From 88d60c32765716289abeb362c44adf6c35c6824c Mon Sep 17 00:00:00 2001 From: Eric Paris Date: Mon, 8 Nov 2010 18:19:22 -0500 Subject: fanotify: remove packed from access response message Since fanotify has decided to be careful about alignment and packing rather than rely on __attribute__((packed)) for multiarch support. Since this attribute isn't doing anything on fanotify_response we just drop it. This does not break API/ABI. Suggested-by: Tvrtko Ursulin Signed-off-by: Eric Paris --- include/linux/fanotify.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'include/linux') diff --git a/include/linux/fanotify.h b/include/linux/fanotify.h index 0f0121467fc4..bdbf9bb29b54 100644 --- a/include/linux/fanotify.h +++ b/include/linux/fanotify.h @@ -96,7 +96,7 @@ struct fanotify_event_metadata { struct fanotify_response { __s32 fd; __u32 response; -} __attribute__ ((packed)); +}; /* Legit userspace responses to a _PERM event */ #define FAN_ALLOW 0x01 -- cgit v1.2.3-71-gd317 From b1085ba80cd2784400a7beec3fda5099198ed01c Mon Sep 17 00:00:00 2001 From: Lino Sanfilippo Date: Fri, 5 Nov 2010 17:05:27 +0100 Subject: fanotify: if set by user unset FMODE_NONOTIFY before fsnotify_perm() is called Unsetting FMODE_NONOTIFY in fsnotify_open() is too late, since fsnotify_perm() is called before. If FMODE_NONOTIFY is set fsnotify_perm() will skip permission checks, so a user can still disable permission checks by setting this flag in an open() call. This patch corrects this by unsetting the flag before fsnotify_perm is called. Signed-off-by: Lino Sanfilippo Signed-off-by: Eric Paris --- fs/namei.c | 3 +++ include/linux/fsnotify.h | 3 --- 2 files changed, 3 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/fs/namei.c b/fs/namei.c index 5362af9b7372..4ff7ca530533 100644 --- a/fs/namei.c +++ b/fs/namei.c @@ -1748,6 +1748,9 @@ struct file *do_filp_open(int dfd, const char *pathname, if (!(open_flag & O_CREAT)) mode = 0; + /* Must never be set by userspace */ + open_flag &= ~FMODE_NONOTIFY; + /* * O_SYNC is implemented as __O_SYNC|O_DSYNC. As many places only * check for O_DSYNC if the need any syncing at all we enforce it's diff --git a/include/linux/fsnotify.h b/include/linux/fsnotify.h index 5c185fa27089..b10bcdeaef76 100644 --- a/include/linux/fsnotify.h +++ b/include/linux/fsnotify.h @@ -235,9 +235,6 @@ static inline void fsnotify_open(struct file *file) if (S_ISDIR(inode->i_mode)) mask |= FS_ISDIR; - /* FMODE_NONOTIFY must never be set from user */ - file->f_mode &= ~FMODE_NONOTIFY; - fsnotify_parent(path, NULL, mask); fsnotify(inode, mask, path, FSNOTIFY_EVENT_PATH, NULL, 0); } -- cgit v1.2.3-71-gd317 From 09e5f14e57c70f9d357862bb56e57026c51092a1 Mon Sep 17 00:00:00 2001 From: Lino Sanfilippo Date: Fri, 19 Nov 2010 10:58:07 +0100 Subject: fanotify: on group destroy allow all waiters to bypass permission check When fanotify_release() is called, there may still be processes waiting for access permission. Currently only processes for which an event has already been queued into the groups access list will be woken up. Processes for which no event has been queued will continue to sleep and thus cause a deadlock when fsnotify_put_group() is called. Furthermore there is a race allowing further processes to be waiting on the access wait queue after wake_up (if they arrive before clear_marks_by_group() is called). This patch corrects this by setting a flag to inform processes that the group is about to be destroyed and thus not to wait for access permission. [additional changelog from eparis] Lets think about the 4 relevant code paths from the PoV of the 'operator' 'listener' 'responder' and 'closer'. Where operator is the process doing an action (like open/read) which could require permission. Listener is the task (or in this case thread) slated with reading from the fanotify file descriptor. The 'responder' is the thread responsible for responding to access requests. 'Closer' is the thread attempting to close the fanotify file descriptor. The 'operator' is going to end up in: fanotify_handle_event() get_response_from_access() (THIS BLOCKS WAITING ON USERSPACE) The 'listener' interesting code path fanotify_read() copy_event_to_user() prepare_for_access_response() (THIS CREATES AN fanotify_response_event) The 'responder' code path: fanotify_write() process_access_response() (REMOVE A fanotify_response_event, SET RESPONSE, WAKE UP 'operator') The 'closer': fanotify_release() (SUPPOSED TO CLEAN UP THE REST OF THIS MESS) What we have today is that in the closer we remove all of the fanotify_response_events and set a bit so no more response events are ever created in prepare_for_access_response(). The bug is that we never wake all of the operators up and tell them to move along. You fix that in fanotify_get_response_from_access(). You also fix other operators which haven't gotten there yet. So I agree that's a good fix. [/additional changelog from eparis] [remove additional changes to minimize patch size] [move initialization so it was inside CONFIG_FANOTIFY_PERMISSION] Signed-off-by: Lino Sanfilippo Signed-off-by: Eric Paris --- fs/notify/fanotify/fanotify.c | 6 +++++- fs/notify/fanotify/fanotify_user.c | 5 +++-- include/linux/fsnotify_backend.h | 2 +- 3 files changed, 9 insertions(+), 4 deletions(-) (limited to 'include/linux') diff --git a/fs/notify/fanotify/fanotify.c b/fs/notify/fanotify/fanotify.c index b04f88eed09e..f35794b97e8e 100644 --- a/fs/notify/fanotify/fanotify.c +++ b/fs/notify/fanotify/fanotify.c @@ -92,7 +92,11 @@ static int fanotify_get_response_from_access(struct fsnotify_group *group, pr_debug("%s: group=%p event=%p\n", __func__, group, event); - wait_event(group->fanotify_data.access_waitq, event->response); + wait_event(group->fanotify_data.access_waitq, event->response || + atomic_read(&group->fanotify_data.bypass_perm)); + + if (!event->response) /* bypass_perm set */ + return 0; /* userspace responded, convert to something usable */ spin_lock(&event->lock); diff --git a/fs/notify/fanotify/fanotify_user.c b/fs/notify/fanotify/fanotify_user.c index 480434c5ee5f..01fffe62a2d4 100644 --- a/fs/notify/fanotify/fanotify_user.c +++ b/fs/notify/fanotify/fanotify_user.c @@ -200,7 +200,7 @@ static int prepare_for_access_response(struct fsnotify_group *group, mutex_lock(&group->fanotify_data.access_mutex); - if (group->fanotify_data.bypass_perm) { + if (atomic_read(&group->fanotify_data.bypass_perm)) { mutex_unlock(&group->fanotify_data.access_mutex); kmem_cache_free(fanotify_response_event_cache, re); event->response = FAN_ALLOW; @@ -390,7 +390,7 @@ static int fanotify_release(struct inode *ignored, struct file *file) mutex_lock(&group->fanotify_data.access_mutex); - group->fanotify_data.bypass_perm = true; + atomic_inc(&group->fanotify_data.bypass_perm); list_for_each_entry_safe(re, lre, &group->fanotify_data.access_list, list) { pr_debug("%s: found group=%p re=%p event=%p\n", __func__, group, @@ -703,6 +703,7 @@ SYSCALL_DEFINE2(fanotify_init, unsigned int, flags, unsigned int, event_f_flags) mutex_init(&group->fanotify_data.access_mutex); init_waitqueue_head(&group->fanotify_data.access_waitq); INIT_LIST_HEAD(&group->fanotify_data.access_list); + atomic_set(&group->fanotify_data.bypass_perm, 0); #endif switch (flags & FAN_ALL_CLASS_BITS) { case FAN_CLASS_NOTIF: diff --git a/include/linux/fsnotify_backend.h b/include/linux/fsnotify_backend.h index 0a68f924f06f..7380763595d3 100644 --- a/include/linux/fsnotify_backend.h +++ b/include/linux/fsnotify_backend.h @@ -166,7 +166,7 @@ struct fsnotify_group { struct mutex access_mutex; struct list_head access_list; wait_queue_head_t access_waitq; - bool bypass_perm; /* protected by access_mutex */ + atomic_t bypass_perm; #endif /* CONFIG_FANOTIFY_ACCESS_PERMISSIONS */ int f_flags; unsigned int max_marks; -- cgit v1.2.3-71-gd317 From e9a3854fd4ff3907e6c200a3980e19365ee695e9 Mon Sep 17 00:00:00 2001 From: Lino Sanfilippo Date: Wed, 24 Nov 2010 18:22:09 +0100 Subject: fanotify: Introduce FAN_NOFD FAN_NOFD is used in fanotify events that do not provide an open file descriptor (like the overflow_event). Signed-off-by: Lino Sanfilippo Signed-off-by: Eric Paris --- include/linux/fanotify.