From 426d31071ac476ea62c62656b242930c17b58c00 Mon Sep 17 00:00:00 2001 From: Paul Bolle Date: Sat, 7 Aug 2010 12:30:03 +0200 Subject: fix printk typo 'faild' Signed-off-by: Paul Bolle Signed-off-by: Jiri Kosina --- kernel/trace/trace_kprobe.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index f52b5f50299..58716e73e2a 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -490,7 +490,7 @@ static int register_trace_probe(struct trace_probe *tp) } ret = register_probe_event(tp); if (ret) { - pr_warning("Faild to register probe event(%d)\n", ret); + pr_warning("Failed to register probe event(%d)\n", ret); goto end; } -- cgit v1.2.2 From 6038f373a3dc1f1c26496e60b6c40b164716f07e Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Sun, 15 Aug 2010 18:52:59 +0200 Subject: llseek: automatically add .llseek fop All file_operations should get a .llseek operation so we can make nonseekable_open the default for future file operations without a .llseek pointer. The three cases that we can automatically detect are no_llseek, seq_lseek and default_llseek. For cases where we can we can automatically prove that the file offset is always ignored, we use noop_llseek, which maintains the current behavior of not returning an error from a seek. New drivers should normally not use noop_llseek but instead use no_llseek and call nonseekable_open at open time. Existing drivers can be converted to do the same when the maintainer knows for certain that no user code relies on calling seek on the device file. The generated code is often incorrectly indented and right now contains comments that clarify for each added line why a specific variant was chosen. In the version that gets submitted upstream, the comments will be gone and I will manually fix the indentation, because there does not seem to be a way to do that using coccinelle. Some amount of new code is currently sitting in linux-next that should get the same modifications, which I will do at the end of the merge window. Many thanks to Julia Lawall for helping me learn to write a semantic patch that does all this. ===== begin semantic patch ===== // This adds an llseek= method to all file operations, // as a preparation for making no_llseek the default. // // The rules are // - use no_llseek explicitly if we do nonseekable_open // - use seq_lseek for sequential files // - use default_llseek if we know we access f_pos // - use noop_llseek if we know we don't access f_pos, // but we still want to allow users to call lseek // @ open1 exists @ identifier nested_open; @@ nested_open(...) { <+... nonseekable_open(...) ...+> } @ open exists@ identifier open_f; identifier i, f; identifier open1.nested_open; @@ int open_f(struct inode *i, struct file *f) { <+... ( nonseekable_open(...) | nested_open(...) ) ...+> } @ read disable optional_qualifier exists @ identifier read_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; expression E; identifier func; @@ ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off) { <+... ( *off = E | *off += E | func(..., off, ...) | E = *off ) ...+> } @ read_no_fpos disable optional_qualifier exists @ identifier read_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; @@ ssize_t read_f(struct file *f, char *p, size_t s, loff_t *off) { ... when != off } @ write @ identifier write_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; expression E; identifier func; @@ ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off) { <+... ( *off = E | *off += E | func(..., off, ...) | E = *off ) ...+> } @ write_no_fpos @ identifier write_f; identifier f, p, s, off; type ssize_t, size_t, loff_t; @@ ssize_t write_f(struct file *f, const char *p, size_t s, loff_t *off) { ... when != off } @ fops0 @ identifier fops; @@ struct file_operations fops = { ... }; @ has_llseek depends on fops0 @ identifier fops0.fops; identifier llseek_f; @@ struct file_operations fops = { ... .llseek = llseek_f, ... }; @ has_read depends on fops0 @ identifier fops0.fops; identifier read_f; @@ struct file_operations fops = { ... .read = read_f, ... }; @ has_write depends on fops0 @ identifier fops0.fops; identifier write_f; @@ struct file_operations fops = { ... .write = write_f, ... }; @ has_open depends on fops0 @ identifier fops0.fops; identifier open_f; @@ struct file_operations fops = { ... .open = open_f, ... }; // use no_llseek if we call nonseekable_open //////////////////////////////////////////// @ nonseekable1 depends on !has_llseek && has_open @ identifier fops0.fops; identifier nso ~= "nonseekable_open"; @@ struct file_operations fops = { ... .open = nso, ... +.llseek = no_llseek, /* nonseekable */ }; @ nonseekable2 depends on !has_llseek @ identifier fops0.fops; identifier open.open_f; @@ struct file_operations fops = { ... .open = open_f, ... +.llseek = no_llseek, /* open uses nonseekable */ }; // use seq_lseek for sequential files ///////////////////////////////////// @ seq depends on !has_llseek @ identifier fops0.fops; identifier sr ~= "seq_read"; @@ struct file_operations fops = { ... .read = sr, ... +.llseek = seq_lseek, /* we have seq_read */ }; // use default_llseek if there is a readdir /////////////////////////////////////////// @ fops1 depends on !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier readdir_e; @@ // any other fop is used that changes pos struct file_operations fops = { ... .readdir = readdir_e, ... +.llseek = default_llseek, /* readdir is present */ }; // use default_llseek if at least one of read/write touches f_pos ///////////////////////////////////////////////////////////////// @ fops2 depends on !fops1 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier read.read_f; @@ // read fops use offset struct file_operations fops = { ... .read = read_f, ... +.llseek = default_llseek, /* read accesses f_pos */ }; @ fops3 depends on !