md/raid1: split out two sub-functions from sync_request_write
authorNeilBrown <neilb@suse.de>
Wed, 11 May 2011 04:40:44 +0000 (14:40 +1000)
committerNeilBrown <neilb@suse.de>
Wed, 11 May 2011 04:40:44 +0000 (14:40 +1000)
sync_request_write is too big and too deep.
So split out two self-contains bits of functionality into separate
function.

Signed-off-by: NeilBrown <neilb@suse.de>
drivers/md/raid1.c

index f0b0c79b3899bce3d4ef0d3c8ee084e6d9acc414..7fd7a4d6f5c7f1931efb548e7cae8447241a89f7 100644 (file)
@@ -1181,194 +1181,213 @@ static void end_sync_write(struct bio *bio, int error)
        }
 }
 
-static void sync_request_write(mddev_t *mddev, r1bio_t *r1_bio)
+static int fix_sync_read_error(r1bio_t *r1_bio)
 {
+       /* Try some synchronous reads of other devices to get
+        * good data, much like with normal read errors.  Only
+        * read into the pages we already have so we don't
+        * need to re-issue the read request.
+        * We don't need to freeze the array, because being in an
+        * active sync request, there is no normal IO, and
+        * no overlapping syncs.
+        */
+       mddev_t *mddev = r1_bio->mddev;
        conf_t *conf = mddev->private;
-       int i;
-       int disks = conf->raid_disks;
-       struct bio *bio, *wbio;
+       struct bio *bio = r1_bio->bios[r1_bio->read_disk];
+       sector_t sect = r1_bio->sector;
+       int sectors = r1_bio->sectors;
+       int idx = 0;
 
-       bio = r1_bio->bios[r1_bio->read_disk];
-
-
-       if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery)) {
-               /* We have read all readable devices.  If we haven't
-                * got the block, then there is no hope left.
-                * If we have, then we want to do a comparison
-                * and skip the write if everything is the same.
-                * If any blocks failed to read, then we need to
-                * attempt an over-write
-                */
-               int primary;
-               if (!test_bit(R1BIO_Uptodate, &r1_bio->state)) {
-                       for (i=0; i<mddev->raid_disks; i++)
-                               if (r1_bio->bios[i]->bi_end_io == end_sync_read)
-                                       md_error(mddev, conf->mirrors[i].rdev);
+       while(sectors) {
+               int s = sectors;
+               int d = r1_bio->read_disk;
+               int success = 0;
+               mdk_rdev_t *rdev;
 
