Merge remote-tracking branch 'remotes/kevin/tags/for-upstream' into staging

Block patches for 2.3

# gpg: Signature made Tue Mar 10 13:03:17 2015 GMT using RSA key ID C88F2FD6
# gpg: Good signature from "Kevin Wolf <kwolf@redhat.com>"

* remotes/kevin/tags/for-upstream: (73 commits)
  MAINTAINERS: Add jcody as blockjobs, block devices maintainer
  iotests: add O_DIRECT alignment probing test
  block/raw-posix: fix launching with failed disks
  MAINTAINERS: Add jsnow as IDE maintainer
  sheepdog: Fix misleading error messages in sd_snapshot_create()
  Add testcase for scsi-hd devices without drive property
  scsi-hd: fix property unset case
  block/vdi: Add locking for parallel requests
  iotests: Drop vpc from 004's and 104's format list
  iotests: Remove 006
  iotests: Fix 051's reference output
  virtio-blk: Remove the stale FIXME comment
  tests: Check QVIRTIO_F_ANY_LAYOUT flag in virtio-blk test
  libqos: Solve bug in interrupt checking when using MSIX in virtio-pci.c
  sheepdog: fix confused return values
  qtest/ahci: add fragmented dma test
  qtest/ahci: Add PIO and LBA48 tests
  qtest/ahci: Add DMA test variants
  libqos/ahci: add ahci command helpers
  qtest/ahci: Add a macro bootup routine
  ...

Signed-off-by: Peter Maydell <peter.maydell@linaro.org>

19 files changed, 218 insertions, 163 deletions

diff --git a/hw/block/block.c b/hw/block/block.c
index a625773..f7243e5 100644
--- a/hw/block/block.c
+++ b/hw/block/block.c
@@ -25,6 +25,30 @@ void blkconf_serial(BlockConf *conf, char **serial)
     }
 }
 
+void blkconf_blocksizes(BlockConf *conf)
+{
+    BlockBackend *blk = conf->blk;
+    BlockSizes blocksizes;
+    int backend_ret;
+
+    backend_ret = blk_probe_blocksizes(blk, &blocksizes);
+    /* fill in detected values if they are not defined via qemu command line */
+    if (!conf->physical_block_size) {
+        if (!backend_ret) {
+           conf->physical_block_size = blocksizes.phys;
+        } else {
+            conf->physical_block_size = BDRV_SECTOR_SIZE;
+        }
+    }
+    if (!conf->logical_block_size) {
+        if (!backend_ret) {
+            conf->logical_block_size = blocksizes.log;
+        } else {
+            conf->logical_block_size = BDRV_SECTOR_SIZE;
+        }
+    }
+}
+
 void blkconf_geometry(BlockConf *conf, int *ptrans,
                       unsigned cyls_max, unsigned heads_max, unsigned secs_max,
                       Error **errp)
diff --git a/hw/block/hd-geometry.c b/hw/block/hd-geometry.c
index 6fcf74d..b187878 100644
--- a/hw/block/hd-geometry.c
+++ b/hw/block/hd-geometry.c
@@ -121,8 +121,16 @@ void hd_geometry_guess(BlockBackend *blk,
                        int *ptrans)
 {
     int cylinders, heads, secs, translation;
+    HDGeometry geo;
 
-    if (guess_disk_lchs(blk, &cylinders, &heads, &secs) < 0) {
+    /* Try to probe the backing device geometry, otherwise fallback
+       to the old logic. (as of 12/2014 probing only succeeds on DASDs) */
+    if (blk_probe_geometry(blk, &geo) == 0) {
+        *pcyls = geo.cylinders;
+        *psecs = geo.sectors;
+        *pheads = geo.heads;
+        translation = BIOS_ATA_TRANSLATION_NONE;
+    } else if (guess_disk_lchs(blk, &cylinders, &heads, &secs) < 0) {
         /* no LCHS guess: use a standard physical disk geometry  */
         guess_chs_for_size(blk, pcyls, pheads, psecs);
         translation = hd_bios_chs_auto_trans(*pcyls, *pheads, *psecs);
diff --git a/hw/block/nvme.c b/hw/block/nvme.c
index ce079ae..0f3dfb9 100644
--- a/hw/block/nvme.c
+++ b/hw/block/nvme.c
@@ -765,6 +765,7 @@ static int nvme_init(PCIDevice *pci_dev)
     if (!n->serial) {
         return -1;
     }
+    blkconf_blocksizes(&n->conf);
 
     pci_conf = pci_dev->config;
     pci_conf[PCI_INTERRUPT_PIN] = 1;
diff --git a/hw/block/virtio-blk.c b/hw/block/virtio-blk.c
index 1e5b918..000c38d 100644
--- a/hw/block/virtio-blk.c
+++ b/hw/block/virtio-blk.c
@@ -201,6 +201,7 @@ static int virtio_blk_handle_scsi_req(VirtIOBlockReq *req)
 #ifdef __linux__
     int i;
     VirtIOBlockIoctlReq *ioctl_req;
+    BlockAIOCB *acb;
 #endif
 