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier write.write_f; @@ // write fops use offset struct file_operations fops = { ... .write = write_f, ... + .llseek = default_llseek, /* write accesses f_pos */ }; // Use noop_llseek if neither read nor write accesses f_pos /////////////////////////////////////////////////////////// @ fops4 depends on !fops1 && !fops2 && !fops3 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier read_no_fpos.read_f; identifier write_no_fpos.write_f; @@ // write fops use offset struct file_operations fops = { ... .write = write_f, .read = read_f, ... +.llseek = noop_llseek, /* read and write both use no f_pos */ }; @ depends on has_write && !has_read && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier write_no_fpos.write_f; @@ struct file_operations fops = { ... .write = write_f, ... +.llseek = noop_llseek, /* write uses no f_pos */ }; @ depends on has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; identifier read_no_fpos.read_f; @@ struct file_operations fops = { ... .read = read_f, ... +.llseek = noop_llseek, /* read uses no f_pos */ }; @ depends on !has_read && !has_write && !fops1 && !fops2 && !has_llseek && !nonseekable1 && !nonseekable2 && !seq @ identifier fops0.fops; @@ struct file_operations fops = { ... +.llseek = noop_llseek, /* no read or write fn */ }; ===== End semantic patch ===== Signed-off-by: Arnd Bergmann Cc: Julia Lawall Cc: Christoph Hellwig --- kernel/trace/blktrace.c | 2 ++ kernel/trace/ftrace.c | 2 ++ kernel/trace/ring_buffer.c | 1 + kernel/trace/trace_events.c | 6 ++++++ kernel/trace/trace_stack.c | 1 + 5 files changed, 12 insertions(+) (limited to 'kernel/trace') diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 959f8d6c8cc..2d5f3a75731 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -326,6 +326,7 @@ static const struct file_operations blk_dropped_fops = { .owner = THIS_MODULE, .open = blk_dropped_open, .read = blk_dropped_read, + .llseek = default_llseek, }; static int blk_msg_open(struct inode *inode, struct file *filp) @@ -365,6 +366,7 @@ static const struct file_operations blk_msg_fops = { .owner = THIS_MODULE, .open = blk_msg_open, .write = blk_msg_write, + .llseek = noop_llseek, }; /* diff --git a/kernel/trace/ftrace.c b/kernel/trace/ftrace.c index fa7ece649fe..5e1ad476309 100644 --- a/kernel/trace/ftrace.c +++ b/kernel/trace/ftrace.c @@ -800,6 +800,7 @@ static const struct file_operations ftrace_profile_fops = { .open = tracing_open_generic, .read = ftrace_profile_read, .write = ftrace_profile_write, + .llseek = default_llseek, }; /* used to initialize the real stat files */ @@ -2632,6 +2633,7 @@ static const struct file_operations ftrace_graph_fops = { .read = seq_read, .write = ftrace_graph_write, .release = ftrace_graph_release, + .llseek = seq_lseek, }; #endif /* CONFIG_FUNCTION_GRAPH_TRACER */ diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 492197e2f86..3aea966d16d 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -3965,6 +3965,7 @@ static const struct file_operations rb_simple_fops = { .open = tracing_open_generic, .read = rb_simple_read, .write = rb_simple_write, + .llseek = default_llseek, }; diff --git a/kernel/trace/trace_events.c b/kernel/trace/trace_events.c index 4c758f14632..0369c5e0998 100644 --- a/kernel/trace/trace_events.c +++ b/kernel/trace/trace_events.c @@ -951,6 +951,7 @@ static const struct file_operations ftrace_enable_fops = { .open = tracing_open_generic, .read = event_enable_read, .write = event_enable_write, + .llseek = default_llseek, }; static const struct file_operations ftrace_event_format_fops = { @@ -963,29 +964,34 @@ static const struct file_operations ftrace_event_format_fops = { static const struct file_operations ftrace_event_id_fops = { .open = tracing_open_generic, .read = event_id_read, + .llseek = default_llseek, }; static const struct file_operations ftrace_event_filter_fops = { .open = tracing_open_generic, .read = event_filter_read, .write = event_filter_write, + .llseek = default_llseek, }; static const struct file_operations ftrace_subsystem_filter_fops = { .open = tracing_open_generic, .read = subsystem_filter_read, .write = subsystem_filter_write, + .llseek = default_llseek, }; static const struct file_operations ftrace_system_enable_fops = { .open = tracing_open_generic, .read = system_enable_read, .write = system_enable_write, + .llseek = default_llseek, }; static const struct file_operations ftrace_show_header_fops = { .open = tracing_open_generic, .read = show_header, + .llseek = default_llseek, }; static struct dentry *event_trace_events_dir(void) diff --git a/kernel/trace/trace_stack.c b/kernel/trace/trace_stack.c index a6b7e0e0f3e..4c5dead0c23 100644 --- a/kernel/trace/trace_stack.c +++ b/kernel/trace/trace_stack.c @@ -195,6 +195,7 @@ static const struct file_operations stack_max_size_fops = { .open = tracing_open_generic, .read = stack_max_size_read, .write = stack_max_size_write, + .llseek = default_llseek, }; static void * -- cgit v1.2.2 From 01b284f9b6d51cc3f3bcf3b49f16d2601d3ca22d Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 17 Sep 2010 20:39:22 +0200 Subject: blktrace: remove the big kernel lock According to Jens, this code does not need the BKL at all, it is sufficiently serialized by