Re: [PATCH 01/14] block: separate failfast into multiple bits.

Mike Christie <michaelc@xxxxxxxxxxx> · Tue, 02 Sep 2008 11:59:18 -0500

Grant Grundler wrote:
On Tue, Sep 2, 2008 at 9:05 AM, Mike Anderson
<andmike@xxxxxxxxxxxxxxxxxx> wrote:
From: Mike Christie <michaelc@xxxxxxxxxxx>

Multipath is best at handling transport errors. If it gets a device
error then there is not much the multipath layer can do. It will just
access the same device but from a different path.

This patch breaks up failfast into device, transport and driver errors.

Is there any document that describes what those errors are for each
class of transport?



Not yet. For SCSI I was still trying to classify the host byte errors, 
because drivers are using them differently. I had sent patches in the 
thread here
http://marc.info/?l=linux-scsi&m=121918956332584&w=2
 that just start syncing up the transport errors for SCSI by adding 
some new transport host byte errors and converting drivers and transport 
classes to them.




This great work though...I'm looking forward to a storage subsystem
where each level
can cooperate with the ones above it.

The multipath layers (md and dm mutlipath) only ask the lower levels to
fast fail transport errors. The user of failfast, read ahead, will ask
to fast fail on all errors.

Note that blk_noretry_request will return true if any failfast bit
is set. This allows drivers that do not support the multipath failfast
bits to continue to fail on any failfast error like before. Drivers
like scsi that are able to fail fast specific errors can check
for the specific fail fast type. In the next patch I will convert
scsi.

Signed-off-by: Mike Christie <michaelc@xxxxxxxxxxx>
Cc: Jens Axboe <jens.axboe@xxxxxxxxxx>
Cc: Alasdair G Kergon <agk@xxxxxxxxxx>
Cc: Neil Brown <neilb@xxxxxxx>
Cc: Martin Schwidefsky <schwidefsky@xxxxxxxxxx>
Signed-off-by: Mike Anderson <andmike@xxxxxxxxxxxxxxxxxx>
---
 block/blk-core.c                            |   11 +++++++++--
 drivers/md/dm-mpath.c                       |    2 +-
 drivers/md/multipath.c                      |    4 ++--
 drivers/s390/block/dasd_diag.c              |    2 +-
 drivers/s390/block/dasd_eckd.c              |    2 +-
 drivers/s390/block/dasd_fba.c               |    2 +-
 drivers/scsi/device_handler/scsi_dh_alua.c  |    3 ++-
 drivers/scsi/device_handler/scsi_dh_emc.c   |    3 ++-
 drivers/scsi/device_handler/scsi_dh_hp_sw.c |    6 ++++--
 drivers/scsi/device_handler/scsi_dh_rdac.c  |    3 ++-
 drivers/scsi/scsi_transport_spi.c           |    4 +++-
 include/linux/bio.h                         |   26 +++++++++++++++++---------
 include/linux/blkdev.h                      |   15 ++++++++++++---
 13 files changed, 57 insertions(+), 26 deletions(-)

diff --git a/block/blk-core.c b/block/blk-core.c
index 4889eb8..f3c29d0 100644
--- a/block/blk-core.c
+++ b/block/blk-core.c
@@ -1073,8 +1073,15 @@ void init_request_from_bio(struct request *req, struct bio *bio)
       /*
        * inherit FAILFAST from bio (for read-ahead, and explicit FAILFAST)
        */
-       if (bio_rw_ahead(bio) || bio_failfast(bio))
-               req->cmd_flags |= REQ_FAILFAST;
+       if (bio_rw_ahead(bio))
+               req->cmd_flags |= (REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+                                  REQ_FAILFAST_DRIVER);
+       if (bio_failfast_dev(bio))
+               req->cmd_flags |= REQ_FAILFAST_DEV;
+       if (bio_failfast_transport(bio))
+               req->cmd_flags |= REQ_FAILFAST_TRANSPORT;
+       if (bio_failfast_driver(bio))
+               req->cmd_flags |= REQ_FAILFAST_DRIVER;

This is open source.
Why can't something like this be done?
    req->cmd_flags |= bio_failfast_flags(bio);


We used to do that when it was just 1 bit for failfast, and it kept 
getting messed up. I fixed it once before and Tomo or someone else had 
to fix it again later when it got messed up again, so I thought this is 
more clear and would prevent future screw ups.




I'm assuming the REQ_FAILFAST_* flags can be equated to the same bits
in the bio layer.