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/linux') diff --git a/include/linux/fanotify.h b/include/linux/fanotify.h index bdbf9bb29b54..c73224315aee 100644 --- a/include/linux/fanotify.h +++ b/include/linux/fanotify.h @@ -101,6 +101,8 @@ struct fanotify_response { /* Legit userspace responses to a _PERM event */ #define FAN_ALLOW 0x01 #define FAN_DENY 0x02 +/* No fd set in event */ +#define FAN_NOFD -1 /* Helper functions to deal with fanotify_event_metadata buffers */ #define FAN_EVENT_METADATA_LEN (sizeof(struct fanotify_event_metadata)) -- cgit v1.2.3-71-gd317 From 5167695753c63444a9e6cbbef136200a16c7a225 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 7 Dec 2010 14:18:20 +0100 Subject: perf: Fix duplicate events with multiple-pmu vs software events Because the multi-pmu bits can share contexts between struct pmu instances we could get duplicate events by iterating the pmu list. Signed-off-by: Peter Zijlstra Signed-off-by: Thomas Gleixner LKML-Reference: Signed-off-by: Ingo Molnar --- include/linux/perf_event.h | 1 + kernel/perf_event.c | 35 +++++++++++++++++++++++++++++------ 2 files changed, 30 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/perf_event.h b/include/linux/perf_event.h index de2c41758e29..4f1279e105ee 100644 --- a/include/linux/perf_event.h +++ b/include/linux/perf_event.h @@ -887,6 +887,7 @@ struct perf_cpu_context { int exclusive; struct list_head rotation_list; int jiffies_interval; + struct pmu *active_pmu; }; struct perf_output_handle { diff --git a/kernel/perf_event.c b/kernel/perf_event.c index eac7e3364335..7b870174c56d 100644 --- a/kernel/perf_event.c +++ b/kernel/perf_event.c @@ -3824,6 +3824,8 @@ static void perf_event_task_event(struct perf_task_event *task_event) rcu_read_lock(); list_for_each_entry_rcu(pmu, &pmus, entry) { cpuctx = get_cpu_ptr(pmu->pmu_cpu_context); + if (cpuctx->active_pmu != pmu) + goto next; perf_event_task_ctx(&cpuctx->ctx, task_event); ctx = task_event->task_ctx; @@ -3959,6 +3961,8 @@ static void perf_event_comm_event(struct perf_comm_event *comm_event) rcu_read_lock(); list_for_each_entry_rcu(pmu, &pmus, entry) { cpuctx = get_cpu_ptr(pmu->pmu_cpu_context); + if (cpuctx->active_pmu != pmu) + goto next; perf_event_comm_ctx(&cpuctx->ctx, comm_event); ctxn = pmu->task_ctx_nr; @@ -4144,6 +4148,8 @@ got_name: rcu_read_lock(); list_for_each_entry_rcu(pmu, &pmus, entry) { cpuctx = get_cpu_ptr(pmu->pmu_cpu_context); + if (cpuctx->active_pmu != pmu) + goto next; perf_event_mmap_ctx(&cpuctx->ctx, mmap_event, vma->vm_flags & VM_EXEC); @@ -5145,20 +5151,36 @@ static void *find_pmu_context(int ctxn) return NULL; } -static void free_pmu_context(void * __percpu cpu_context) +static void update_pmu_context(struct pmu *pmu, struct pmu *old_pmu) { - struct pmu *pmu; + int cpu; + + for_each_possible_cpu(cpu) { + struct perf_cpu_context *cpuctx; + + cpuctx = per_cpu_ptr(pmu->pmu_cpu_context, cpu); + + if (cpuctx->active_pmu == old_pmu) + cpuctx->active_pmu = pmu; + } +} + +static void free_pmu_context(struct pmu *pmu) +{ + struct pmu *i; mutex_lock(&pmus_lock); /* * Like a real lame refcount. */ - list_for_each_entry(pmu, &pmus, entry) { - if (pmu->pmu_cpu_context == cpu_context) + list_for_each_entry(i, &pmus, entry) { + if (i->pmu_cpu_context == pmu->pmu_cpu_context) { + update_pmu_context(i, pmu); goto out; + } } - free_percpu(cpu_context); + free_percpu(pmu->pmu_cpu_context); out: mutex_unlock(&pmus_lock); } @@ -5190,6 +5212,7 @@ int perf_pmu_register(struct pmu *pmu) cpuctx->ctx.pmu = pmu; cpuctx->jiffies_interval = 1; INIT_LIST_HEAD(&cpuctx->rotation_list); + cpuctx->active_pmu = pmu; } got_cpu_context: @@ -5241,7 +5264,7 @@ void perf_pmu_unregister(struct pmu *pmu) synchronize_rcu(); free_percpu(pmu->pmu_disable_count); - free_pmu_context(pmu->pmu_cpu_context); + free_pmu_context(pmu); } struct pmu *perf_init_event(struct perf_event *event) -- cgit v1.2.3-71-gd317 From 0f004f5a696a9434b7214d0d3cbd0525ee77d428 Mon Sep 17 00:00:00 2001 From: Peter Zijlstra Date: Tue, 30 Nov 2010 19:48:45 +0100 Subject: sched: Cure more NO_HZ load average woes There's a long-running regression that proved difficult to fix and which is hitting certain people and is rather annoying in its effects. Damien reported that after 74f5187ac8 (sched: Cure load average vs NO_HZ woes) his load average is unnaturally high, he also noted that even with that patch reverted the load avgerage numbers are not correct. The problem is that the previous patch only solved half the NO_HZ problem, it addressed the part of going into NO_HZ mode, not of comming out of NO_HZ mode. This patch implements that missing half. When comming out of NO_HZ mode there are two important things to take care of: - Folding the pending idle delta into the global active count. - Correctly aging the averages for the idle-duration. So with this patch the NO_HZ interaction should be complete and behaviour between CONFIG_NO_HZ=[yn] should be equivalent. Furthermore, this patch slightly changes the load average computation by adding a rounding term to the fixed point multiplication. Reported-by: Damien Wyart Reported-by: Tim McGrath Tested-by: Damien Wyart Tested-by: Orion Poplawski Tested-by: Kyle McMartin Signed-off-by: Peter Zijlstra Cc: stable@kernel.org Cc: Chase Douglas LKML-Reference: <1291129145.32004.874.camel@laptop> Signed-off-by: Ingo Molnar --- include/linux/sched.h | 2 +- kernel/sched.c | 150 ++++++++++++++++++++++++++++++++++++++++++++++---- kernel/timer.c | 2 +- 3 files changed, 141 insertions(+), 13 deletions(-) (limited to 'include/linux') diff --git a/include/linux/sched.h b/include/linux/sched.h index 2c79e921a68b..223874538b33 100644 --- a/include/linux/sched.h +++ b/include/linux/sched.h @@ -143,7 +143,7 @@ extern unsigned long nr_iowait_cpu(int cpu); extern unsigned long this_cpu_load(void); -extern void calc_global_load(void); +extern void calc_global_load(unsigned long ticks); extern unsigned long get_parent_ip(unsigned long addr); diff --git a/kernel/sched.c b/kernel/sched.c index dc91a4d09ac3..6b7c26a1a097 100644 --- a/kernel/sched.c +++ b/kernel/sched.c @@ -3119,6 +3119,15 @@ static long calc_load_fold_active(struct rq *this_rq) return delta; } +static unsigned long +calc_load(unsigned long load, unsigned long exp, unsigned long active) +{ + load *= exp; + load += active * (FIXED_1 - exp); + load += 1UL << (FSHIFT - 1); + return load >> FSHIFT; +} + #ifdef CONFIG_NO_HZ /* * For NO_HZ we delay the active fold to the next LOAD_FREQ update. @@ -3148,6 +3157,128 @@ static long calc_load_fold_idle(void) return delta; } + +/** + * fixed_power_int - compute: x^n, in O(log n) time + * + * @x: base of the power + * @frac_bits: fractional bits of @x + * @n: power to raise @x to. + * + * By exploiting the relation between the definition of the natural power + * function: x^n := x*x*...