-                       md_done_sync(mddev, r1_bio->sectors, 1);
+               if (s > (PAGE_SIZE>>9))
+                       s = PAGE_SIZE >> 9;
+               do {
+                       if (r1_bio->bios[d]->bi_end_io == end_sync_read) {
+                               /* No rcu protection needed here devices
+                                * can only be removed when no resync is
+                                * active, and resync is currently active
+                                */
+                               rdev = conf->mirrors[d].rdev;
+                               if (sync_page_io(rdev,
+                                                sect,
+                                                s<<9,
+                                                bio->bi_io_vec[idx].bv_page,
+                                                READ, false)) {
+                                       success = 1;
+                                       break;
+                               }
+                       }
+                       d++;
+                       if (d == conf->raid_disks)
+                               d = 0;
+               } while (!success && d != r1_bio->read_disk);
+
+               if (success) {
+                       int start = d;
+                       /* write it back and re-read */
+                       set_bit(R1BIO_Uptodate, &r1_bio->state);
+                       while (d != r1_bio->read_disk) {
+                               if (d == 0)
+                                       d = conf->raid_disks;
+                               d--;
+                               if (r1_bio->bios[d]->bi_end_io != end_sync_read)
+                                       continue;
+                               rdev = conf->mirrors[d].rdev;
+                               atomic_add(s, &rdev->corrected_errors);
+                               if (sync_page_io(rdev,
+                                                sect,
+                                                s<<9,
+                                                bio->bi_io_vec[idx].bv_page,
+                                                WRITE, false) == 0)
+                                       md_error(mddev, rdev);
+                       }
+                       d = start;
+                       while (d != r1_bio->read_disk) {
+                               if (d == 0)
+                                       d = conf->raid_disks;
+                               d--;
+                               if (r1_bio->bios[d]->bi_end_io != end_sync_read)
+                                       continue;
+                               rdev = conf->mirrors[d].rdev;
+                               if (sync_page_io(rdev,
+                                                sect,
+                                                s<<9,
+                                                bio->bi_io_vec[idx].bv_page,
+                                                READ, false) == 0)
+                                       md_error(mddev, rdev);
+                       }
+               } else {
+                       char b[BDEVNAME_SIZE];
+                       /* Cannot read from anywhere, array is toast */
+                       md_error(mddev, conf->mirrors[r1_bio->read_disk].rdev);
+                       printk(KERN_ALERT "md/raid1:%s: %s: unrecoverable I/O read error"
+                              " for block %llu\n",
+                              mdname(mddev),
+                              bdevname(bio->bi_bdev, b),
+                              (unsigned long long)r1_bio->sector);
+                       md_done_sync(mddev, r1_bio->sectors, 0);
                        put_buf(r1_bio);
-                       return;
+                       return 0;
                }
-               for (primary=0; primary<mddev->raid_disks; primary++)
-                       if (r1_bio->bios[primary]->bi_end_io == end_sync_read &&
-                           test_bit(BIO_UPTODATE, &r1_bio->bios[primary]->bi_flags)) {
-                               r1_bio->bios[primary]->bi_end_io = NULL;
-                               rdev_dec_pending(conf->mirrors[primary].rdev, mddev);
-                               break;
-                       }
-               r1_bio->read_disk = primary;
+               sectors -= s;
+               sect += s;
+               idx ++;
+       }
+       return 1;
+}
+
+static int process_checks(r1bio_t *r1_bio)
+{
+       /* We have read all readable devices.  If we haven't
+        * got the block, then there is no hope left.
+        * If we have, then we want to do a comparison
+        * and skip the write if everything is the same.
+        * If any blocks failed to read, then we need to
+        * attempt an over-write
+        */
+       mddev_t *mddev = r1_bio->mddev;
+       conf_t *conf = mddev->private;
+       int primary;
+       int i;
+
+       if (!test_bit(R1BIO_Uptodate, &r1_bio->state)) {
                for (i=0; i<mddev->raid_disks; i++)
-                       if (r1_bio->bios[i]->bi_end_io == end_sync_read) {
-                               int j;
-                               int vcnt = r1_bio->sectors >> (PAGE_SHIFT- 9);
-                               struct bio *pbio = r1_bio->bios[primary];
-                               struct bio *sbio = r1_bio->bios[i];
-
-                               if (test_bit(BIO_UPTODATE, &sbio->bi_flags)) {
-                                       for (j = vcnt; j-- ; ) {
-                                               struct page *p, *s;
-                                               p = pbio->bi_io_vec[j].bv_page;
-                                               s = sbio->bi_io_vec[j].bv_page;
-                                               if (memcmp(page_address(p),
-                                                          page_address(s),
-                                                          PAGE_SIZE))
-                                                       break;
-                                       }
-                               } else
-                                       j = 0;
-                               if (j >= 0)
-                                       mddev->resync_mismatches += r1_bio->sectors;
-                               if (j < 0 || (test_bit(MD_RECOVERY_CHECK, &mddev->recovery)
-                                             && test_bit(BIO_UPTODATE, &sbio->bi_flags))) {
-                                       sbio->bi_end_io = NULL;
-                                       rdev_dec_pending(conf->mirrors[i].rdev, mddev);
-                               } else {
-                                       /* fixup the bio for reuse */
-                                       int size;
-                                       sbio->bi_vcnt = vcnt;
-                                       sbio->bi_size = r1_bio->sectors << 9;
-                                       sbio->bi_idx = 0;
-                                       sbio->bi_phys_segments = 0;
-                                       sbio->bi_flags &= ~(BIO_POOL_MASK - 1);
-                                       sbio->bi_flags |= 1 << BIO_UPTODATE;
-                                       sbio->bi_next = NULL;
-                                       sbio->bi_sector = r1_bio->sector +
-                                               conf->mirrors[i].rdev->data_offset;
-                                       sbio->bi_bdev = conf->mirrors[i].rdev->bdev;
-                                       size = sbio->bi_size;
-                                       for (j = 0; j < vcnt ; j++) {
-                                               struct bio_vec *bi;
-                                               bi = &sbio->bi_io_vec[j];
-                                               bi->bv_offset = 0;
-                                               if (size > PAGE_SIZE)
-                                                       bi->bv_len = PAGE_SIZE;
-                                               else
-                                                       bi->bv_len = size;
-                                               size -= PAGE_SIZE;
-                                               memcpy(page_address(bi->bv_page),
-                                                      page_address(pbio->bi_io_vec[j].bv_page),
-                                                      PAGE_SIZE);
-                                       }
+                       if (r1_bio->bios[i]->bi_end_io == end_sync_read)
+                               md_error(mddev, conf->mirrors[i].rdev);
 