     /*
@@ -278,8 +279,13 @@ static int virtio_blk_handle_scsi_req(VirtIOBlockReq *req)
     ioctl_req->hdr.sbp = elem->in_sg[elem->in_num - 3].iov_base;
     ioctl_req->hdr.mx_sb_len = elem->in_sg[elem->in_num - 3].iov_len;
 
-    blk_aio_ioctl(blk->blk, SG_IO, &ioctl_req->hdr,
-                  virtio_blk_ioctl_complete, ioctl_req);
+    acb = blk_aio_ioctl(blk->blk, SG_IO, &ioctl_req->hdr,
+                        virtio_blk_ioctl_complete, ioctl_req);
+    if (!acb) {
+        g_free(ioctl_req);
+        status = VIRTIO_BLK_S_UNSUPP;
+        goto fail;
+    }
     return -EINPROGRESS;
 #else
     abort();
@@ -591,12 +597,6 @@ static void virtio_blk_handle_output(VirtIODevice *vdev, VirtQueue *vq)
     if (mrb.num_reqs) {
         virtio_blk_submit_multireq(s->blk, &mrb);
     }
-
-    /*
-     * FIXME: Want to check for completions before returning to guest mode,
-     * so cached reads and writes are reported as quickly as possible. But
-     * that should be done in the generic block layer.
-     */
 }
 
 static void virtio_blk_dma_restart_bh(void *opaque)
@@ -884,6 +884,7 @@ static void virtio_blk_device_realize(DeviceState *dev, Error **errp)
         error_propagate(errp, err);
         return;
     }
+    blkconf_blocksizes(&conf->conf);
 
     virtio_init(vdev, "virtio-blk", VIRTIO_ID_BLOCK,
                 sizeof(struct virtio_blk_config));
diff --git a/hw/core/qdev-properties.c b/hw/core/qdev-properties.c
index 5a4e4d5..570d5f0 100644
--- a/hw/core/qdev-properties.c
+++ b/hw/core/qdev-properties.c
@@ -580,7 +580,8 @@ static void set_blocksize(Object *obj, Visitor *v, void *opaque,
         error_propagate(errp, local_err);
         return;
     }
-    if (value < min || value > max) {
+    /* value of 0 means "unset" */
+    if (value && (value < min || value > max)) {
         error_set(errp, QERR_PROPERTY_VALUE_OUT_OF_RANGE,
                   dev->id?:"", name, (int64_t)value, min, max);
         return;
diff --git a/hw/ide/ahci.c b/hw/ide/ahci.c
index 5651372..e1ae36f 100644
--- a/hw/ide/ahci.c
+++ b/hw/ide/ahci.c
@@ -1160,6 +1160,11 @@ static void ahci_start_dma(IDEDMA *dma, IDEState *s,
     dma_cb(s, 0);
 }
 
+static void ahci_restart_dma(IDEDMA *dma)
+{
+    /* Nothing to do, ahci_start_dma already resets s->io_buffer_offset.  */
+}
+
 /**
  * Called in DMA R/W chains to read the PRDT, utilizing ahci_populate_sglist.
  * Not currently invoked by PIO R/W chains,
@@ -1226,12 +1231,6 @@ static int ahci_dma_rw_buf(IDEDMA *dma, int is_write)
     return 1;
 }
 
-static int ahci_dma_set_unit(IDEDMA *dma, int unit)
-{
-    /* only a single unit per link */
-    return 0;
-}
-
 static void ahci_cmd_done(IDEDMA *dma)
 {
     AHCIDevice *ad = DO_UPCAST(AHCIDevice, dma, dma);
@@ -1252,19 +1251,14 @@ static void ahci_irq_set(void *opaque, int n, int level)
 {
 }
 
-static void ahci_dma_restart_cb(void *opaque, int running, RunState state)
-{
-}
-
 static const IDEDMAOps ahci_dma_ops = {
     .start_dma = ahci_start_dma,
+    .restart_dma = ahci_restart_dma,
     .start_transfer = ahci_start_transfer,
     .prepare_buf = ahci_dma_prepare_buf,
     .commit_buf = ahci_commit_buf,
     .rw_buf = ahci_dma_rw_buf,
-    .set_unit = ahci_dma_set_unit,
     .cmd_done = ahci_cmd_done,
-    .restart_cb = ahci_dma_restart_cb,
 };
 
 void ahci_init(AHCIState *s, DeviceState *qdev, AddressSpace *as, int ports)
@@ -1294,6 +1288,7 @@ void ahci_init(AHCIState *s, DeviceState *qdev, AddressSpace *as, int ports)
         ad->port_no = i;
         ad->port.dma = &ad->dma;
         ad->port.dma->ops = &ahci_dma_ops;
+        ide_register_restart_cb(&ad->port);
     }
 }
 