*x (x multiplied by itself for n times), and + * the binary encoding of numbers used by computers: n := \Sum n_i * 2^i, + * (where: n_i \elem {0, 1}, the binary vector representing n), + * we find: x^n := x^(\Sum n_i * 2^i) := \Prod x^(n_i * 2^i), which is + * of course trivially computable in O(log_2 n), the length of our binary + * vector. + */ +static unsigned long +fixed_power_int(unsigned long x, unsigned int frac_bits, unsigned int n) +{ + unsigned long result = 1UL << frac_bits; + + if (n) for (;;) { + if (n & 1) { + result *= x; + result += 1UL << (frac_bits - 1); + result >>= frac_bits; + } + n >>= 1; + if (!n) + break; + x *= x; + x += 1UL << (frac_bits - 1); + x >>= frac_bits; + } + + return result; +} + +/* + * a1 = a0 * e + a * (1 - e) + * + * a2 = a1 * e + a * (1 - e) + * = (a0 * e + a * (1 - e)) * e + a * (1 - e) + * = a0 * e^2 + a * (1 - e) * (1 + e) + * + * a3 = a2 * e + a * (1 - e) + * = (a0 * e^2 + a * (1 - e) * (1 + e)) * e + a * (1 - e) + * = a0 * e^3 + a * (1 - e) * (1 + e + e^2) + * + * ... + * + * an = a0 * e^n + a * (1 - e) * (1 + e + ... + e^n-1) [1] + * = a0 * e^n + a * (1 - e) * (1 - e^n)/(1 - e) + * = a0 * e^n + a * (1 - e^n) + * + * [1] application of the geometric series: + * + * n 1 - x^(n+1) + * S_n := \Sum x^i = ------------- + * i=0 1 - x + */ +static unsigned long +calc_load_n(unsigned long load, unsigned long exp, + unsigned long active, unsigned int n) +{ + + return calc_load(load, fixed_power_int(exp, FSHIFT, n), active); +} + +/* + * NO_HZ can leave us missing all per-cpu ticks calling + * calc_load_account_active(), but since an idle CPU folds its delta into + * calc_load_tasks_idle per calc_load_account_idle(), all we need to do is fold + * in the pending idle delta if our idle period crossed a load cycle boundary. + * + * Once we've updated the global active value, we need to apply the exponential + * weights adjusted to the number of cycles missed. + */ +static void calc_global_nohz(unsigned long ticks) +{ + long delta, active, n; + + if (time_before(jiffies, calc_load_update)) + return; + + /* + * If we crossed a calc_load_update boundary, make sure to fold + * any pending idle changes, the respective CPUs might have + * missed the tick driven calc_load_account_active() update + * due to NO_HZ. + */ + delta = calc_load_fold_idle(); + if (delta) + atomic_long_add(delta, &calc_load_tasks); + + /* + * If we were idle for multiple load cycles, apply them. + */ + if (ticks >= LOAD_FREQ) { + n = ticks / LOAD_FREQ; + + active = atomic_long_read(&calc_load_tasks); + active = active > 0 ? active * FIXED_1 : 0; + + avenrun[0] = calc_load_n(avenrun[0], EXP_1, active, n); + avenrun[1] = calc_load_n(avenrun[1], EXP_5, active, n); + avenrun[2] = calc_load_n(avenrun[2], EXP_15, active, n); + + calc_load_update += n * LOAD_FREQ; + } + + /* + * Its possible the remainder of the above division also crosses + * a LOAD_FREQ period, the regular check in calc_global_load() + * which comes after this will take care of that. + * + * Consider us being 11 ticks before a cycle completion, and us + * sleeping for 4*LOAD_FREQ + 22 ticks, then the above code will + * age us 4 cycles, and the test in calc_global_load() will + * pick up the final one. + */ +} #else static void calc_load_account_idle(struct rq *this_rq) { @@ -3157,6 +3288,10 @@ static inline long calc_load_fold_idle(void) { return 0; } + +static void calc_global_nohz(unsigned long ticks) +{ +} #endif /** @@ -3174,24 +3309,17 @@ void get_avenrun(unsigned long *loads, unsigned long offset, int shift) loads[2] = (avenrun[2] + offset) << shift; } -static unsigned long -calc_load(unsigned long load, unsigned long exp, unsigned long active) -{ - load *= exp; - load += active * (FIXED_1 - exp); - return load >> FSHIFT; -} - /* * calc_load - update the avenrun load estimates 10 ticks after the * CPUs have updated calc_load_tasks. */ -void calc_global_load(void) +void calc_global_load(unsigned long ticks) { - unsigned long upd = calc_load_update + 10; long active; - if (time_before(jiffies, upd)) + calc_global_nohz(ticks); + + if (time_before(jiffies, calc_load_update + 10)) return; active = atomic_long_read(&calc_load_tasks); diff --git a/kernel/timer.c b/kernel/timer.c index 68a9ae7679b7..7bd715fda974 100644 --- a/kernel/timer.c +++ b/kernel/timer.c @@ -1319,7 +1319,7 @@ void do_timer(unsigned long ticks) { jiffies_64 += ticks; update_wall_time(); - calc_global_load(); + calc_global_load(ticks); } #ifdef __ARCH_WANT_SYS_ALARM -- cgit v1.2.3-71-gd317 From 53dde5f385bc56e312f78b7cb25ffaf8efd4735d Mon Sep 17 00:00:00 2001 From: Suresh Siddha Date: Tue, 16 Nov 2010 13:23:50 -0800 Subject: bootmem: Add alloc_bootmem_align() Add an alloc_bootmem_align() interface to allocate bootmem with specified alignment. This is necessary to be able to allocate the xsave area in a subsequent patch. Signed-off-by: Suresh Siddha LKML-Reference: <20101116212441.977574826@sbsiddha-MOBL3.sc.intel.com> Acked-by: H. Peter Anvin Signed-off-by: H. Peter Anvin Cc: --- include/linux/bootmem.h | 2 ++ 1 file changed, 2 insertions(+) (limited to 'include/linux') diff --git a/include/linux/bootmem.h b/include/linux/bootmem.h index 266ab9291232..499dfe982a0e 100644 --- a/include/linux/bootmem.h +++ b/include/linux/bootmem.h @@ -105,6 +105,8 @@ extern void *__alloc_bootmem_low_node(pg_data_t *pgdat, #define alloc_bootmem(x) \ __alloc_bootmem(x, SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS)) +#define alloc_bootmem_align(x, align) \ + __alloc_bootmem(x, align, __pa(MAX_DMA_ADDRESS)) #define alloc_bootmem_nopanic(x) \ __alloc_bootmem_nopanic(x, SMP_CACHE_BYTES, __pa(MAX_DMA_ADDRESS)) #define alloc_bootmem_pages(x) \ -- cgit v1.2.3-71-gd317 From ab4e0192196b8d4e43a3945742d4996da934a86f Mon Sep 17 00:00:00 2001 From: Dmitry Torokhov Date: Tue, 14 Dec 2010 23:53:21 -0800 Subject: Input: define separate EVIOCGKEYCODE_V2/EVIOCSKEYCODE_V2 The desire to keep old names for the EVIOCGKEYCODE/EVIOCSKEYCODE while extending them to support large scancodes was a mistake. While we tried to keep ABI intact (and we succeeded in doing that, programs compiled on older kernels will work on newer ones) there is still a problem with recompiling existing software with newer kernel headers. New kernel headers will supply updated ioctl numbers and kernel will expect that userspace will use struct input_keymap_entry to set and retrieve keymap data. But since the names of ioctls are still the same userspace will happily compile even if not adjusted to make use of the new structure and will start miraculously fail in the field. To avoid this issue let's revert EVIOCGKEYCODE/EVIOCSKEYCODE definitions and add EVIOCGKEYCODE_V2/EVIOCSKEYCODE_V2 so that userspace can explicitly select the style of ioctls it wants to employ. Reviewed-by: Henrik Rydberg Acked-by: Jarod Wilson Acked-by: Mauro Carvalho Chehab Signed-off-by: Dmitry Torokhov --- drivers/input/evdev.c | 113 ++++++++++++++++++++++++++------------------------ include/linux/input.h | 6 ++- 2 files changed, 62 insertions(+), 57 deletions(-) (limited to 'include/linux') diff --git a/drivers/input/evdev.c b/drivers/input/evdev.c index e3f7fc6f9565..68f09a868434 100644 --- a/drivers/input/evdev.c +++ b/drivers/input/evdev.c @@ -534,76 +534,73 @@ static int handle_eviocgbit(struct input_dev *dev, } #undef