-                               }
-                       }
+               md_done_sync(mddev, r1_bio->sectors, 1);
+               put_buf(r1_bio);
+               return -1;
        }
-       if (!test_bit(R1BIO_Uptodate, &r1_bio->state)) {
-               /* ouch - failed to read all of that.
-                * Try some synchronous reads of other devices to get
-                * good data, much like with normal read errors.  Only
-                * read into the pages we already have so we don't
-                * need to re-issue the read request.
-                * We don't need to freeze the array, because being in an
-                * active sync request, there is no normal IO, and
-                * no overlapping syncs.
-                */
-               sector_t sect = r1_bio->sector;
-               int sectors = r1_bio->sectors;
-               int idx = 0;
-
-               while(sectors) {
-                       int s = sectors;
-                       int d = r1_bio->read_disk;
-                       int success = 0;
-                       mdk_rdev_t *rdev;
-
-                       if (s > (PAGE_SIZE>>9))
-                               s = PAGE_SIZE >> 9;
-                       do {
-                               if (r1_bio->bios[d]->bi_end_io == end_sync_read) {
-                                       /* No rcu protection needed here devices
-                                        * can only be removed when no resync is
-                                        * active, and resync is currently active
-                                        */
-                                       rdev = conf->mirrors[d].rdev;
-                                       if (sync_page_io(rdev,
-                                                        sect,
-                                                        s<<9,
-                                                        bio->bi_io_vec[idx].bv_page,
-                                                        READ, false)) {
-                                               success = 1;
+       for (primary=0; primary<mddev->raid_disks; primary++)
+               if (r1_bio->bios[primary]->bi_end_io == end_sync_read &&
+                   test_bit(BIO_UPTODATE, &r1_bio->bios[primary]->bi_flags)) {
+                       r1_bio->bios[primary]->bi_end_io = NULL;
+                       rdev_dec_pending(conf->mirrors[primary].rdev, mddev);
+                       break;
+               }
+       r1_bio->read_disk = primary;
+       for (i=0; i<mddev->raid_disks; i++)
+               if (r1_bio->bios[i]->bi_end_io == end_sync_read) {
+                       int j;
+                       int vcnt = r1_bio->sectors >> (PAGE_SHIFT- 9);
+                       struct bio *pbio = r1_bio->bios[primary];
+                       struct bio *sbio = r1_bio->bios[i];
+
+                       if (test_bit(BIO_UPTODATE, &sbio->bi_flags)) {
+                               for (j = vcnt; j-- ; ) {
+                                       struct page *p, *s;
+                                       p = pbio->bi_io_vec[j].bv_page;
+                                       s = sbio->bi_io_vec[j].bv_page;
+                                       if (memcmp(page_address(p),
+                                                  page_address(s),
+                                                  PAGE_SIZE))
                                                break;
-                                       }
-                               }
-                               d++;
-                               if (d == conf->raid_disks)
-                                       d = 0;
-                       } while (!success && d != r1_bio->read_disk);
-
-                       if (success) {
-                               int start = d;
-                               /* write it back and re-read */
-                               set_bit(R1BIO_Uptodate, &r1_bio->state);
-                               while (d != r1_bio->read_disk) {
-                                       if (d == 0)
-                                               d = conf->raid_disks;
-                                       d--;
-                                       if (r1_bio->bios[d]->bi_end_io != end_sync_read)
-                                               continue;
-                                       rdev = conf->mirrors[d].rdev;
-                                       atomic_add(s, &rdev->corrected_errors);
-                                       if (sync_page_io(rdev,
-                                                        sect,
-                                                        s<<9,
-                                                        bio->bi_io_vec[idx].bv_page,