       /*
        * REQ_BARRIER implies no merging, but lets make it explicit
diff --git a/drivers/md/dm-mpath.c b/drivers/md/dm-mpath.c
index 71dd65a..b48e201 100644
--- a/drivers/md/dm-mpath.c
+++ b/drivers/md/dm-mpath.c
@@ -827,7 +827,7 @@ static int multipath_map(struct dm_target *ti, struct bio *bio,
       dm_bio_record(&mpio->details, bio);

       map_context->ptr = mpio;
-       bio->bi_rw |= (1 << BIO_RW_FAILFAST);
+       bio->bi_rw |= (1 << BIO_RW_FAILFAST_TRANSPORT);
       r = map_io(m, bio, mpio, 0);
       if (r < 0 || r == DM_MAPIO_REQUEUE)
               mempool_free(mpio, m->mpio_pool);
diff --git a/drivers/md/multipath.c b/drivers/md/multipath.c
index c4779cc..2426201 100644
--- a/drivers/md/multipath.c
+++ b/drivers/md/multipath.c
@@ -172,7 +172,7 @@ static int multipath_make_request (struct request_queue *q, struct bio * bio)
       mp_bh->bio = *bio;
       mp_bh->bio.bi_sector += multipath->rdev->data_offset;
       mp_bh->bio.bi_bdev = multipath->rdev->bdev;
-       mp_bh->bio.bi_rw |= (1 << BIO_RW_FAILFAST);
+       mp_bh->bio.bi_rw |= (1 << BIO_RW_FAILFAST_TRANSPORT);
       mp_bh->bio.bi_end_io = multipath_end_request;
       mp_bh->bio.bi_private = mp_bh;
       generic_make_request(&mp_bh->bio);
@@ -398,7 +398,7 @@ static void multipathd (mddev_t *mddev)
                       *bio = *(mp_bh->master_bio);
                       bio->bi_sector += conf->multipaths[mp_bh->path].rdev->data_offset;
                       bio->bi_bdev = conf->multipaths[mp_bh->path].rdev->bdev;
-                       bio->bi_rw |= (1 << BIO_RW_FAILFAST);
+                       bio->bi_rw |= (1 << BIO_RW_FAILFAST_TRANSPORT);
                       bio->bi_end_io = multipath_end_request;
                       bio->bi_private = mp_bh;
                       generic_make_request(bio);
diff --git a/drivers/s390/block/dasd_diag.c b/drivers/s390/block/dasd_diag.c
index 85fcb43..7844461 100644
--- a/drivers/s390/block/dasd_diag.c
+++ b/drivers/s390/block/dasd_diag.c
@@ -544,7 +544,7 @@ static struct dasd_ccw_req *dasd_diag_build_cp(struct dasd_device *memdev,
       }
       cqr->retries = DIAG_MAX_RETRIES;
       cqr->buildclk = get_clock();
-       if (req->cmd_flags & REQ_FAILFAST)
+       if (blk_noretry_request(req))
               set_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags);
       cqr->startdev = memdev;
       cqr->memdev = memdev;
diff --git a/drivers/s390/block/dasd_eckd.c b/drivers/s390/block/dasd_eckd.c
index 773b3fe..b11a221 100644
--- a/drivers/s390/block/dasd_eckd.c
+++ b/drivers/s390/block/dasd_eckd.c
@@ -1683,7 +1683,7 @@ static struct dasd_ccw_req *dasd_eckd_build_cp(struct dasd_device *startdev,
                       recid++;
               }
       }
-       if (req->cmd_flags & REQ_FAILFAST)
+       if (blk_noretry_request(req))
               set_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags);
       cqr->startdev = startdev;
       cqr->memdev = startdev;
diff --git a/drivers/s390/block/dasd_fba.c b/drivers/s390/block/dasd_fba.c
index aa0c533..115e032 100644
--- a/drivers/s390/block/dasd_fba.c
+++ b/drivers/s390/block/dasd_fba.c
@@ -355,7 +355,7 @@ static struct dasd_ccw_req *dasd_fba_build_cp(struct dasd_device * memdev,
                       recid++;
               }
       }
-       if (req->cmd_flags & REQ_FAILFAST)
+       if (blk_noretry_request(req))
               set_bit(DASD_CQR_FLAGS_FAILFAST, &cqr->flags);
       cqr->startdev = memdev;
       cqr->memdev = memdev;
diff --git a/drivers/scsi/device_handler/scsi_dh_alua.c b/drivers/scsi/device_handler/scsi_dh_alua.c
index 994da56..6bc55a6 100644
--- a/drivers/scsi/device_handler/scsi_dh_alua.c
+++ b/drivers/scsi/device_handler/scsi_dh_alua.c
@@ -109,7 +109,8 @@ static struct request *get_alua_req(struct scsi_device *sdev,
       }