@@ -1333,6 +1328,7 @@ static const VMStateDescription vmstate_ahci_device = {
     .version_id = 1,
     .fields = (VMStateField[]) {
         VMSTATE_IDE_BUS(port, AHCIDevice),
+        VMSTATE_IDE_DRIVE(port.ifs[0], AHCIDevice),
         VMSTATE_UINT32(port_state, AHCIDevice),
         VMSTATE_UINT32(finished, AHCIDevice),
         VMSTATE_UINT32(port_regs.lst_addr, AHCIDevice),
@@ -1371,16 +1367,23 @@ static int ahci_state_post_load(void *opaque, int version_id)
         map_page(s->as, &ad->res_fis,
                  ((uint64_t)pr->fis_addr_hi << 32) | pr->fis_addr, 256);
         /*
-         * All pending i/o should be flushed out on a migrate. However,
-         * we might not have cleared the busy_slot since this is done
-         * in a bh. Also, issue i/o against any slots that are pending.
+         * If an error is present, ad->busy_slot will be valid and not -1.
+         * In this case, an operation is waiting to resume and will re-check
+         * for additional AHCI commands to execute upon completion.
+         *
+         * In the case where no error was present, busy_slot will be -1,
+         * and we should check to see if there are additional commands waiting.
          */
-        if ((ad->busy_slot != -1) &&
-            !(ad->port.ifs[0].status & (BUSY_STAT|DRQ_STAT))) {
-            pr->cmd_issue &= ~(1 << ad->busy_slot);
-            ad->busy_slot = -1;
+        if (ad->busy_slot == -1) {
+            check_cmd(s, i);
+        } else {
+            /* We are in the middle of a command, and may need to access
+             * the command header in guest memory again. */
+            if (ad->busy_slot < 0 || ad->busy_slot >= AHCI_MAX_CMDS) {
+                return -1;
+            }
+            ad->cur_cmd = &((AHCICmdHdr *)ad->lst)[ad->busy_slot];
         }
-        check_cmd(s, i);
     }
 
     return 0;
diff --git a/hw/ide/atapi.c b/hw/ide/atapi.c
index 1bf8b34..950e311 100644
--- a/hw/ide/atapi.c
+++ b/hw/ide/atapi.c
@@ -252,7 +252,6 @@ static void ide_atapi_cmd_reply(IDEState *s, int size, int max_size)
     s->packet_transfer_size = size;
     s->io_buffer_size = size;    /* dma: send the reply data as one chunk */
     s->elementary_transfer_size = 0;
-    s->io_buffer_index = 0;
 
     if (s->atapi_dma) {
         block_acct_start(blk_get_stats(s->blk), &s->acct, size,
@@ -261,6 +260,7 @@ static void ide_atapi_cmd_reply(IDEState *s, int size, int max_size)
         ide_start_dma(s, ide_atapi_cmd_read_dma_cb);
     } else {
         s->status = READY_STAT | SEEK_STAT;
+        s->io_buffer_index = 0;
         ide_atapi_cmd_reply_end(s);
     }
 }
@@ -368,7 +368,6 @@ static void ide_atapi_cmd_read_dma(IDEState *s, int lba, int nb_sectors,
 {
     s->lba = lba;
     s->packet_transfer_size = nb_sectors * sector_size;
-    s->io_buffer_index = 0;
     s->io_buffer_size = 0;
     s->cd_sector_size = sector_size;
 
diff --git a/hw/ide/cmd646.c b/hw/ide/cmd646.c
index dafa9ec..66fb9d9 100644
--- a/hw/ide/cmd646.c
+++ b/hw/ide/cmd646.c
@@ -368,8 +368,7 @@ static void pci_cmd646_ide_realize(PCIDevice *dev, Error **errp)
 
         bmdma_init(&d->bus[i], &d->bmdma[i], d);
         d->bmdma[i].bus = &d->bus[i];
-        qemu_add_vm_change_state_handler(d->bus[i].dma->ops->restart_cb,
-                                         &d->bmdma[i].dma);
+        ide_register_restart_cb(&d->bus[i]);
     }
 
     vmstate_register(DEVICE(dev), 0, &vmstate_ide_pci, d);
diff --git a/hw/ide/core.c b/hw/ide/core.c
index ac3f015..ef52f35 100644
--- a/hw/ide/core.c
+++ b/hw/ide/core.c
@@ -561,6 +561,8 @@ static bool ide_sect_range_ok(IDEState *s,
     return true;
 }
 
+static void ide_sector_read(IDEState *s);
+
 static void ide_sector_read_cb(void *opaque, int ret)
 {
     IDEState *s = opaque;
@@ -595,7 +597,7 @@ static void ide_sector_read_cb(void *opaque, int ret)
     s->io_buffer_offset += 512 * n;
 }
 
