Re: [PATCH] kmemleak: record accurate early log buffer count and report when exceeded

Thanks a lot for the correction. On 2015/4/15 23:13, Catalin Marinas wrote: > (I see you corrected the LKML address; I replied to your early patch) > > On Wed, Apr 15, 2015 at 12:49:23PM +0100, Wang Kai wrote: >> In log_early function, crt_early_log should also count once when >> 'crt_early_log >

Re: [PATCH] kmemleak: record accurate early log buffer count and report when exceeded

(I see you corrected the LKML address; I replied to your early patch) On Wed, Apr 15, 2015 at 12:49:23PM +0100, Wang Kai wrote: > In log_early function, crt_early_log should also count once when > 'crt_early_log >= ARRAY_SIZE(early_log)'. Otherwise the reported > count from kmemleak_init is one le

Re: [PATCH] kmemleak: record accurate early log buffer count and report when exceeded

BTW, you misspelled the LKML address (I fixed it: s/\.or/\.org/) On Wed, Apr 15, 2015 at 12:44:52PM +0100, Wang Kai wrote: > In log_early function, crt_early_log should also count once when > 'crt_early_log >= ARRAY_SIZE(early_log)'. Otherwise the reported > count from kmemleak_init is one less th

[PATCH] kmemleak: record accurate early log buffer count and report when exceeded

In log_early function, crt_early_log should also count once when 'crt_early_log >= ARRAY_SIZE(early_log)'. Otherwise the reported count from kmemleak_init is one less than 'actual number'. Then, in kmemleak_init, if early_log buffer size equal actual number, kmemleak will init sucessful, so change