[v2,1/2] block/nvme: add support for write zeros
diff mbox series

Message ID 20190913133627.28450-2-mlevitsk@redhat.com
State New
Headers show
Series
  • block/nvme: add support for write zeros and discard
Related show

Commit Message

Maxim Levitsky Sept. 13, 2019, 1:36 p.m. UTC
Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>
---
 block/nvme.c         | 72 +++++++++++++++++++++++++++++++++++++++++++-
 block/trace-events   |  1 +
 include/block/nvme.h | 19 +++++++++++-
 3 files changed, 90 insertions(+), 2 deletions(-)

Comments

John Snow Sept. 18, 2019, 8:22 p.m. UTC | #1
On 9/13/19 9:36 AM, Maxim Levitsky wrote:
> Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com>

It'd still be nice to have a commit message...

> ---

Or here, what changed from V1.

>  block/nvme.c         | 72 +++++++++++++++++++++++++++++++++++++++++++-
>  block/trace-events   |  1 +
>  include/block/nvme.h | 19 +++++++++++-
>  3 files changed, 90 insertions(+), 2 deletions(-)
> 
> diff --git a/block/nvme.c b/block/nvme.c
> index 5be3a39b63..d95265fae4 100644
> --- a/block/nvme.c
> +++ b/block/nvme.c
> @@ -111,6 +111,8 @@ typedef struct {
>      uint64_t max_transfer;
>      bool plugged;
>  
> +    bool supports_write_zeroes;
> +
>      CoMutex dma_map_lock;
>      CoQueue dma_flush_queue;
>  
> @@ -421,6 +423,7 @@ static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
>      NvmeIdNs *idns;
>      NvmeLBAF *lbaf;
>      uint8_t *resp;
> +    uint16_t oncs;
>      int r;
>      uint64_t iova;
>      NvmeCmd cmd = {
> @@ -458,6 +461,9 @@ static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
>      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;
> @@ -470,6 +476,12 @@ static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
>      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;
> @@ -764,6 +776,8 @@ static int nvme_file_open(BlockDriverState *bs, QDict *options, int flags,
>      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);
> @@ -792,7 +806,6 @@ static int nvme_file_open(BlockDriverState *bs, QDict *options, int flags,
>              goto fail;
>          }
>      }
> -    bs->supported_write_flags = BDRV_REQ_FUA;
>      return 0;
>  fail:
>      nvme_close(bs);
> @@ -1086,6 +1099,60 @@ static coroutine_fn int nvme_co_flush(BlockDriverState *bs)
>  }
>  
>  
> +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)
>  {
> @@ -1190,6 +1257,9 @@ static BlockDriver bdrv_nvme = {
>  
>      .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,
>  
> diff --git a/block/trace-events b/block/trace-events
> index 04209f058d..651aa461d5 100644
> --- a/block/trace-events
> +++ b/block/trace-events
> @@ -149,6 +149,7 @@ nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6,
>  nvme_handle_event(void *s) "s %p"
>  nvme_poll_cb(void *s) "s %p"
>  nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset %"PRId64" bytes %"PRId64" flags %d niov %d"
> +nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset %"PRId64" bytes %"PRId64" flags %d"
>  nvme_qiov_unaligned(const void *qiov, int n, void *base, size_t size, int align) "qiov %p n %d base %p size 0x%zx align 0x%x"
>  nvme_prw_buffered(void *s, uint64_t offset, uint64_t bytes, int niov, int is_write) "s %p offset %"PRId64" bytes %"PRId64" niov %d is_write %d"
>  nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s %p is_write %d offset %"PRId64" bytes %"PRId64" ret %d"
> diff --git a/include/block/nvme.h b/include/block/nvme.h
> index 3ec8efcc43..33304c5a65 100644
> --- a/include/block/nvme.h
> +++ b/include/block/nvme.h
> @@ -653,12 +653,29 @@ typedef struct NvmeIdNs {
>      uint8_t     mc;
>      uint8_t     dpc;
>      uint8_t     dps;
> -    uint8_t     res30[98];
> +
> +    uint8_t     nmic;
> +    uint8_t     rescap;
> +    uint8_t     fpi;
> +    uint8_t     dlfeat;
> +
> +    uint8_t     res34[94];
>      NvmeLBAF    lbaf[16];
>      uint8_t     res192[192];
>      uint8_t     vs[3712];
>  } NvmeIdNs;
>  
> +
> +/*Deallocate Logical Block Features*/
> +#define NVME_ID_NS_DLFEAT_GUARD_CRC(dlfeat)       ((dlfeat) & 0x10)
> +#define NVME_ID_NS_DLFEAT_WRITE_ZEROES(dlfeat)    ((dlfeat) & 0x08)
> +
> +#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR(dlfeat)     ((dlfeat) & 0x7)
> +#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR_UNDEFINED   0
> +#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR_ZEROES       1
> +#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR_ONES        2

ragged, but can be squished in on commit.