-void ide_sector_read(IDEState *s)
+static void ide_sector_read(IDEState *s)
 {
     int64_t sector_num;
     int n;
@@ -646,6 +648,9 @@ static void dma_buf_commit(IDEState *s, uint32_t tx_bytes)
 void ide_set_inactive(IDEState *s, bool more)
 {
     s->bus->dma->aiocb = NULL;
+    s->bus->retry_unit = -1;
+    s->bus->retry_sector_num = 0;
+    s->bus->retry_nsector = 0;
     if (s->bus->dma->ops->set_inactive) {
         s->bus->dma->ops->set_inactive(s->bus->dma, more);
     }
@@ -666,7 +671,7 @@ static int ide_handle_rw_error(IDEState *s, int error, int op)
     BlockErrorAction action = blk_get_error_action(s->blk, is_read, error);
 
     if (action == BLOCK_ERROR_ACTION_STOP) {
-        s->bus->dma->ops->set_unit(s->bus->dma, s->unit);
+        assert(s->bus->retry_unit == s->unit);
         s->bus->error_status = op;
     } else if (action == BLOCK_ERROR_ACTION_REPORT) {
         if (op & IDE_RETRY_DMA) {
@@ -679,7 +684,7 @@ static int ide_handle_rw_error(IDEState *s, int error, int op)
     return action != BLOCK_ERROR_ACTION_IGNORE;
 }
 
-void ide_dma_cb(void *opaque, int ret)
+static void ide_dma_cb(void *opaque, int ret)
 {
     IDEState *s = opaque;
     int n;
@@ -777,7 +782,6 @@ eot:
 static void ide_sector_start_dma(IDEState *s, enum ide_dma_cmd dma_cmd)
 {
     s->status = READY_STAT | SEEK_STAT | DRQ_STAT | BUSY_STAT;
-    s->io_buffer_index = 0;
     s->io_buffer_size = 0;
     s->dma_cmd = dma_cmd;
 
@@ -799,11 +803,17 @@ static void ide_sector_start_dma(IDEState *s, enum ide_dma_cmd dma_cmd)
 
 void ide_start_dma(IDEState *s, BlockCompletionFunc *cb)
 {
+    s->io_buffer_index = 0;
+    s->bus->retry_unit = s->unit;
+    s->bus->retry_sector_num = ide_get_sector(s);
+    s->bus->retry_nsector = s->nsector;
     if (s->bus->dma->ops->start_dma) {
         s->bus->dma->ops->start_dma(s->bus->dma, s, cb);
     }
 }
 
+static void ide_sector_write(IDEState *s);
+
 static void ide_sector_write_timer_cb(void *opaque)
 {
     IDEState *s = opaque;
@@ -863,7 +873,7 @@ static void ide_sector_write_cb(void *opaque, int ret)
     }
 }
 
-void ide_sector_write(IDEState *s)
+static void ide_sector_write(IDEState *s)
 {
     int64_t sector_num;
     int n;
@@ -917,7 +927,7 @@ static void ide_flush_cb(void *opaque, int ret)
     ide_set_irq(s->bus);
 }
 
-void ide_flush_cache(IDEState *s)
+static void ide_flush_cache(IDEState *s)
 {
     if (s->blk == NULL) {
         ide_flush_cb(s, 0);
@@ -2314,22 +2324,101 @@ static int ide_nop_int(IDEDMA *dma, int x)
     return 0;
 }
 
-static int32_t ide_nop_int32(IDEDMA *dma, int x)
+static void ide_nop(IDEDMA *dma)
 {
-    return 0;
 }
 
-static void ide_nop_restart(void *opaque, int x, RunState y)
+static int32_t ide_nop_int32(IDEDMA *dma, int x)
 {
+    return 0;
 }
 
 static const IDEDMAOps ide_dma_nop_ops = {
     .prepare_buf    = ide_nop_int32,
+    .restart_dma    = ide_nop,
     .rw_buf         = ide_nop_int,
-    .set_unit       = ide_nop_int,
-    .restart_cb     = ide_nop_restart,
 };
 
