From 0b5f9c005def154f9c21f9be0223b65b50d54368 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Thu, 29 Mar 2012 15:38:30 +1030 Subject: remove references to cpu_*_map in arch/ This has been obsolescent for a while; time for the final push. In adjacent context, replaced old cpus_* with cpumask_*. Signed-off-by: Rusty Russell Acked-by: David S. Miller (arch/sparc) Acked-by: Chris Metcalf (arch/tile) Cc: user-mode-linux-devel@lists.sourceforge.net Cc: Russell King Cc: linux-arm-kernel@lists.infradead.org Cc: Richard Kuo Cc: linux-hexagon@vger.kernel.org Cc: Ralf Baechle Cc: linux-mips@linux-mips.org Cc: Kyle McMartin Cc: Helge Deller Cc: sparclinux@vger.kernel.org --- arch/tile/kernel/setup.c | 6 +++--- 1 file changed, 3 insertions(+), 3 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/setup.c b/arch/tile/kernel/setup.c index 5f85d8b34db..5124093b2e1 100644 --- a/arch/tile/kernel/setup.c +++ b/arch/tile/kernel/setup.c @@ -1186,7 +1186,7 @@ static void __init setup_cpu_maps(void) sizeof(cpu_lotar_map)); if (rc < 0) { pr_err("warning: no HV_INQ_TILES_LOTAR; using AVAIL\n"); - cpu_lotar_map = cpu_possible_map; + cpu_lotar_map = *cpu_possible_mask; } #if CHIP_HAS_CBOX_HOME_MAP() @@ -1196,9 +1196,9 @@ static void __init setup_cpu_maps(void) sizeof(hash_for_home_map)); if (rc < 0) early_panic("hv_inquire_tiles(HFH_CACHE) failed: rc %d\n", rc); - cpumask_or(&cpu_cacheable_map, &cpu_possible_map, &hash_for_home_map); + cpumask_or(&cpu_cacheable_map, cpu_possible_mask, &hash_for_home_map); #else - cpu_cacheable_map = cpu_possible_map; + cpu_cacheable_map = *cpu_possible_mask; #endif } -- cgit v1.2.3-70-g09d2 From 5f054e31c63be774bf1ce252f20d56012a00f8a5 Mon Sep 17 00:00:00 2001 From: Rusty Russell Date: Thu, 29 Mar 2012 15:38:31 +1030 Subject: documentation: remove references to cpu_*_map. This has been obsolescent for a while, fix documentation and misc comments. Signed-off-by: Rusty Russell --- Documentation/cgroups/cpusets.txt | 2 +- Documentation/cpu-hotplug.txt | 22 +++++++++++----------- arch/alpha/kernel/smp.c | 2 +- arch/ia64/kernel/acpi.c | 2 +- arch/mips/cavium-octeon/smp.c | 2 +- arch/mips/pmc-sierra/yosemite/smp.c | 2 +- arch/mips/sibyte/bcm1480/smp.c | 2 +- arch/tile/kernel/setup.c | 2 +- arch/x86/xen/enlighten.c | 2 +- init/Kconfig | 4 ++-- kernel/cpuset.c | 10 +++++----- 11 files changed, 26 insertions(+), 26 deletions(-) (limited to 'arch/tile/kernel') diff --git a/Documentation/cgroups/cpusets.txt b/Documentation/cgroups/cpusets.txt index 5c51ed406d1..cefd3d8bbd1 100644 --- a/Documentation/cgroups/cpusets.txt +++ b/Documentation/cgroups/cpusets.txt @@ -217,7 +217,7 @@ and name space for cpusets, with a minimum of additional kernel code. The cpus and mems files in the root (top_cpuset) cpuset are read-only. The cpus file automatically tracks the value of -cpu_online_map using a CPU hotplug notifier, and the mems file +cpu_online_mask using a CPU hotplug notifier, and the mems file automatically tracks the value of node_states[N_HIGH_MEMORY]--i.e., nodes with memory--using the cpuset_track_online_nodes() hook. diff --git a/Documentation/cpu-hotplug.txt b/Documentation/cpu-hotplug.txt index a20bfd415e4..66ef8f35613 100644 --- a/Documentation/cpu-hotplug.txt +++ b/Documentation/cpu-hotplug.txt @@ -47,7 +47,7 @@ maxcpus=n Restrict boot time cpus to n. Say if you have 4 cpus, using other cpus later online, read FAQ's for more info. additional_cpus=n (*) Use this to limit hotpluggable cpus. This option sets - cpu_possible_map = cpu_present_map + additional_cpus + cpu_possible_mask = cpu_present_mask + additional_cpus cede_offline={"off","on"} Use this option to disable/enable putting offlined processors to an extended H_CEDE state on @@ -64,11 +64,11 @@ should only rely on this to count the # of cpus, but *MUST* not rely on the apicid values in those tables for disabled apics. In the event BIOS doesn't mark such hot-pluggable cpus as disabled entries, one could use this parameter "additional_cpus=x" to represent those cpus in the -cpu_possible_map. +cpu_possible_mask. possible_cpus=n [s390,x86_64] use this to set hotpluggable cpus. This option sets possible_cpus bits in - cpu_possible_map. Thus keeping the numbers of bits set + cpu_possible_mask. Thus keeping the numbers of bits set constant even if the machine gets rebooted. CPU maps and such @@ -76,7 +76,7 @@ CPU maps and such [More on cpumaps and primitive to manipulate, please check include/linux/cpumask.h that has more descriptive text.] -cpu_possible_map: Bitmap of possible CPUs that can ever be available in the +cpu_possible_mask: Bitmap of possible CPUs that can ever be available in the system. This is used to allocate some boot time memory for per_cpu variables that aren't designed to grow/shrink as CPUs are made available or removed. Once set during boot time discovery phase, the map is static, i.e no bits @@ -84,13 +84,13 @@ are added or removed anytime. Trimming it accurately for your system needs upfront can save some boot time memory. See below for how we use heuristics in x86_64 case to keep this under check. -cpu_online_map: Bitmap of all CPUs currently online. Its set in __cpu_up() +cpu_online_mask: Bitmap of all CPUs currently online. Its set in __cpu_up() after a cpu is available for kernel scheduling and ready to receive interrupts from devices. Its cleared when a cpu is brought down using __cpu_disable(), before which all OS services including interrupts are migrated to another target CPU. -cpu_present_map: Bitmap of CPUs currently present in the system. Not all +cpu_present_mask: Bitmap of CPUs currently present in the system. Not all of them may be online. When physical hotplug is processed by the relevant subsystem (e.g ACPI) can change and new bit either be added or removed from the map depending on the event is hot-add/hot-remove. There are currently @@ -99,22 +99,22 @@ at which time hotplug is disabled. You really dont need to manipulate any of the system cpu maps. They should be read-only for most use. When setting up per-cpu resources almost always use -cpu_possible_map/for_each_possible_cpu() to iterate. +cpu_possible_mask/for_each_possible_cpu() to iterate. Never use anything other than cpumask_t to represent bitmap of CPUs. #include - for_each_possible_cpu - Iterate over cpu_possible_map - for_each_online_cpu - Iterate over cpu_online_map - for_each_present_cpu - Iterate over cpu_present_map + for_each_possible_cpu - Iterate over cpu_possible_mask + for_each_online_cpu - Iterate over cpu_online_mask + for_each_present_cpu - Iterate over cpu_present_mask for_each_cpu_mask(x,mask) - Iterate over some random collection of cpu mask. #include get_online_cpus() and put_online_cpus(): The above calls are used to inhibit cpu hotplug operations. While the -cpu_hotplug.refcount is non zero, the cpu_online_map will not change. +cpu_hotplug.refcount is non zero, the cpu_online_mask will not change. If you merely need to avoid cpus going away, you could also use preempt_disable() and preempt_enable() for those sections. Just remember the critical section cannot call any diff --git a/arch/alpha/kernel/smp.c b/arch/alpha/kernel/smp.c index 4087a569b43..50d438db1f6 100644 --- a/arch/alpha/kernel/smp.c +++ b/arch/alpha/kernel/smp.c @@ -450,7 +450,7 @@ setup_smp(void) smp_num_probed = 1; } - printk(KERN_INFO "SMP: %d CPUs probed -- cpu_present_map = %lx\n", + printk(KERN_INFO "SMP: %d CPUs probed -- cpu_present_mask = %lx\n", smp_num_probed, cpumask_bits(cpu_present_mask)[0]); } diff --git a/arch/ia64/kernel/acpi.c b/arch/ia64/kernel/acpi.c index ac795d311f4..6f38b6120d9 100644 --- a/arch/ia64/kernel/acpi.c +++ b/arch/ia64/kernel/acpi.c @@ -839,7 +839,7 @@ static __init int setup_additional_cpus(char *s) early_param("additional_cpus", setup_additional_cpus); /* - * cpu_possible_map should be static, it cannot change as CPUs + * cpu_possible_mask should be static, it cannot change as CPUs * are onlined, or offlined. The reason is per-cpu data-structures * are allocated by some modules at init time, and dont expect to * do this dynamically on cpu arrival/departure. diff --git a/arch/mips/cavium-octeon/smp.c b/arch/mips/cavium-octeon/smp.c index ef56d8a0215..97e7ce9b50e 100644 --- a/arch/mips/cavium-octeon/smp.c +++ b/arch/mips/cavium-octeon/smp.c @@ -78,7 +78,7 @@ static inline void octeon_send_ipi_mask(const struct cpumask *mask, } /** - * Detect available CPUs, populate cpu_possible_map + * Detect available CPUs, populate cpu_possible_mask */ static void octeon_smp_hotplug_setup(void) { diff --git a/arch/mips/pmc-sierra/yosemite/smp.c b/arch/mips/pmc-sierra/yosemite/smp.c index 00a0d2e90d0..b71fae23104 100644 --- a/arch/mips/pmc-sierra/yosemite/smp.c +++ b/arch/mips/pmc-sierra/yosemite/smp.c @@ -146,7 +146,7 @@ static void __cpuinit yos_boot_secondary(int cpu, struct task_struct *idle) } /* - * Detect available CPUs, populate cpu_possible_map before smp_init + * Detect available CPUs, populate cpu_possible_mask before smp_init * * We don't want to start the secondary CPU yet nor do we have a nice probing * feature in PMON so we just assume presence of the secondary core. diff --git a/arch/mips/sibyte/bcm1480/smp.c