bd_mutex. Signed-off-by: Arnd Bergmann Cc: Jens Axboe Cc: Steven Rostedt --- kernel/trace/blktrace.c | 14 +++----------- 1 file changed, 3 insertions(+), 11 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index 959f8d6c8cc..5328e8779d4 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -23,7 +23,6 @@ #include #include #include -#include #include #include @@ -639,7 +638,6 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg) if (!q) return -ENXIO; - lock_kernel(); mutex_lock(&bdev->bd_mutex); switch (cmd) { @@ -667,7 +665,6 @@ int blk_trace_ioctl(struct block_device *bdev, unsigned cmd, char __user *arg) } mutex_unlock(&bdev->bd_mutex); - unlock_kernel(); return ret; } @@ -1652,10 +1649,9 @@ static ssize_t sysfs_blk_trace_attr_show(struct device *dev, struct block_device *bdev; ssize_t ret = -ENXIO; - lock_kernel(); bdev = bdget(part_devt(p)); if (bdev == NULL) - goto out_unlock_kernel; + goto out; q = blk_trace_get_queue(bdev); if (q == NULL) @@ -1683,8 +1679,7 @@ out_unlock_bdev: mutex_unlock(&bdev->bd_mutex); out_bdput: bdput(bdev); -out_unlock_kernel: - unlock_kernel(); +out: return ret; } @@ -1714,11 +1709,10 @@ static ssize_t sysfs_blk_trace_attr_store(struct device *dev, ret = -ENXIO; - lock_kernel(); p = dev_to_part(dev); bdev = bdget(part_devt(p)); if (bdev == NULL) - goto out_unlock_kernel; + goto out; q = blk_trace_get_queue(bdev); if (q == NULL) @@ -1753,8 +1747,6 @@ out_unlock_bdev: mutex_unlock(&bdev->bd_mutex); out_bdput: bdput(bdev); -out_unlock_kernel: - unlock_kernel(); out: return ret ? ret : count; } -- cgit v1.2.2 From 747e94ae3d1b4c9bf5380e569f614eb9040b79e7 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 8 Oct 2010 13:51:48 -0400 Subject: ring-buffer: Make write slow path out of line Gcc inlines the slow path of the ring buffer write which can hurt performance. This patch simply forces the slow path function rb_move_tail() to always be a function. The ring_buffer_benchmark module with reader_disabled=1 shows that this patch changes the time to record an event from 135 ns to 132 ns. (3 ns or 2.22% improvement) Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index bca96377fd4..0b88df849a5 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1823,7 +1823,10 @@ rb_reset_tail(struct ring_buffer_per_cpu *cpu_buffer, local_sub(length, &tail_page->write); } -static struct ring_buffer_event * +/* + * This is the slow path, force gcc not to inline it. + */ +static noinline struct ring_buffer_event * rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer, unsigned long length, unsigned long tail, struct buffer_page *tail_page, u64 *ts) @@ -1943,7 +1946,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, tail = write - length; /* See if we shot pass the end of this buffer page */ - if (write > BUF_PAGE_SIZE) + if (unlikely(write > BUF_PAGE_SIZE)) return rb_move_tail(cpu_buffer, length, tail, tail_page, ts); -- cgit v1.2.2 From e8bc43e84fada397af1b677b07dbf26e6ac78fcc Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 20 Oct 2010 10:58:02 -0400 Subject: ring-buffer: Pass timestamp by value and not by reference The original code for the ring buffer had locations that modified the timestamp and that change was used by the callers. Now, the timestamp is not reused by the callers and there is no reason to pass it by reference. By changing the call to pass by value, lets gcc optimize the code a bit more where it can store the timestamp in a register and not worry about updating the reference. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 20 ++++++++++---------- 1 file changed, 10 insertions(+), 10 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 0b88df849a5..c8ce6bde7fa 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -1829,7 +1829,7 @@ rb_reset_tail(struct ring_buffer_per_cpu *cpu_buffer, static noinline struct ring_buffer_event * rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer, unsigned long length, unsigned long tail, - struct buffer_page *tail_page, u64 *ts) + struct buffer_page *tail_page, u64 ts) { struct buffer_page *commit_page = cpu_buffer->commit_page; struct ring_buffer *buffer = cpu_buffer->buffer; @@ -1912,8 +1912,8 @@ rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer, * Nested commits always have zero deltas, so * just reread the time stamp */ - *ts = rb_time_stamp(buffer); - next_page->page->time_stamp = *ts; + ts = rb_time_stamp(buffer); + next_page->page->time_stamp = ts; } out_again: @@ -1932,7 +1932,7 @@ rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer, static struct ring_buffer_event * __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, - unsigned type, unsigned long length, u64 *ts) + unsigned type, unsigned long length, u64 ts) { struct buffer_page *tail_page; struct ring_buffer_event *event; @@ -1965,7 +1965,7 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, * its timestamp. */ if (!tail) - tail_page->page->time_stamp = *ts; + tail_page->page->time_stamp = ts; return event; } @@ -2008,7 +2008,7 @@ rb_try_to_discard(struct ring_buffer_per_cpu *cpu_buffer, static int rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, - u64 *ts, u64 *delta) + u64 ts, u64 *delta) { struct ring_buffer_event *event; int ret; @@ -2016,7 +2016,7 @@ rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, WARN_ONCE(*delta > (1ULL << 59), KERN_WARNING "Delta way too big! %llu ts=%llu write stamp = %llu\n", (unsigned long long)*delta, - (unsigned long long)*ts, + (unsigned long long)ts, (unsigned long long)cpu_buffer->write_stamp); /* @@ -2051,7 +2051,7 @@ rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, event->array[0] = 0; } } - cpu_buffer->write_stamp = *ts; + cpu_buffer->write_stamp = ts; /* let the caller know this was the commit */ ret = 1; } else { @@ -2175,7 +2175,7 @@ rb_reserve_next_event(struct ring_buffer *buffer, delta = diff; if (unlikely(test_time_stamp(delta))) { - commit = rb_add_time_stamp(cpu_buffer, &ts, &delta); + commit = rb_add_time_stamp(cpu_buffer, ts, &delta); if (commit == -EBUSY) goto out_fail; @@ -2187,7 +2187,7 @@ rb_reserve_next_event(struct ring_buffer *buffer, } get_event: - event = __rb_reserve_next(cpu_buffer, 0, length, &ts); + event = __rb_reserve_next(cpu_buffer, 0, length, ts); if (unlikely(PTR_ERR(event) == -EAGAIN)) goto again; -- cgit v1.2.2 From f25106aeab7408394b9dd707e5ecf557e269c723 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 20 Oct 2010 12:40:12 -0400 Subject: ring-buffer: Pass delta by value and not by reference The delta between events is passed to the timestamp code by reference and the timestamp code will reset the value. But it can be reset from the caller. No need to pass it in by reference. By changing the call to pass by value, lets gcc optimize the code a bit more where it can store the delta in a register and not worry about updating the reference. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 16 ++++++++-------- 1 file changed, 8 insertions(+), 8 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index c8ce6bde7fa..3af77cd47f2 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2008,14 +2008,14 @@ rb_try_to_discard(struct ring_buffer_per_cpu *cpu_buffer, static int rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, - u64 ts, u64 *delta) + u64 ts, u64 delta) { struct ring_buffer_event *event; int ret; - WARN_ONCE(*delta > (1ULL << 59), + WARN_ONCE(delta > (1ULL << 59), KERN_WARNING "Delta way too big! %llu ts=%llu write stamp = %llu\n", - (unsigned long long)*delta, + (unsigned long long)delta, (unsigned long long)ts, (unsigned long long)cpu_buffer->write_stamp); @@ -2041,8 +2041,8 @@ rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, * and if we can't just make it zero. */ if (rb_event_index(event)) { - event->time_delta = *delta & TS_MASK; - event->array[0] = *delta >> TS_SHIFT; + event->time_delta = delta & TS_MASK; + event->array[0] = delta >> TS_SHIFT; } else { /* try to discard, since we do not need this */ if (!rb_try_to_discard(cpu_buffer, event)) { @@ -2064,8 +2064,6 @@ rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, ret = 0; } - *delta = 0; - return ret; } @@ -2175,7 +2173,9 @@ rb_reserve_next_event(struct ring_buffer *buffer, delta = diff; if (unlikely(test_time_stamp(delta))) { - commit = rb_add_time_stamp(cpu_buffer, ts, &delta); + commit = rb_add_time_stamp(cpu_buffer, ts, delta); + delta = 0; + if (commit == -EBUSY) goto out_fail; -- cgit v1.2.2 From 69d1b839f7eee347e357b3f6cce7f630cc6ff93d Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Thu, 7 Oct 2010 18:18:05 -0400 Subject: ring-buffer: Bind time extend and data events together When the time between two timestamps is greater than 2^27 nanosecs (~134 ms) a time extend event is added that extends the time difference to 59 bits (~18 years). This is due to events only having a 27 bit field to store time. Currently this time extend is a separate event. We add it just before the event data that is being written to the buffer. But before the event data is committed, the event data can also be discarded (as with the case of filters). But because the time extend has already been committed, it will stay in the buffer. If lots of events are being filtered and no event is being written, then every 134ms a time extend can be added to the buffer without any data attached. To keep from filling the entire buffer with time extends, a time extend will never be the first event in a page because the page timestamp can be used. Time extends can only fill the rest of a page with some data at the beginning. This patch binds the time extend with the data. The difference here is that the time extend is not committed before the data is added. Instead, when a time extend is needed, the space reserved on the ring buffer is the time extend + the data event size. The time extend is added to the first part of the reserved block and the data is added to the second. The time extend event is passed back to the reserver, but since the reserver also uses a function to find the data portion of the reserved block, no changes to the ring buffer interface need to be made. When a commit is discarded, we now remove both the time extend and the event. With this approach no more than one time extend can be in the buffer in a row. Data must always follow a time extend. Thanks to Mathieu Desnoyers for suggesting this idea. Suggested-by: Mathieu Desnoyers Cc: Thomas Gleixner Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 266 ++++++++++++++++++++++++--------------------- 1 file changed, 142 insertions(+), 124 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index 3af77cd47f2..f50f43107e9 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -224,6 +224,9 @@ enum { RB_LEN_TIME_STAMP = 16, }; +#define skip_time_extend(event) \ + ((struct ring_buffer_event *)((char *)event + RB_LEN_TIME_EXTEND)) + static inline int rb_null_event(struct ring_buffer_event *event) { return event->type_len == RINGBUF_TYPE_PADDING && !event->time_delta; @@ -248,8 +251,12 @@ rb_event_data_length(struct ring_buffer_event *event) return length + RB_EVNT_HDR_SIZE; } -/* inline