!pcch_virt_addr) { result = -1; - goto pcch_null; + goto out; } result = pcc_get_offset(cpu); if (result) { dprintk("init: PCCP evaluation failed\n"); - goto free; + goto out; } policy->max = policy->cpuinfo.max_freq = @@ -556,14 +556,15 @@ static int pcc_cpufreq_cpu_init(struct cpufreq_policy *policy) ioread32(&pcch_hdr->minimum_frequency) * 1000; policy->cur = pcc_get_freq(cpu); + if (!policy->cur) { + dprintk("init: Unable to get current CPU frequency\n"); + result = -EINVAL; + goto out; + } + dprintk("init: policy->max is %d, policy->min is %d\n", policy->max, policy->min); - - return 0; -free: - pcc_clear_mapping(); - free_percpu(pcc_cpu_info); -pcch_null: +out: return result; } -- cgit v1.2.3 From 3581ced3b6ac289b5cd31663b34914a7347186a6 Mon Sep 17 00:00:00 2001 From: Borislav Petkov Date: Thu, 8 Jul 2010 17:55:30 +0200 Subject: [CPUFREQ] powernow-k8: Limit Pstate transition latency check The Pstate transition latency check was added for broken F10h BIOSen which wrongly contain a value of 0 for transition and bus master latency. Fam11h and later, however, (will) have similar transition latency so extend that behavior for them too. Signed-off-by: Borislav Petkov Signed-off-by: Dave Jones --- arch/x86/kernel/cpu/cpufreq/powernow-k8.c | 11 +++++------ 1 file changed, 5 insertions(+), 6 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c index 7ec2123838e6..3e90cce3dc8b 100644 --- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c +++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c @@ -1023,13 +1023,12 @@ static int get_transition_latency(struct powernow_k8_data *data) } if (max_latency == 0) { /* - * Fam 11h always returns 0 as transition latency. - * This is intended and means "very fast". While cpufreq core - * and governors currently can handle that gracefully, better - * set it to 1 to avoid problems in the future. - * For all others it's a BIOS bug. + * Fam 11h and later may return 0 as transition latency. This + * is intended and means "very fast". While cpufreq core and + * governors currently can handle that gracefully, better set it + * to 1 to avoid problems in the future. */ - if (boot_cpu_data.x86 != 0x11) + if (boot_cpu_data.x86 < 0x11) printk(KERN_ERR FW_WARN PFX "Invalid zero transition " "latency\n"); max_latency = 1; -- cgit v1.2.3 From ba773f7c510c0b252145933926c636c439889207 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Wed, 28 Jul 2010 19:10:30 -0500 Subject: x86,kgdb: Fix hw breakpoint regression HW breakpoints events stopped working correctly with kgdb as a result of commit: 018cbffe6819f6f8db20a0a3acd9bab9bfd667e4 (Merge commit 'v2.6.33' into perf/core). The regression occurred because the behavior changed for setting NOTIFY_STOP as the return value to the die notifier if the breakpoint was known to the HW breakpoint API. Because kgdb is using the HW breakpoint API to register HW breakpoints slots, it must also now implement the overflow_handler call back else kgdb does not get to see the events from the die notifier. The kgdb_ll_trap function will be changed to be general purpose code which can allow an easy way to implement the hw_breakpoint API overflow call back. Signed-off-by: Jason Wessel Acked-by: Dongdong Deng Acked-by: Frederic Weisbecker --- arch/x86/kernel/kgdb.c | 9 +++++++-- 1 file changed, 7 insertions(+), 2 deletions(-) (limited to 'arch/x86/kernel') diff --git a/arch/x86/kernel/kgdb.c b/arch/x86/kernel/kgdb.c index 4f4af75b9482..01ab17ae2ae7 100644 --- a/arch/x86/kernel/kgdb.c +++ b/arch/x86/kernel/kgdb.c @@ -572,7 +572,6 @@ static int __kgdb_notify(struct die_args *args, unsigned long cmd) return NOTIFY_STOP; } -#ifdef CONFIG_KGDB_LOW_LEVEL_TRAP int kgdb_ll_trap(int cmd, const char *str, struct pt_regs *regs, long err, int trap, int sig) { @@ -590,7 +589,6 @@ int kgdb_ll_trap(int cmd, const char *str, return __kgdb_notify(&args, cmd); } -#endif /* CONFIG_KGDB_LOW_LEVEL_TRAP */ static int kgdb_notify(struct notifier_block *self, unsigned long cmd, void *ptr) @@ -625,6 +623,12 @@ int kgdb_arch_init(void) return register_die_notifier(&kgdb_notifier); } +static void kgdb_hw_overflow_handler(struct perf_event *event, int nmi, + struct perf_sample_data *data, struct pt_regs *regs) +{ + kgdb_ll_trap(DIE_DEBUG, "debug", regs, 0, 0, SIGTRAP); +} + void kgdb_arch_late(void) { int i, cpu; @@ -655,6 +659,7 @@ void kgdb_arch_late(void) for_each_online_cpu(cpu) { pevent = per_cpu_ptr(breakinfo[i].pev, cpu); pevent[0]->hw.sample_period = 1; + pevent[0]->overflow_handler = kgdb_hw_overflow_handler; if (pevent[0]->destroy != NULL) { pevent[0]->destroy = NULL; release_bp_slot(*pevent); -- cgit v1.2.3