b/arch/mips/sibyte/bcm1480/smp.c index 63d2211e616..de88e22694a 100644 --- a/arch/mips/sibyte/bcm1480/smp.c +++ b/arch/mips/sibyte/bcm1480/smp.c @@ -138,7 +138,7 @@ static void __cpuinit bcm1480_boot_secondary(int cpu, struct task_struct *idle) /* * Use CFE to find out how many CPUs are available, setting up - * cpu_possible_map and the logical/physical mappings. + * cpu_possible_mask and the logical/physical mappings. * XXXKW will the boot CPU ever not be physical 0? * * Common setup before any secondaries are started diff --git a/arch/tile/kernel/setup.c b/arch/tile/kernel/setup.c index 5124093b2e1..92a94f4920a 100644 --- a/arch/tile/kernel/setup.c +++ b/arch/tile/kernel/setup.c @@ -1100,7 +1100,7 @@ EXPORT_SYMBOL(hash_for_home_map); /* * cpu_cacheable_map lists all the cpus whose caches the hypervisor can - * flush on our behalf. It is set to cpu_possible_map OR'ed with + * flush on our behalf. It is set to cpu_possible_mask OR'ed with * hash_for_home_map, and it is what should be passed to * hv_flush_remote() to flush all caches. Note that if there are * dedicated hypervisor driver tiles that have authorized use of their diff --git a/arch/x86/xen/enlighten.c b/arch/x86/xen/enlighten.c index b132ade26f7..4f51bebac02 100644 --- a/arch/x86/xen/enlighten.c +++ b/arch/x86/xen/enlighten.c @@ -967,7 +967,7 @@ void xen_setup_shared_info(void) xen_setup_mfn_list_list(); } -/* This is called once we have the cpu_possible_map */ +/* This is called once we have the cpu_possible_mask */ void xen_setup_vcpu_info_placement(void) { int cpu; diff --git a/init/Kconfig b/init/Kconfig index 72f33faca44..6cfd71d0646 100644 --- a/init/Kconfig +++ b/init/Kconfig @@ -1414,8 +1414,8 @@ endif # MODULES config INIT_ALL_POSSIBLE bool help - Back when each arch used to define their own cpu_online_map and - cpu_possible_map, some of them chose to initialize cpu_possible_map + Back when each arch used to define their own cpu_online_mask and + cpu_possible_mask, some of them chose to initialize cpu_possible_mask with all 1s, and others with all 0s. When they were centralised, it was better to provide this option than to break all the archs and have several arch maintainers pursuing me down dark alleys. diff --git a/kernel/cpuset.c b/kernel/cpuset.c index 1010cc61931..eedeebe64b1 100644 --- a/kernel/cpuset.c +++ b/kernel/cpuset.c @@ -270,11 +270,11 @@ static struct file_system_type cpuset_fs_type = { * are online. If none are online, walk up the cpuset hierarchy * until we find one that does have some online cpus. If we get * all the way to the top and still haven't found any online cpus, - * return cpu_online_map. Or if passed a NULL cs from an exit'ing - * task, return cpu_online_map. + * return cpu_online_mask. Or if passed a NULL cs from an exit'ing + * task, return cpu_online_mask. * * One way or another, we guarantee to return some non-empty subset - * of cpu_online_map. + * of cpu_online_mask. * * Call with callback_mutex held. */ @@ -867,7 +867,7 @@ static int update_cpumask(struct cpuset *cs, struct cpuset *trialcs, int retval; int is_load_balanced; - /* top_cpuset.cpus_allowed tracks cpu_online_map; it's read-only */ + /* top_cpuset.cpus_allowed tracks cpu_online_mask; it's read-only */ if (cs == &top_cpuset) return -EACCES; @@ -2149,7 +2149,7 @@ void __init cpuset_init_smp(void) * * Description: Returns the cpumask_var_t cpus_allowed of the cpuset * attached to the specified @tsk. Guaranteed to return some non-empty - * subset of cpu_online_map, even if this means going outside the + * subset of cpu_online_mask, even if this means going outside the * tasks cpuset. **/ -- cgit v1.2.3-70-g09d2 From 34f2c0ac111aaf090c7d2432dab532640870733a Mon Sep 17 00:00:00 2001 From: Paul Gortmaker Date: Sun, 1 Apr 2012 16:38:46 -0400 Subject: tile: fix multiple build failures from system.h dismantle Commit bd119c69239322caafdb64517a806037d0d0c70a "Disintegrate asm/system.h for Tile" created the asm/switch_to.h file, but did not add an include of it to all its users. Also, commit b4816afa3986704d1404fc48e931da5135820472 "Move the asm-generic/system.h xchg() implementation to asm-generic/cmpxchg.h" introduced the concept of asm/cmpxchg.h but the tile arch never got one. Fork the cmpxchg content out of the asm/atomic.h file to create one. Acked-by: David Howells Signed-off-by: Paul Gortmaker Signed-off-by: Chris Metcalf --- arch/tile/include/asm/atomic.h | 50 ++-------------------------- arch/tile/include/asm/cmpxchg.h | 73 +++++++++++++++++++++++++++++++++++++++++ arch/tile/kernel/process.c | 1 + arch/tile/kernel/stack.c | 1 + 4 files changed, 77 insertions(+), 48 deletions(-) create mode 100644 arch/tile/include/asm/cmpxchg.h (limited to 'arch/tile/kernel') diff --git a/arch/tile/include/asm/atomic.h b/arch/tile/include/asm/atomic.h index bb696da5d7c..f2461429a4a 100644 --- a/arch/tile/include/asm/atomic.h +++ b/arch/tile/include/asm/atomic.h @@ -17,6 +17,8 @@ #ifndef _ASM_TILE_ATOMIC_H #define _ASM_TILE_ATOMIC_H +#include + #ifndef __ASSEMBLY__ #include @@ -121,54 +123,6 @@ static inline int atomic_read(const atomic_t *v) */ #define atomic_add_negative(i, v) (atomic_add_return((i), (v)) < 0) -/* Nonexistent functions intended to cause link errors. */ -extern unsigned long __xchg_called_with_bad_pointer(void); -extern unsigned long __cmpxchg_called_with_bad_pointer(void); - -#define xchg(ptr, x) \ - ({ \ - typeof(*(ptr)) __x; \ - switch (sizeof(*(ptr))) { \ - case 4: \ - __x = (typeof(__x))(typeof(__x-__x))atomic_xchg( \ - (atomic_t *)(ptr), \ - (u32)(typeof((x)-(x)))(x)); \ - break; \ - case 8: \ - __x = (typeof(__x))(typeof(__x-__x))atomic64_xchg( \ - (atomic64_t *)(ptr), \ - (u64)(typeof((x)-(x)))(x)); \ - break; \ - default: \ - __xchg_called_with_bad_pointer(); \ - } \ - __x; \ - }) - -#define cmpxchg(ptr, o, n) \ - ({ \ - typeof(*(ptr)) __x; \ - switch (sizeof(*(ptr))) { \ - case 4: \ - __x = (typeof(__x))(typeof(__x-__x))atomic_cmpxchg( \ - (atomic_t *)(ptr), \ - (u32)(typeof((o)-(o)))(o), \ - (u32)(typeof((n)-(n)))(n)); \ - break; \ - case 8: \ - __x = (typeof(__x))(typeof(__x-__x))atomic64_cmpxchg( \ - (atomic64_t *)(ptr), \ - (u64)(typeof((o)-(o)))(o), \ - (u64)(typeof((n)-(n)))(n)); \ - break; \ - default: \ - __cmpxchg_called_with_bad_pointer(); \ - } \ - __x; \ - }) - -#define tas(ptr) (xchg((ptr), 1)) - #endif /* __ASSEMBLY__ */ #ifndef __tilegx__ diff --git a/arch/tile/include/asm/cmpxchg.h b/arch/tile/include/asm/cmpxchg.h new file mode 100644 index 00000000000..276f067e364 --- /dev/null +++ b/arch/tile/include/asm/cmpxchg.h @@ -0,0 +1,73 @@ +/* + * cmpxchg.h -- forked from asm/atomic.h with this copyright: + * + * Copyright 2010 Tilera Corporation. All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License + * as published by the Free Software Foundation, version 2. + * + * This program is distributed in the hope that it will be useful, but + * WITHOUT ANY WARRANTY; without even the implied warranty of + * MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE, GOOD TITLE or + * NON INFRINGEMENT. See the GNU General Public License for + * more details. + * + */ + +#ifndef _ASM_TILE_CMPXCHG_H +#define _ASM_TILE_CMPXCHG_H + +#ifndef __ASSEMBLY__ + +/* Nonexistent functions intended to cause link errors. */ +extern unsigned long __xchg_called_with_bad_pointer(void); +extern unsigned long __cmpxchg_called_with_bad_pointer(void); + +#define xchg(ptr, x) \ + ({ \ + typeof(*(ptr)) __x; \ + switch (sizeof(*(ptr))) { \ + case 4: \ + __x = (typeof(__x))(typeof(__x-__x))atomic_xchg( \ + (atomic_t *)(ptr), \ + (u32)(typeof((x)-(x)))(x)); \ + break; \ + case 8: \ + __x = (typeof(__x))(typeof(__x-__x))atomic64_xchg( \ + (atomic64_t *)(ptr), \ + (u64)(typeof((x)-(x)))(x)); \ + break; \ + default: \ + __xchg_called_with_bad_pointer(); \ + } \ + __x; \ + }) + +#define cmpxchg(ptr, o, n) \ + ({ \ + typeof(*(ptr)) __x; \ + switch (sizeof(*(ptr))) { \ + case 4: \ + __x = (typeof(__x))(typeof(__x-__x))atomic_cmpxchg( \ + (atomic_t *)(ptr), \ + (u32)(typeof((o)-(o)))(o), \ + (u32)(typeof((n)-(n)))(n)); \ + break; \ + case 8: \ + __x = (typeof(__x))(typeof(__x-__x))atomic64_cmpxchg( \ + (atomic64_t *)(ptr), \ + (u64)(typeof((o)-(o)))(o), \ + (u64)(typeof((n)-(n)))(n)); \ + break; \ + default: \ + __cmpxchg_called_with_bad_pointer(); \ + } \ + __x; \ + }) + +#define tas(ptr) (xchg((ptr), 1)) + +#endif /* __ASSEMBLY__ */ + +#endif /* _ASM_TILE_CMPXCHG_H */ diff --git a/arch/tile/kernel/process.c b/arch/tile/kernel/process.c index 30caecac94d..ee01b1c683e 100644 --- a/arch/tile/kernel/process.c +++ b/arch/tile/kernel/process.c @@ -28,6 +28,7 @@ #include #include #include +#include #include #include #include diff --git a/arch/tile/kernel/stack.c b/arch/tile/kernel/stack.c index 37ee4d037e0..0be6b0109ce 100644 --- a/arch/tile/kernel/stack.c +++ b/arch/tile/kernel/stack.c @@ -25,6 +25,7 @@ #include #include #include +#include #include #include #include -- cgit v1.2.3-70-g09d2 From b287f69676a34a9fc341de4d79a9c74e1959dec6 Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 14:02:52 -0400 Subject: arch/tile: avoid false corrupt frame warning in early boot With lockstat we can end up trying to get a backtrace before "high_memory" is initialized, so don't worry about range testing if it is zero. Signed-off-by: Chris Metcalf --- arch/tile/kernel/process.