From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-7.0 required=3.0 tests=HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_PATCH,MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS autolearn=ham autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id 04B5BC43381 for ; Thu, 14 Feb 2019 14:30:43 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id C5B8A2229F for ; Thu, 14 Feb 2019 14:30:42 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S2439205AbfBNOal (ORCPT ); Thu, 14 Feb 2019 09:30:41 -0500 Received: from mx2.suse.de ([195.135.220.15]:43304 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-FAIL) by vger.kernel.org with ESMTP id S2436709AbfBNOaj (ORCPT ); Thu, 14 Feb 2019 09:30:39 -0500 X-Virus-Scanned: by amavisd-new at test-mx.suse.de Received: from relay2.suse.de (unknown [195.135.220.254]) by mx1.suse.de (Postfix) with ESMTP id E261DB03B; Thu, 14 Feb 2019 14:30:35 +0000 (UTC) Subject: Re: [PATCH 1/4] btrfs: Refactor cow_file_range_async To: linux-btrfs@vger.kernel.org References: <20190213155529.31500-1-nborisov@suse.com> <20190213155529.31500-2-nborisov@suse.com> Cc: Josef Bacik From: Nikolay Borisov Openpgp: preference=signencrypt Autocrypt: addr=nborisov@suse.com; prefer-encrypt=mutual; keydata= mQINBFiKBz4BEADNHZmqwhuN6EAzXj9SpPpH/nSSP8YgfwoOqwrP+JR4pIqRK0AWWeWCSwmZ T7g+RbfPFlmQp+EwFWOtABXlKC54zgSf+uulGwx5JAUFVUIRBmnHOYi/lUiE0yhpnb1KCA7f u/W+DkwGerXqhhe9TvQoGwgCKNfzFPZoM+gZrm+kWv03QLUCr210n4cwaCPJ0Nr9Z3c582xc bCUVbsjt7BN0CFa2BByulrx5xD9sDAYIqfLCcZetAqsTRGxM7LD0kh5WlKzOeAXj5r8DOrU2 GdZS33uKZI/kZJZVytSmZpswDsKhnGzRN1BANGP8sC+WD4eRXajOmNh2HL4P+meO1TlM3GLl EQd2shHFY0qjEo7wxKZI1RyZZ5AgJnSmehrPCyuIyVY210CbMaIKHUIsTqRgY5GaNME24w7h TyyVCy2qAM8fLJ4Vw5bycM/u5xfWm7gyTb9V1TkZ3o1MTrEsrcqFiRrBY94Rs0oQkZvunqia c+NprYSaOG1Cta14o94eMH271Kka/reEwSZkC7T+o9hZ4zi2CcLcY0DXj0qdId7vUKSJjEep c++s8ncFekh1MPhkOgNj8pk17OAESanmDwksmzh1j12lgA5lTFPrJeRNu6/isC2zyZhTwMWs k3LkcTa8ZXxh0RfWAqgx/ogKPk4ZxOXQEZetkEyTFghbRH2BIwARAQABtCNOaWtvbGF5IEJv cmlzb3YgPG5ib3Jpc292QHN1c2UuY29tPokCOAQTAQIAIgUCWIo48QIbAwYLCQgHAwIGFQgC CQoLBBYCAwECHgECF4AACgkQcb6CRuU/KFc0eg/9GLD3wTQz9iZHMFbjiqTCitD7B6dTLV1C ddZVlC8Hm/TophPts1bWZORAmYIihHHI1EIF19+bfIr46pvfTu0yFrJDLOADMDH+Ufzsfy2v HSqqWV/nOSWGXzh8bgg/ncLwrIdEwBQBN9SDS6aqsglagvwFD91UCg/TshLlRxD5BOnuzfzI Leyx2c6YmH7Oa1R4MX9Jo79SaKwdHt2yRN3SochVtxCyafDlZsE/efp21pMiaK1HoCOZTBp5 VzrIP85GATh18pN7YR9CuPxxN0V6IzT7IlhS4Jgj0NXh6vi1DlmKspr+FOevu4RVXqqcNTSS E2rycB2v6cttH21UUdu/0FtMBKh+rv8+yD49FxMYnTi1jwVzr208vDdRU2v7Ij/TxYt/v4O8 V+jNRKy5Fevca/1xroQBICXsNoFLr10X5IjmhAhqIH8Atpz/89ItS3+HWuE4BHB6RRLM0gy8 T7rN6ja+KegOGikp/VTwBlszhvfLhyoyjXI44Tf3oLSFM+8+qG3B7MNBHOt60CQlMkq0fGXd mm4xENl/SSeHsiomdveeq7cNGpHi6i6ntZK33XJLwvyf00PD7tip/GUj0Dic/ZUsoPSTF/mG EpuQiUZs8X2xjK/AS/l3wa4Kz2tlcOKSKpIpna7V1+CMNkNzaCOlbv7QwprAerKYywPCoOSC 7P25Ag0EWIoHPgEQAMiUqvRBZNvPvki34O/dcTodvLSyOmK/MMBDrzN8Cnk302XfnGlW/YAQ csMWISKKSpStc6tmD+2Y0z9WjyRqFr3EGfH1RXSv9Z1vmfPzU42jsdZn667UxrRcVQXUgoKg QYx055Q2FdUeaZSaivoIBD9WtJq/66UPXRRr4H/+Y5FaUZx+gWNGmBT6a0S/GQnHb9g3nonD jmDKGw+YO4P6aEMxyy3k9PstaoiyBXnzQASzdOi39BgWQuZfIQjN0aW+Dm8kOAfT5i/yk59h