for ring buffer fast paths */ -static unsigned +/* + * Return the length of the given event. Will return + * the length of the time extend if the event is a + * time extend. + */ +static inline unsigned rb_event_length(struct ring_buffer_event *event) { switch (event->type_len) { @@ -274,13 +281,41 @@ rb_event_length(struct ring_buffer_event *event) return 0; } +/* + * Return total length of time extend and data, + * or just the event length for all other events. + */ +static inline unsigned +rb_event_ts_length(struct ring_buffer_event *event) +{ + unsigned len = 0; + + if (event->type_len == RINGBUF_TYPE_TIME_EXTEND) { + /* time extends include the data event after it */ + len = RB_LEN_TIME_EXTEND; + event = skip_time_extend(event); + } + return len + rb_event_length(event); +} + /** * ring_buffer_event_length - return the length of the event * @event: the event to get the length of + * + * Returns the size of the data load of a data event. + * If the event is something other than a data event, it + * returns the size of the event itself. With the exception + * of a TIME EXTEND, where it still returns the size of the + * data load of the data event after it. */ unsigned ring_buffer_event_length(struct ring_buffer_event *event) { - unsigned length = rb_event_length(event); + unsigned length; + + if (event->type_len == RINGBUF_TYPE_TIME_EXTEND) + event = skip_time_extend(event); + + length = rb_event_length(event); if (event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX) return length; length -= RB_EVNT_HDR_SIZE; @@ -294,6 +329,8 @@ EXPORT_SYMBOL_GPL(ring_buffer_event_length); static void * rb_event_data(struct ring_buffer_event *event) { + if (event->type_len == RINGBUF_TYPE_TIME_EXTEND) + event = skip_time_extend(event); BUG_ON(event->type_len > RINGBUF_TYPE_DATA_TYPE_LEN_MAX); /* If length is in len field, then array[0] has the data */ if (event->type_len) @@ -1546,6 +1583,25 @@ static void rb_inc_iter(struct ring_buffer_iter *iter) iter->head = 0; } +/* Slow path, do not inline */ +static noinline struct ring_buffer_event * +rb_add_time_stamp(struct ring_buffer_event *event, u64 delta) +{ + event->type_len = RINGBUF_TYPE_TIME_EXTEND; + + /* Not the first event on the page? */ + if (rb_event_index(event)) { + event->time_delta = delta & TS_MASK; + event->array[0] = delta >> TS_SHIFT; + } else { + /* nope, just zero it */ + event->time_delta = 0; + event->array[0] = 0; + } + + return skip_time_extend(event); +} + /** * ring_buffer_update_event - update event type and data * @event: the even to update @@ -1558,28 +1614,31 @@ static void rb_inc_iter(struct ring_buffer_iter *iter) * data field. */ static void -rb_update_event(struct ring_buffer_event *event, - unsigned type, unsigned length) +rb_update_event(struct ring_buffer_per_cpu *cpu_buffer, + struct ring_buffer_event *event, unsigned length, + int add_timestamp, u64 delta) { - event->type_len = type; - - switch (type) { - - case RINGBUF_TYPE_PADDING: - case RINGBUF_TYPE_TIME_EXTEND: - case RINGBUF_TYPE_TIME_STAMP: - break; + /* Only a commit updates the timestamp */ + if (unlikely(!rb_event_is_commit(cpu_buffer, event))) + delta = 0; - case 0: - length -= RB_EVNT_HDR_SIZE; - if (length > RB_MAX_SMALL_DATA || RB_FORCE_8BYTE_ALIGNMENT) - event->array[0] = length; - else - event->type_len = DIV_ROUND_UP(length, RB_ALIGNMENT); - break; - default: - BUG(); + /* + * If we need to add a timestamp, then we + * add it to the start of the resevered space. + */ + if (unlikely(add_timestamp)) { + event = rb_add_time_stamp(event, delta); + length -= RB_LEN_TIME_EXTEND; + delta = 0; } + + event->time_delta = delta; + length -= RB_EVNT_HDR_SIZE; + if (length > RB_MAX_SMALL_DATA || RB_FORCE_8BYTE_ALIGNMENT) { + event->type_len = 0; + event->array[0] = length; + } else + event->type_len = DIV_ROUND_UP(length, RB_ALIGNMENT); } /* @@ -1932,12 +1991,21 @@ rb_move_tail(struct ring_buffer_per_cpu *cpu_buffer, static struct ring_buffer_event * __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, - unsigned type, unsigned long length, u64 ts) + unsigned long length, u64 ts, + u64 delta, int add_timestamp) { struct buffer_page *tail_page; struct ring_buffer_event *event; unsigned long tail, write; + /* + * If the time delta since the last event is too big to + * hold in the time field of the event, then we append a + * TIME EXTEND event ahead of the data event. + */ + if (unlikely(add_timestamp)) + length += RB_LEN_TIME_EXTEND; + tail_page = cpu_buffer->tail_page; write = local_add_return(length, &tail_page->write); @@ -1954,11 +2022,9 @@ __rb_reserve_next(struct ring_buffer_per_cpu *cpu_buffer, event = __rb_page_index(tail_page, tail); kmemcheck_annotate_bitfield(event, bitfield); - rb_update_event(event, type, length); + rb_update_event(cpu_buffer, event, length, add_timestamp, delta); - /* The passed in type is zero for DATA */ - if (likely(!type)) - local_inc(&tail_page->entries); + local_inc(&tail_page->entries); /* * If this is the first commit on the page, then update @@ -1980,7 +2046,7 @@ rb_try_to_discard(struct ring_buffer_per_cpu *cpu_buffer, unsigned long addr; new_index = rb_event_index(event); - old_index = new_index + rb_event_length(event); + old_index = new_index + rb_event_ts_length(event); addr = (unsigned long)event; addr &= PAGE_MASK; @@ -2006,67 +2072,6 @@ rb_try_to_discard(struct ring_buffer_per_cpu *cpu_buffer, return 0; } -static int -rb_add_time_stamp(struct ring_buffer_per_cpu *cpu_buffer, - u64 ts, u64 delta) -{ - struct ring_buffer_event *event; - int