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/process.c b/arch/tile/kernel/process.c index ee01b1c683e..2d5ef617bb3 100644 --- a/arch/tile/kernel/process.c +++ b/arch/tile/kernel/process.c @@ -286,7 +286,7 @@ struct task_struct *validate_current(void) static struct task_struct corrupt = { .comm = "" }; struct task_struct *tsk = current; if (unlikely((unsigned long)tsk < PAGE_OFFSET || - (void *)tsk > high_memory || + (high_memory && (void *)tsk > high_memory) || ((unsigned long)tsk & (__alignof__(*tsk) - 1)) != 0)) { pr_err("Corrupt 'current' %p (sp %#lx)\n", tsk, stack_pointer); tsk = &corrupt; -- cgit v1.2.3-70-g09d2 From 5f639fdcd8c186c8128c616e94a7e7b159c968ae Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 14:06:14 -0400 Subject: arch/tile: various bugs in stack backtracer Fix a long-standing bug in the stack backtracer where we would print garbage to the console instead of kernel function names, if the kernel wasn't built with symbol support (e.g. mboot). Make sure to tag every line of userspace backtrace output if we actually have the mmap_sem, since that way if there's no tag, we know that it's because we couldn't trylock the semaphore. Stop doing a TLB flush and examining page tables during backtrace. Instead, just trust that __copy_from_user_inatomic() will properly fault and return a failure, which it should do in all cases. Fix a latent bug where the backtracer would directly examine a signal context in user space, rather than copying it safely to kernel memory first. This meant that a race with another thread could potentially have caused a kernel panic. Guard against unaligned sp when trying to restart backtrace at an interrupt or signal handler point in the kernel backtracer. Report kernel symbolic information for the call instruction rather than for the following instruction. We still report the actual numeric address corresponding to the instruction after the call, for the sake of consistency with the normal expectations for stack backtracers. Signed-off-by: Chris Metcalf --- arch/tile/include/asm/stack.h | 1 - arch/tile/kernel/stack.c | 231 ++++++++++++++++++++---------------------- 2 files changed, 112 insertions(+), 120 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/include/asm/stack.h b/arch/tile/include/asm/stack.h index 4d97a2db932..0e9d382a2d4 100644 --- a/arch/tile/include/asm/stack.h +++ b/arch/tile/include/asm/stack.h @@ -25,7 +25,6 @@ struct KBacktraceIterator { BacktraceIterator it; struct task_struct *task; /* task we are backtracing */ - pte_t *pgtable; /* page table for user space access */ int end; /* iteration complete. */ int new_context; /* new context is starting */ int profile; /* profiling, so stop on async intrpt */ diff --git a/arch/tile/kernel/stack.c b/arch/tile/kernel/stack.c index 0be6b0109ce..b2f44c28dda 100644 --- a/arch/tile/kernel/stack.c +++ b/arch/tile/kernel/stack.c @@ -21,9 +21,10 @@ #include #include #include +#include +#include #include #include -#include #include #include #include @@ -45,72 +46,23 @@ static int in_kernel_stack(struct KBacktraceIterator *kbt, unsigned long sp) return sp >= kstack_base && sp < kstack_base + THREAD_SIZE; } -/* Is address valid for reading? */ -static int valid_address(struct KBacktraceIterator *kbt, unsigned long address) -{ - HV_PTE *l1_pgtable = kbt->pgtable; - HV_PTE *l2_pgtable; - unsigned long pfn; - HV_PTE pte; - struct page *page; - - if (l1_pgtable == NULL) - return 0; /* can't read user space in other tasks */ - -#ifdef CONFIG_64BIT - /* Find the real l1_pgtable by looking in the l0_pgtable. */ - pte = l1_pgtable[HV_L0_INDEX(address)]; - if (!hv_pte_get_present(pte)) - return 0; - pfn = hv_pte_get_pfn(pte); - if (pte_huge(pte)) { - if (!pfn_valid(pfn)) { - pr_err("L0 huge page has bad pfn %#lx\n", pfn); - return 0; - } - return hv_pte_get_present(pte) && hv_pte_get_readable(pte); - } - page = pfn_to_page(pfn); - BUG_ON(PageHighMem(page)); /* No HIGHMEM on 64-bit. */ - l1_pgtable = (HV_PTE *)pfn_to_kaddr(pfn); -#endif - pte = l1_pgtable[HV_L1_INDEX(address)]; - if (!hv_pte_get_present(pte)) - return 0; - pfn = hv_pte_get_pfn(pte); - if (pte_huge(pte)) { - if (!pfn_valid(pfn)) { - pr_err("huge page has bad pfn %#lx\n", pfn); - return 0; - } - return hv_pte_get_present(pte) && hv_pte_get_readable(pte); - } - - page = pfn_to_page(pfn); - if (PageHighMem(page)) { - pr_err("L2 page table not in LOWMEM (%#llx)\n", - HV_PFN_TO_CPA(pfn)); - return 0; - } - l2_pgtable = (HV_PTE *)pfn_to_kaddr(pfn); - pte = l2_pgtable[HV_L2_INDEX(address)]; - return hv_pte_get_present(pte) && hv_pte_get_readable(pte); -} - /* Callback for backtracer; basically a glorified memcpy */ static bool read_memory_func(void *result, unsigned long address, unsigned int size, void *vkbt) { int retval; struct KBacktraceIterator *kbt = (struct KBacktraceIterator *)vkbt; + + if (address == 0) + return 0; if (__kernel_text_address(address)) { /* OK to read kernel code. */ } else if (address >= PAGE_OFFSET) { /* We only tolerate kernel-space reads of this task's stack */ if (!in_kernel_stack(kbt, address)) return 0; - } else if (!valid_address(kbt, address)) { - return 0; /* invalid user-space address */ + } else if (!kbt->is_current) { + return 0; /* can't read from other user address spaces */ } pagefault_disable(); retval = __copy_from_user_inatomic(result, @@ -128,6 +80,8 @@ static struct pt_regs *valid_fault_handler(struct KBacktraceIterator* kbt) unsigned long sp = kbt->it.sp; struct pt_regs *p; + if (sp % sizeof(long) != 0) + return NULL; if (!in_kernel_stack(kbt, sp)) return NULL; if (!in_kernel_stack(kbt, sp + C_ABI_SAVE_AREA_SIZE + PTREGS_SIZE-1)) @@ -170,27 +124,27 @@ static int is_sigreturn(unsigned long pc) } /* Return a pt_regs pointer for a valid signal handler frame */ -static struct pt_regs *valid_sigframe(struct KBacktraceIterator* kbt) +static struct pt_regs *valid_sigframe(struct KBacktraceIterator* kbt, + struct rt_sigframe* kframe) { BacktraceIterator *b = &kbt->it; - if (b->pc == VDSO_BASE) { - struct rt_sigframe *frame; - unsigned long sigframe_top = - b->sp + sizeof(struct rt_sigframe) - 1; - if (!valid_address(kbt, b->sp) || - !valid_address(kbt, sigframe_top)) { - if (kbt->verbose) - pr_err(" (odd signal: sp %#lx?)\n", - (unsigned long)(b->sp)); + if (b->pc == VDSO_BASE && b->sp < PAGE_OFFSET && + b->sp % sizeof(long) == 0) { + int retval; + pagefault_disable(); + retval = __copy_from_user_inatomic( + kframe, (void __user __force *)b->sp, + sizeof(*kframe)); + pagefault_enable(); + if (retval != 0 || + (unsigned int)(kframe->info.si_signo) >= _NSIG) return NULL; - } - frame = (struct rt_sigframe *)b->sp; if (kbt->verbose) { pr_err(" \n", - frame->info.si_signo); + kframe->info.si_signo); } - return (struct pt_regs *)&frame->uc.uc_mcontext; + return (struct pt_regs *)&kframe->uc.uc_mcontext; } return NULL; } @@ -203,10 +157,11 @@ static int KBacktraceIterator_is_sigreturn(struct KBacktraceIterator *kbt) static int KBacktraceIterator_restart(struct KBacktraceIterator *kbt) { struct pt_regs *p; + struct rt_sigframe kframe; p = valid_fault_handler(kbt); if (p == NULL) - p = valid_sigframe(kbt); + p = valid_sigframe(kbt, &kframe); if (p == NULL) return 0; backtrace_init(&kbt->it, read_memory_func, kbt, @@ -266,41 +221,19 @@ void KBacktraceIterator_init(struct KBacktraceIterator *kbt, /* * Set up callback information. We grab the kernel stack base - * so we will allow reads of that address range, and if we're - * asking about the current process we grab the page table - * so we can check user accesses before trying to read them. - * We flush the TLB to avoid any weird skew issues. + * so we will allow reads of that address range. */ - is_current = (t == NULL); + is_current = (t == NULL || t == current); kbt->is_current = is_current; if (is_current) t = validate_current(); kbt->task = t; - kbt->pgtable = NULL; kbt->verbose = 0; /* override in caller if desired */ kbt->profile = 0; /* override in caller if desired */ kbt->end = KBT_ONGOING; - kbt->new_context = 0; - if (is_current) { - HV_PhysAddr pgdir_pa = hv_inquire_context().page_table; - if (pgdir_pa == (unsigned long)swapper_pg_dir - PAGE_OFFSET) { - /* - * Not just an optimization: this also allows - * this to work at all before va/pa mappings - * are set up. - */ - kbt->pgtable = swapper_pg_dir; - } else { - struct page *page = pfn_to_page(PFN_DOWN(pgdir_pa)); - if (!PageHighMem(page)) - kbt->pgtable = __va(pgdir_pa); - else - pr_err("page table not in LOWMEM" - " (%#llx)\n", pgdir_pa); - } - local_flush_tlb_all(); + kbt->new_context = 1; + if (is_current) validate_stack(regs); - } if (regs == NULL) { if (is_current || t->state == TASK_RUNNING) { @@ -346,6 +279,78 @@ void KBacktraceIterator_next(struct KBacktraceIterator *kbt) } EXPORT_SYMBOL(KBacktraceIterator_next); +static void describe_addr(struct KBacktraceIterator *kbt, + unsigned long address, + int have_mmap_sem, char *buf, size_t bufsize) +{ + struct vm_area_struct *vma; + size_t namelen, remaining; + unsigned long size, offset, adjust; + char *p, *modname; + const char *name; + int rc; + + /* + * Look one byte back for every caller frame (i.e. those that + * aren't a new context) so we look up symbol data for the + * call itself, not the following instruction, which may be on + * a different line (or in a different function). + */ + adjust = !kbt->new_context; + address -= adjust; + + if (address >= PAGE_OFFSET) { + /* Handle kernel symbols. */ + BUG_ON(bufsize < KSYM_NAME_LEN); + name = kallsyms_lookup(address, &size, &offset, + &modname, buf); + if (name == NULL) { + buf[0] = '\0'; + return; + } + namelen = strlen(buf); + remaining = (bufsize - 1) - namelen; + p = buf + namelen; + rc = snprintf(p, remaining, "+%#lx/%#lx ", + offset + adjust, size); + if (modname && rc < remaining) + snprintf(p + rc, remaining - rc, "[%s] ", modname); + buf[bufsize-1] = '\0'; + return; + } + + /* If we don't have the mmap_sem, we can't show any more info. */ + buf[0] = '\0'; + if (!have_mmap_sem) + return; + + /* Find vma info. */ + vma = find_vma(kbt->task->mm, address); + if (vma == NULL || address < vma->vm_start) { + snprintf(buf, bufsize, "[unmapped address] "); + return; + } + + if (vma->vm_file) { + char *s; + p = d_path(&vma->vm_file->f_path, buf, bufsize); + if (IS_ERR(p)) + p = "?"; + s = strrchr(p, '/'); + if (s) + p = s+1; + } else { + p = "anon"; + } + + /* Generate a string description of the vma info. */ + namelen = strlen(p); + remaining = (bufsize - 1) - namelen; + memmove(buf, p, namelen); + snprintf(buf + namelen, remaining, "[%lx+%lx] ", + vma->vm_start, vma->vm_end - vma->vm_start); +} + /* * This method wraps the backtracer's more generic support. * It is only invoked from the architecture-specific code; show_stack() @@ -354,6 +359,7 @@ EXPORT_SYMBOL(KBacktraceIterator_next); void tile_show_stack(struct KBacktraceIterator *kbt, int headers) { int i; + int have_mmap_sem = 0; if (headers) { /* @@ -370,31 +376,16 @@ void tile_show_stack(struct KBacktraceIterator *kbt, int headers) kbt->verbose = 1; i = 0; for (; !KBacktraceIterator_end(kbt); KBacktraceIterator_next(kbt)) { - char *modname; - const char *name; - unsigned long address = kbt->it.pc; - unsigned long offset, size; char namebuf[KSYM_NAME_LEN+100]; + unsigned long address = kbt->it.pc; - if (address >= PAGE_OFFSET) - name = kallsyms_lookup(address, &size, &offset, - &modname, namebuf); - else - name = NULL; - - if (!name) - namebuf[0] = '\0'; - else { - size_t namelen = strlen(namebuf); - size_t remaining = (sizeof(namebuf) - 1) - namelen; - char *p = namebuf + namelen; - int rc = snprintf(p, remaining, "+%#lx/%#lx ", - offset, size); - if (modname && rc < remaining) - snprintf(p + rc, remaining - rc, - "[%s] ", modname); - namebuf[sizeof(namebuf)-1] = '\0'; - } + /* Try to acquire the mmap_sem as we pass into userspace. */ + if (address < PAGE_OFFSET && !have_mmap_sem && kbt->task->mm) + have_mmap_sem = + down_read_trylock(&kbt->task->mm->mmap_sem); + + describe_addr(kbt, address, have_mmap_sem, + namebuf, sizeof(namebuf)); pr_err(" frame %d: 0x%lx %s(sp 0x%lx)\n", i++, address, namebuf, (unsigned long)(kbt->it.sp)); @@ -409,6 +400,8 @@ void tile_show_stack(struct KBacktraceIterator *kbt, int headers) pr_err("Stack dump stopped; next frame identical to this one\n"); if (headers) pr_err("Stack dump complete\n"); + if (have_mmap_sem) + up_read(&kbt->task->mm->mmap_sem); } EXPORT_SYMBOL(tile_show_stack); -- cgit v1.2.3-70-g09d2 From e17235382dbb05f70146e141e4b780fd069050dc Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 14:52:00 -0400 Subject: arch/tile: work around a hardware issue with the return-address stack In certain circumstances we need to do a bunch of jump-and-link instructions to fill the hardware return-address stack with nonzero values. Signed-off-by: Chris Metcalf --- arch/tile/include/asm/traps.h | 6 +++++- arch/tile/kernel/intvec_64.S | 12 ++++++++++++ arch/tile/kernel/traps.c | 6 +++++- 3 files changed, 22 insertions(+), 2 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/include/asm/traps.h b/arch/tile/include/asm/traps.h index 5f20f920f93..e28c3df4176 100644 --- a/arch/tile/include/asm/traps.h +++ b/arch/tile/include/asm/traps.h @@ -64,7 +64,11 @@ void do_breakpoint(struct pt_regs *, int fault_num); #ifdef __tilegx__ +/* kernel/single_step.c */ void gx_singlestep_handle(struct pt_regs *, int fault_num); + +/* kernel/intvec_64.S */ +void fill_ra_stack(void); #endif -#endif /* _ASM_TILE_SYSCALLS_H */ +#endif /* _ASM_TILE_TRAPS_H */ diff --git a/arch/tile/kernel/intvec_64.S b/arch/tile/kernel/intvec_64.S index 79c93e10ba2..2c181c864ef 100644 --- a/arch/tile/kernel/intvec_64.S +++ b/arch/tile/kernel/intvec_64.S @@ -1156,6 +1156,18 @@ int_unalign: push_extra_callee_saves r0 j do_trap +/* Fill the return address stack with nonzero entries. */ +STD_ENTRY(fill_ra_stack) + { + move r0, lr + jal 1f + } +1: jal 2f +2: jal 3f +3: jal 4f +4: jrp r0 + STD_ENDPROC(fill_ra_stack) + /* Include .intrpt1 array of interrupt vectors */ .section ".intrpt1", "ax" diff --git a/arch/tile/kernel/traps.c b/arch/tile/kernel/traps.c index 2bb6602a1ee..32acfd9e23d 100644 --- a/arch/tile/kernel/traps.c +++ b/arch/tile/kernel/traps.c @@ -289,7 +289,10 @@ void __kprobes do_trap(struct pt_regs *regs, int fault_num, address = regs->pc; break; #ifdef __tilegx__ - case INT_ILL_TRANS: + case INT_ILL_TRANS: { + /* Avoid a hardware erratum with the return address stack. */ + fill_ra_stack(); + signo = SIGSEGV; code = SEGV_MAPERR; if (reason & SPR_ILL_TRANS_REASON__I_STREAM_VA_RMASK) @@ -297,6 +300,7 @@ void __kprobes do_trap(struct pt_regs *regs, int fault_num, else address = 0; /* FIXME: GX: single-step for address */ break; + } #endif default: panic("Unexpected do_trap interrupt number %d", fault_num); -- cgit v1.2.3-70-g09d2 From a714ffff36a581756ec3b001f47e8e5e96a9fa0e Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 15:23:54 -0400 Subject: arch/tile: fix up some minor trap handling issues We now respond to MEM_ERROR traps (e.g. an atomic instruction to non-cacheable memory) with a SIGBUS. We also no longer generate a console crash message if a user process die due to a SIGTRAP. Signed-off-by: Chris Metcalf --- arch/tile/kernel/intvec_64.S | 2 +- arch/tile/kernel/traps.c | 9 +++++++-- 2 files changed, 8 insertions(+), 3 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/intvec_64.S b/arch/tile/kernel/intvec_64.S index 2c181c864ef..005535d108c 100644 --- a/arch/tile/kernel/intvec_64.S +++ b/arch/tile/kernel/intvec_64.S @@ -1178,7 +1178,7 @@ STD_ENTRY(fill_ra_stack) #define do_hardwall_trap bad_intr #endif - int_hand INT_MEM_ERROR, MEM_ERROR, bad_intr + int_hand INT_MEM_ERROR, MEM_ERROR, do_trap int_hand INT_SINGLE_STEP_3, SINGLE_STEP_3, bad_intr #if CONFIG_KERNEL_PL == 2 int_hand INT_SINGLE_STEP_2, SINGLE_STEP_2, gx_singlestep_handle diff --git a/arch/tile/kernel/traps.c b/arch/tile/kernel/traps.c index 32acfd9e23d..73cff814ac5 100644 --- a/arch/tile/kernel/traps.c +++ b/arch/tile/kernel/traps.c @@ -200,7 +200,7 @@ void __kprobes do_trap(struct pt_regs *regs, int fault_num, { siginfo_t info = { 0 }; int signo, code; - unsigned long address; + unsigned long address = 0; bundle_bits instr; /* Re-enable interrupts. */ @@ -223,6 +223,10 @@ void __kprobes do_trap(struct pt_regs *regs, int fault_num, } switch (fault_num) { + case INT_MEM_ERROR: + signo = SIGBUS; + code = BUS_OBJERR; + break; case INT_ILL: if (copy_from_user(&instr, (void __user *)regs->pc, sizeof(instr))) { @@ -312,7 +316,8 @@ void __kprobes do_trap(struct pt_regs *regs, int fault_num, info.si_addr = (void __user *)address; if (signo == SIGILL) info.si_trapno = fault_num; - trace_unhandled_signal("trap", regs, address, signo); + if (signo != SIGTRAP) + trace_unhandled_signal("trap", regs, address, signo); force_sig_info(signo, &info, current); } -- cgit v1.2.3-70-g09d2 From 51bcdf8879f7920946c90087e6160680812a44bd Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 15:29:28 -0400 Subject: arch/tile: fix a couple of comments that needed updating Not associated with any code changes, so I'm just lumping these comment changes into a commit by themselves. Signed-off-by: Chris Metcalf --- arch/tile/kernel/setup.c | 11 +++++++++-- arch/tile/mm/fault.c | 2 +- 2 files changed, 10 insertions(+), 3 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/setup.c b/arch/tile/kernel/setup.c index 5f85d8b34db..023e2e1cf7f 100644 --- a/arch/tile/kernel/setup.c +++ b/arch/tile/kernel/setup.c @@ -913,6 +913,13 @@ void __cpuinit setup_cpu(int boot) #ifdef CONFIG_BLK_DEV_INITRD +/* + * Note that the kernel can potentially support other compression + * techniques than gz, though we don't do so by default. If we ever + * decide to do so we can either look for other filename extensions, + * or just allow a file with this name to be compressed with an + * arbitrary compressor (somewhat counterintuitively). + */ static int __initdata set_initramfs_file; static char __initdata initramfs_file[128] = "initramfs.cpio.gz"; @@ -928,9 +935,9 @@ static int __init setup_initramfs_file(char *str) early_param("initramfs_file", setup_initramfs_file); /* - * We look for an additional "initramfs.cpio.gz" file in the hvfs. + * We look for an "initramfs.cpio.gz" file in the hvfs. * If there is one, we allocate some memory for it and it will be - * unpacked to the initramfs after any built-in initramfs_data. + * unpacked to the initramfs. */ static void __init load_hv_initrd(void) { diff --git a/arch/tile/mm/fault.c b/arch/tile/mm/fault.c index cba30e9547b..cac17c4f2ec 100644 --- a/arch/tile/mm/fault.c +++ b/arch/tile/mm/fault.c @@ -130,7 +130,7 @@ static inline pmd_t *vmalloc_sync_one(pgd_t *pgd, unsigned long address) } /* - * Handle a fault on the vmalloc or module mapping area + * Handle a fault on the vmalloc area. */ static inline int vmalloc_fault(pgd_t *pgd, unsigned long address) { -- cgit v1.2.3-70-g09d2 From 5f220704127ae70db519fabbda4ece649eadac7f Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 15:44:10 -0400 Subject: arch/tile: don't leak kernel memory when we unload modules We were failing to track the memory when we allocated it. Signed-off-by: Chris Metcalf --- arch/tile/kernel/module.