+static void ide_restart_dma(IDEState *s, enum ide_dma_cmd dma_cmd)
+{
+    s->unit = s->bus->retry_unit;
+    ide_set_sector(s, s->bus->retry_sector_num);
+    s->nsector = s->bus->retry_nsector;
+    s->bus->dma->ops->restart_dma(s->bus->dma);
+    s->io_buffer_size = 0;
+    s->dma_cmd = dma_cmd;
+    ide_start_dma(s, ide_dma_cb);
+}
+
+static void ide_restart_bh(void *opaque)
+{
+    IDEBus *bus = opaque;
+    IDEState *s;
+    bool is_read;
+    int error_status;
+
+    qemu_bh_delete(bus->bh);
+    bus->bh = NULL;
+
+    error_status = bus->error_status;
+    if (bus->error_status == 0) {
+        return;
+    }
+
+    s = idebus_active_if(bus);
+    is_read = (bus->error_status & IDE_RETRY_READ) != 0;
+
+    /* The error status must be cleared before resubmitting the request: The
+     * request may fail again, and this case can only be distinguished if the
+     * called function can set a new error status. */
+    bus->error_status = 0;
+
+    if (error_status & IDE_RETRY_DMA) {
+        if (error_status & IDE_RETRY_TRIM) {
+            ide_restart_dma(s, IDE_DMA_TRIM);
+        } else {
+            ide_restart_dma(s, is_read ? IDE_DMA_READ : IDE_DMA_WRITE);
+        }
+    } else if (error_status & IDE_RETRY_PIO) {
+        if (is_read) {
+            ide_sector_read(s);
+        } else {
+            ide_sector_write(s);
+        }
+    } else if (error_status & IDE_RETRY_FLUSH) {
+        ide_flush_cache(s);
+    } else {
+        /*
+         * We've not got any bits to tell us about ATAPI - but
+         * we do have the end_transfer_func that tells us what
+         * we're trying to do.
+         */
+        if (s->end_transfer_func == ide_atapi_cmd) {
+            ide_atapi_dma_restart(s);
+        }
+    }
+}
+
+static void ide_restart_cb(void *opaque, int running, RunState state)
+{
+    IDEBus *bus = opaque;
+
+    if (!running)
+        return;
+
+    if (!bus->bh) {
+        bus->bh = qemu_bh_new(ide_restart_bh, bus);
+        qemu_bh_schedule(bus->bh);
+    }
+}
+
+void ide_register_restart_cb(IDEBus *bus)
+{
+    if (bus->dma->ops->restart_dma) {
+        qemu_add_vm_change_state_handler(ide_restart_cb, bus);
+    }
+}
+
 static IDEDMA ide_dma_nop = {
     .ops = &ide_dma_nop_ops,
     .aiocb = NULL,
@@ -2557,10 +2646,13 @@ const VMStateDescription vmstate_ide_drive = {
 
 static const VMStateDescription vmstate_ide_error_status = {
     .name ="ide_bus/error",
-    .version_id = 1,
+    .version_id = 2,
     .minimum_version_id = 1,
     .fields = (VMStateField[]) {
         VMSTATE_INT32(error_status, IDEBus),
+        VMSTATE_INT64_V(retry_sector_num, IDEBus, 2),
+        VMSTATE_UINT32_V(retry_nsector, IDEBus, 2),
+        VMSTATE_UINT8_V(retry_unit, IDEBus, 2),
         VMSTATE_END_OF_LIST()
     }
 };
diff --git a/hw/ide/internal.h b/hw/ide/internal.h
index ee9a57f..965cc55 100644
--- a/hw/ide/internal.h
+++ b/hw/ide/internal.h
@@ -436,10 +436,9 @@ struct IDEDMAOps {
     DMAInt32Func *prepare_buf;
     DMAu32Func *commit_buf;
     DMAIntFunc *rw_buf;
-    DMAIntFunc *set_unit;
+    DMAVoidFunc *restart_dma;
     DMAStopFunc *set_inactive;
     DMAVoidFunc *cmd_done;
-    DMARestartFunc *restart_cb;
     DMAVoidFunc *reset;
 };
 
@@ -455,6 +454,8 @@ struct IDEBus {
     IDEDevice *master;
     IDEDevice *slave;
     IDEState ifs[2];
+    QEMUBH *bh;
+
     int bus_id;
     int max_units;
     IDEDMA *dma;
@@ -463,6 +464,9 @@ struct IDEBus {
     qemu_irq irq;
 
     int error_status;
+    uint8_t retry_unit;
+    int64_t retry_sector_num;
+    uint32_t retry_nsector;
 };
 
 #define TYPE_IDE_DEVICE "ide-device"
@@ -522,6 +526,9 @@ extern const VMStateDescription vmstate_ide_drive;
 #define VMSTATE_IDE_DRIVES(_field, _state) \
     VMSTATE_STRUCT_ARRAY(_field, _state, 2, 3, vmstate_ide_drive, IDEState)
 
+#define VMSTATE_IDE_DRIVE(_field, _state) \
+    VMSTATE_STRUCT(_field, _state, 1, vmstate_ide_drive, IDEState)
+
 void ide_bus_reset(IDEBus *bus);
 int64_t ide_get_sector(IDEState *s);
 void ide_set_sector(IDEState *s, int64_t sector_num);
@@ -550,12 +557,9 @@ int ide_init_drive(IDEState *s, BlockBackend *blk, IDEDriveKind kind,
                    int chs_trans);
 void ide_init2(IDEBus *bus, qemu_irq irq);
 void ide_init_ioport(IDEBus *bus, ISADevice *isa, int iobase, int iobase2);
+void ide_register_restart_cb(IDEBus *bus);
 
 void ide_exec_cmd(IDEBus *bus, uint32_t val);
-void ide_dma_cb(void *opaque, int ret);
-void ide_sector_write(IDEState *s);
-void ide_sector_read(IDEState *s);
-void ide_flush_cache(IDEState *s);
 
 void ide_transfer_start(IDEState *s, uint8_t *buf, int size,
                         EndTransferFunc *end_transfer_func);
diff --git a/hw/ide/isa.c b/hw/ide/isa.c
index c0c4e1b..9f80503 100644
--- a/hw/ide/isa.c
+++ b/hw/ide/isa.c
@@ -74,7 +74,8 @@ static void isa_ide_realizefn(DeviceState *dev, Error **errp)
     isa_init_irq(isadev, &s->irq, s->isairq);
     ide_init2(&s->bus, s->irq);
     vmstate_register(dev, 0, &vmstate_ide_isa, s);
-};
+    ide_register_restart_cb(&s->bus);
+}
 