OLD_KEY_MAX -static int evdev_handle_get_keycode(struct input_dev *dev, - void __user *p, size_t size) +static int evdev_handle_get_keycode(struct input_dev *dev, void __user *p) { - struct input_keymap_entry ke; + struct input_keymap_entry ke = { + .len = sizeof(unsigned int), + .flags = 0, + }; + int __user *ip = (int __user *)p; int error; - memset(&ke, 0, sizeof(ke)); - - if (size == sizeof(unsigned int[2])) { - /* legacy case */ - int __user *ip = (int __user *)p; + /* legacy case */ + if (copy_from_user(ke.scancode, p, sizeof(unsigned int))) + return -EFAULT; - if (copy_from_user(ke.scancode, p, sizeof(unsigned int))) - return -EFAULT; + error = input_get_keycode(dev, &ke); + if (error) + return error; - ke.len = sizeof(unsigned int); - ke.flags = 0; + if (put_user(ke.keycode, ip + 1)) + return -EFAULT; - error = input_get_keycode(dev, &ke); - if (error) - return error; + return 0; +} - if (put_user(ke.keycode, ip + 1)) - return -EFAULT; +static int evdev_handle_get_keycode_v2(struct input_dev *dev, void __user *p) +{ + struct input_keymap_entry ke; + int error; - } else { - size = min(size, sizeof(ke)); + if (copy_from_user(&ke, p, sizeof(ke))) + return -EFAULT; - if (copy_from_user(&ke, p, size)) - return -EFAULT; + error = input_get_keycode(dev, &ke); + if (error) + return error; - error = input_get_keycode(dev, &ke); - if (error) - return error; + if (copy_to_user(p, &ke, sizeof(ke))) + return -EFAULT; - if (copy_to_user(p, &ke, size)) - return -EFAULT; - } return 0; } -static int evdev_handle_set_keycode(struct input_dev *dev, - void __user *p, size_t size) +static int evdev_handle_set_keycode(struct input_dev *dev, void __user *p) { - struct input_keymap_entry ke; - - memset(&ke, 0, sizeof(ke)); + struct input_keymap_entry ke = { + .len = sizeof(unsigned int), + .flags = 0, + }; + int __user *ip = (int __user *)p; - if (size == sizeof(unsigned int[2])) { - /* legacy case */ - int __user *ip = (int __user *)p; + if (copy_from_user(ke.scancode, p, sizeof(unsigned int))) + return -EFAULT; - if (copy_from_user(ke.scancode, p, sizeof(unsigned int))) - return -EFAULT; + if (get_user(ke.keycode, ip + 1)) + return -EFAULT; - if (get_user(ke.keycode, ip + 1)) - return -EFAULT; + return input_set_keycode(dev, &ke); +} - ke.len = sizeof(unsigned int); - ke.flags = 0; +static int evdev_handle_set_keycode_v2(struct input_dev *dev, void __user *p) +{ + struct input_keymap_entry ke; - } else { - size = min(size, sizeof(ke)); + if (copy_from_user(&ke, p, sizeof(ke))) + return -EFAULT; - if (copy_from_user(&ke, p, size)) - return -EFAULT; - - if (ke.len > sizeof(ke.scancode)) - return -EINVAL; - } + if (ke.len > sizeof(ke.scancode)) + return -EINVAL; return input_set_keycode(dev, &ke); } @@ -669,6 +666,18 @@ static long evdev_do_ioctl(struct file *file, unsigned int cmd, return evdev_grab(evdev, client); else return evdev_ungrab(evdev, client); + + case EVIOCGKEYCODE: + return evdev_handle_get_keycode(dev, p); + + case EVIOCSKEYCODE: + return evdev_handle_set_keycode(dev, p); + + case EVIOCGKEYCODE_V2: + return evdev_handle_get_keycode_v2(dev, p); + + case EVIOCSKEYCODE_V2: + return evdev_handle_set_keycode_v2(dev, p); } size = _IOC_SIZE(cmd); @@ -708,12 +717,6 @@ static long evdev_do_ioctl(struct file *file, unsigned int cmd, return -EFAULT; return error; - - case EVIOC_MASK_SIZE(EVIOCGKEYCODE): - return evdev_handle_get_keycode(dev, p, size); - - case EVIOC_MASK_SIZE(EVIOCSKEYCODE): - return evdev_handle_set_keycode(dev, p, size); } /* Multi-number variable-length handlers */ diff --git a/include/linux/input.h b/include/linux/input.h index a8af21d42bc1..9777668883be 100644 --- a/include/linux/input.h +++ b/include/linux/input.h @@ -104,8 +104,10 @@ struct input_keymap_entry { #define EVIOCGREP _IOR('E', 0x03, unsigned int[2]) /* get repeat settings */ #define EVIOCSREP _IOW('E', 0x03, unsigned int[2]) /* set repeat settings */ -#define EVIOCGKEYCODE _IOR('E', 0x04, struct input_keymap_entry) /* get keycode */ -#define EVIOCSKEYCODE _IOW('E', 0x04, struct input_keymap_entry) /* set keycode */ +#define EVIOCGKEYCODE _IOR('E', 0x04, unsigned int[2]) /* get keycode */ +#define EVIOCGKEYCODE_V2 _IOR('E', 0x04, struct input_keymap_entry) +#define EVIOCSKEYCODE _IOW('E', 0x04, unsigned int[2]) /* set keycode */ +#define EVIOCSKEYCODE_V2 _IOW('E', 0x04, struct input_keymap_entry) #define EVIOCGNAME(len) _IOC(_IOC_READ, 'E', 0x06, len) /* get device name */ #define EVIOCGPHYS(len) _IOC(_IOC_READ, 'E', 0x07, len) /* get physical location */ -- cgit v1.2.3-71-gd317 From 62731fa0c893515dc6cbc3e0a2879a92793c735f Mon Sep 17 00:00:00 2001 From: Alexey Zaytsev Date: Mon, 22 Nov 2010 00:33:03 +0000 Subject: fanotify: split version into version and metadata_len To implement per event type optional headers we are interested in knowing how long the metadata structure is. This patch slits the __u32 version field into a __u8 version and a __u16 metadata_len field (with __u8 left over). This should allow for backwards compat ABI. Signed-off-by: Alexey Zaytsev [rewrote descrtion and changed object sizes and ordering - eparis] Signed-off-by: Eric Paris --- include/linux/fanotify.h | 6 ++++-- 1 file changed, 4 insertions(+), 2 deletions(-) (limited to 'include/linux') diff --git a/include/linux/fanotify.h b/include/linux/fanotify.h index c73224315aee..6c6133f76e16 100644 --- a/include/linux/fanotify.h +++ b/include/linux/fanotify.h @@ -83,11 +83,13 @@ FAN_ALL_PERM_EVENTS |\ FAN_Q_OVERFLOW) -#define FANOTIFY_METADATA_VERSION 2 +#define FANOTIFY_METADATA_VERSION 3 struct fanotify_event_metadata { __u32 event_len; - __u32 vers; + __u8 vers; + __u8 reserved; + __u16 metadata_len; __aligned_u64 mask; __s32 fd; __s32 pid; -- cgit v1.2.3-71-gd317 From f08f5a0add20834d3f3d876dfe08005a5df656db Mon Sep 17 00:00:00 2001 From: "Rafael J. Wysocki" Date: Thu, 16 Dec 2010 17:11:58 +0100 Subject: PM / Runtime: Fix pm_runtime_suspended() There are some situations (e.g. in __pm_generic_call()), where pm_runtime_suspended() is used to decide whether or not to execute a device's (system) ->suspend() callback. The callback is not executed if pm_runtime_suspended() returns true, but it does so for devices that don't even support runtime PM, because the power.disable_depth device field is ignored by it. This leads to problems (i.e. devices are not suspened when they should), so rework pm_runtime_suspended() so that it returns false if the device's power.disable_depth field is different from zero. Signed-off-by: Rafael J. Wysocki Cc: stable@kernel.org --- Documentation/power/runtime_pm.txt | 4 ++-- include/linux/pm_runtime.h | 3 ++- 2 files changed, 4 insertions(+), 3 deletions(-) (limited to 'include/linux') diff --git a/Documentation/power/runtime_pm.txt b/Documentation/power/runtime_pm.txt index 489e9bacd165..41cc7b30d7dd 100644 --- a/Documentation/power/runtime_pm.txt +++ b/Documentation/power/runtime_pm.txt @@ -379,8 +379,8 @@ drivers/base/power/runtime.c and include/linux/pm_runtime.h: zero) bool pm_runtime_suspended(struct device *dev); - - return true if the device's runtime PM status is 'suspended', or false - otherwise + - return true if the device's runtime PM status is 'suspended' and its + 'power.disable_depth' field is equal to zero, or false otherwise void pm_runtime_allow(struct device *dev); - set the power.runtime_auto flag for the device and decrease its usage diff --git a/include/linux/pm_runtime.h b/include/linux/pm_runtime.h