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/module.c b/arch/tile/kernel/module.c index b90ab992567..98d47692010 100644 --- a/arch/tile/kernel/module.c +++ b/arch/tile/kernel/module.c @@ -67,6 +67,8 @@ void *module_alloc(unsigned long size) area = __get_vm_area(size, VM_ALLOC, MEM_MODULE_START, MEM_MODULE_END); if (!area) goto error; + area->nr_pages = npages; + area->pages = pages; if (map_vm_area(area, prot_rwx, &pages)) { vunmap(area->addr); -- cgit v1.2.3-70-g09d2 From bfffe79bc29a9c4c817d5f51590961220e26db1a Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 15:56:18 -0400 Subject: arch/tile: use proper memparse() for "maxmem" options This is more standard and avoids having to remember what units the options actually take. Signed-off-by: Chris Metcalf --- arch/tile/kernel/setup.c | 17 ++++++++--------- 1 file changed, 8 insertions(+), 9 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/setup.c b/arch/tile/kernel/setup.c index 023e2e1cf7f..f3598e7a47f 100644 --- a/arch/tile/kernel/setup.c +++ b/arch/tile/kernel/setup.c @@ -103,13 +103,11 @@ unsigned long __initdata pci_reserve_end_pfn = -1U; static int __init setup_maxmem(char *str) { - long maxmem_mb; - if (str == NULL || strict_strtol(str, 0, &maxmem_mb) != 0 || - maxmem_mb == 0) + unsigned long long maxmem; + if (str == NULL || (maxmem = memparse(str, NULL)) == 0) return -EINVAL; - maxmem_pfn = (maxmem_mb >> (HPAGE_SHIFT - 20)) << - (HPAGE_SHIFT - PAGE_SHIFT); + maxmem_pfn = (maxmem >> HPAGE_SHIFT) << (HPAGE_SHIFT - PAGE_SHIFT); pr_info("Forcing RAM used to no more than %dMB\n", maxmem_pfn >> (20 - PAGE_SHIFT)); return 0; @@ -119,14 +117,15 @@ early_param("maxmem", setup_maxmem); static int __init setup_maxnodemem(char *str) { char *endp; - long maxnodemem_mb, node; + unsigned long long maxnodemem; + long node; node = str ? simple_strtoul(str, &endp, 0) : INT_MAX; - if (node >= MAX_NUMNODES || *endp != ':' || - strict_strtol(endp+1, 0, &maxnodemem_mb) != 0) + if (node >= MAX_NUMNODES || *endp != ':') return -EINVAL; - maxnodemem_pfn[node] = (maxnodemem_mb >> (HPAGE_SHIFT - 20)) << + maxnodemem = memparse(endp+1, NULL); + maxnodemem_pfn[node] = (maxnodemem >> HPAGE_SHIFT) << (HPAGE_SHIFT - PAGE_SHIFT); pr_info("Forcing RAM used on node %ld to no more than %dMB\n", node, maxnodemem_pfn[node] >> (20 - PAGE_SHIFT)); -- cgit v1.2.3-70-g09d2 From 8c92ba6c327ee5089dec1e92eaa82927bee63d6d Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 15:57:18 -0400 Subject: arch/tile: add "nop" after "nap" to help GX idle power draw This avoids the hardware istream prefetcher doing unnecessary work. Signed-off-by: Chris Metcalf --- arch/tile/kernel/entry.S | 2 ++ arch/tile/kernel/smp.c | 2 +- 2 files changed, 3 insertions(+), 1 deletion(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/entry.S b/arch/tile/kernel/entry.S index 431e9ae6048..ec91568df88 100644 --- a/arch/tile/kernel/entry.S +++ b/arch/tile/kernel/entry.S @@ -85,6 +85,7 @@ STD_ENTRY(cpu_idle_on_new_stack) /* Loop forever on a nap during SMP boot. */ STD_ENTRY(smp_nap) nap + nop /* avoid provoking the icache prefetch with a jump */ j smp_nap /* we are not architecturally guaranteed not to exit nap */ jrp lr /* clue in the backtracer */ STD_ENDPROC(smp_nap) @@ -105,5 +106,6 @@ STD_ENTRY(_cpu_idle) .global _cpu_idle_nap _cpu_idle_nap: nap + nop /* avoid provoking the icache prefetch with a jump */ jrp lr STD_ENDPROC(_cpu_idle) diff --git a/arch/tile/kernel/smp.c b/arch/tile/kernel/smp.c index a44e103c5a6..7b6df8c2770 100644 --- a/arch/tile/kernel/smp.c +++ b/arch/tile/kernel/smp.c @@ -103,7 +103,7 @@ static void smp_stop_cpu_interrupt(void) set_cpu_online(smp_processor_id(), 0); arch_local_irq_disable_all(); for (;;) - asm("nap"); + asm("nap; nop"); } /* This function calls the 'stop' function on all other CPUs in the system. */ -- cgit v1.2.3-70-g09d2 From cb210ee3a81afab7c64777635cc18899a2bdd9a5 Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 15:59:11 -0400 Subject: arch/tile: implement panic_smp_self_stop() This allows the later-panicking tiles to wait in a lower power state until they get interrupted with an smp_send_stop(). Signed-off-by: Chris Metcalf --- arch/tile/kernel/smp.c | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/smp.c b/arch/tile/kernel/smp.c index 7b6df8c2770..91da0f72195 100644 --- a/arch/tile/kernel/smp.c +++ b/arch/tile/kernel/smp.c @@ -113,6 +113,12 @@ void smp_send_stop(void) send_IPI_allbutself(MSG_TAG_STOP_CPU); } +/* On panic, just wait; we may get an smp_send_stop() later on. */ +void panic_smp_self_stop(void) +{ + while (1) + asm("nap; nop"); +} /* * Dispatch code called from hv_message_intr() for HV_MSG_TILE hv messages. -- cgit v1.2.3-70-g09d2 From 2858f856021340f3730fa8639dd520a2e4331f7f Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Thu, 29 Mar 2012 16:11:09 -0400 Subject: arch/tile: fix single-stepping over swint1 instructions on tilegx If we are single-stepping and make a syscall, we call ptrace_notify() explicitly on the return path back to user space, since we are returning to a pc value set artificially to the next instruction, and otherwise we won't register that we stepped over the syscall instruction (swint1). Signed-off-by: Chris Metcalf --- arch/tile/kernel/intvec_64.S | 21 ++++++++++++++++++--- 1 file changed, 18 insertions(+), 3 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/intvec_64.S b/arch/tile/kernel/intvec_64.S index 005535d108c..fdff17c70cc 100644 --- a/arch/tile/kernel/intvec_64.S +++ b/arch/tile/kernel/intvec_64.S @@ -22,6 +22,7 @@ #include #include #include +#include #include #include #include @@ -1039,11 +1040,25 @@ handle_syscall: /* Do syscall trace again, if requested. */ ld r30, r31 - andi r30, r30, _TIF_SYSCALL_TRACE - beqzt r30, 1f + andi r0, r30, _TIF_SYSCALL_TRACE + { + andi r0, r30, _TIF_SINGLESTEP + beqzt r0, 1f + } jal do_syscall_trace FEEDBACK_REENTER(handle_syscall) -1: j .Lresume_userspace /* jump into middle of interrupt_return */ + andi r0, r30, _TIF_SINGLESTEP + +1: beqzt r0, 2f + + /* Single stepping -- notify ptrace. */ + { + movei r0, SIGTRAP + jal ptrace_notify + } + FEEDBACK_REENTER(handle_syscall) + +2: j .Lresume_userspace /* jump into middle of interrupt_return */ .Lcompat_syscall: /* -- cgit v1.2.3-70-g09d2 From cdd8e16feba87a3fc2bb1885d36f895a2a3288bf Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Fri, 30 Mar 2012 16:24:41 -0400 Subject: arch/tile: return SIGBUS for addresses that are unaligned AND invalid Previously we were returning SIGSEGV in this case. It seems cleaner to return SIGBUS since the hardware figures out alignment traps before TLB violations, so SIGBUS is the "more correct" signal. Signed-off-by: Chris Metcalf --- arch/tile/kernel/single_step.c | 31 +++++++++++++++++++------------ 1 file changed, 19 insertions(+), 12 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/single_step.c b/arch/tile/kernel/single_step.c index bc1eb586e24..9efbc1391b3 100644 --- a/arch/tile/kernel/single_step.c +++ b/arch/tile/kernel/single_step.c @@ -153,6 +153,25 @@ static tile_bundle_bits rewrite_load_store_unaligned( if (((unsigned long)addr % size) == 0) return bundle; + /* + * Return SIGBUS with the unaligned address, if requested. + * Note that we return SIGBUS even for completely invalid addresses + * as long as they are in fact unaligned; this matches what the + * tilepro hardware would be doing, if it could provide us with the + * actual bad address in an SPR, which it doesn't. + */ + if (unaligned_fixup == 0) { + siginfo_t info = { + .si_signo = SIGBUS, + .si_code = BUS_ADRALN, + .si_addr = addr + }; + trace_unhandled_signal("unaligned trap", regs, + (unsigned long)addr, SIGBUS); + force_sig_info(info.si_signo, &info, current); + return (tilepro_bundle_bits) 0; + } + #ifndef __LITTLE_ENDIAN # error We assume little-endian representation with copy_xx_user size 2 here #endif @@ -192,18 +211,6 @@ static tile_bundle_bits rewrite_load_store_unaligned( return (tile_bundle_bits) 0; } - if (unaligned_fixup == 0) { - siginfo_t info = { - .si_signo = SIGBUS, - .si_code = BUS_ADRALN, - .si_addr = addr - }; - trace_unhandled_signal("unaligned trap", regs, - (unsigned long)addr, SIGBUS); - force_sig_info(info.si_signo, &info, current); - return (tile_bundle_bits) 0; - } - if (unaligned_printk || unaligned_fixup_count == 0) { pr_info("Process %d/%s: PC %#lx: Fixup of" " unaligned %s at %#lx.