 ISADevice *isa_ide_init(ISABus *bus, int iobase, int iobase2, int isairq,
                         DriveInfo *hd0, DriveInfo *hd1)
diff --git a/hw/ide/macio.c b/hw/ide/macio.c
index f6074f2..a009674 100644
--- a/hw/ide/macio.c
+++ b/hw/ide/macio.c
@@ -558,10 +558,6 @@ static int32_t ide_nop_int32(IDEDMA *dma, int x)
     return 0;
 }
 
-static void ide_nop_restart(void *opaque, int x, RunState y)
-{
-}
-
 static void ide_dbdma_start(IDEDMA *dma, IDEState *s,
                             BlockCompletionFunc *cb)
 {
@@ -576,8 +572,6 @@ static const IDEDMAOps dbdma_ops = {
     .start_dma      = ide_dbdma_start,
     .prepare_buf    = ide_nop_int32,
     .rw_buf         = ide_nop_int,
-    .set_unit       = ide_nop_int,
-    .restart_cb     = ide_nop_restart,
 };
 
 static void macio_ide_realizefn(DeviceState *dev, Error **errp)
diff --git a/hw/ide/pci.c b/hw/ide/pci.c
index 913a976..1b3d1c1 100644
--- a/hw/ide/pci.c
+++ b/hw/ide/pci.c
@@ -42,13 +42,10 @@ static void bmdma_start_dma(IDEDMA *dma, IDEState *s,
 {
     BMDMAState *bm = DO_UPCAST(BMDMAState, dma, dma);
 
-    bm->unit = s->unit;
     bm->dma_cb = dma_cb;
     bm->cur_prd_last = 0;
     bm->cur_prd_addr = 0;
     bm->cur_prd_len = 0;
-    bm->sector_num = ide_get_sector(s);
-    bm->nsector = s->nsector;
 
     if (bm->status & BM_STATUS_DMAING) {
         bm->dma_cb(bmdma_active_if(bm), 0);
@@ -163,20 +160,11 @@ static int bmdma_rw_buf(IDEDMA *dma, int is_write)
     return 1;
 }
 
-static int bmdma_set_unit(IDEDMA *dma, int unit)
-{
-    BMDMAState *bm = DO_UPCAST(BMDMAState, dma, dma);
-    bm->unit = unit;
-
-    return 0;
-}
-
 static void bmdma_set_inactive(IDEDMA *dma, bool more)
 {
     BMDMAState *bm = DO_UPCAST(BMDMAState, dma, dma);
 
     bm->dma_cb = NULL;
-    bm->unit = -1;
     if (more) {
         bm->status |= BM_STATUS_DMAING;
     } else {
@@ -184,83 +172,11 @@ static void bmdma_set_inactive(IDEDMA *dma, bool more)
     }
 }
 
-static void bmdma_restart_dma(BMDMAState *bm, enum ide_dma_cmd dma_cmd)
-{
-    IDEState *s = bmdma_active_if(bm);
-
-    ide_set_sector(s, bm->sector_num);
-    s->io_buffer_index = 0;
-    s->io_buffer_size = 0;
-    s->nsector = bm->nsector;
-    s->dma_cmd = dma_cmd;
-    bm->cur_addr = bm->addr;
-    bm->dma_cb = ide_dma_cb;
-    bmdma_start_dma(&bm->dma, s, bm->dma_cb);
-}
-
-/* TODO This should be common IDE code */
-static void bmdma_restart_bh(void *opaque)
-{
-    BMDMAState *bm = opaque;
-    IDEBus *bus = bm->bus;
-    bool is_read;
-    int error_status;
-
-    qemu_bh_delete(bm->bh);
-    bm->bh = NULL;
-
-    if (bm->unit == (uint8_t) -1) {
-        return;
-    }
-
-    is_read = (bus->error_status & IDE_RETRY_READ) != 0;
-
-    /* The error status must be cleared before resubmitting the request: The
-     * request may fail again, and this case can only be distinguished if the
-     * called function can set a new error status. */
-    error_status = bus->error_status;
-    bus->error_status = 0;
-
-    if (error_status & IDE_RETRY_DMA) {
-        if (error_status & IDE_RETRY_TRIM) {
-            bmdma_restart_dma(bm, IDE_DMA_TRIM);
-        } else {
-            bmdma_restart_dma(bm, is_read ? IDE_DMA_READ : IDE_DMA_WRITE);
-        }
-    } else if (error_status & IDE_RETRY_PIO) {
-        if (is_read) {
-            ide_sector_read(bmdma_active_if(bm));
-        } else {
-            ide_sector_write(bmdma_active_if(bm));
-        }
-    } else if (error_status & IDE_RETRY_FLUSH) {
-        ide_flush_cache(bmdma_active_if(bm));
-    } else {
-        IDEState *s = bmdma_active_if(bm);
-
-        /*
-         * We've not got any bits to tell us about ATAPI - but
-         * we do have the end_transfer_func that tells us what
-         * we're trying to do.
-         */
-        if (s->end_transfer_func == ide_atapi_cmd) {
-            ide_atapi_dma_restart(s);
-        }
-    }
-}
-
-static void bmdma_restart_cb(void *opaque, int running, RunState state)
+static void bmdma_restart_dma(IDEDMA *dma)
 {
-    IDEDMA *dma = opaque;
     BMDMAState *bm = DO_UPCAST(BMDMAState, dma, dma);
 