index 3ec2358f8692..d19f1cca7f74 100644 --- a/include/linux/pm_runtime.h +++ b/include/linux/pm_runtime.h @@ -77,7 +77,8 @@ static inline void device_set_run_wake(struct device *dev, bool enable) static inline bool pm_runtime_suspended(struct device *dev) { - return dev->power.runtime_status == RPM_SUSPENDED; + return dev->power.runtime_status == RPM_SUSPENDED + && !dev->power.disable_depth; } static inline void pm_runtime_mark_last_busy(struct device *dev) -- cgit v1.2.3-71-gd317 From 3f84622d7c7818077f5e6cf4b8a0d1b10dc65147 Mon Sep 17 00:00:00 2001 From: Hauke Mehrtens Date: Sat, 27 Nov 2010 19:26:32 +0100 Subject: SSB: Fix nvram_get on BCM47xx platform The nvram_get function was never in the mainline kernel, it only existed in an external OpenWrt patch. Use nvram_getenv function, which is in mainline and use an include instead of an extra function declaration. et0macaddr contains the mac address in text from like 00:11:22:33:44:55. We have to parse it before adding it into macaddr. nvram_parse_macaddr will be merged into asm/mach-bcm47xx/nvram.h through the MIPS git tree and will be available soon. It will not build now without nvram_parse_macaddr, but it hasn't before either. Signed-off-by: Hauke Mehrtens To: linux-mips@linux-mips.org Cc: mb@bu3sch.de Cc: netdev@vger.kernel.org Cc: Hauke Mehrtens Acked-by: Michael Buesch Patchwork: https://patchwork.linux-mips.org/patch/1849/ Signed-off-by: Ralf Baechle --- include/linux/ssb/ssb_driver_gige.h | 17 +++++++++++------ 1 file changed, 11 insertions(+), 6 deletions(-) (limited to 'include/linux') diff --git a/include/linux/ssb/ssb_driver_gige.h b/include/linux/ssb/ssb_driver_gige.h index 942e38736901..eba52a100533 100644 --- a/include/linux/ssb/ssb_driver_gige.h +++ b/include/linux/ssb/ssb_driver_gige.h @@ -96,16 +96,21 @@ static inline bool ssb_gige_must_flush_posted_writes(struct pci_dev *pdev) return 0; } -extern char * nvram_get(const char *name); +#ifdef CONFIG_BCM47XX +#include /* Get the device MAC address */ static inline void ssb_gige_get_macaddr(struct pci_dev *pdev, u8 *macaddr) { -#ifdef CONFIG_BCM47XX - char *res = nvram_get("et0macaddr"); - if (res) - memcpy(macaddr, res, 6); -#endif + char buf[20]; + if (nvram_getenv("et0macaddr", buf, sizeof(buf)) < 0) + return; + nvram_parse_macaddr(buf, macaddr); } +#else +static inline void ssb_gige_get_macaddr(struct pci_dev *pdev, u8 *macaddr) +{ +} +#endif extern int ssb_gige_pcibios_plat_dev_init(struct ssb_device *sdev, struct pci_dev *pdev); -- cgit v1.2.3-71-gd317 From e692cb668fdd5a712c6ed2a2d6f2a36ee83997b4 Mon Sep 17 00:00:00 2001 From: "Martin K. Petersen" Date: Wed, 1 Dec 2010 19:41:49 +0100 Subject: block: Deprecate QUEUE_FLAG_CLUSTER and use queue_limits instead When stacking devices, a request_queue is not always available. This forced us to have a no_cluster flag in the queue_limits that could be used as a carrier until the request_queue had been set up for a metadevice. There were several problems with that approach. First of all it was up to the stacking device to remember to set queue flag after stacking had completed. Also, the queue flag and the queue limits had to be kept in sync at all times. We got that wrong, which could lead to us issuing commands that went beyond the max scatterlist limit set by the driver. The proper fix is to avoid having two flags for tracking the same thing. We deprecate QUEUE_FLAG_CLUSTER and use the queue limit directly in the block layer merging functions. The queue_limit 'no_cluster' is turned into 'cluster' to avoid double negatives and to ease stacking. Clustering defaults to being enabled as before. The queue flag logic is removed from the stacking function, and explicitly setting the cluster flag is no longer necessary in DM and MD. Reported-by: Ed Lin Signed-off-by: Martin K. Petersen Acked-by: Mike Snitzer Cc: stable@kernel.org Signed-off-by: Jens Axboe --- block/blk-merge.c | 6 +++--- block/blk-settings.c | 25 ++----------------------- block/blk-sysfs.c | 2 +- drivers/md/dm-table.c | 5 ----- drivers/md/md.c | 3 --- drivers/scsi/scsi_lib.c | 3 +-- include/linux/blkdev.h | 9 ++++++--- 7 files changed, 13 insertions(+), 40 deletions(-) (limited to 'include/linux') diff --git a/block/blk-merge.c b/block/blk-merge.c index 77b7c26df6b5..74bc4a768f32 100644 --- a/block/blk-merge.c +++ b/block/blk-merge.c @@ -21,7 +21,7 @@ static unsigned int __blk_recalc_rq_segments(struct request_queue *q, return 0; fbio = bio; - cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags); + cluster = blk_queue_cluster(q); seg_size = 0; nr_phys_segs = 0; for_each_bio(bio) { @@ -87,7 +87,7 @@ EXPORT_SYMBOL(blk_recount_segments); static int blk_phys_contig_segment(struct request_queue *q, struct bio *bio, struct bio *nxt) { - if (!test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags)) + if (!blk_queue_cluster(q)) return 0; if (bio->bi_seg_back_size + nxt->bi_seg_front_size > @@ -123,7 +123,7 @@ int blk_rq_map_sg(struct request_queue *q, struct request *rq, int nsegs, cluster; nsegs = 0; - cluster = test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags); + cluster = blk_queue_cluster(q); /* * for each bio in rq diff --git a/block/blk-settings.c b/block/blk-settings.c index 701859fb9647..e55f5fc4ca22 100644 --- a/block/blk-settings.c +++ b/block/blk-settings.c @@ -126,7 +126,7 @@ void blk_set_default_limits(struct queue_limits *lim) lim->alignment_offset = 0; lim->io_opt = 0; lim->misaligned = 0; - lim->no_cluster = 0; + lim->cluster = 1; } EXPORT_SYMBOL(blk_set_default_limits); @@ -464,15 +464,6 @@ EXPORT_SYMBOL(blk_queue_io_opt); void blk_queue_stack_limits(struct request_queue *t, struct request_queue *b) { blk_stack_limits(&t->limits, &b->limits, 0); - - if (!t->queue_lock) - WARN_ON_ONCE(1); - else if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) { - unsigned long flags; - spin_lock_irqsave(t->queue_lock, flags); - queue_flag_clear(QUEUE_FLAG_CLUSTER, t); - spin_unlock_irqrestore(t->queue_lock, flags); - } } EXPORT_SYMBOL(blk_queue_stack_limits); @@ -545,7 +536,7 @@ int blk_stack_limits(struct queue_limits *t, struct queue_limits *b, t->io_min = max(t->io_min, b->io_min); t->io_opt = lcm(t->io_opt, b->io_opt); - t->no_cluster |= b->no_cluster; + t->cluster &= b->cluster; t->discard_zeroes_data &= b->discard_zeroes_data; /* Physical block size a multiple of the logical block size? */ @@ -641,7 +632,6 @@ void disk_stack_limits(struct gendisk *disk, struct block_device *bdev, sector_t offset) { struct request_queue *t = disk->queue; - struct request_queue *b = bdev_get_queue(bdev); if (bdev_stack_limits(&t->limits, bdev, offset >> 9) < 0) { char top[BDEVNAME_SIZE], bottom[BDEVNAME_SIZE]; @@ -652,17 +642,6 @@ void disk_stack_limits(struct gendisk *disk, struct block_device *bdev, printk(KERN_NOTICE "%s: Warning: Device %s is misaligned\n", top, bottom); } - - if (!t->queue_lock) - WARN_ON_ONCE(1); - else if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) { - unsigned long flags; - - spin_lock_irqsave(t->queue_lock, flags); - if (!test_bit(QUEUE_FLAG_CLUSTER, &b->queue_flags)) - queue_flag_clear(QUEUE_FLAG_CLUSTER, t); - spin_unlock_irqrestore(t->queue_lock, flags); - } } EXPORT_SYMBOL(disk_stack_limits); diff --git a/block/blk-sysfs.c b/block/blk-sysfs.c index 013457f47fdc..41fb69150b4d 100644 --- a/block/blk-sysfs.c +++ b/block/blk-sysfs.c @@ -119,7 +119,7 @@ static ssize_t queue_max_integrity_segments_show(struct