\n", -- cgit v1.2.3-70-g09d2 From e1d5c0195075abaa45cd04ca397dbeaa0d18c490 Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Fri, 30 Mar 2012 16:29:06 -0400 Subject: arch/tile: avoid accidentally unmasking NMI-type interrupt accidentally The return path as we reload registers and core state requires that r30 hold a boolean indicating whether we are returning from an NMI, but in a couple of cases we weren't setting this properly, with the result that we could accidentally unmask the NMI interrupt(s), which could cause confusion. Now we set r30 in every place where we jump into the interrupt return path. Signed-off-by: Chris Metcalf --- arch/tile/kernel/intvec_32.S | 24 ++++++++++++++++++++---- arch/tile/kernel/intvec_64.S | 19 ++++++++++++++++--- 2 files changed, 36 insertions(+), 7 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/intvec_32.S b/arch/tile/kernel/intvec_32.S index aecc8ed5f39..5d56a1ef5ba 100644 --- a/arch/tile/kernel/intvec_32.S +++ b/arch/tile/kernel/intvec_32.S @@ -799,6 +799,10 @@ handle_interrupt: * This routine takes a boolean in r30 indicating if this is an NMI. * If so, we also expect a boolean in r31 indicating whether to * re-enable the oprofile interrupts. + * + * Note that .Lresume_userspace is jumped to directly in several + * places, and we need to make sure r30 is set correctly in those + * callers as well. */ STD_ENTRY(interrupt_return) /* If we're resuming to kernel space, don't check thread flags. */ @@ -1237,7 +1241,10 @@ handle_syscall: bzt r30, 1f jal do_syscall_trace FEEDBACK_REENTER(handle_syscall) -1: j .Lresume_userspace /* jump into middle of interrupt_return */ +1: { + movei r30, 0 /* not an NMI */ + j .Lresume_userspace /* jump into middle of interrupt_return */ + } .Linvalid_syscall: /* Report an invalid syscall back to the user program */ @@ -1246,7 +1253,10 @@ handle_syscall: movei r28, -ENOSYS } sw r29, r28 - j .Lresume_userspace /* jump into middle of interrupt_return */ + { + movei r30, 0 /* not an NMI */ + j .Lresume_userspace /* jump into middle of interrupt_return */ + } STD_ENDPROC(handle_syscall) /* Return the address for oprofile to suppress in backtraces. */ @@ -1262,7 +1272,10 @@ STD_ENTRY(ret_from_fork) jal sim_notify_fork jal schedule_tail FEEDBACK_REENTER(ret_from_fork) - j .Lresume_userspace /* jump into middle of interrupt_return */ + { + movei r30, 0 /* not an NMI */ + j .Lresume_userspace /* jump into middle of interrupt_return */ + } STD_ENDPROC(ret_from_fork) /* @@ -1376,7 +1389,10 @@ handle_ill: jal send_sigtrap /* issue a SIGTRAP */ FEEDBACK_REENTER(handle_ill) - j .Lresume_userspace /* jump into middle of interrupt_return */ + { + movei r30, 0 /* not an NMI */ + j .Lresume_userspace /* jump into middle of interrupt_return */ + } .Ldispatch_normal_ill: { diff --git a/arch/tile/kernel/intvec_64.S b/arch/tile/kernel/intvec_64.S index fdff17c70cc..49d9d662168 100644 --- a/arch/tile/kernel/intvec_64.S +++ b/arch/tile/kernel/intvec_64.S @@ -606,6 +606,10 @@ handle_interrupt: * This routine takes a boolean in r30 indicating if this is an NMI. * If so, we also expect a boolean in r31 indicating whether to * re-enable the oprofile interrupts. + * + * Note that .Lresume_userspace is jumped to directly in several + * places, and we need to make sure r30 is set correctly in those + * callers as well. */ STD_ENTRY(interrupt_return) /* If we're resuming to kernel space, don't check thread flags. */ @@ -1058,7 +1062,10 @@ handle_syscall: } FEEDBACK_REENTER(handle_syscall) -2: j .Lresume_userspace /* jump into middle of interrupt_return */ +2: { + movei r30, 0 /* not an NMI */ + j .Lresume_userspace /* jump into middle of interrupt_return */ + } .Lcompat_syscall: /* @@ -1092,7 +1099,10 @@ handle_syscall: movei r28, -ENOSYS } st r29, r28 - j .Lresume_userspace /* jump into middle of interrupt_return */ + { + movei r30, 0 /* not an NMI */ + j .Lresume_userspace /* jump into middle of interrupt_return */ + } STD_ENDPROC(handle_syscall) /* Return the address for oprofile to suppress in backtraces. */ @@ -1108,7 +1118,10 @@ STD_ENTRY(ret_from_fork) jal sim_notify_fork jal schedule_tail FEEDBACK_REENTER(ret_from_fork) - j .Lresume_userspace + { + movei r30, 0 /* not an NMI */ + j .Lresume_userspace /* jump into middle of interrupt_return */ + } STD_ENDPROC(ret_from_fork) /* Various stub interrupt handlers and syscall handlers */ -- cgit v1.2.3-70-g09d2 From d1640130cda146ed925f12434bfe579ee7d80a1c Mon Sep 17 00:00:00 2001 From: "Srivatsa S. Bhat" Date: Thu, 22 Mar 2012 16:59:11 +0530 Subject: tile/CPU hotplug: Add missing call to notify_cpu_starting() The scheduler depends on receiving the CPU_STARTING notification, without which we end up into a lot of trouble. So add the missing call to notify_cpu_starting() in the bringup code. Signed-off-by: Srivatsa S. Bhat Signed-off-by: Chris Metcalf --- arch/tile/kernel/smpboot.c | 2 ++ 1 file changed, 2 insertions(+) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/smpboot.c b/arch/tile/kernel/smpboot.c index b949edcec20..172aef7d315 100644 --- a/arch/tile/kernel/smpboot.c +++ b/arch/tile/kernel/smpboot.c @@ -196,6 +196,8 @@ void __cpuinit online_secondary(void) /* This must be done before setting cpu_online_mask */ wmb(); + notify_cpu_starting(smp_processor_id()); + /* * We need to hold call_lock, so there is no inconsistency * between the time smp_call_function() determines number of -- cgit v1.2.3-70-g09d2 From e72d5c7e9c831f6e393c71dcd62acafbac2b58d0 Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Wed, 11 Apr 2012 12:45:20 -0400 Subject: arch/tile: avoid unused variable warning in proc.c for tilegx Until we push the unaligned access support for tilegx, it's silly to have arch/tile/kernel/proc.c generate a warning about an unused variable. Extend the #ifdef to cover all the code and data for now. Signed-off-by: Chris Metcalf --- arch/tile/kernel/proc.c | 4 +--- 1 file changed, 1 insertion(+), 3 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/proc.c b/arch/tile/kernel/proc.c index 7a932704640..446a7f52cc1 100644 --- a/arch/tile/kernel/proc.c +++ b/arch/tile/kernel/proc.c @@ -146,7 +146,6 @@ static ctl_table unaligned_table[] = { }, {} }; -#endif static struct ctl_path tile_path[] = { { .procname = "tile" }, @@ -155,10 +154,9 @@ static struct ctl_path tile_path[] = { static int __init proc_sys_tile_init(void) { -#ifndef __tilegx__ /* FIXME: GX: no support for unaligned access yet */ register_sysctl_paths(tile_path, unaligned_table); -#endif return 0; } arch_initcall(proc_sys_tile_init); +#endif -- cgit v1.2.3-70-g09d2 From 6be5ceb02e98eaf6cfc4f8b12a896d04023f340d Mon Sep 17 00:00:00 2001 From: Linus Torvalds Date: Fri, 20 Apr 2012 17:13:58 -0700 Subject: VM: add "vm_mmap()" helper function This continues the theme started with vm_brk() and vm_munmap(): vm_mmap() does the same thing as do_mmap(), but additionally does the required VM locking. This uninlines (and rewrites it to be clearer) do_mmap(), which sadly duplicates it in mm/mmap.c and mm/nommu.c. But that way we don't have to export our internal do_mmap_pgoff() function. Some day we hopefully don't have to export do_mmap() either, if all modular users can become the simpler vm_mmap() instead. We're actually very close to that already, with the notable exception of the (broken) use in i810, and a couple of stragglers in binfmt_elf. Signed-off-by: Linus Torvalds --- arch/tile/kernel/single_step.c | 4 +--- arch/x86/ia32/ia32_aout.c | 12 +++--------- arch/x86/kvm/x86.c | 4 +--- drivers/gpu/drm/drm_bufs.c | 12 ++++-------- drivers/gpu/drm/exynos/exynos_drm_gem.c | 4 +--- drivers/gpu/drm/i810/i810_dma.c | 1 + drivers/gpu/drm/i915/i915_gem.c | 4 +--- fs/binfmt_aout.c | 12 +++--------- fs/binfmt_elf.c | 8 ++------ fs/binfmt_elf_fdpic.c | 18 ++++-------------- fs/binfmt_flat.c | 12 +++--------- fs/binfmt_som.c | 12 +++--------- include/linux/mm.h | 23 ++++++----------------- mm/mmap.c | 29 +++++++++++++++++++++++++++-- mm/nommu.c | 29 +++++++++++++++++++++++++++-- 15 files changed, 87 insertions(+), 97 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/kernel/single_step.c b/arch/tile/kernel/single_step.c index 9efbc1391b3..89529c9f060 100644 --- a/arch/tile/kernel/single_step.c +++ b/arch/tile/kernel/single_step.c @@ -346,12 +346,10 @@ void single_step_once(struct pt_regs *regs) } /* allocate a cache line of writable, executable memory */ - down_write(¤t->mm->mmap_sem); - buffer = (void __user *) do_mmap(NULL, 0, 64, + buffer = (void __user *) vm_mmap(NULL, 0, 64, PROT_EXEC | PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANONYMOUS, 0); - up_write(¤t->mm->mmap_sem); if (IS_ERR((void __force *)buffer)) { kfree(state); diff --git a/arch/x86/ia32/ia32_aout.c b/arch/x86/ia32/ia32_aout.c index b6817ee9033..4824fb45560 100644 --- a/arch/x86/ia32/ia32_aout.c +++ b/arch/x86/ia32/ia32_aout.c @@ -379,26 +379,22 @@ static int load_aout_binary(struct linux_binprm *bprm, struct pt_regs *regs) goto