Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1311677 > unrolled thread
| Started by | Jaewon Kim <jaewon31.kim@gmail.com> |
|---|---|
| First post | 2016-01-18 18:00 +0100 |
| Last post | 2016-01-18 18:00 +0100 |
| Articles | 1 — 1 participant |
Back to article view | Back to linux.kernel
[RFC] ratelimit: fix time interval by setting right start time Jaewon Kim <jaewon31.kim@gmail.com> - 2016-01-18 18:00 +0100
| From | Jaewon Kim <jaewon31.kim@gmail.com> |
|---|---|
| Date | 2016-01-18 18:00 +0100 |
| Subject | [RFC] ratelimit: fix time interval by setting right start time |
| Message-ID | <qSljT-3pb-57@gated-at.bofh.it> |
rs->begin should be initialized to the current jiffies if the begin was 0. Even when the time interval passes, the current resetting logic sets the begin back to 0. So next interval starts from the next call rather than from the current resetting call. This incurrs improper suppression. For example: B will be suppressed althouth 3 seconds passed. DEFINE_RATELIMIT_STATE(limit, HZ, 1); __ratelimit(&limit); /* A */ sleep(3); __ratelimit(&limit); /* B */ Signed-off-by: Jaewon Kim <jaewon31.kim@gmail.com> --- lib/ratelimit.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/lib/ratelimit.c b/lib/ratelimit.c index 40e03ea..2c5de86 100644 --- a/lib/ratelimit.c +++ b/lib/ratelimit.c @@ -49,7 +49,7 @@ int ___ratelimit(struct ratelimit_state *rs, const char *func) if (rs->missed) printk(KERN_WARNING "%s: %d callbacks suppressed\n", func, rs->missed); - rs->begin = 0; + rs->begin = jiffies; rs->printed = 0; rs->missed = 0; } -- 1.9.1
Back to top | Article view | linux.kernel
csiph-web