       rq->cmd_type = REQ_TYPE_BLOCK_PC;
-       rq->cmd_flags |= REQ_FAILFAST | REQ_NOMERGE;
+       rq->cmd_flags |= REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+                        REQ_FAILFAST_DRIVER | REQ_NOMERGE;
       rq->retries = ALUA_FAILOVER_RETRIES;
       rq->timeout = ALUA_FAILOVER_TIMEOUT;

diff --git a/drivers/scsi/device_handler/scsi_dh_emc.c b/drivers/scsi/device_handler/scsi_dh_emc.c
index b9d23e9..64a56e5 100644
--- a/drivers/scsi/device_handler/scsi_dh_emc.c
+++ b/drivers/scsi/device_handler/scsi_dh_emc.c
@@ -304,7 +304,8 @@ static struct request *get_req(struct scsi_device *sdev, int cmd,

       rq->cmd[4] = len;
       rq->cmd_type = REQ_TYPE_BLOCK_PC;
-       rq->cmd_flags |= REQ_FAILFAST;
+       rq->cmd_flags |= REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+                        REQ_FAILFAST_DRIVER;
       rq->timeout = CLARIION_TIMEOUT;
       rq->retries = CLARIION_RETRIES;

diff --git a/drivers/scsi/device_handler/scsi_dh_hp_sw.c b/drivers/scsi/device_handler/scsi_dh_hp_sw.c
index a6a4ef3..08ba1ce 100644
--- a/drivers/scsi/device_handler/scsi_dh_hp_sw.c
+++ b/drivers/scsi/device_handler/scsi_dh_hp_sw.c
@@ -112,7 +112,8 @@ static int hp_sw_tur(struct scsi_device *sdev, struct hp_sw_dh_data *h)
               return SCSI_DH_RES_TEMP_UNAVAIL;

       req->cmd_type = REQ_TYPE_BLOCK_PC;
-       req->cmd_flags |= REQ_FAILFAST;
+       req->cmd_flags |= REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+                         REQ_FAILFAST_DRIVER;
       req->cmd_len = COMMAND_SIZE(TEST_UNIT_READY);
       memset(req->cmd, 0, MAX_COMMAND_SIZE);
       req->cmd[0] = TEST_UNIT_READY;
@@ -205,7 +206,8 @@ static int hp_sw_start_stop(struct scsi_device *sdev, struct hp_sw_dh_data *h)
               return SCSI_DH_RES_TEMP_UNAVAIL;

       req->cmd_type = REQ_TYPE_BLOCK_PC;
-       req->cmd_flags |= REQ_FAILFAST;
+       req->cmd_flags |= REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+                         REQ_FAILFAST_DRIVER;
       req->cmd_len = COMMAND_SIZE(START_STOP);
       memset(req->cmd, 0, MAX_COMMAND_SIZE);
       req->cmd[0] = START_STOP;
diff --git a/drivers/scsi/device_handler/scsi_dh_rdac.c b/drivers/scsi/device_handler/scsi_dh_rdac.c
index 2dee69d..c504afe 100644
--- a/drivers/scsi/device_handler/scsi_dh_rdac.c
+++ b/drivers/scsi/device_handler/scsi_dh_rdac.c
@@ -228,7 +228,8 @@ static struct request *get_rdac_req(struct scsi_device *sdev,
       memset(rq->cmd, 0, BLK_MAX_CDB);

       rq->cmd_type = REQ_TYPE_BLOCK_PC;
-       rq->cmd_flags |= REQ_FAILFAST | REQ_NOMERGE;
+       rq->cmd_flags |= REQ_FAILFAST_DEV | REQ_FAILFAST_TRANSPORT |
+                        REQ_FAILFAST_DRIVER;

Was dropping the REQ_NOMERGE intentional?

Yes, and no.

Everywhere else it seems REQ_FAILFAST was simply replaced with the
three new flag bits.

We do not need it, because blk_execute_rq_nowait sets it for us, so I 
dropped it when I first made the patches a long time ago. It was a 
different patch for a while, but I think it got merged with this one a 
while back by accident (the alua NOMERGE setting did not get dropped 
when it could have, because it was added to scsi-misc at a different time).

Thanks for the review.
--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html