beyond_if; } - down_write(¤t->mm->mmap_sem); - error = do_mmap(bprm->file, N_TXTADDR(ex), ex.a_text, + error = vm_mmap(bprm->file, N_TXTADDR(ex), ex.a_text, PROT_READ | PROT_EXEC, MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE | MAP_32BIT, fd_offset); - up_write(¤t->mm->mmap_sem); if (error != N_TXTADDR(ex)) { send_sig(SIGKILL, current, 0); return error; } - down_write(¤t->mm->mmap_sem); - error = do_mmap(bprm->file, N_DATADDR(ex), ex.a_data, + error = vm_mmap(bprm->file, N_DATADDR(ex), ex.a_data, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE | MAP_32BIT, fd_offset + ex.a_text); - up_write(¤t->mm->mmap_sem); if (error != N_DATADDR(ex)) { send_sig(SIGKILL, current, 0); return error; @@ -482,12 +478,10 @@ static int load_aout_library(struct file *file) goto out; } /* Now use mmap to map the library into memory. */ - down_write(¤t->mm->mmap_sem); - error = do_mmap(file, start_addr, ex.a_text + ex.a_data, + error = vm_mmap(file, start_addr, ex.a_text + ex.a_data, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_32BIT, N_TXTOFF(ex)); - up_write(¤t->mm->mmap_sem); retval = error; if (error != start_addr) goto out; diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c index 8beb9ce7936..1457be305fb 100644 --- a/arch/x86/kvm/x86.c +++ b/arch/x86/kvm/x86.c @@ -6336,13 +6336,11 @@ int kvm_arch_prepare_memory_region(struct kvm *kvm, if (npages && !old.rmap) { unsigned long userspace_addr; - down_write(¤t->mm->mmap_sem); - userspace_addr = do_mmap(NULL, 0, + userspace_addr = vm_mmap(NULL, 0, npages * PAGE_SIZE, PROT_READ | PROT_WRITE, map_flags, 0); - up_write(¤t->mm->mmap_sem); if (IS_ERR((void *)userspace_addr)) return PTR_ERR((void *)userspace_addr); diff --git a/drivers/gpu/drm/drm_bufs.c b/drivers/gpu/drm/drm_bufs.c index 30372f7b2d4..348b367debe 100644 --- a/drivers/gpu/drm/drm_bufs.c +++ b/drivers/gpu/drm/drm_bufs.c @@ -1510,8 +1510,8 @@ int drm_freebufs(struct drm_device *dev, void *data, * \param arg pointer to a drm_buf_map structure. * \return zero on success or a negative number on failure. * - * Maps the AGP, SG or PCI buffer region with do_mmap(), and copies information - * about each buffer into user space. For PCI buffers, it calls do_mmap() with + * Maps the AGP, SG or PCI buffer region with vm_mmap(), and copies information + * about each buffer into user space. For PCI buffers, it calls vm_mmap() with * offset equal to 0, which drm_mmap() interpretes as PCI buffers and calls * drm_mmap_dma(). */ @@ -1553,18 +1553,14 @@ int drm_mapbufs(struct drm_device *dev, void *data, retcode = -EINVAL; goto done; } - down_write(¤t->mm->mmap_sem); - virtual = do_mmap(file_priv->filp, 0, map->size, + virtual = vm_mmap(file_priv->filp, 0, map->size, PROT_READ | PROT_WRITE, MAP_SHARED, token); - up_write(¤t->mm->mmap_sem); } else { - down_write(¤t->mm->mmap_sem); - virtual = do_mmap(file_priv->filp, 0, dma->byte_count, + virtual = vm_mmap(file_priv->filp, 0, dma->byte_count, PROT_READ | PROT_WRITE, MAP_SHARED, 0); - up_write(¤t->mm->mmap_sem); } if (virtual > -1024UL) { /* Real error */ diff --git a/drivers/gpu/drm/exynos/exynos_drm_gem.c b/drivers/gpu/drm/exynos/exynos_drm_gem.c index 26d51979116..392ce71ed6a 100644 --- a/drivers/gpu/drm/exynos/exynos_drm_gem.c +++ b/drivers/gpu/drm/exynos/exynos_drm_gem.c @@ -581,10 +581,8 @@ int exynos_drm_gem_mmap_ioctl(struct drm_device *dev, void *data, obj->filp->f_op = &exynos_drm_gem_fops; obj->filp->private_data = obj; - down_write(¤t->mm->mmap_sem); - addr = do_mmap(obj->filp, 0, args->size, + addr = vm_mmap(obj->filp, 0, args->size, PROT_READ | PROT_WRITE, MAP_SHARED, 0); - up_write(¤t->mm->mmap_sem); drm_gem_object_unreference_unlocked(obj); diff --git a/drivers/gpu/drm/i810/i810_dma.c b/drivers/gpu/drm/i810/i810_dma.c index b85337f06fb..a4ba453b3d2 100644 --- a/drivers/gpu/drm/i810/i810_dma.c +++ b/drivers/gpu/drm/i810/i810_dma.c @@ -129,6 +129,7 @@ static int i810_map_buffer(struct drm_buf *buf, struct drm_file *file_priv) if (buf_priv->currently_mapped == I810_BUF_MAPPED) return -EINVAL; + /* This is all entirely broken */ down_write(¤t->mm->mmap_sem); old_fops = file_priv->filp->f_op; file_priv->filp->f_op = &i810_buffer_fops; diff --git a/drivers/gpu/drm/i915/i915_gem.c b/drivers/gpu/drm/i915/i915_gem.c index 0e3c6acde95..0d1e4b7b4b9 100644 --- a/drivers/gpu/drm/i915/i915_gem.c +++ b/drivers/gpu/drm/i915/i915_gem.c @@ -1087,11 +1087,9 @@ i915_gem_mmap_ioctl(struct drm_device *dev, void *data, if (obj == NULL) return -ENOENT; - down_write(¤t->mm->mmap_sem); - addr = do_mmap(obj->filp, 0, args->size, + addr = vm_mmap(obj->filp, 0, args->size, PROT_READ | PROT_WRITE, MAP_SHARED, args->offset); - up_write(¤t->mm->mmap_sem); drm_gem_object_unreference_unlocked(obj); if (IS_ERR((void *)addr)) return addr; diff --git a/fs/binfmt_aout.c b/fs/binfmt_aout.c index 88527492b91..d146e181d10 100644 --- a/fs/binfmt_aout.c +++ b/fs/binfmt_aout.c @@ -319,24 +319,20 @@ static int load_aout_binary(struct linux_binprm * bprm, struct pt_regs * regs) goto beyond_if; } - down_write(¤t->mm->mmap_sem); - error = do_mmap(bprm->file, N_TXTADDR(ex), ex.a_text, + error = vm_mmap(bprm->file, N_TXTADDR(ex), ex.a_text, PROT_READ | PROT_EXEC, MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE, fd_offset); - up_write(¤t->mm->mmap_sem); if (error != N_TXTADDR(ex)) { send_sig(SIGKILL, current, 0); return error; } - down_write(¤t->mm->mmap_sem); - error = do_mmap(bprm->file, N_DATADDR(ex), ex.a_data, + error = vm_mmap(bprm->file, N_DATADDR(ex), ex.a_data, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE | MAP_EXECUTABLE, fd_offset + ex.a_text); - up_write(¤t->mm->mmap_sem); if (error != N_DATADDR(ex)) { send_sig(SIGKILL, current, 0); return error; @@ -417,12 +413,10 @@ static int load_aout_library(struct file *file) goto out; } /* Now use mmap to map the library into memory. */ - down_write(¤t->mm->mmap_sem); - error = do_mmap(file, start_addr, ex.a_text + ex.a_data, + error = vm_mmap(file, start_addr, ex.a_text + ex.a_data, PROT_READ | PROT_WRITE | PROT_EXEC, MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE, N_TXTOFF(ex)); - up_write(¤t->mm->mmap_sem); retval = error; if (error != start_addr) goto out; diff --git a/fs/binfmt_elf.c b/fs/binfmt_elf.c index 0708a0bf0ba..16f73541707 100644 --- a/fs/binfmt_elf.c +++ b/fs/binfmt_elf.c @@ -958,10 +958,8 @@ static int load_elf_binary(struct linux_binprm *bprm, struct pt_regs *regs) and some applications "depend" upon this behavior. Since we do not have the power to recompile these, we emulate the SVr4 behavior. Sigh. */ - down_write(¤t->mm->mmap_sem); - error = do_mmap(NULL, 0, PAGE_SIZE, PROT_READ | PROT_EXEC, + error = vm_mmap(NULL, 0, PAGE_SIZE, PROT_READ | PROT_EXEC, MAP_FIXED | MAP_PRIVATE, 0); - up_write(¤t->mm->mmap_sem); } #ifdef ELF_PLAT_INIT @@ -1046,8 +1044,7 @@ static int load_elf_library(struct file *file) eppnt++; /* Now use mmap to map the library into memory. */ - down_write(¤t->mm->mmap_sem); - error = do_mmap(file, + error = vm_mmap(file, ELF_PAGESTART(eppnt->p_vaddr), (eppnt->p_filesz + ELF_PAGEOFFSET(eppnt->p_vaddr)), @@ -1055,7 +1052,6 @@ static int load_elf_library(struct file *file) MAP_FIXED | MAP_PRIVATE | MAP_DENYWRITE, (eppnt->p_offset - ELF_PAGEOFFSET(eppnt->p_vaddr))); - up_write(¤t->mm->mmap_sem); if (error != ELF_PAGESTART(eppnt->p_vaddr)) goto out_free_ph; diff --git a/fs/binfmt_elf_fdpic.c b/fs/binfmt_elf_fdpic.c index 9bd5612a822..d390a0fffc6 100644 --- a/fs/binfmt_elf_fdpic.c +++ b/fs/binfmt_elf_fdpic.c @@ -390,21 +390,17 @@ static int load_elf_fdpic_binary(struct linux_binprm *bprm, (executable_stack == EXSTACK_DEFAULT && VM_STACK_FLAGS & VM_EXEC)) stack_prot |= PROT_EXEC; - down_write(¤t->mm->mmap_sem); - current->mm->start_brk = do_mmap(NULL, 0, stack_size, stack_prot, + current->mm->start_brk = vm_mmap(NULL, 0, stack_size, stack_prot, MAP_PRIVATE | MAP_ANONYMOUS | MAP_UNINITIALIZED | MAP_GROWSDOWN, 0); if (IS_ERR_VALUE(current->mm->start_brk)) { - up_write(¤t->mm->mmap_sem); retval = current->mm->start_brk; current->mm->start_brk = 0; goto error_kill; } - up_write(¤t->mm->mmap_sem); - current->mm->brk = current->mm->start_brk; current->mm->context.end_brk = current->mm->start_brk; current->mm->context.end_brk += @@ -955,10 +951,8 @@ static int elf_fdpic_map_file_constdisp_on_uclinux( if (params->flags & ELF_FDPIC_FLAG_EXECUTABLE) mflags |= MAP_EXECUTABLE; - down_write(&mm->mmap_sem); - maddr = do_mmap(NULL, load_addr, top - base, + maddr = vm_mmap(NULL, load_addr, top - base, PROT_READ | PROT_WRITE | PROT_EXEC, mflags, 0); - up_write(&mm->mmap_sem); if (IS_ERR_VALUE(maddr)) return (int) maddr; @@ -1096,10 +1090,8 @@ static int elf_fdpic_map_file_by_direct_mmap(struct elf_fdpic_params *params, /* create the mapping */ disp = phdr->p_vaddr & ~PAGE_MASK; - down_write(&mm->mmap_sem); - maddr = do_mmap(file, maddr, phdr->p_memsz + disp, prot, flags, + maddr = vm_mmap(file, maddr, phdr->p_memsz + disp, prot, flags, phdr->p_offset - disp); - up_write(&mm->mmap_sem); kdebug("mmap[%d] sz=%lx pr=%x fl=%x of=%lx --> %08lx", loop, phdr->p_memsz + disp, prot, flags, @@ -1143,10 +1135,8 @@ static int elf_fdpic_map_file_by_direct_mmap(struct elf_fdpic_params *params, unsigned long