From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from cn.fujitsu.com ([59.151.112.132]:34316 "EHLO heian.cn.fujitsu.com" rhost-flags-OK-FAIL-OK-FAIL) by vger.kernel.org with ESMTP id S1750939AbdBTGcQ (ORCPT ); Mon, 20 Feb 2017 01:32:16 -0500 Subject: Re: [PATCH 4/7] Btrfs: handle operations for device replace separately To: Liu Bo , References: <1487381301-865-1-git-send-email-bo.li.liu@oracle.com> <1487381301-865-5-git-send-email-bo.li.liu@oracle.com> CC: David Sterba From: Qu Wenruo Message-ID: <1d1b2dae-e063-9f83-e545-bf01ea52ff42@cn.fujitsu.com> Date: Mon, 20 Feb 2017 14:30:56 +0800 MIME-Version: 1.0 In-Reply-To: <1487381301-865-5-git-send-email-bo.li.liu@oracle.com> Content-Type: text/plain; charset="utf-8"; format=flowed Sender: linux-btrfs-owner@vger.kernel.org List-ID: At 02/18/2017 09:28 AM, Liu Bo wrote: > Since this part is mostly self-independent, this moves it to a separate > function. Just code move and modified part looks good to me. Reviewed-by: Qu Wenruo Thanks, Qu > > Signed-off-by: Liu Bo > --- > fs/btrfs/volumes.c | 179 +++++++++++++++++++++++++++++------------------------ > 1 file changed, 98 insertions(+), 81 deletions(-) > > diff --git a/fs/btrfs/volumes.c b/fs/btrfs/volumes.c > index f62efc7..c434489 100644 > --- a/fs/btrfs/volumes.c > +++ b/fs/btrfs/volumes.c > @@ -5545,6 +5545,100 @@ static int get_extra_mirror_from_replace(struct btrfs_fs_info *fs_info, > return ret; > } > > +static void handle_ops_on_dev_replace(enum btrfs_map_op op, > + struct btrfs_bio **bbio_ret, > + struct btrfs_dev_replace *dev_replace, > + int *num_stripes_ret, int *max_errors_ret) > +{ > + struct btrfs_bio *bbio = *bbio_ret; > + u64 srcdev_devid = dev_replace->srcdev->devid; > + int tgtdev_indexes = 0; > + int num_stripes = *num_stripes_ret; > + int max_errors = *max_errors_ret; > + int i; > + > + if (op == BTRFS_MAP_WRITE) { > + int index_where_to_add; > + > + /* > + * duplicate the write operations while the dev replace > + * procedure is running. Since the copying of the old disk to > + * the new disk takes place at run time while the filesystem is > + * mounted writable, the regular write operations to the old > + * disk have to be duplicated to go to the new disk as well. > + * > + * Note that device->missing is handled by the caller, and that > + * the write to the old disk is already set up in the stripes > + * array. > + */ > + index_where_to_add = num_stripes; > + for (i = 0; i < num_stripes; i++) { > + if (bbio->stripes[i].dev->devid == srcdev_devid) { > + /* write to new disk, too */ > + struct btrfs_bio_stripe *new = > + bbio->stripes + index_where_to_add; > + struct btrfs_bio_stripe *old = > + bbio->stripes + i; > + > + new->physical = old->physical; > + new->length = old->length; > + new->dev = dev_replace->tgtdev; > + bbio->tgtdev_map[i] = index_where_to_add; > + index_where_to_add++; > + max_errors++; > + tgtdev_indexes++; > + } > + } > + num_stripes = index_where_to_add; > + } else if (op == BTRFS_MAP_GET_READ_MIRRORS) { > + int index_srcdev = 0; > + int found = 0; > + u64 physical_of_found = 0; > + > + /* > + * During the dev-replace procedure, the target drive can also > + * be used to read data in case it is needed to repair a corrupt > + * block elsewhere. This is possible if the requested area is > + * left of the left cursor. In this area, the target drive is a > + * full copy of the source drive. > + */ > + for (i = 0; i < num_stripes; i++) { > + if (bbio->stripes[i].dev->devid == srcdev_devid) { > + /* > + * In case of DUP, in order to keep it simple, > + * only add the mirror with the lowest physical > + * address > + */ > + if (found && > + physical_of_found <= > + bbio->stripes[i].physical) > + continue; > + index_srcdev = i; > + found = 1; > + physical_of_found = bbio->stripes[i].physical; > + } > + } > + if (found) { > + struct btrfs_bio_stripe *tgtdev_stripe = > + bbio->stripes + num_stripes; > + > + tgtdev_stripe->physical = physical_of_found; > + tgtdev_stripe->length = > + bbio->stripes[index_srcdev].length; > + tgtdev_stripe->dev = dev_replace->tgtdev; > + bbio->tgtdev_map[index_srcdev] = num_stripes; > + > + tgtdev_indexes++; > + num_stripes++; > + } > + } > + > + *num_stripes_ret = num_stripes; > + *max_errors_ret = max_errors; > + bbio->num_tgtdevs = tgtdev_indexes; > + *bbio_ret = bbio; > +} > + > static int __btrfs_map_block(struct btrfs_fs_info *fs_info, > enum btrfs_map_op op, > u64 logical, u64 *length, > @@ -5825,86 +5919,10 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, > if (bbio->raid_map) > sort_parity_stripes(bbio, num_stripes); > > - tgtdev_indexes = 0; > - if (dev_replace_is_ongoing && op == BTRFS_MAP_WRITE && > - dev_replace->tgtdev != NULL) { > - int index_where_to_add; > - u64 srcdev_devid = dev_replace->srcdev->devid; > - > - /* > - * duplicate the write operations while the dev replace > - * procedure is running. Since the copying of the old disk > - * to the new disk takes place at run time while the > - * filesystem is mounted writable, the regular write > - * operations to the old disk have to be duplicated to go > - * to the new disk as well. > - * Note that device->missing is handled by the caller, and > - * that the write to the old disk is already set up in the > - * stripes array. > - */ > - index_where_to_add = num_stripes; > - for (i = 0; i < num_stripes; i++) { > - if (bbio->stripes[i].dev->devid == srcdev_devid) { > - /* write to new disk, too */ > - struct btrfs_bio_stripe *new = > - bbio->stripes + index_where_to_add; > - struct btrfs_bio_stripe *old = > - bbio->stripes + i; > - > - new->physical = old->physical; > - new->length = old->length; > - new->dev = dev_replace->tgtdev; > - bbio->tgtdev_map[i] = index_where_to_add; > - index_where_to_add++; > - max_errors++; > - tgtdev_indexes++; > - } > - } > - num_stripes = index_where_to_add; > - } else if (dev_replace_is_ongoing && > - op == BTRFS_MAP_GET_READ_MIRRORS && > - dev_replace->tgtdev != NULL) { > - u64 srcdev_devid = dev_replace->srcdev->devid; > - int index_srcdev = 0; > - int found = 0; > - u64 physical_of_found = 0; > - > - /* > - * During the dev-replace procedure, the target drive can > - * also be used to read data in case it is needed to repair > - * a corrupt block elsewhere. This is possible if the > - * requested area is left of the left cursor. In this area, > - * the target drive is a full copy of the source drive. > - */ > - for (i = 0; i < num_stripes; i++) { > - if (bbio->stripes[i].dev->devid == srcdev_devid) { > - /* > - * In case of DUP, in order to keep it > - * simple, only add the mirror with the > - * lowest physical address > - */ > - if (found && > - physical_of_found <= > - bbio->stripes[i].physical) > - continue; > - index_srcdev = i; > - found = 1; > - physical_of_found = bbio->stripes[i].physical; > - } > - } > - if (found) { > - struct btrfs_bio_stripe *tgtdev_stripe = > - bbio->stripes + num_stripes; > - > - tgtdev_stripe->physical = physical_of_found; > - tgtdev_stripe->length = > - bbio->stripes[index_srcdev].length; > - tgtdev_stripe->dev = dev_replace->tgtdev; > - bbio->tgtdev_map[index_srcdev] = num_stripes; > - > - tgtdev_indexes++; > - num_stripes++; > - } > + if (dev_replace_is_ongoing && dev_replace->tgtdev != NULL && > + (op == BTRFS_MAP_WRITE || op == BTRFS_MAP_GET_READ_MIRRORS)) { > + handle_ops_on_dev_replace(op, &bbio, dev_replace, &num_stripes, > + &max_errors); > } > > *bbio_ret = bbio; > @@ -5912,7 +5930,6 @@ static int __btrfs_map_block(struct btrfs_fs_info *fs_info, > bbio->num_stripes = num_stripes; > bbio->max_errors = max_errors; > bbio->mirror_num = mirror_num; > - bbio->num_tgtdevs = tgtdev_indexes; > > /* > * this is the case that REQ_READ && dev_replace_is_ongoing && >