There is a bug in 2.4 kernels which affects INN. Can somebody answer the question related to mmap? On this server I'm still using 1.7.x+insync. -- ciao, Marco -- Attached file included as plaintext by Listar -- Return-Path: Delivered-To: md@wonderland.linux.it Received: by wonderland.linux.it (Postfix/Md, from userid 10) id 45D4C17DD0; Fri, 15 Sep 2000 14:51:12 +0200 (CEST) Received: from ils with UUCP (rmail); Fri, 15 Sep 2000 14:51:12 +0200 Received: by giano.linux.it (Postfix, from userid 10) id 9A3A23765; Fri, 15 Sep 2000 14:25:16 +0200 (CEST) Received: from math.psu.edu (leibniz.math.psu.edu [146.186.130.2]) by spock.linux.it (Postfix) with ESMTP id 3B034196DF for ; Fri, 15 Sep 2000 14:21:12 +0200 (CEST) Received: from weyl.math.psu.edu (weyl.math.psu.edu [146.186.130.226]) by math.psu.edu (8.9.3/8.9.3) with ESMTP id IAA29390; Fri, 15 Sep 2000 08:21:03 -0400 (EDT) Received: from localhost (viro@localhost) by weyl.math.psu.edu (8.9.3/8.9.3) with ESMTP id IAA04247; Fri, 15 Sep 2000 08:21:02 -0400 (EDT) X-Authentication-Warning: weyl.math.psu.edu: viro owned process doing -bs Date: Fri, 15 Sep 2000 08:21:02 -0400 (EDT) From: Alexander Viro To: "Marco d'Itri" Cc: tytso@mit.edu, torvalds@transmeta.com, linux-kernel@vger.kernel.org Subject: Re: Update Linux 2.4 Status/TODO list In-Reply-To: <20000915131540.A1485@wonderland.linux.it> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII On Fri, 15 Sep 2000, Marco d'Itri wrote: > On Sep 14, Alexander Viro wrote: > > >> - the whole active file is now 100% identical to the saved copy > >Ugh... How about relevant subset of strace? > I tried doing that for Andrea but I think it's not useful, active is > mmapped and strace shows nothing interesting. mmapped... Does mmap() go past the end of file? > >> Right now it happened after the daily expire run: I stopped INN and the > >> file on disk changed to the copy I saved before expire started. > >Wait a minute. I don't believe in on-disk file being restored by magic, > >but I could believe in page(s) being never written to disk and giving the > >impression of "update that doesn't stick". You have a file shorter than > Sure, this is what I meant. > But what about the mtime which does not get updated? Well, _that_ makes perfect sense - after all, you update mtime when you write the data and that is the part that didn't happen. > >BTW, how does test8+patch to block_truncate_page() behave? > I missed it, do you still have the patch? Sure. Apply to fs/buffer.c: @@ -1758,13 +1758,14 @@ pos += blocksize; } + err = 0; + if (!buffer_mapped(bh)) { + get_block(inode, iblock, bh, 0); + if (!buffer_mapped(bh)) + goto unlock; + } + if (!buffer_uptodate(bh)) { - err = 0; - if (!buffer_mapped(bh)) { - get_block(inode, iblock, bh, 0); - if (!buffer_mapped(bh)) - goto unlock; - } err = -EIO; bh->b_end_io = end_buffer_io_sync; ll_rw_block(READ, 1, &bh);