-    if (!running)
-        return;
-
-    if (!bm->bh) {
-        bm->bh = qemu_bh_new(bmdma_restart_bh, &bm->dma);
-        qemu_bh_schedule(bm->bh);
-    }
+    bm->cur_addr = bm->addr;
 }
 
 static void bmdma_cancel(BMDMAState *bm)
@@ -286,8 +202,6 @@ static void bmdma_reset(IDEDMA *dma)
     bm->cur_prd_last = 0;
     bm->cur_prd_addr = 0;
     bm->cur_prd_len = 0;
-    bm->sector_num = 0;
-    bm->nsector = 0;
 }
 
 static void bmdma_irq(void *opaque, int n, int level)
@@ -404,6 +318,9 @@ static void ide_bmdma_pre_save(void *opaque)
     BMDMAState *bm = opaque;
     uint8_t abused_bits = BM_MIGRATION_COMPAT_STATUS_BITS;
 
+    bm->migration_retry_unit = bm->bus->retry_unit;
+    bm->migration_retry_sector_num = bm->bus->retry_sector_num;
+    bm->migration_retry_nsector = bm->bus->retry_nsector;
     bm->migration_compat_status =
         (bm->status & ~abused_bits) | (bm->bus->error_status & abused_bits);
 }
@@ -420,6 +337,11 @@ static int ide_bmdma_post_load(void *opaque, int version_id)
         bm->status = bm->migration_compat_status & ~abused_bits;
         bm->bus->error_status |= bm->migration_compat_status & abused_bits;
     }
+    if (bm->bus->error_status) {
+        bm->bus->retry_sector_num = bm->migration_retry_sector_num;
+        bm->bus->retry_nsector = bm->migration_retry_nsector;
+        bm->bus->retry_unit = bm->migration_retry_unit;
+    }
 
     return 0;
 }
@@ -456,9 +378,9 @@ static const VMStateDescription vmstate_bmdma = {
         VMSTATE_UINT8(cmd, BMDMAState),
         VMSTATE_UINT8(migration_compat_status, BMDMAState),
         VMSTATE_UINT32(addr, BMDMAState),
-        VMSTATE_INT64(sector_num, BMDMAState),
-        VMSTATE_UINT32(nsector, BMDMAState),
-        VMSTATE_UINT8(unit, BMDMAState),
+        VMSTATE_INT64(migration_retry_sector_num, BMDMAState),
+        VMSTATE_UINT32(migration_retry_nsector, BMDMAState),
+        VMSTATE_UINT8(migration_retry_unit, BMDMAState),
         VMSTATE_END_OF_LIST()
     },
     .subsections = (VMStateSubsection []) {
@@ -482,7 +404,7 @@ static int ide_pci_post_load(void *opaque, int version_id)
     for(i = 0; i < 2; i++) {
         /* current versions always store 0/1, but older version
            stored bigger values. We only need last bit */
-        d->bmdma[i].unit &= 1;
+        d->bmdma[i].migration_retry_unit &= 1;
         ide_bmdma_post_load(&d->bmdma[i], -1);
     }
 
@@ -523,9 +445,8 @@ static const struct IDEDMAOps bmdma_ops = {
     .start_dma = bmdma_start_dma,
     .prepare_buf = bmdma_prepare_buf,
     .rw_buf = bmdma_rw_buf,
-    .set_unit = bmdma_set_unit,
+    .restart_dma = bmdma_restart_dma,
     .set_inactive = bmdma_set_inactive,
-    .restart_cb = bmdma_restart_cb,
     .reset = bmdma_reset,
 };
 
diff --git a/hw/ide/pci.h b/hw/ide/pci.h
index 2e9314a..0f2d4b9 100644
--- a/hw/ide/pci.h
+++ b/hw/ide/pci.h
@@ -22,18 +22,18 @@ typedef struct BMDMAState {
     uint32_t cur_prd_last;
     uint32_t cur_prd_addr;
     uint32_t cur_prd_len;
-    uint8_t unit;
     BlockCompletionFunc *dma_cb;
-    int64_t sector_num;
-    uint32_t nsector;
     MemoryRegion addr_ioport;
     MemoryRegion extra_io;
-    QEMUBH *bh;
     qemu_irq irq;
 
