From: Steven Rostedt (Google) rostedt@goodmis.org
[ Upstream commit 31029a8b2c7e656a0289194ef16415050ae4c4ac ]
The function ring_buffer_nr_dirty_pages() was created to find out how many pages are filled in the ring buffer. There's two running counters. One is incremented whenever a new page is touched (pages_touched) and the other is whenever a page is read (pages_read). The dirty count is the number touched minus the number read. This is used to determine if a blocked task should be woken up if the percentage of the ring buffer it is waiting for is hit.
The problem is that it does not take into account dropped pages (when the new writes overwrite pages that were not read). And then the dirty pages will always be greater than the percentage.
This makes the "buffer_percent" file inaccurate, as the number of dirty pages end up always being larger than the percentage, event when it's not and this causes user space to be woken up more than it wants to be.
Add a new counter to keep track of lost pages, and include that in the accounting of dirty pages so that it is actually accurate.
Link: https://lkml.kernel.org/r/20221021123013.55fb6055@gandalf.local.home
Fixes: 2c2b0a78b3739 ("ring-buffer: Add percentage of ring buffer full to wake up reader") Signed-off-by: Steven Rostedt (Google) rostedt@goodmis.org Signed-off-by: Sasha Levin sashal@kernel.org --- kernel/trace/ring_buffer.c | 12 ++++++++++++ 1 file changed, 12 insertions(+)
diff --git a/kernel/trace/ring_buffer.c b/kernel/trace/ring_buffer.c index f1dd405b98a5..c4234430afea 100644 --- a/kernel/trace/ring_buffer.c +++ b/kernel/trace/ring_buffer.c @@ -477,6 +477,7 @@ struct ring_buffer_per_cpu { local_t committing; local_t commits; local_t pages_touched; + local_t pages_lost; local_t pages_read; long last_pages_touch; size_t shortest_full; @@ -544,10 +545,18 @@ size_t ring_buffer_nr_pages(struct ring_buffer *buffer, int cpu) size_t ring_buffer_nr_dirty_pages(struct ring_buffer *buffer, int cpu) { size_t read; + size_t lost; size_t cnt;
read = local_read(&buffer->buffers[cpu]->pages_read); + lost = local_read(&buffer->buffers[cpu]->pages_lost); cnt = local_read(&buffer->buffers[cpu]->pages_touched); + + if (WARN_ON_ONCE(cnt < lost)) + return 0; + + cnt -= lost; + /* The reader can read an empty page, but not more than that */ if (cnt < read) { WARN_ON_ONCE(read > cnt + 1); @@ -1599,6 +1608,7 @@ rb_remove_pages(struct ring_buffer_per_cpu *cpu_buffer, unsigned long nr_pages) */ local_add(page_entries, &cpu_buffer->overrun); local_sub(BUF_PAGE_SIZE, &cpu_buffer->entries_bytes); + local_inc(&cpu_buffer->pages_lost); }
/* @@ -2023,6 +2033,7 @@ rb_handle_head_page(struct ring_buffer_per_cpu *cpu_buffer, */ local_add(entries, &cpu_buffer->overrun); local_sub(BUF_PAGE_SIZE, &cpu_buffer->entries_bytes); + local_inc(&cpu_buffer->pages_lost);
/* * The entries will be zeroed out when we move the @@ -4475,6 +4486,7 @@ rb_reset_cpu(struct ring_buffer_per_cpu *cpu_buffer) local_set(&cpu_buffer->committing, 0); local_set(&cpu_buffer->commits, 0); local_set(&cpu_buffer->pages_touched, 0); + local_set(&cpu_buffer->pages_lost, 0); local_set(&cpu_buffer->pages_read, 0); cpu_buffer->last_pages_touch = 0; cpu_buffer->shortest_full = 0;