Date: Wed, 17 Aug 2005 08:36:04 -0400 From: John Baldwin <jhb@FreeBSD.org> To: freebsd-arch@FreeBSD.org Cc: gnn@FreeBSD.org Subject: Re: Special schedulers, one CPU only kernel, one only userland Message-ID: <200508170836.05861.jhb@FreeBSD.org> In-Reply-To: <m2pssdjoob.wl%gnn@neville-neil.com> References: <42F9ECF2.8080809@freebsd.org> <200508161716.44597.jhb@FreeBSD.org> <m2pssdjoob.wl%gnn@neville-neil.com>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tuesday 16 August 2005 10:13 pm, gnn@freebsd.org wrote:
> At Tue, 16 Aug 2005 16:22:41 -0700,
>
> luigi wrote:
> > ok just for the records, the issue i had in mind is the release/acquire
> > the mutex around some code that might cause a deadlock, not for the
> > mutex ops per se, but for the need to make sure that the data
> > structure is consistent before releasing the lock, and rechecking
> > the state afterwards. Basically:
> >
> > 1 mtx_lock(&foo)
> > 2 .. work on obj_foo
> > 3 .. make obj_foo consistent
> > 4 mtx_unlock(&foo)
> > 5 f()
> > 6 mtx_lock(&foo)
> > 7 .. revalidate state of obj_foo
> > 8 .. more work
> >
> > where f() is the call that might sleep or cause a deadlock.
> > In cases where f() has a low probability of blocking, we could
> > save a bit of work at runtime by writing the code as below:
> >
> > 1 mtx_lock(&foo)
> > 2 .. work on obj_foo
> > if (try_f_without_blocking() =3D=3D EWOULDBLOCK) {
> > 3 .. make obj_foo consistent
> > 4 mtx_unlock(&foo)
> > 5 f()
> > 6 mtx_lock(&foo)
> > 7 .. revalidate state of obj_foo
> > }
> > 8 .. more work
> >
> > (where try_f_without_blocking() is a version of f() that returns
> > EWOULDBLOCK in case the 'other' lock is busy).
> > Here maybe we would benefit by some support (macros, whatever) that
> > permits us to specify in a compact way what to do around f() should
> > it become blocking.
> >
> > On the other hand, maybe the instances of code as the one above
> > are so rare that there is hardly any need for that.
>
> Correct me if I'm wrong but I suspect you're thinking of cases such
> as:
>
> RT_UNLOCK(rt); /* XXX workaround LOR */
> gwrt =3D rtalloc1(gate, 1, 0);
> RT_LOCK(rt);
>
> in the routing/networking code. A quick, and by no means exhaustive,
> check of CURRENT with cscope and Emacs looking for these turned up 3.
> It may still be something to look at, but perhaps mroe from a point of
> view of cleaning up our APIs to not do this kind of jiggery pokery,
> rather than inventing a new locking paradigm, which is fraught with
> peril.
>
> This isn't exhaustive though, if others can show this kind of thing
> being done a lot, or becoming a idiom in our code, then there is more
> cause for concern.
>
> Just my 2 yen :-)
Agreed. I'd rather that we fix our APIs and try to organize the workflow t=
o=20
minimize the number of times that a lock has to be dropped and then=20
reacquired.
=2D-=20
John Baldwin <jhb@FreeBSD.org> =A0<>< =A0http://www.FreeBSD.org/~jhb/
"Power Users Use the Power to Serve" =A0=3D =A0http://www.FreeBSD.org
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200508170836.05861.jhb>