VV6v3NLHBjHVw9kHli3jwvsizIX9X2W8tb1SefaVxqvqO1132AO8V9CbE1DcVT8fzICvGi42 FoV/k0QOGwq+LmLf0t04Q0csEl+h69ZcqeBSQcIMm/Ir+NorfCr6HjrB6lW7giBkQl6hhomn l1mtDP6MTdbyYzEiBFcwQD4terc7S/8ELRRybWQHQp7sxQM/Lnuhs77MgY/e6c5AVWnMKd/z MKm4ru7A8+8gdHeydrRQSWDaVbfy3Hup0Ia76J9FaolnjB8YLUOJPdhI2vbvNCQ2ipxw3Y3c KhVIpGYqwdvFIiz0Fej7wnJICIrpJs/+XLQHyqcmERn3s/iWwBpeogrx2Lf8AGezqnv9woq7 OSoWlwXDJiUdaqPEB/HmGfqoRRN20jx+OOvuaBMPAPb+aKJyle8zABEBAAGJAh8EGAECAAkF AliKBz4CGwwACgkQcb6CRuU/KFdacg/+M3V3Ti9JYZEiIyVhqs+yHb6NMI1R0kkAmzsGQ1jU zSQUz9AVMR6T7v2fIETTT/f5Oout0+Hi9cY8uLpk8CWno9V9eR/B7Ifs2pAA8lh2nW43FFwp IDiSuDbH6oTLmiGCB206IvSuaQCp1fed8U6yuqGFcnf0ZpJm/sILG2ECdFK9RYnMIaeqlNQm iZicBY2lmlYFBEaMXHoy+K7nbOuizPWdUKoKHq+tmZ3iA+qL5s6Qlm4trH28/fPpFuOmgP8P K+7LpYLNSl1oQUr+WlqilPAuLcCo5Vdl7M7VFLMq4xxY/dY99aZx0ZJQYFx0w/6UkbDdFLzN upT7NIN68lZRucImffiWyN7CjH23X3Tni8bS9ubo7OON68NbPz1YIaYaHmnVQCjDyDXkQoKC R82Vf9mf5slj0Vlpf+/Wpsv/TH8X32ajva37oEQTkWNMsDxyw3aPSps6MaMafcN7k60y2Wk/ TCiLsRHFfMHFY6/lq/c0ZdOsGjgpIK0G0z6et9YU6MaPuKwNY4kBdjPNBwHreucrQVUdqRRm RcxmGC6ohvpqVGfhT48ZPZKZEWM+tZky0mO7bhZYxMXyVjBn4EoNTsXy1et9Y1dU3HVJ8fod 5UqrNrzIQFbdeM0/JqSLrtlTcXKJ7cYFa9ZM2AP7UIN9n1UWxq+OPY9YMOewVfYtL8M= Message-ID: <673f2523-c863-b298-971d-30fb0497e3f8@suse.com> Date: Thu, 14 Feb 2019 16:30:34 +0200 User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.2.1 MIME-Version: 1.0 In-Reply-To: <20190213155529.31500-2-nborisov@suse.com> Content-Type: text/plain; charset=utf-8 Content-Language: en-US Content-Transfer-Encoding: 8bit Sender: linux-btrfs-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: linux-btrfs@vger.kernel.org On 13.02.19 г. 17:55 ч., Nikolay Borisov wrote: > This commit changes the implementation of cow_file_range_async in order > to get rid of the BUG_ON in the middle of the loop. Additionally it > reworks the inner loop in the hopes of making it more understandable. > > Main change is that the number of chunks required to handle the given > range is calculated before going into the loop and the logic of the loop > just iterates the chunk count. Furthermore, the way memory is allocated > is reworked and now the code does a single kmalloc with enough space to > handle all chunks. Depending on whether compression is enabled or not > chunks are either 1 (in non-compress case) or the range divided by 512k. > > Signed-off-by: Nikolay Borisov > --- > fs/btrfs/inode.c | 75 +++++++++++++++++++++++++++++++----------------- > 1 file changed, 49 insertions(+), 26 deletions(-) > > diff --git a/fs/btrfs/inode.c b/fs/btrfs/inode.c > index 05bbfd02ea49..c2180f0a5f42 100644 > --- a/fs/btrfs/inode.c > +++ b/fs/btrfs/inode.c > @@ -375,6 +375,7 @@ struct async_cow { > unsigned int write_flags; > struct list_head extents; > struct btrfs_work work; > + atomic_t *pending; > }; > > static noinline int add_async_extent(struct async_cow *cow, > @@ -1181,7 +1182,12 @@ static noinline void async_cow_free(struct btrfs_work *work) > async_cow = container_of(work, struct async_cow, work); > if (async_cow->inode) > btrfs_add_delayed_iput(async_cow->inode); > - kfree(async_cow); > + /* > + * Since the pointer