request_queue *q, char * static ssize_t queue_max_segment_size_show(struct request_queue *q, char *page) { - if (test_bit(QUEUE_FLAG_CLUSTER, &q->queue_flags)) + if (blk_queue_cluster(q)) return queue_var_show(queue_max_segment_size(q), (page)); return queue_var_show(PAGE_CACHE_SIZE, (page)); diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c index 90267f8d64ee..e2da1912a2cb 100644 --- a/drivers/md/dm-table.c +++ b/drivers/md/dm-table.c @@ -1131,11 +1131,6 @@ void dm_table_set_restrictions(struct dm_table *t, struct request_queue *q, */ q->limits = *limits; - if (limits->no_cluster) - queue_flag_clear_unlocked(QUEUE_FLAG_CLUSTER, q); - else - queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, q); - if (!dm_table_supports_discards(t)) queue_flag_clear_unlocked(QUEUE_FLAG_DISCARD, q); else diff --git a/drivers/md/md.c b/drivers/md/md.c index 84c46a161927..52694d29663d 100644 --- a/drivers/md/md.c +++ b/drivers/md/md.c @@ -4296,9 +4296,6 @@ static int md_alloc(dev_t dev, char *name) goto abort; mddev->queue->queuedata = mddev; - /* Can be unlocked because the queue is new: no concurrency */ - queue_flag_set_unlocked(QUEUE_FLAG_CLUSTER, mddev->queue); - blk_queue_make_request(mddev->queue, md_make_request); disk = alloc_disk(1 << shift); diff --git a/drivers/scsi/scsi_lib.c b/drivers/scsi/scsi_lib.c index eafeeda6e194..9d7ba07dc5ef 100644 --- a/drivers/scsi/scsi_lib.c +++ b/drivers/scsi/scsi_lib.c @@ -1642,9 +1642,8 @@ struct request_queue *__scsi_alloc_queue(struct Scsi_Host *shost, blk_queue_max_segment_size(q, dma_get_max_seg_size(dev)); - /* New queue, no concurrency on queue_flags */ if (!shost->use_clustering) - queue_flag_clear_unlocked(QUEUE_FLAG_CLUSTER, q); + q->limits.cluster = 0; /* * set a reasonable default alignment on word boundaries: the diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index aae86fd10c4f..95aeeeb49e8b 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -250,7 +250,7 @@ struct queue_limits { unsigned char misaligned; unsigned char discard_misaligned; - unsigned char no_cluster; + unsigned char cluster; signed char discard_zeroes_data; }; @@ -380,7 +380,6 @@ struct request_queue #endif }; -#define QUEUE_FLAG_CLUSTER 0 /* cluster several segments into 1 */ #define QUEUE_FLAG_QUEUED 1 /* uses generic tag queueing */ #define QUEUE_FLAG_STOPPED 2 /* queue is stopped */ #define QUEUE_FLAG_SYNCFULL 3 /* read queue has been filled */ @@ -403,7 +402,6 @@ struct request_queue #define QUEUE_FLAG_SECDISCARD 19 /* supports SECDISCARD */ #define QUEUE_FLAG_DEFAULT ((1 << QUEUE_FLAG_IO_STAT) | \ - (1 << QUEUE_FLAG_CLUSTER) | \ (1 << QUEUE_FLAG_STACKABLE) | \ (1 << QUEUE_FLAG_SAME_COMP) | \ (1 << QUEUE_FLAG_ADD_RANDOM)) @@ -510,6 +508,11 @@ static inline void queue_flag_clear(unsigned int flag, struct request_queue *q) #define rq_data_dir(rq) ((rq)->cmd_flags & 1) +static inline unsigned int blk_queue_cluster(struct request_queue *q) +{ + return q->limits.cluster; +} + /* * We regard a request as sync, if either a read or a sync write */ -- cgit v1.2.3-71-gd317 From 72d4cd9f38b5ed96b75df4c622be25e1c2648dd3 Mon Sep 17 00:00:00 2001 From: Mike Snitzer Date: Fri, 17 Dec 2010 08:34:20 +0100 Subject: block: max hardware sectors limit wrapper Implement blk_limits_max_hw_sectors() and make blk_queue_max_hw_sectors() a wrapper around it. DM needs this to avoid setting queue_limits' max_hw_sectors and max_sectors directly. dm_set_device_limits() now leverages blk_limits_max_hw_sectors() logic to establish the appropriate max_hw_sectors minimum (PAGE_SIZE). Fixes issue where DM was incorrectly setting max_sectors rather than max_hw_sectors (which caused dm_merge_bvec()'s max_hw_sectors check to be ineffective). Signed-off-by: Mike Snitzer Cc: stable@kernel.org Acked-by: Martin K. Petersen Signed-off-by: Jens Axboe --- block/blk-settings.c | 26 ++++++++++++++++++++------ drivers/md/dm-table.c | 5 ++--- include/linux/blkdev.h | 1 + 3 files changed, 23 insertions(+), 9 deletions(-) (limited to 'include/linux') diff --git a/block/blk-settings.c b/block/blk-settings.c index e55f5fc4ca22..36c8c1f2af18 100644 --- a/block/blk-settings.c +++ b/block/blk-settings.c @@ -229,8 +229,8 @@ void blk_queue_bounce_limit(struct request_queue *q, u64 dma_mask) EXPORT_SYMBOL(blk_queue_bounce_limit); /** - * blk_queue_max_hw_sectors - set max sectors for a request for this queue - * @q: the request queue for the device + * blk_limits_max_hw_sectors - set hard and soft limit of max sectors for request + * @limits: the queue limits * @max_hw_sectors: max hardware sectors in the usual 512b unit * * Description: @@ -244,7 +244,7 @@ EXPORT_SYMBOL(blk_queue_bounce_limit); * per-device basis in /sys/block//queue/max_sectors_kb. * The soft limit can not exceed max_hw_sectors. **/ -void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_sectors) +void blk_limits_max_hw_sectors(struct queue_limits *limits, unsigned int max_hw_sectors) { if ((max_hw_sectors << 9) < PAGE_CACHE_SIZE) { max_hw_sectors = 1 << (PAGE_CACHE_SHIFT - 9); @@ -252,9 +252,23 @@ void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_secto __func__, max_hw_sectors); } - q->limits.max_hw_sectors = max_hw_sectors; - q->limits.max_sectors = min_t(unsigned int, max_hw_sectors, - BLK_DEF_MAX_SECTORS); + limits->max_hw_sectors = max_hw_sectors; + limits->max_sectors = min_t(unsigned int, max_hw_sectors, + BLK_DEF_MAX_SECTORS); +} +EXPORT_SYMBOL(blk_limits_max_hw_sectors); + +/** + * blk_queue_max_hw_sectors - set max sectors for a request for this queue + * @q: the request queue for the device + * @max_hw_sectors: max hardware sectors in the usual 512b unit + * + * Description: + * See description for blk_limits_max_hw_sectors(). + **/ +void blk_queue_max_hw_sectors(struct request_queue *q, unsigned int max_hw_sectors) +{ + blk_limits_max_hw_sectors(&q->limits, max_hw_sectors); } EXPORT_SYMBOL(blk_queue_max_hw_sectors); diff --git a/drivers/md/dm-table.c b/drivers/md/dm-table.c index e2da1912a2cb..4d705cea0f8c 100644 --- a/drivers/md/dm-table.c +++ b/drivers/md/dm-table.c @@ -517,9 +517,8 @@ int dm_set_device_limits(struct dm_target *ti, struct dm_dev *dev, */ if (q->merge_bvec_fn && !ti->type->merge) - limits->max_sectors = - min_not_zero(limits->max_sectors, - (unsigned int) (PAGE_SIZE >> 9)); + blk_limits_max_hw_sectors(limits, + (unsigned int) (PAGE_SIZE >> 9)); return 0; } EXPORT_SYMBOL_GPL(dm_set_device_limits); diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h index 95aeeeb49e8b..36ab42c9bb99 100644 --- a/include/linux/blkdev.h +++ b/include/linux/blkdev.h @@ -808,6 +808,7 @@ extern struct request_queue *blk_init_allocated_queue(struct request_queue *, extern void blk_cleanup_queue(struct request_queue *); extern void blk_queue_make_request(struct request_queue *, make_request_fn *); extern void blk_queue_bounce_limit(struct request_queue *, u64); +extern void blk_limits_max_hw_sectors(struct queue_limits *, unsigned int); extern void blk_queue_max_hw_sectors(struct request_queue *, unsigned int); extern void blk_queue_max_segments(struct request_queue *, unsigned short); extern void blk_queue_max_segment_size(struct request_queue *, unsigned int); -- cgit v1.2.3-71-gd317 From b6aa5901c7a2bd90d0b6b9866300d2648b2568f3 Mon Sep 17 00:00:00 2001 From: Henry C Chang Date: Wed, 15 Dec 2010 20:45:41 -0800 Subject: ceph: mark user pages dirty on direct-io reads For read operation, we have to set the argument _write_ of get_user_pages