From: Eric Sandeen Subject: Re: "data=writeback" and TRIM don't get along Date: Wed, 07 Apr 2010 23:37:36 -0500 Message-ID: <4BBD5D90.4090203@redhat.com> References: <4BBD285B.9000603@gmail.com> <4BBD2FDF.4040407@redhat.com> <4BBD3365.90306@gmail.com> <4BBD5740.4070101@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: linux-ext4@vger.kernel.org To: Nebojsa Trpkovic Return-path: Received: from mx1.redhat.com ([209.132.183.28]:16521 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750746Ab0DHEhl (ORCPT ); Thu, 8 Apr 2010 00:37:41 -0400 In-Reply-To: <4BBD5740.4070101@redhat.com> Sender: linux-ext4-owner@vger.kernel.org List-ID: Eric Sandeen wrote: > I'll have to think about the right way to do this... it seems pretty > convoluted to me right now. > Something like this probably works, but I really REALLY would not test it on an important filesystem. :) I'm not sure it's a good idea to discard it before returning it to the prealloc pool, because it may well get re-used again quickly.... not sure if that's helpful. Just a note, I think eventually we may move to more of a batch discard in the background, because these little discards are actually quite inefficient on the hardware we've tested so far. -Eric p.s. really. Don't test this with important data. I haven't tested it at all yet. Index: linux-2.6/fs/ext4/mballoc.c =================================================================== --- linux-2.6.orig/fs/ext4/mballoc.c +++ linux-2.6/fs/ext4/mballoc.c @@ -4602,6 +4606,8 @@ do_more: mb_clear_bits(bitmap_bh->b_data, bit, count); ext4_mb_free_metadata(handle, &e4b, new_entry); } else { + ext4_fsblk_t discard_block; + /* need to update group_info->bb_free and bitmap * with group lock held. generate_buddy look at * them with group lock_held @@ -4609,6 +4615,11 @@ do_more: ext4_lock_group(sb, block_group); mb_clear_bits(bitmap_bh->b_data, bit, count); mb_free_blocks(inode, &e4b, bit, count); + discard_block = bit + + ext4_group_first_block_no(sb, block_group); + trace_ext4_discard_blocks(sb, + (unsigned long long)discard_block, count); + sb_issue_discard(sb, discard_block, count); ext4_mb_return_to_preallocation(inode, &e4b, block, count); }