> +
> +
>  #define NVME_ID_NS_NSFEAT_THIN(nsfeat)      ((nsfeat & 0x1))
>  #define NVME_ID_NS_FLBAS_EXTENDED(flbas)    ((flbas >> 4) & 0x1)
>  #define NVME_ID_NS_FLBAS_INDEX(flbas)       ((flbas & 0xf))
> 

more or less, looks OK as far as I can tell, but there's a bit of
benefit-of-doubt going on for the exact mechanisms of NVME.

I pointed out some sections in the NVME spec that can be used to help
review this patch last time; your commit message should mention some of
these sections ideally so that constants and registers can be more
quickly verified.


Reviewed-by: John Snow <jsnow@redhat.com>

Patch
diff mbox series

diff --git a/block/nvme.c b/block/nvme.c
index 5be3a39b63..d95265fae4 100644
--- a/block/nvme.c
+++ b/block/nvme.c
@@ -111,6 +111,8 @@  typedef struct {
     uint64_t max_transfer;
     bool plugged;
 
+    bool supports_write_zeroes;
+
     CoMutex dma_map_lock;
     CoQueue dma_flush_queue;
 
@@ -421,6 +423,7 @@  static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
     NvmeIdNs *idns;
     NvmeLBAF *lbaf;
     uint8_t *resp;
+    uint16_t oncs;
     int r;
     uint64_t iova;
     NvmeCmd cmd = {
@@ -458,6 +461,9 @@  static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
     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;
@@ -470,6 +476,12 @@  static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp)
     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;
@@ -764,6 +776,8 @@  static int nvme_file_open(BlockDriverState *bs, QDict *options, int flags,
     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);
@@ -792,7 +806,6 @@  static int nvme_file_open(BlockDriverState *bs, QDict *options, int flags,
             goto fail;
         }
     }
-    bs->supported_write_flags = BDRV_REQ_FUA;
     return 0;
 fail:
     nvme_close(bs);
@@ -1086,6 +1099,60 @@  static coroutine_fn int nvme_co_flush(BlockDriverState *bs)
 }
 
 
+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)
 {
@@ -1190,6 +1257,9 @@  static BlockDriver bdrv_nvme = {
 
     .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,
 
diff --git a/block/trace-events b/block/trace-events
index 04209f058d..651aa461d5 100644
--- a/block/trace-events
+++ b/block/trace-events
@@ -149,6 +149,7 @@  nvme_submit_command_raw(int c0, int c1, int c2, int c3, int c4, int c5, int c6,
 nvme_handle_event(void *s) "s %p"
 nvme_poll_cb(void *s) "s %p"
 nvme_prw_aligned(void *s, int is_write, uint64_t offset, uint64_t bytes, int flags, int niov) "s %p is_write %d offset %"PRId64" bytes %"PRId64" flags %d niov %d"
+nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p offset %"PRId64" bytes %"PRId64" flags %d"
 nvme_qiov_unaligned(const void *qiov, int n, void *base, size_t size, int align) "qiov %p n %d base %p size 0x%zx align 0x%x"
 nvme_prw_buffered(void *s, uint64_t offset, uint64_t bytes, int niov, int is_write) "s %p offset %"PRId64" bytes %"PRId64" niov %d is_write %d"
 nvme_rw_done(void *s, int is_write, uint64_t offset, uint64_t bytes, int ret) "s %p is_write %d offset %"PRId64" bytes %"PRId64" ret %d"
diff --git a/include/block/nvme.h b/include/block/nvme.h
index 3ec8efcc43..33304c5a65 100644
--- a/include/block/nvme.h
+++ b/include/block/nvme.h
@@ -653,12 +653,29 @@  typedef struct NvmeIdNs {
     uint8_t     mc;
     uint8_t     dpc;
     uint8_t     dps;
-    uint8_t     res30[98];
+
+    uint8_t     nmic;
+    uint8_t     rescap;
+    uint8_t     fpi;
+    uint8_t     dlfeat;
+
+    uint8_t     res34[94];
     NvmeLBAF    lbaf[16];
     uint8_t     res192[192];
     uint8_t     vs[3712];
 } NvmeIdNs;
 
+
+/*Deallocate Logical Block Features*/
+#define NVME_ID_NS_DLFEAT_GUARD_CRC(dlfeat)       ((dlfeat) & 0x10)
+#define NVME_ID_NS_DLFEAT_WRITE_ZEROES(dlfeat)    ((dlfeat) & 0x08)
+
+#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR(dlfeat)     ((dlfeat) & 0x7)
+#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR_UNDEFINED   0
+#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR_ZEROES       1
+#define NVME_ID_NS_DLFEAT_READ_BEHAVIOR_ONES        2
+
+
 #define NVME_ID_NS_NSFEAT_THIN(nsfeat)      ((nsfeat & 0x1))
 #define NVME_ID_NS_FLBAS_EXTENDED(flbas)    ((flbas >> 4) & 0x1)
 #define NVME_ID_NS_FLBAS_INDEX(flbas)       ((flbas & 0xf))