xmaddr; flags |= MAP_FIXED | MAP_ANONYMOUS; - down_write(&mm->mmap_sem); - xmaddr = do_mmap(NULL, xaddr, excess - excess1, + xmaddr = vm_mmap(NULL, xaddr, excess - excess1, prot, flags, 0); - up_write(&mm->mmap_sem); kdebug("mmap[%d] " " ad=%lx sz=%lx pr=%x fl=%x of=0 --> %08lx", diff --git a/fs/binfmt_flat.c b/fs/binfmt_flat.c index 024d20ee3ca..6b2daf99fab 100644 --- a/fs/binfmt_flat.c +++ b/fs/binfmt_flat.c @@ -542,10 +542,8 @@ static int load_flat_file(struct linux_binprm * bprm, */ DBG_FLT("BINFMT_FLAT: ROM mapping of file (we hope)\n"); - down_write(¤t->mm->mmap_sem); - textpos = do_mmap(bprm->file, 0, text_len, PROT_READ|PROT_EXEC, + textpos = vm_mmap(bprm->file, 0, text_len, PROT_READ|PROT_EXEC, MAP_PRIVATE|MAP_EXECUTABLE, 0); - up_write(¤t->mm->mmap_sem); if (!textpos || IS_ERR_VALUE(textpos)) { if (!textpos) textpos = (unsigned long) -ENOMEM; @@ -556,10 +554,8 @@ static int load_flat_file(struct linux_binprm * bprm, len = data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long); len = PAGE_ALIGN(len); - down_write(¤t->mm->mmap_sem); - realdatastart = do_mmap(0, 0, len, + realdatastart = vm_mmap(0, 0, len, PROT_READ|PROT_WRITE|PROT_EXEC, MAP_PRIVATE, 0); - up_write(¤t->mm->mmap_sem); if (realdatastart == 0 || IS_ERR_VALUE(realdatastart)) { if (!realdatastart) @@ -603,10 +599,8 @@ static int load_flat_file(struct linux_binprm * bprm, len = text_len + data_len + extra + MAX_SHARED_LIBS * sizeof(unsigned long); len = PAGE_ALIGN(len); - down_write(¤t->mm->mmap_sem); - textpos = do_mmap(0, 0, len, + textpos = vm_mmap(0, 0, len, PROT_READ | PROT_EXEC | PROT_WRITE, MAP_PRIVATE, 0); - up_write(¤t->mm->mmap_sem); if (!textpos || IS_ERR_VALUE(textpos)) { if (!textpos) diff --git a/fs/binfmt_som.c b/fs/binfmt_som.c index e4fc746629a..4517aaff61b 100644 --- a/fs/binfmt_som.c +++ b/fs/binfmt_som.c @@ -147,10 +147,8 @@ static int map_som_binary(struct file *file, code_size = SOM_PAGEALIGN(hpuxhdr->exec_tsize); current->mm->start_code = code_start; current->mm->end_code = code_start + code_size; - down_write(¤t->mm->mmap_sem); - retval = do_mmap(file, code_start, code_size, prot, + retval = vm_mmap(file, code_start, code_size, prot, flags, SOM_PAGESTART(hpuxhdr->exec_tfile)); - up_write(¤t->mm->mmap_sem); if (retval < 0 && retval > -1024) goto out; @@ -158,20 +156,16 @@ static int map_som_binary(struct file *file, data_size = SOM_PAGEALIGN(hpuxhdr->exec_dsize); current->mm->start_data = data_start; current->mm->end_data = bss_start = data_start + data_size; - down_write(¤t->mm->mmap_sem); - retval = do_mmap(file, data_start, data_size, + retval = vm_mmap(file, data_start, data_size, prot | PROT_WRITE, flags, SOM_PAGESTART(hpuxhdr->exec_dfile)); - up_write(¤t->mm->mmap_sem); if (retval < 0 && retval > -1024) goto out; som_brk = bss_start + SOM_PAGEALIGN(hpuxhdr->exec_bsize); current->mm->start_brk = current->mm->brk = som_brk; - down_write(¤t->mm->mmap_sem); - retval = do_mmap(NULL, bss_start, som_brk - bss_start, + retval = vm_mmap(NULL, bss_start, som_brk - bss_start, prot | PROT_WRITE, MAP_FIXED | MAP_PRIVATE, 0); - up_write(¤t->mm->mmap_sem); if (retval > 0 || retval < -1024) retval = 0; out: diff --git a/include/linux/mm.h b/include/linux/mm.h index cb61950a3aa..86a692c3b23 100644 --- a/include/linux/mm.h +++ b/include/linux/mm.h @@ -1393,31 +1393,20 @@ extern int install_special_mapping(struct mm_struct *mm, extern unsigned long get_unmapped_area(struct file *, unsigned long, unsigned long, unsigned long, unsigned long); -extern unsigned long do_mmap_pgoff(struct file *file, unsigned long addr, - unsigned long len, unsigned long prot, - unsigned long flag, unsigned long pgoff); extern unsigned long mmap_region(struct file *file, unsigned long addr, unsigned long len, unsigned long flags, vm_flags_t vm_flags, unsigned long pgoff); - -static inline unsigned long do_mmap(struct file *file, unsigned long addr, - unsigned long len, unsigned long prot, - unsigned long flag, unsigned long offset) -{ - unsigned long ret = -EINVAL; - if ((offset + PAGE_ALIGN(len)) < offset) - goto out; - if (!(offset & ~PAGE_MASK)) - ret = do_mmap_pgoff(file, addr, len, prot, flag, offset >> PAGE_SHIFT); -out: - return ret; -} - +extern unsigned long do_mmap(struct file *, unsigned long, + unsigned long, unsigned long, + unsigned long, unsigned long); extern int do_munmap(struct mm_struct *, unsigned long, size_t); /* These take the mm semaphore themselves */ extern unsigned long vm_brk(unsigned long, unsigned long); extern int vm_munmap(struct mm_struct *, unsigned long, size_t); +extern unsigned long vm_mmap(struct file *, unsigned long, + unsigned long, unsigned long, + unsigned long, unsigned long); /* truncate.c */ extern void truncate_inode_pages(struct address_space *, loff_t); diff --git a/mm/mmap.c b/mm/mmap.c index 4af45f519f1..b38b47ef1f7 100644 --- a/mm/mmap.c +++ b/mm/mmap.c @@ -953,7 +953,7 @@ static inline unsigned long round_hint_to_min(unsigned long hint) * The caller must hold down_write(¤t->mm->mmap_sem). */ -unsigned long do_mmap_pgoff(struct file *file, unsigned long addr, +static unsigned long do_mmap_pgoff(struct file *file, unsigned long addr, unsigned long len, unsigned long prot, unsigned long flags, unsigned long pgoff) { @@ -1089,7 +1089,32 @@ unsigned long do_mmap_pgoff(struct file *file, unsigned long addr, return mmap_region(file, addr, len, flags, vm_flags, pgoff); } -EXPORT_SYMBOL(do_mmap_pgoff); + +unsigned long do_mmap(struct file *file, unsigned long addr, + unsigned long len, unsigned long prot, + unsigned long flag, unsigned long offset) +{ + if (unlikely(offset + PAGE_ALIGN(len) < offset)) + return -EINVAL; + if (unlikely(offset & ~PAGE_MASK)) + return -EINVAL; + return do_mmap_pgoff(file, addr, len, prot, flag, offset >> PAGE_SHIFT); +} +EXPORT_SYMBOL(do_mmap); + +unsigned long vm_mmap(struct file *file, unsigned long addr, + unsigned long len, unsigned long prot, + unsigned long flag, unsigned long offset) +{ + unsigned long ret; + struct mm_struct *mm = current->mm; + + down_write(&mm->mmap_sem); + ret = do_mmap(file, addr, len, prot, flag, offset); + up_write(&mm->mmap_sem); + return ret; +} +EXPORT_SYMBOL(vm_mmap); SYSCALL_DEFINE6(mmap_pgoff, unsigned long, addr, unsigned long, len, unsigned long, prot, unsigned long, flags, diff --git a/mm/nommu.c b/mm/nommu.c index 11a69b22bd4..dd00383be2d 100644 --- a/mm/nommu.c +++ b/mm/nommu.c @@ -1233,7 +1233,7 @@ enomem: /* * handle mapping creation for uClinux */ -unsigned long do_mmap_pgoff(struct file *file, +static unsigned long do_mmap_pgoff(struct file *file, unsigned long addr, unsigned long len, unsigned long prot, @@ -1470,7 +1470,32 @@ error_getting_region: show_free_areas(0); return -ENOMEM; } -EXPORT_SYMBOL(do_mmap_pgoff); + +unsigned long do_mmap(struct file *file, unsigned long addr, + unsigned long len, unsigned long prot, + unsigned long flag, unsigned long offset) +{ + if (unlikely(offset + PAGE_ALIGN(len) < offset)) + return -EINVAL; + if (unlikely(offset & ~PAGE_MASK)) + return -EINVAL; + return do_mmap_pgoff(file, addr, len, prot, flag, offset >> PAGE_SHIFT); +} +EXPORT_SYMBOL(do_mmap); + +unsigned long vm_mmap(struct file *file, unsigned long addr, + unsigned long len, unsigned long prot, + unsigned long flag, unsigned long offset) +{ + unsigned long ret; + struct mm_struct *mm = current->mm; + + down_write(&mm->mmap_sem); + ret = do_mmap(file, addr, len, prot, flag, offset); + up_write(&mm->mmap_sem); + return ret; +} +EXPORT_SYMBOL(vm_mmap); SYSCALL_DEFINE6(mmap_pgoff, unsigned long, addr, unsigned long, len, unsigned long, prot, unsigned long, flags, -- cgit v1.2.3-70-g09d2 From 05ef1b79d46347f94d9a78214cc745046c03e45a Mon Sep 17 00:00:00 2001 From: Chris Metcalf Date: Wed, 25 Apr 2012 12:45:26 -0400 Subject: arch/tile: fix a couple of functions that should be __init They were marked __devinit by mistake, causing some warnings at link time. Signed-off-by: Chris Metcalf --- arch/tile/include/asm/pci.h | 4 ++-- arch/tile/kernel/pci.c | 4 ++-- 2 files changed, 4 insertions(+), 4 deletions(-) (limited to 'arch/tile/kernel') diff --git a/arch/tile/include/asm/pci.h b/arch/tile/include/asm/pci.h index 5d5a635530b..32e6cbe8dff 100644 --- a/arch/tile/include/asm/pci.h +++ b/arch/tile/include/asm/pci.h @@ -47,8 +47,8 @@ struct pci_controller { */ #define PCI_DMA_BUS_IS_PHYS 1 -int __devinit tile_pci_init(void); -int __devinit pcibios_init(void); +int __init tile_pci_init(void); +int __init pcibios_init(void); static inline void pci_iounmap(struct pci_dev *dev, void __iomem *addr) {} diff --git a/arch/tile/kernel/pci.c b/arch/tile/kernel/pci.c index a1bb59eecc1..b56d12bf590 100644 --- a/arch/tile/kernel/pci.c +++ b/arch/tile/kernel/pci.c @@ -141,7 +141,7 @@ static int __devinit tile_init_irqs(int controller_id, * * Returns the number of controllers discovered. */ -int __devinit tile_pci_init(void) +int __init tile_pci_init(void) { int i; @@ -287,7 +287,7 @@ static void __devinit fixup_read_and_payload_sizes(void) * The controllers have been set up by the time we get here, by a call to * tile_pci_init. */ -int __devinit pcibios_init(void) +int __init pcibios_init(void) { int i; -- cgit v1.2.3-70-g09d2