From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail.kernel.org ([198.145.29.136]:43007 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751735AbcDZOXG convert rfc822-to-8bit (ORCPT ); Tue, 26 Apr 2016 10:23:06 -0400 Received: from mail.kernel.org (localhost [127.0.0.1]) by mail.kernel.org (Postfix) with ESMTP id 7B43D20142 for ; Tue, 26 Apr 2016 14:23:04 +0000 (UTC) Received: from mail-ig0-f178.google.com (mail-ig0-f178.google.com [209.85.213.178]) (using TLSv1.2 with cipher AES128-GCM-SHA256 (128/128 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id 2286D20138 for ; Tue, 26 Apr 2016 14:23:03 +0000 (UTC) Received: by mail-ig0-f178.google.com with SMTP id bi2so98211331igb.0 for ; Tue, 26 Apr 2016 07:23:03 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1461546072-11154-1-git-send-email-fdmanana@kernel.org> Date: Tue, 26 Apr 2016 15:23:02 +0100 Message-ID: Subject: Re: [PATCH 0/3] Fixes for races in relocation and avoid start and wait for unrelated IO From: Filipe Manana To: =?UTF-8?Q?Holger_Hoffst=C3=A4tte?= Cc: "linux-btrfs@vger.kernel.org" Content-Type: text/plain; charset=UTF-8 Sender: linux-btrfs-owner@vger.kernel.org List-ID: On Tue, Apr 26, 2016 at 2:42 PM, Holger Hoffstätte wrote: > On Mon, Apr 25, 2016 at 3:01 AM, wrote: >> The following patches fix 2 hard to hit races in relocation that make its >> first phase (MOVE_DATA_EXTENTS) miss extents, triggers a warning in the >> second phase (UPDATE_DATA_PTRS) and leaves metadata in an invalid state >> (file extent items pointing to areas corresponding to the deleted block >> group), leading to a BUG_ON() when attempting to read those extents after >> the relocation finishes. > > Never saw this particular race/error, but decided to give these > patches a workout > to see whether they cause any new or unrelated problems. > > Continuous rebalancing (full, partial) for ~30m while unpacking and > deleting kernel > trees on a 16GB tmpfs-backed loopback device did not cause any problem; > balance just cruises along at (sometimes) up to ~1GB/s and does its thing. > Finally btrfs check also found nothing wrong. > > Not sure if this qualifies as testing, but anyway: > > Tested-by: Holger Hoffstaette Thanks. I've now realized I sent the wrong version and patch 1 shouldn't be in the set. I'll resend a v2. > > cheers, > Holger