to 'pending' is at the beginning of the array of > + * async_cow's, freeing it ensures the whole array has been freed. > + */ > + if (atomic_dec_and_test(async_cow->pending)) > + kfree(async_cow->pending); > } > > static int cow_file_range_async(struct inode *inode, struct page *locked_page, > @@ -1193,42 +1199,59 @@ static int cow_file_range_async(struct inode *inode, struct page *locked_page, > struct async_cow *async_cow; > unsigned long nr_pages; > u64 cur_end; > + u64 num_chunks = DIV_ROUND_UP(end - start, SZ_512K); > + int i; > + bool should_compress; > + atomic_t *p; > + > > clear_extent_bit(&BTRFS_I(inode)->io_tree, start, end, EXTENT_LOCKED, > 1, 0, NULL); In case of error I wonder shouldn't the cleared range really be within page_Start-page_end. Because subsequent pages which are being written back could follow this whole region? > - while (start < end) { > - async_cow = kmalloc(sizeof(*async_cow), GFP_NOFS); > - BUG_ON(!async_cow); /* -ENOMEM */ > - /* > - * igrab is called higher up in the call chain, take only the > - * lightweight reference for the callback lifetime > - */ > - ihold(inode); > - async_cow->inode = inode; > - async_cow->fs_info = fs_info; > - async_cow->locked_page = locked_page; > - async_cow->start = start; > - async_cow->write_flags = write_flags; > - > - if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS && > - !btrfs_test_opt(fs_info, FORCE_COMPRESS)) > - cur_end = end; > - else > - cur_end = min(end, start + SZ_512K - 1); > > - async_cow->end = cur_end; > - INIT_LIST_HEAD(&async_cow->extents); > + if (BTRFS_I(inode)->flags & BTRFS_INODE_NOCOMPRESS && > + !btrfs_test_opt(fs_info, FORCE_COMPRESS)) { > + num_chunks = 1; > + should_compress = false; > + } else { > + should_compress = true; > + } > + > + /* Layout is [atomic_t][async_cow1][async_cowN].... */ > + async_cow = kmalloc(sizeof(atomic_t) + num_chunks*sizeof(*async_cow), > + GFP_NOFS); > + if (!async_cow) > + return -ENOMEM; > + > + p = (atomic_t *)async_cow; > + async_cow = (struct async_cow *)((char *)async_cow + sizeof(atomic_t)); > + atomic_set(p, num_chunks); > + > + for (i = 0; i < num_chunks; i++) { > > - btrfs_init_work(&async_cow->work, > + if (should_compress) > + cur_end = min(end, start + SZ_512K - 1); > + else > + cur_end = end; > + > + ihold(inode); > + async_cow[i].pending= p; > + async_cow[i].inode = inode; > + async_cow[i].start = start; > + async_cow[i].end = cur_end; > + async_cow[i].fs_info = fs_info; > + async_cow[i].locked_page = locked_page; > + async_cow[i].write_flags = write_flags; > + INIT_LIST_HEAD(&async_cow[i].extents); > + > + btrfs_init_work(&async_cow[i].work, > btrfs_delalloc_helper, > async_cow_start, async_cow_submit, > async_cow_free); > > - nr_pages = (cur_end - start + PAGE_SIZE) >> > - PAGE_SHIFT; > + nr_pages = DIV_ROUND_UP(cur_end - start, PAGE_SIZE); > atomic_add(nr_pages, &fs_info->async_delalloc_pages); > > - btrfs_queue_work(fs_info->delalloc_workers, &async_cow->work); > + btrfs_queue_work(fs_info->delalloc_workers, &async_cow[i].work); > > *nr_written += nr_pages; > start = cur_end + 1; >