From mboxrd@z Thu Jan 1 00:00:00 1970 From: Oliver Neukum Subject: Re: Re: [PATCH] Remove process freezer from suspend to RAM pathway Date: Thu, 5 Jul 2007 22:53:34 +0200 Message-ID: <200707052253.34817.oliver@neukum.org> References: <18059.10554.955290.148535@cargo.ozlabs.ibm.com> <200707052249.45495.oliver@neukum.org> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: quoted-printable Return-path: In-Reply-To: <200707052249.45495.oliver@neukum.org> Content-Disposition: inline List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: linux-pm-bounces@lists.linux-foundation.org Errors-To: linux-pm-bounces@lists.linux-foundation.org To: linux-pm@lists.linux-foundation.org Cc: mjg59@srcf.ucam.org, johannes@sipsolutions.net, linux-kernel@vger.kernel.org, pavel@ucw.cz, Miklos Szeredi List-Id: linux-pm@vger.kernel.org Am Donnerstag, 5. Juli 2007 schrieb Oliver Neukum: > Am Donnerstag, 5. Juli 2007 schrieb Miklos Szeredi: > > > > Yes, fuse could handle being frozen there. =C2=A0However that wou= ld only > > > > solve part of the problem: an operation waiting for a reply could= be > > > > holding a VFS mutex and some other task may be blocked on that mu= tex. > > > >=20 > > > > How would you solve freezing those tasks? > > >=20 > > > OK, you made me reach for literatur on theoretical computer science. > > >=20 > > > IMHO the range of actions a fuse server is inherently limited. > > > You must never ever block on a lock one of your clients is holding.= In > > > this case the limitation is not influenced by the freezer. > >=20 > > Obviously. =C2=A0But I wasn't about the server trying to acquire a lo= ck > > held by a client. =C2=A0I was talking about a client trying to acquir= e a > > lock held by _another_ client. > >=20 > > If this coincides with the server (or some other task which the serve= r > > is depending on) being frozen before the clients, the freezer has a > > problem. >=20 > True, but that case can only happen if servers are frozen before client= s. > You don't need a full dependency graph. A simple set sequence of two > classes of tasks will do. Any replying to myself. A deadlock here is not fatal. You can and will timeout in the freezer and can try again. REegards Oliver