[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20170823060332.GC22976@X58A-UD3R>
Date: Wed, 23 Aug 2017 15:03:32 +0900
From: Byungchul Park <byungchul.park@....com>
To: Peter Zijlstra <peterz@...radead.org>
Cc: Dave Chinner <david@...morbit.com>, mingo@...nel.org,
linux-kernel@...r.kernel.org, kernel-team@....com,
Arnaldo Carvalho de Melo <acme@...nel.org>,
Tejun Heo <tj@...nel.org>, Boqun Feng <boqun.feng@...il.com>
Subject: Re: [PATCH v3 1/3] lockdep: Make LOCKDEP_CROSSRELEASE configs all
part of PROVE_LOCKING
On Wed, Aug 23, 2017 at 11:12:17AM +0900, Byungchul Park wrote:
> > > We have to detect dependecies if it exists, even in the following case:
> > >
> > > oooooooiiiiiiiiiiiiiiiiiii.........
> > > |<- range for commit ->|
> > >
> > > where
> > > o: acquisition outside of each work,
> > > i: acquisition inside of each work,
> > >
> > > With yours, we can never detect dependecies wrt 'o'.
> >
> > There really shouldn't be any o's when you call
>
> There can be any o's.
I meant it's very possible for 'o's to exist. And we have to, of course,
consider them wrt dependencies. No reason we should ignore them.
> > crossrelease_hist_start(XHLOCK_PROC), it should denote the bottom of a
Honestly, I prefer another naming like XHLOCK_WORK or XHLOCK_SYSCALL over
XHLOCK_PROC, since the functions are for special contexts e.g. works.
But I thought XHLOCK_PROC is not bad because XHLOCK_WORK and
XHLOCK_SYSCALL does never conflict to each other and they anyway run as
normal process contexts.
Remind they are for special contexts.
Powered by blists - more mailing lists