ret; - - WARN_ONCE(delta > (1ULL << 59), - KERN_WARNING "Delta way too big! %llu ts=%llu write stamp = %llu\n", - (unsigned long long)delta, - (unsigned long long)ts, - (unsigned long long)cpu_buffer->write_stamp); - - /* - * The delta is too big, we to add a - * new timestamp. - */ - event = __rb_reserve_next(cpu_buffer, - RINGBUF_TYPE_TIME_EXTEND, - RB_LEN_TIME_EXTEND, - ts); - if (!event) - return -EBUSY; - - if (PTR_ERR(event) == -EAGAIN) - return -EAGAIN; - - /* Only a commited time event can update the write stamp */ - if (rb_event_is_commit(cpu_buffer, event)) { - /* - * If this is the first on the page, then it was - * updated with the page itself. Try to discard it - * and if we can't just make it zero. - */ - if (rb_event_index(event)) { - event->time_delta = delta & TS_MASK; - event->array[0] = delta >> TS_SHIFT; - } else { - /* try to discard, since we do not need this */ - if (!rb_try_to_discard(cpu_buffer, event)) { - /* nope, just zero it */ - event->time_delta = 0; - event->array[0] = 0; - } - } - cpu_buffer->write_stamp = ts; - /* let the caller know this was the commit */ - ret = 1; - } else { - /* Try to discard the event */ - if (!rb_try_to_discard(cpu_buffer, event)) { - /* Darn, this is just wasted space */ - event->time_delta = 0; - event->array[0] = 0; - } - ret = 0; - } - - return ret; -} - static void rb_start_commit(struct ring_buffer_per_cpu *cpu_buffer) { local_inc(&cpu_buffer->committing); @@ -2111,9 +2116,9 @@ rb_reserve_next_event(struct ring_buffer *buffer, unsigned long length) { struct ring_buffer_event *event; - u64 ts, delta = 0; - int commit = 0; + u64 ts, delta; int nr_loops = 0; + int add_timestamp; rb_start_commit(cpu_buffer); @@ -2134,6 +2139,9 @@ rb_reserve_next_event(struct ring_buffer *buffer, length = rb_calculate_event_length(length); again: + add_timestamp = 0; + delta = 0; + /* * We allow for interrupts to reenter here and do a trace. * If one does, it will cause this original code to loop @@ -2172,33 +2180,24 @@ rb_reserve_next_event(struct ring_buffer *buffer, delta = diff; if (unlikely(test_time_stamp(delta))) { - - commit = rb_add_time_stamp(cpu_buffer, ts, delta); - delta = 0; - - if (commit == -EBUSY) - goto out_fail; - - if (commit == -EAGAIN) - goto again; - - RB_WARN_ON(cpu_buffer, commit < 0); + WARN_ONCE(delta > (1ULL << 59), + KERN_WARNING "Delta way too big! %llu ts=%llu write stamp = %llu\n", + (unsigned long long)delta, + (unsigned long long)ts, + (unsigned long long)cpu_buffer->write_stamp); + add_timestamp = 1; } } get_event: - event = __rb_reserve_next(cpu_buffer, 0, length, ts); + event = __rb_reserve_next(cpu_buffer, length, ts, + delta, add_timestamp); if (unlikely(PTR_ERR(event) == -EAGAIN)) goto again; if (!event) goto out_fail; - if (!rb_event_is_commit(cpu_buffer, event)) - delta = 0; - - event->time_delta = delta; - return event; out_fail: @@ -2311,12 +2310,28 @@ static void rb_update_write_stamp(struct ring_buffer_per_cpu *cpu_buffer, struct ring_buffer_event *event) { + u64 delta; + /* * The event first in the commit queue updates the * time stamp. */ - if (rb_event_is_commit(cpu_buffer, event)) - cpu_buffer->write_stamp += event->time_delta; + if (rb_event_is_commit(cpu_buffer, event)) { + /* + * A commit event that is first on a page + * updates the write timestamp with the page stamp + */ + if (!rb_event_index(event)) + cpu_buffer->write_stamp = + cpu_buffer->commit_page->page->time_stamp; + else if (event->type_len == RINGBUF_TYPE_TIME_EXTEND) { + delta = event->array[0]; + delta <<= TS_SHIFT; + delta += event->time_delta; + cpu_buffer->write_stamp += delta; + } else + cpu_buffer->write_stamp += event->time_delta; + } } static void rb_commit(struct ring_buffer_per_cpu *cpu_buffer, @@ -2356,6 +2371,9 @@ EXPORT_SYMBOL_GPL(ring_buffer_unlock_commit); static inline void rb_event_discard(struct ring_buffer_event *event) { + if (event->type_len == RINGBUF_TYPE_TIME_EXTEND) + event = skip_time_extend(event); + /* array[0] holds the actual length for the discarded event */ event->array[0] = rb_event_data_length(event) - RB_EVNT_HDR_SIZE; event->type_len = RINGBUF_TYPE_PADDING; @@ -3043,12 +3061,12 @@ rb_buffer_peek(struct ring_buffer_per_cpu *cpu_buffer, u64 *ts, again: /* - * We repeat when a timestamp is encountered. It is possible - * to get multiple timestamps from an interrupt entering just - * as one timestamp is about to be written, or from discarded - * commits. The most that we can have is the number on a single page. + * We repeat when a time extend is encountered. + * Since the time extend is always attached to a data event, + * we should never loop more than once. + * (We never hit the following condition more than twice). */ - if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE)) + if (RB_WARN_ON(cpu_buffer, ++nr_loops > 2)) return NULL; reader = rb_get_reader_page(cpu_buffer); @@ -3124,14 +3142,12 @@ rb_iter_peek(struct ring_buffer_iter *iter, u64 *ts) return NULL; /* - * We repeat when a timestamp is encountered. - * We can get multiple timestamps by nested interrupts or also - * if filtering is on (discarding commits). Since discarding - * commits can be frequent we can get a lot of timestamps. - * But we limit them by not adding timestamps if they begin - * at the start of a page. + * We repeat when a time extend is encountered. + * Since the time extend is always attached to a data event, + * we should never loop more than once. + * (We never hit the following condition more than twice). */ - if (RB_WARN_ON(cpu_buffer, ++nr_loops > RB_TIMESTAMPS_PER_PAGE)) + if (RB_WARN_ON(cpu_buffer, ++nr_loops > 2)) return NULL; if (rb_per_cpu_empty(cpu_buffer)) @@ -3829,7 +3845,8 @@ int ring_buffer_read_page(struct ring_buffer *buffer, if (len > (commit - read)) len = (commit - read); - size = rb_event_length(event); + /* Always keep the time extend and data together */ + size = rb_event_ts_length(event); if (len < size) goto out_unlock; @@ -3851,7 +3868,8 @@ int ring_buffer_read_page(struct ring_buffer *buffer, break; event = rb_reader_event(cpu_buffer); - size = rb_event_length(event); + /* Always keep the time extend and data together */ + size = rb_event_ts_length(event); } while (len > size); /* update bpage */ -- cgit v1.2.2 From 140ff89127c74b1b1c1b0152a36ea3720ccf6bc3 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Fri, 8 Oct 2010 10:50:30 -0400 Subject: ring-buffer: Remove condition to add timestamp in fast path There's a condition to check if we should add a time extend or not in the fast path. But this condition is racey (in the sense that we can add a unnecessary time extend, but nothing that can break anything). We later check if the time or event time delta should be zero or have real data in it (not racey), making this first check redundant. This check may help save space once in a while, but really is not worth the hassle to try to save some space that happens at most 134 ms at a time. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 28 ++++++---------------------- 1 file changed, 6 insertions(+), 22 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index f50f43107e9..d9f3e7a8213 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2119,6 +2119,7 @@ rb_reserve_next_event(struct ring_buffer *buffer, u64 ts, delta; int nr_loops = 0; int add_timestamp; + u64 diff; rb_start_commit(cpu_buffer); @@ -2155,29 +2156,13 @@ rb_reserve_next_event(struct ring_buffer *buffer, goto out_fail; ts = rb_time_stamp(cpu_buffer->buffer); + diff = ts - cpu_buffer->write_stamp; - /* - * Only the first commit can update the timestamp. - * Yes there is a race here. If an interrupt comes in - * just after the conditional and it traces too, then it - * will also check the deltas. More than one timestamp may - * also be made. But only the entry that did the actual - * commit will be something other than zero. - */ - if (likely(cpu_buffer->tail_page == cpu_buffer->commit_page && - rb_page_write(cpu_buffer->tail_page) == - rb_commit_index(cpu_buffer))) { - u64 diff; - - diff = ts - cpu_buffer->write_stamp; - - /* make sure this diff is calculated here */ - barrier(); - - /* Did the write stamp get updated already? */ - if (unlikely(ts < cpu_buffer->write_stamp)) - goto get_event; + /* make sure this diff is calculated here */ + barrier(); + /* Did the write stamp get updated already? */ + if (likely(ts >= cpu_buffer->write_stamp)) { delta = diff; if (unlikely(test_time_stamp(delta))) { WARN_ONCE(delta > (1ULL << 59), @@ -2189,7 +2174,6 @@ rb_reserve_next_event(struct ring_buffer *buffer, } } - get_event: event = __rb_reserve_next(cpu_buffer, length, ts, delta, add_timestamp); if (unlikely(PTR_ERR(event) == -EAGAIN)) -- cgit v1.2.2 From d9abde2138e0a00a0d7e44676928efa0ef629d48 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 19 Oct 2010 13:17:08 -0400 Subject: ring-buffer: Micro-optimize with some strategic inlining By using inline and noinline, we are able to make the fast path of recording an event 4% faster. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 23 +++++++++++++++-------- 1 file changed, 15 insertions(+), 8 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index d9f3e7a8213..f5007d0d932 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -2078,7 +2078,7 @@ static void rb_start_commit(struct ring_buffer_per_cpu *cpu_buffer) local_inc(&cpu_buffer->commits); } -static void rb_end_commit(struct ring_buffer_per_cpu *cpu_buffer) +static inline void rb_end_commit(struct ring_buffer_per_cpu *cpu_buffer) { unsigned long commits; @@ -2193,13 +2193,9 @@ rb_reserve_next_event(struct ring_buffer *buffer, #define TRACE_RECURSIVE_DEPTH 16 -static int trace_recursive_lock(void) +/* Keep this code out of the fast path cache */ +static noinline void trace_recursive_fail(void) { - current->trace_recursion++; - - if (likely(current->trace_recursion < TRACE_RECURSIVE_DEPTH)) - return 0; - /* Disable all tracing before we do anything else */ tracing_off_permanent(); @@ -2211,10 +2207,21 @@ static int trace_recursive_lock(void) in_nmi()); WARN_ON_ONCE(1); +} + +static inline int trace_recursive_lock(void) +{ + current->trace_recursion++; + + if (likely(current->trace_recursion < TRACE_RECURSIVE_DEPTH)) + return 0; + + trace_recursive_fail(); + return -1; } -static void trace_recursive_unlock(void) +static inline void trace_recursive_unlock(void) { WARN_ON_ONCE(!current->trace_recursion); -- cgit v1.2.2 From b8b2663bd7c9da04ac804659b9f617c199d0252c Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Tue, 19 Oct 2010 13:23:25 -0400 Subject: ring-buffer: Remove unused macro RB_TIMESTAMPS_PER_PAGE With the binding of time extends to events we no longer need to use the macro RB_TIMESTAMPS_PER_PAGE. Remove it. Signed-off-by: Steven Rostedt --- kernel/trace/ring_buffer.c | 3 --- 1 file changed, 3 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index f5007d0d932..ad25490f8b4 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -441,9 +441,6 @@ static inline int test_time_stamp(u64 delta) /* Max payload is BUF_PAGE_SIZE - header (8bytes) */ #define BUF_MAX_DATA_SIZE (BUF_PAGE_SIZE - (sizeof(u32) * 2)) -/* Max number of timestamps that can fit on a page */ -#define RB_TIMESTAMPS_PER_PAGE (BUF_PAGE_SIZE / RB_LEN_TIME_EXTEND) - int ring_buffer_print_page_header(struct trace_seq *s) { struct buffer_data_page field; -- cgit v1.2.2 From dd49a38cf30944be27892c10b1c0e5b3fa73bcb2 Mon Sep 17 00:00:00 2001 From: Steven Rostedt Date: Wed, 20 Oct 2010 21:51:26 -0400 Subject: tracing: Do not limit the size of the number of CPU buffers The tracing per_cpu buffers were limited to 999 CPUs for a mear savings in stack space of a char array. Up the array to 30 characters which is more than enough to hold a 64 bit number. Reported-by: Robin Holt Suggested-by: Ingo Molnar Signed-off-by: Steven Rostedt --- kernel/trace/trace.c | 8 ++------ 1 file changed, 2 insertions(+), 6 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace.c b/kernel/trace/trace.c index 001bcd2ccf4..82d9b8106cd 100644 --- a/kernel/trace/trace.c +++ b/kernel/trace/trace.c @@ -3996,13 +3996,9 @@ static void tracing_init_debugfs_percpu(long cpu) { struct dentry *d_percpu = tracing_dentry_percpu(); struct dentry *d_cpu; - /* strlen(cpu) + MAX(log10(cpu)) + '\0' */ - char cpu_dir[7]; + char cpu_dir[30]; /* 30 characters should be more than enough */ - if (cpu > 999 || cpu < 0) - return; - - sprintf(cpu_dir, "cpu%ld", cpu); + snprintf(cpu_dir, 30, "cpu%ld", cpu); d_cpu = debugfs_create_dir(cpu_dir, d_percpu); if (!d_cpu) { pr_warning("Could not create debugfs '%s' entry\n", cpu_dir); -- cgit v1.2.2 From e3bda3ac33d3bf3e5a4049e2cabe82d3caaffc26 Mon Sep 17 00:00:00 2001 From: Jason Wessel Date: Mon, 11 Oct 2010 10:20:14 -0500 Subject: kdb,ftdump: Remove reference to internal kdb include Now that include/linux/kdb.h properly exports all the functions required to dynamically add a kdb shell command, the reference to the private kdb header can be removed. Signed-off-by: Jason Wessel --- kernel/trace/trace_kdb.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_kdb.c b/kernel/trace/trace_kdb.c index 7b8ecd751d9..3c5c5dfea0b 100644 --- a/kernel/trace/trace_kdb.c +++ b/kernel/trace/trace_kdb.c @@ -13,7 +13,6 @@ #include #include -#include "../debug/kdb/kdb_private.h" #include "trace.h" #include "trace_output.h" -- cgit v1.2.2 From aa7b250c252cc8e6b1daf0e1eada5eba42a1a68d Mon Sep 17 00:00:00 2001 From: Joe Perches Date: Thu, 21 Oct 2010 22:17:19 -0700 Subject: tracing: Fix 'faild' -> 'failed' typo Signed-off-by: Joe Perches Cc: Steven Rostedt Cc: Frederic Weisbecker Cc: Jiri Kosina LKML-Reference: Signed-off-by: Ingo Molnar --- kernel/trace/trace_kprobe.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index 544301d29de..b8d2852baa4 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -648,7 +648,7 @@ static int register_trace_probe(struct trace_probe *tp) } ret = register_probe_event(tp); if (ret) { - pr_warning("Faild to register probe event(%d)\n", ret); + pr_warning("Failed to register probe event(%d)\n", ret); goto end; } -- cgit v1.2.2 From 61d8e11e519ee7912ab59610fba1aaf08e3c1d84 Mon Sep 17 00:00:00 2001 From: Zimny Lech Date: Wed, 27 Oct 2010 15:34:53 -0700 Subject: Remove duplicate includes from many files Signed-off-by: Zimny Lech Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- kernel/trace/trace_kprobe.c | 1 - 1 file changed, 1 deletion(-) (limited to 'kernel/trace') diff --git a/kernel/trace/trace_kprobe.c b/kernel/trace/trace_kprobe.c index b8d2852baa4..2dec9bcde8b 100644 --- a/kernel/trace/trace_kprobe.c +++ b/kernel/trace/trace_kprobe.c @@ -31,7 +31,6 @@ #include #include #include -#include #include #include "trace.h" -- cgit v1.2.2 From 02e031cbc843b010e72fcc05c76113c688b2860f Mon Sep 17 00:00:00 2001 From: Christoph Hellwig Date: Wed, 10 Nov 2010 14:54:09 +0100 Subject: block: remove REQ_HARDBARRIER REQ_HARDBARRIER is dead now, so remove the leftovers. What's left at this point is: - various checks inside the block layer. - sanity checks in bio based drivers. - now unused bio_empty_barrier helper. - Xen blockfront use of BLKIF_OP_WRITE_BARRIER - it's dead for a while, but Xen really needs to sort out it's barrier situaton. - setting of ordered tags in uas - dead code copied from old scsi drivers. - scsi different retry for barriers - it's dead and should have been removed when flushes were converted to FS requests. - blktrace handling of barriers - removed. Someone who knows blktrace better should add support for REQ_FLUSH and REQ_FUA, though. Signed-off-by: Christoph Hellwig Signed-off-by: Jens Axboe --- kernel/trace/blktrace.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'kernel/trace') diff --git a/kernel/trace/blktrace.c b/kernel/trace/blktrace.c index bc251ed6672..7b8ec028154 100644 --- a/kernel/trace/blktrace.c +++ b/kernel/trace/blktrace.c @@ -168,7 +168,6 @@ static int act_log_check(struct blk_trace *bt, u32 what, sector_t sector, static const u32 ddir_act[2] = { BLK_TC_ACT(BLK_TC_READ), BLK_TC_ACT(BLK_TC_WRITE) }; -#define BLK_TC_HARDBARRIER BLK_TC_BARRIER #define BLK_TC_RAHEAD BLK_TC_AHEAD /* The ilog2() calls fall out because they're constant */ @@ -196,7 +195,6 @@ static void __blk_add_trace(struct blk_trace *bt, sector_t sector, int bytes, return; what |= ddir_act[rw & WRITE]; - what |= MASK_TC_BIT(rw, HARDBARRIER); what |= MASK_TC_BIT(rw, SYNC); what |= MASK_TC_BIT(rw, RAHEAD); what |= MASK_TC_BIT(rw, META); @@ -1807,8 +1805,6 @@ void blk_fill_rwbs(char *rwbs, u32 rw, int bytes) if (rw & REQ_RAHEAD) rwbs[i++] = 'A'; - if (rw & REQ_HARDBARRIER) - rwbs[i++] = 'B'; if (rw & REQ_SYNC) rwbs[i++] = 'S'; if (rw & REQ_META) -- cgit v1.2.2