to 1 since we will write data to pages. Also, we need to SetPageDirty before releasing these pages. Signed-off-by: Henry C Chang Signed-off-by: Sage Weil --- fs/ceph/file.c | 8 ++++---- include/linux/ceph/libceph.h | 6 ++++-- net/ceph/pagevec.c | 11 +++++++---- 3 files changed, 15 insertions(+), 10 deletions(-) (limited to 'include/linux') diff --git a/fs/ceph/file.c b/fs/ceph/file.c index e860d8f1bb45..7d0e4a82d898 100644 --- a/fs/ceph/file.c +++ b/fs/ceph/file.c @@ -384,7 +384,7 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data, if (file->f_flags & O_DIRECT) { num_pages = calc_pages_for((unsigned long)data, len); - pages = ceph_get_direct_page_vector(data, num_pages); + pages = ceph_get_direct_page_vector(data, num_pages, true); } else { num_pages = calc_pages_for(off, len); pages = ceph_alloc_page_vector(num_pages, GFP_NOFS); @@ -413,7 +413,7 @@ static ssize_t ceph_sync_read(struct file *file, char __user *data, done: if (file->f_flags & O_DIRECT) - ceph_put_page_vector(pages, num_pages); + ceph_put_page_vector(pages, num_pages, true); else ceph_release_page_vector(pages, num_pages); dout("sync_read result %d\n", ret); @@ -522,7 +522,7 @@ more: return -ENOMEM; if (file->f_flags & O_DIRECT) { - pages = ceph_get_direct_page_vector(data, num_pages); + pages = ceph_get_direct_page_vector(data, num_pages, false); if (IS_ERR(pages)) { ret = PTR_ERR(pages); goto out; @@ -572,7 +572,7 @@ more: } if (file->f_flags & O_DIRECT) - ceph_put_page_vector(pages, num_pages); + ceph_put_page_vector(pages, num_pages, false); else if (file->f_flags & O_SYNC) ceph_release_page_vector(pages, num_pages); diff --git a/include/linux/ceph/libceph.h b/include/linux/ceph/libceph.h index 9e76d35670d2..72c72bfccb88 100644 --- a/include/linux/ceph/libceph.h +++ b/include/linux/ceph/libceph.h @@ -227,8 +227,10 @@ extern int ceph_open_session(struct ceph_client *client); extern void ceph_release_page_vector(struct page **pages, int num_pages); extern struct page **ceph_get_direct_page_vector(const char __user *data, - int num_pages); -extern void ceph_put_page_vector(struct page **pages, int num_pages); + int num_pages, + bool write_page); +extern void ceph_put_page_vector(struct page **pages, int num_pages, + bool dirty); extern void ceph_release_page_vector(struct page **pages, int num_pages); extern struct page **ceph_alloc_page_vector(int num_pages, gfp_t flags); extern int ceph_copy_user_to_page_vector(struct page **pages, diff --git a/net/ceph/pagevec.c b/net/ceph/pagevec.c index ac34feeb2b3a..01947a5d03b7 100644 --- a/net/ceph/pagevec.c +++ b/net/ceph/pagevec.c @@ -13,7 +13,7 @@ * build a vector of user pages */ struct page **ceph_get_direct_page_vector(const char __user *data, - int num_pages) + int num_pages, bool write_page) { struct page **pages; int rc; @@ -24,7 +24,7 @@ struct page **ceph_get_direct_page_vector(const char __user *data, down_read(¤t->mm->mmap_sem); rc = get_user_pages(current, current->mm, (unsigned long)data, - num_pages, 0, 0, pages, NULL); + num_pages, write_page, 0, pages, NULL); up_read(¤t->mm->mmap_sem); if (rc < 0) goto fail; @@ -36,12 +36,15 @@ fail: } EXPORT_SYMBOL(ceph_get_direct_page_vector); -void ceph_put_page_vector(struct page **pages, int num_pages) +void ceph_put_page_vector(struct page **pages, int num_pages, bool dirty) { int i; - for (i = 0; i < num_pages; i++) + for (i = 0; i < num_pages; i++) { + if (dirty) + set_page_dirty_lock(pages[i]); put_page(pages[i]); + } kfree(pages); } EXPORT_SYMBOL(ceph_put_page_vector); -- cgit v1.2.3-71-gd317 From c0f5ac5426f7fd82b23dd5c6a1e633b290294a08 Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Thu, 16 Dec 2010 10:38:41 -0700 Subject: Revert "resources: support allocating space within a region from the top down" This reverts commit e7f8567db9a7f6b3151b0b275e245c1cef0d9c70. Acked-by: H. Peter Anvin Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- Documentation/kernel-parameters.txt | 5 -- include/linux/ioport.h | 1 - kernel/resource.c | 98 ++----------------------------------- 3 files changed, 4 insertions(+), 100 deletions(-) (limited to 'include/linux') diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index cdd2a6e8a3b7..8b61c9360999 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -2175,11 +2175,6 @@ and is between 256 and 4096 characters. It is defined in the file reset_devices [KNL] Force drivers to reset the underlying device during initialization. - resource_alloc_from_bottom - Allocate new resources from the beginning of available - space, not the end. If you need to use this, please - report a bug. - resume= [SWSUSP] Specify the partition device for software suspend diff --git a/include/linux/ioport.h b/include/linux/ioport.h index d377ea815d45..b22790268b64 100644 --- a/include/linux/ioport.h +++ b/include/linux/ioport.h @@ -112,7 +112,6 @@ struct resource_list { /* PC/ISA/whatever - the normal PC address spaces: IO and memory */ extern struct resource ioport_resource; extern struct resource iomem_resource; -extern int resource_alloc_from_bottom; extern struct resource *request_resource_conflict(struct resource *root, struct resource *new); extern int request_resource(struct resource *root, struct resource *new); diff --git a/kernel/resource.c b/kernel/resource.c index 9fad33efd0db..560659f7baef 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -40,23 +40,6 @@ EXPORT_SYMBOL(iomem_resource); static DEFINE_RWLOCK(resource_lock); -/* - * By default, we allocate free space bottom-up. The architecture can request - * top-down by clearing this flag. The user can override the architecture's - * choice with the "resource_alloc_from_bottom" kernel boot option, but that - * should only be a debugging tool. - */ -int resource_alloc_from_bottom = 1; - -static __init int setup_alloc_from_bottom(char *s) -{ - printk(KERN_INFO - "resource: allocating from bottom-up; please report a bug\n"); - resource_alloc_from_bottom = 1; - return 0; -} -early_param("resource_alloc_from_bottom", setup_alloc_from_bottom); - static void *r_next(struct seq_file *m, void *v, loff_t *pos) { struct resource *p = v; @@ -396,75 +379,8 @@ static bool resource_contains(struct resource *res1, struct resource *res2) return res1->start <= res2->start && res1->end >= res2->end; } -/* - * Find the resource before "child" in the sibling list of "root" children. - */ -static struct resource *find_sibling_prev(struct resource *root, struct resource *child) -{ - struct resource *this; - - for (this = root->child; this; this = this->sibling) - if (this->sibling == child) - return this; - - return NULL; -} - -/* - * Find empty slot in the resource tree given range and alignment. - * This version allocates from the end of the root resource first. - */ -static int find_resource_from_top(struct resource *root, struct resource *new, - resource_size_t size, resource_size_t min, - resource_size_t max, resource_size_t align, - resource_size_t (*alignf)(void *, - const struct resource *, - resource_size_t, - resource_size_t), - void *alignf_data) -{ - struct resource *this; - struct resource tmp, avail, alloc; - - tmp.start = root->end; - tmp.end = root->end; - - this = find_sibling_prev(root, NULL); - for (;;) { - if (this) { - if (this->end < root->end) - tmp.start = this->end + 1; - } else - tmp.start = root->start; - - resource_clip(&tmp, min, max); - - /* Check for overflow after ALIGN() */ - avail = *new; - avail.start = ALIGN(tmp.start, align); - avail.end = tmp.end; - if (avail.start >= tmp.start) { - alloc.start = alignf(alignf_data, &avail, size, align); - alloc.end = alloc.start + size - 1; - if (resource_contains(&avail, &alloc)) { - new->start = alloc.start; - new->end = alloc.end; - return 0; - } - } - - if (!this || this->start == root->start) - break; - - tmp.end = this->start - 1; - this = find_sibling_prev(root, this); - } - return -EBUSY; -} - /* * Find empty slot in the resource tree given range and alignment. - * This version allocates from the beginning of the root resource first. */ static int find_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, @@ -480,15 +396,14 @@ static int find_resource(struct resource *root, struct resource *new, tmp.start = root->start; /* - * Skip past an allocated resource that starts at 0, since the - * assignment of this->start - 1 to tmp->end below would cause an - * underflow. + * Skip past an allocated resource that starts at 0, since the assignment + * of this->start - 1 to tmp->end below would cause an underflow. */ if (this && this->start == 0) { tmp.start = this->end + 1; this = this->sibling; } - for (;;) { + for(;;) { if (this) tmp.end = this->start - 1; else @@ -509,10 +424,8 @@ static int find_resource(struct resource *root, struct resource *new, return 0; } } - if (!this) break; - tmp.start = this->end + 1; this = this->sibling; } @@ -545,10 +458,7 @@ int allocate_resource(struct resource *root, struct resource *new, alignf = simple_align_resource; write_lock(&resource_lock); - if (resource_alloc_from_bottom) - err = find_resource(root, new, size, min, max, align, alignf, alignf_data); - else - err = find_resource_from_top(root, new, size, min, max, align, alignf, alignf_data); + err = find_resource(root, new, size, min, max, align, alignf, alignf_data); if (err >= 0 && __request_resource(root, new)) err = -EBUSY; write_unlock(&resource_lock); -- cgit v1.2.3-71-gd317 From fcb119183c73bf0781009713f303e28b1fb13d3e Mon Sep 17 00:00:00 2001 From: Bjorn Helgaas Date: Thu, 16 Dec 2010 10:38:46 -0700 Subject: resources: add arch hook for preventing allocation in reserved areas This adds arch_remove_reservations(), which an arch can implement if it needs to protect part of the address space from allocation. Sometimes that can be done by just putting a region in the resource tree, but there are cases where that doesn't work well. For example, x86 BIOS E820 reservations are not related to devices, so they may overlap part of, all of, or more than a device resource, so they may not end up at the correct spot in the resource tree. Acked-by: H. Peter Anvin Signed-off-by: Bjorn Helgaas Signed-off-by: Jesse Barnes --- include/linux/ioport.h | 1 + kernel/resource.c | 6 ++++++ 2 files changed, 7 insertions(+) (limited to 'include/linux') diff --git a/include/linux/ioport.h b/include/linux/ioport.h index b22790268b64..e9bb22cba764 100644 --- a/include/linux/ioport.h +++ b/include/linux/ioport.h @@ -123,6 +123,7 @@ extern void reserve_region_with_split(struct resource *root, extern struct resource *insert_resource_conflict(struct resource *parent, struct resource *new); extern int insert_resource(struct resource *parent, struct resource *new); extern void insert_resource_expand_to_fit(struct resource *root, struct resource *new); +extern void arch_remove_reservations(struct resource *avail); extern int allocate_resource(struct resource *root, struct resource *new, resource_size_t size, resource_size_t min, resource_size_t max, resource_size_t align, diff --git a/kernel/resource.c b/kernel/resource.c index 560659f7baef..798e2fae2a06 100644 --- a/kernel/resource.c +++ b/kernel/resource.c @@ -357,6 +357,10 @@ int __weak page_is_ram(unsigned long pfn) return walk_system_ram_range(pfn, 1, NULL, __is_ram) == 1; } +void __weak arch_remove_reservations(struct resource *avail) +{ +} + static resource_size_t simple_align_resource(void *data, const struct resource *avail, resource_size_t size, @@ -394,6 +398,7 @@ static int find_resource(struct resource *root, struct resource *new, struct resource *this = root->child; struct resource tmp = *new, avail, alloc; + tmp.flags = new->flags; tmp.start = root->start; /* * Skip past an allocated resource that starts at 0, since the assignment @@ -410,6 +415,7 @@ static int find_resource(struct resource *root, struct resource *new, tmp.end = root->end; resource_clip(&tmp, min, max); + arch_remove_reservations(&tmp); /* Check for overflow after ALIGN() */ avail = *new; -- cgit v1.2.3-71-gd317 From b8da46d3d55807037b58f14621a0949f18053bde Mon Sep 17 00:00:00 2001 From: Nicolas Pitre Date: Mon, 20 Dec 2010 00:29:32 -0500 Subject: clarify a usage constraint for cnt32_to_63() The cnt32_to_63 algorithm relies on proper counter data evaluation ordering to work properly. This was missing from the provided documentation. Let's augment the documentation with the missing usage constraint and fix the only instance that got it wrong. Signed-off-by: Nicolas Pitre Acked-by: David Howells Signed-off-by: Linus Torvalds --- arch/mn10300/kernel/time.c | 10 +++------- include/linux/cnt32_to_63.h | 20 +++++++++++++++++++- 2 files changed, 22 insertions(+), 8 deletions(-) (limited to 'include/linux') diff --git a/arch/mn10300/kernel/time.c b/arch/mn10300/kernel/time.c index f860a340acc9..75da468090b9 100644 --- a/arch/mn10300/kernel/time.c +++ b/arch/mn10300/kernel/time.c @@ -40,21 +40,17 @@ unsigned long long sched_clock(void) unsigned long long ll; unsigned l[2]; } tsc64, result; - unsigned long tsc, tmp; + unsigned long tmp; unsigned product[3]; /* 96-bit intermediate value */ /* cnt32_to_63() is not safe with preemption */ preempt_disable(); - /* read the TSC value - */ - tsc = get_cycles(); - - /* expand to 64-bits. + /* expand the tsc to 64-bits. * - sched_clock() must be called once a minute or better or the * following will go horribly wrong - see cnt32_to_63() */ - tsc64.ll = cnt32_to_63(tsc) & 0x7fffffffffffffffULL; + tsc64.ll = cnt32_to_63(get_cycles()) & 0x7fffffffffffffffULL; preempt_enable(); diff --git a/include/linux/cnt32_to_63.h b/include/linux/cnt32_to_63.h index 7605fdd1eb65..e3d8bf26e5eb 100644 --- a/include/linux/cnt32_to_63.h +++ b/include/linux/cnt32_to_63.h @@ -61,13 +61,31 @@ union cnt32_to_63 { * * 2) this code must not be preempted for a duration longer than the * 32-bit counter half period minus the longest period between two - * calls to this code. + * calls to this code; * * Those requirements ensure proper update to the state bit in memory. * This is usually not a problem in practice, but if it is then a kernel * timer should be scheduled to manage for this code to be executed often * enough. * + * And finally: + * + * 3) the cnt_lo argument must be seen as a globally incrementing value, + * meaning that it should be a direct reference to the counter data which + * can be evaluated according to a specific ordering within the macro, + * and not the result of a previous evaluation stored in a variable. + * + * For example, this is wrong: + * + * u32 partial = get_hw_count(); + * u64 full = cnt32_to_63(partial); + * return full; + * + * This is fine: + * + * u64 full = cnt32_to_63(get_hw_count()); + * return full; + * * Note that the top bit (bit 63) in the returned value should be considered * as garbage. It is not cleared here because callers are likely to use a * multiplier on the returned value which can get rid of the top bit -- cgit v1.2.3-71-gd317