block/nvme: add support for write zeros
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
Message-id: 20190913133627.28450-2-mlevitsk@redhat.com
Reviewed-by: John Snow <jsnow@redhat.com>
Signed-off-by: Max Reitz <mreitz@redhat.com>
diff --git a/block/nvme.c b/block/nvme.c
index 910872e..0ac38c2 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -112,6 +112,8 @@
uint64_t max_transfer;
bool plugged;
+ bool supports_write_zeroes;
+
CoMutex dma_map_lock;
CoQueue dma_flush_queue;
@@ -423,6 +425,7 @@
NvmeIdNs *idns;
NvmeLBAF *lbaf;
uint8_t *resp;
+ uint16_t oncs;
int r;
uint64_t iova;
NvmeCmd cmd = {
@@ -460,6 +463,9 @@
s->max_transfer = MIN_NON_ZERO(s->max_transfer,
s->page_size / sizeof(uint64_t) * s->page_size);
+ oncs = le16_to_cpu(idctrl->oncs);
+ s->supports_write_zeroes = !!(oncs & NVME_ONCS_WRITE_ZEROS);
+
memset(resp, 0, 4096);
cmd.cdw10 = 0;
@@ -472,6 +478,12 @@
s->nsze = le64_to_cpu(idns->nsze);
lbaf = &idns->lbaf[NVME_ID_NS_FLBAS_INDEX(idns->flbas)];
+ if (NVME_ID_NS_DLFEAT_WRITE_ZEROES(idns->dlfeat) &&
+ NVME_ID_NS_DLFEAT_READ_BEHAVIOR(idns->dlfeat) ==
+ NVME_ID_NS_DLFEAT_READ_BEHAVIOR_ZEROES) {
+ bs->supported_write_flags |= BDRV_REQ_MAY_UNMAP;
+ }
+
if (lbaf->ms) {
error_setg(errp, "Namespaces with metadata are not yet supported");
goto out;
@@ -766,6 +778,8 @@
int ret;
BDRVNVMeState *s = bs->opaque;
+ bs->supported_write_flags = BDRV_REQ_FUA;
+
opts = qemu_opts_create(&runtime_opts, NULL, 0, &error_abort);
qemu_opts_absorb_qdict(opts, options, &error_abort);
device = qemu_opt_get(opts, NVME_BLOCK_OPT_DEVICE);
@@ -794,7 +808,6 @@
goto fail;
}
}
- bs->supported_write_flags = BDRV_REQ_FUA;
return 0;
fail:
nvme_close(bs);
@@ -1088,6 +1101,60 @@
}
+static coroutine_fn int nvme_co_pwrite_zeroes(BlockDriverState *bs,
+ int64_t offset,
+ int bytes,
+ BdrvRequestFlags flags)
+{
+ BDRVNVMeState *s = bs->opaque;
+ NVMeQueuePair *ioq = s->queues[1];
+ NVMeRequest *req;
+
+ uint32_t cdw12 = ((bytes >> s->blkshift) - 1) & 0xFFFF;
+
+ if (!s->supports_write_zeroes) {
+ return -ENOTSUP;
+ }
+
+ NvmeCmd cmd = {
+ .opcode = NVME_CMD_WRITE_ZEROS,
+ .nsid = cpu_to_le32(s->nsid),
+ .cdw10 = cpu_to_le32((offset >> s->blkshift) & 0xFFFFFFFF),
+ .cdw11 = cpu_to_le32(((offset >> s->blkshift) >> 32) & 0xFFFFFFFF),
+ };
+
+ NVMeCoData data = {
+ .ctx = bdrv_get_aio_context(bs),
+ .ret = -EINPROGRESS,
+ };
+
+ if (flags & BDRV_REQ_MAY_UNMAP) {
+ cdw12 |= (1 << 25);
+ }
+
+ if (flags & BDRV_REQ_FUA) {
+ cdw12 |= (1 << 30);
+ }
+
+ cmd.cdw12 = cpu_to_le32(cdw12);
+
+ trace_nvme_write_zeroes(s, offset, bytes, flags);
+ assert(s->nr_queues > 1);
+ req = nvme_get_free_req(ioq);
+ assert(req);
+
+ nvme_submit_command(s, ioq, req, &cmd, nvme_rw_cb, &data);
+
+ data.co = qemu_coroutine_self();
+ while (data.ret == -EINPROGRESS) {
+ qemu_coroutine_yield();
+ }
+
+ trace_nvme_rw_done(s, true, offset, bytes, data.ret);
+ return data.ret;
+}
+
+
static int nvme_reopen_prepare(BDRVReopenState *reopen_state,
BlockReopenQueue *queue, Error **errp)
{
@@ -1192,6 +1259,9 @@
.bdrv_co_preadv = nvme_co_preadv,
.bdrv_co_pwritev = nvme_co_pwritev,
+
+ .bdrv_co_pwrite_zeroes = nvme_co_pwrite_zeroes,
+
.bdrv_co_flush_to_disk = nvme_co_flush,
.bdrv_reopen_prepare = nvme_reopen_prepare,