-                                                        WRITE, false) == 0)
-                                               md_error(mddev, rdev);
-                               }
-                               d = start;
-                               while (d != r1_bio->read_disk) {
-                                       if (d == 0)
-                                               d = conf->raid_disks;
-                                       d--;
-                                       if (r1_bio->bios[d]->bi_end_io != end_sync_read)
-                                               continue;
-                                       rdev = conf->mirrors[d].rdev;
-                                       if (sync_page_io(rdev,
-                                                        sect,
-                                                        s<<9,
-                                                        bio->bi_io_vec[idx].bv_page,
-                                                        READ, false) == 0)
-                                               md_error(mddev, rdev);
                                }
+                       } else
+                               j = 0;
+                       if (j >= 0)
+                               mddev->resync_mismatches += r1_bio->sectors;
+                       if (j < 0 || (test_bit(MD_RECOVERY_CHECK, &mddev->recovery)
+                                     && test_bit(BIO_UPTODATE, &sbio->bi_flags))) {
+                               sbio->bi_end_io = NULL;
+                               rdev_dec_pending(conf->mirrors[i].rdev, mddev);
                        } else {
-                               char b[BDEVNAME_SIZE];
-                               /* Cannot read from anywhere, array is toast */
-                               md_error(mddev, conf->mirrors[r1_bio->read_disk].rdev);
-                               printk(KERN_ALERT "md/raid1:%s: %s: unrecoverable I/O read error"
-                                      " for block %llu\n",
-                                      mdname(mddev),
-                                      bdevname(bio->bi_bdev, b),
-                                      (unsigned long long)r1_bio->sector);
-                               md_done_sync(mddev, r1_bio->sectors, 0);
-                               put_buf(r1_bio);
-                               return;
+                               /* fixup the bio for reuse */
+                               int size;
+                               sbio->bi_vcnt = vcnt;
+                               sbio->bi_size = r1_bio->sectors << 9;
+                               sbio->bi_idx = 0;
+                               sbio->bi_phys_segments = 0;
+                               sbio->bi_flags &= ~(BIO_POOL_MASK - 1);
+                               sbio->bi_flags |= 1 << BIO_UPTODATE;
+                               sbio->bi_next = NULL;
+                               sbio->bi_sector = r1_bio->sector +
+                                       conf->mirrors[i].rdev->data_offset;
+                               sbio->bi_bdev = conf->mirrors[i].rdev->bdev;
+                               size = sbio->bi_size;
+                               for (j = 0; j < vcnt ; j++) {
+                                       struct bio_vec *bi;
+                                       bi = &sbio->bi_io_vec[j];
+                                       bi->bv_offset = 0;
+                                       if (size > PAGE_SIZE)
+                                               bi->bv_len = PAGE_SIZE;
+                                       else
+                                               bi->bv_len = size;
+                                       size -= PAGE_SIZE;
+                                       memcpy(page_address(bi->bv_page),
+                                              page_address(pbio->bi_io_vec[j].bv_page),
+                                              PAGE_SIZE);
+                               }
+
                        }
-                       sectors -= s;
-                       sect += s;
-                       idx ++;
                }
-       }
+       return 0;
+}
+
+static void sync_request_write(mddev_t *mddev, r1bio_t *r1_bio)
+{
+       conf_t *conf = mddev->private;
+       int i;
+       int disks = conf->raid_disks;
+       struct bio *bio, *wbio;
+
+       bio = r1_bio->bios[r1_bio->read_disk];
+
 
+       if (test_bit(MD_RECOVERY_REQUESTED, &mddev->recovery))
+               if (process_checks(r1_bio) < 0)
+                       return;
+
+       if (!test_bit(R1BIO_Uptodate, &r1_bio->state))
+               /* ouch - failed to read all of that. */
+               if (!fix_sync_read_error(r1_bio))
+                       return;
        /*
         * schedule writes
         */