     /* Bit 0-2 and 7:   BM status register
      * Bit 3-6:         bus->error_status */
     uint8_t migration_compat_status;
+    uint8_t migration_retry_unit;
+    int64_t migration_retry_sector_num;
+    uint32_t migration_retry_nsector;
+
     struct PCIIDEState *pci_dev;
 } BMDMAState;
 
@@ -62,8 +62,8 @@ typedef struct PCIIDEState {
 
 static inline IDEState *bmdma_active_if(BMDMAState *bmdma)
 {
-    assert(bmdma->unit != (uint8_t)-1);
-    return bmdma->bus->ifs + bmdma->unit;
+    assert(bmdma->bus->retry_unit != (uint8_t)-1);
+    return bmdma->bus->ifs + bmdma->bus->retry_unit;
 }
 
 
diff --git a/hw/ide/piix.c b/hw/ide/piix.c
index ab93084..adb6649 100644
--- a/hw/ide/piix.c
+++ b/hw/ide/piix.c
@@ -143,8 +143,7 @@ static void pci_piix_init_ports(PCIIDEState *d) {
 
         bmdma_init(&d->bus[i], &d->bmdma[i], d);
         d->bmdma[i].bus = &d->bus[i];
-        qemu_add_vm_change_state_handler(d->bus[i].dma->ops->restart_cb,
-                                         &d->bmdma[i].dma);
+        ide_register_restart_cb(&d->bus[i]);
     }
 }
 
diff --git a/hw/ide/qdev.c b/hw/ide/qdev.c
index b4103fa..788b361 100644
--- a/hw/ide/qdev.c
+++ b/hw/ide/qdev.c
@@ -163,6 +163,7 @@ static int ide_dev_initfn(IDEDevice *dev, IDEDriveKind kind)
         return -1;
     }
 
+    blkconf_blocksizes(&dev->conf);
     if (dev->conf.logical_block_size != 512) {
         error_report("logical_block_size must be 512 for IDE");
         return -1;
diff --git a/hw/ide/via.c b/hw/ide/via.c
index 7a5151c..e2da9ef 100644
--- a/hw/ide/via.c
+++ b/hw/ide/via.c
@@ -166,8 +166,7 @@ static void vt82c686b_init_ports(PCIIDEState *d) {
 
         bmdma_init(&d->bus[i], &d->bmdma[i], d);
         d->bmdma[i].bus = &d->bus[i];
-        qemu_add_vm_change_state_handler(d->bus[i].dma->ops->restart_cb,
-                                         &d->bmdma[i].dma);
+        ide_register_restart_cb(&d->bus[i]);
     }
 }
 
diff --git a/hw/scsi/scsi-disk.c b/hw/scsi/scsi-disk.c
index f65618d..54d71f4 100644
--- a/hw/scsi/scsi-disk.c
+++ b/hw/scsi/scsi-disk.c
@@ -2251,6 +2251,7 @@ static void scsi_realize(SCSIDevice *dev, Error **errp)
     }
 
     blkconf_serial(&s->qdev.conf, &s->serial);
+    blkconf_blocksizes(&s->qdev.conf);
     if (dev->type == TYPE_DISK) {
         blkconf_geometry(&dev->conf, NULL, 65535, 255, 255, &err);
         if (err) {
@@ -2290,6 +2291,12 @@ static void scsi_realize(SCSIDevice *dev, Error **errp)
 static void scsi_hd_realize(SCSIDevice *dev, Error **errp)
 {
     SCSIDiskState *s = DO_UPCAST(SCSIDiskState, qdev, dev);
+    /* can happen for devices without drive. The error message for missing
+     * backend will be issued in scsi_realize
+     */
+    if (s->qdev.conf.blk) {
+        blkconf_blocksizes(&s->qdev.conf);
+    }
     s->qdev.blocksize = s->qdev.conf.logical_block_size;
     s->qdev.type = TYPE_DISK;
     if (!s->product) {
diff --git a/hw/usb/dev-storage.c b/hw/usb/dev-storage.c
index 65d9aa6..8edc314 100644
--- a/hw/usb/dev-storage.c
+++ b/hw/usb/dev-storage.c
@@ -611,6 +611,7 @@ static void usb_msd_realize_storage(USBDevice *dev, Error **errp)
     }
 
     blkconf_serial(&s->conf, &dev->serial);
+    blkconf_blocksizes(&s->conf);
 
     /*
      * Hack alert: this pretends to be a block device, but it's really