* [Qemu-devel] [PATCH v2 0/2] block/nvme: add support for write zeros and discard @ 2019-09-13 13:36 Maxim Levitsky 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 1/2] block/nvme: add support for write zeros Maxim Levitsky ` (2 more replies) 0 siblings, 3 replies; 8+ messages in thread From: Maxim Levitsky @ 2019-09-13 13:36 UTC (permalink / raw) To: qemu-devel Cc: Fam Zheng, Kevin Wolf, qemu-block, Max Reitz, Keith Busch, Paolo Bonzini, Maxim Levitsky, John Snow This is the second part of the patches I prepared for this driver back when I worked on mdev-nvme. V2: addressed review feedback, no major changes Best regards, Maxim Levitsky Maxim Levitsky (2): block/nvme: add support for write zeros block/nvme: add support for discard block/nvme.c | 155 ++++++++++++++++++++++++++++++++++++++++++- block/trace-events | 3 + include/block/nvme.h | 19 +++++- 3 files changed, 175 insertions(+), 2 deletions(-) -- 2.17.2 ^ permalink raw reply [flat|nested] 8+ messages in thread
* [Qemu-devel] [PATCH v2 1/2] block/nvme: add support for write zeros 2019-09-13 13:36 [Qemu-devel] [PATCH v2 0/2] block/nvme: add support for write zeros and discard Maxim Levitsky @ 2019-09-13 13:36 ` Maxim Levitsky 2019-09-18 20:22 ` John Snow 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 2/2] block/nvme: add support for discard Maxim Levitsky 2019-10-28 10:35 ` [PATCH v2 0/2] block/nvme: add support for write zeros and discard Max Reitz 2 siblings, 1 reply; 8+ messages in thread From: Maxim Levitsky @ 2019-09-13 13:36 UTC (permalink / raw) To: qemu-devel Cc: Fam Zheng, Kevin Wolf, qemu-block, Max Reitz, Keith Busch, Paolo Bonzini, Maxim Levitsky, John Snow 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(-) 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)) -- 2.17.2 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [Qemu-devel] [PATCH v2 1/2] block/nvme: add support for write zeros 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 1/2] block/nvme: add support for write zeros Maxim Levitsky @ 2019-09-18 20:22 ` John Snow 0 siblings, 0 replies; 8+ messages in thread From: John Snow @ 2019-09-18 20:22 UTC (permalink / raw) To: Maxim Levitsky, qemu-devel Cc: Fam Zheng, Kevin Wolf, qemu-block, Max Reitz, Keith Busch, Paolo Bonzini 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> ^ permalink raw reply [flat|nested] 8+ messages in thread
* [Qemu-devel] [PATCH v2 2/2] block/nvme: add support for discard 2019-09-13 13:36 [Qemu-devel] [PATCH v2 0/2] block/nvme: add support for write zeros and discard Maxim Levitsky 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 1/2] block/nvme: add support for write zeros Maxim Levitsky @ 2019-09-13 13:36 ` Maxim Levitsky 2019-09-18 20:24 ` John Snow 2019-10-28 10:35 ` [PATCH v2 0/2] block/nvme: add support for write zeros and discard Max Reitz 2 siblings, 1 reply; 8+ messages in thread From: Maxim Levitsky @ 2019-09-13 13:36 UTC (permalink / raw) To: qemu-devel Cc: Fam Zheng, Kevin Wolf, qemu-block, Max Reitz, Keith Busch, Paolo Bonzini, Maxim Levitsky, John Snow Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com> --- block/nvme.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++ block/trace-events | 2 ++ 2 files changed, 85 insertions(+) diff --git a/block/nvme.c b/block/nvme.c index d95265fae4..c17edd6aae 100644 --- a/block/nvme.c +++ b/block/nvme.c @@ -112,6 +112,7 @@ typedef struct { bool plugged; bool supports_write_zeroes; + bool supports_discard; CoMutex dma_map_lock; CoQueue dma_flush_queue; @@ -463,6 +464,7 @@ static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp) oncs = le16_to_cpu(idctrl->oncs); s->supports_write_zeroes = !!(oncs & NVME_ONCS_WRITE_ZEROS); + s->supports_discard = !!(oncs & NVME_ONCS_DSM); memset(resp, 0, 4096); @@ -1153,6 +1155,86 @@ static coroutine_fn int nvme_co_pwrite_zeroes(BlockDriverState *bs, } +static int coroutine_fn nvme_co_pdiscard(BlockDriverState *bs, + int64_t offset, + int bytes) +{ + BDRVNVMeState *s = bs->opaque; + NVMeQueuePair *ioq = s->queues[1]; + NVMeRequest *req; + NvmeDsmRange *buf; + QEMUIOVector local_qiov; + int ret; + + NvmeCmd cmd = { + .opcode = NVME_CMD_DSM, + .nsid = cpu_to_le32(s->nsid), + .cdw10 = cpu_to_le32(0), /*number of ranges - 0 based*/ + .cdw11 = cpu_to_le32(1 << 2), /*deallocate bit*/ + }; + + NVMeCoData data = { + .ctx = bdrv_get_aio_context(bs), + .ret = -EINPROGRESS, + }; + + if (!s->supports_discard) { + return -ENOTSUP; + } + + assert(s->nr_queues > 1); + + buf = qemu_try_blockalign0(bs, s->page_size); + if (!buf) { + return -ENOMEM; + } + + buf->nlb = cpu_to_le32(bytes >> s->blkshift); + buf->slba = cpu_to_le64(offset >> s->blkshift); + buf->cattr = 0; + + qemu_iovec_init(&local_qiov, 1); + qemu_iovec_add(&local_qiov, buf, 4096); + + req = nvme_get_free_req(ioq); + assert(req); + + qemu_co_mutex_lock(&s->dma_map_lock); + ret = nvme_cmd_map_qiov(bs, &cmd, req, &local_qiov); + qemu_co_mutex_unlock(&s->dma_map_lock); + + if (ret) { + req->busy = false; + goto out; + } + + trace_nvme_dsm(s, offset, bytes); + + nvme_submit_command(s, ioq, req, &cmd, nvme_rw_cb, &data); + + data.co = qemu_coroutine_self(); + while (data.ret == -EINPROGRESS) { + qemu_coroutine_yield(); + } + + qemu_co_mutex_lock(&s->dma_map_lock); + ret = nvme_cmd_unmap_qiov(bs, &local_qiov); + qemu_co_mutex_unlock(&s->dma_map_lock); + + if (ret) { + goto out; + } + + ret = data.ret; + trace_nvme_dsm_done(s, offset, bytes, ret); +out: + qemu_iovec_destroy(&local_qiov); + qemu_vfree(buf); + return ret; + +} + + static int nvme_reopen_prepare(BDRVReopenState *reopen_state, BlockReopenQueue *queue, Error **errp) { @@ -1259,6 +1341,7 @@ static BlockDriver bdrv_nvme = { .bdrv_co_pwritev = nvme_co_pwritev, .bdrv_co_pwrite_zeroes = nvme_co_pwrite_zeroes, + .bdrv_co_pdiscard = nvme_co_pdiscard, .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 651aa461d5..c61553b4b8 100644 --- a/block/trace-events +++ b/block/trace-events @@ -153,6 +153,8 @@ nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p off 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" +nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset %"PRId64" bytes %"PRId64"" +nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset %"PRId64" bytes %"PRId64" ret %d" nvme_dma_map_flush(void *s) "s %p" nvme_free_req_queue_wait(void *q) "q %p" nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d" -- 2.17.2 ^ permalink raw reply related [flat|nested] 8+ messages in thread
* Re: [Qemu-devel] [PATCH v2 2/2] block/nvme: add support for discard 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 2/2] block/nvme: add support for discard Maxim Levitsky @ 2019-09-18 20:24 ` John Snow 0 siblings, 0 replies; 8+ messages in thread From: John Snow @ 2019-09-18 20:24 UTC (permalink / raw) To: Maxim Levitsky, qemu-devel Cc: Fam Zheng, Kevin Wolf, qemu-block, Max Reitz, Keith Busch, Paolo Bonzini On 9/13/19 9:36 AM, Maxim Levitsky wrote: > Signed-off-by: Maxim Levitsky <mlevitsk@redhat.com> Same comments as 1/2; but not worth holding anything up. We'll find out from users if there are problems, but I wish we had a nicer way to test it. Reviewed-by: John Snow <jsnow@redhat.com> > --- > block/nvme.c | 83 ++++++++++++++++++++++++++++++++++++++++++++++ > block/trace-events | 2 ++ > 2 files changed, 85 insertions(+) > > diff --git a/block/nvme.c b/block/nvme.c > index d95265fae4..c17edd6aae 100644 > --- a/block/nvme.c > +++ b/block/nvme.c > @@ -112,6 +112,7 @@ typedef struct { > bool plugged; > > bool supports_write_zeroes; > + bool supports_discard; > > CoMutex dma_map_lock; > CoQueue dma_flush_queue; > @@ -463,6 +464,7 @@ static void nvme_identify(BlockDriverState *bs, int namespace, Error **errp) > > oncs = le16_to_cpu(idctrl->oncs); > s->supports_write_zeroes = !!(oncs & NVME_ONCS_WRITE_ZEROS); > + s->supports_discard = !!(oncs & NVME_ONCS_DSM); > > memset(resp, 0, 4096); > > @@ -1153,6 +1155,86 @@ static coroutine_fn int nvme_co_pwrite_zeroes(BlockDriverState *bs, > } > > > +static int coroutine_fn nvme_co_pdiscard(BlockDriverState *bs, > + int64_t offset, > + int bytes) > +{ > + BDRVNVMeState *s = bs->opaque; > + NVMeQueuePair *ioq = s->queues[1]; > + NVMeRequest *req; > + NvmeDsmRange *buf; > + QEMUIOVector local_qiov; > + int ret; > + > + NvmeCmd cmd = { > + .opcode = NVME_CMD_DSM, > + .nsid = cpu_to_le32(s->nsid), > + .cdw10 = cpu_to_le32(0), /*number of ranges - 0 based*/ > + .cdw11 = cpu_to_le32(1 << 2), /*deallocate bit*/ > + }; > + > + NVMeCoData data = { > + .ctx = bdrv_get_aio_context(bs), > + .ret = -EINPROGRESS, > + }; > + > + if (!s->supports_discard) { > + return -ENOTSUP; > + } > + > + assert(s->nr_queues > 1); > + > + buf = qemu_try_blockalign0(bs, s->page_size); > + if (!buf) { > + return -ENOMEM; > + } > + > + buf->nlb = cpu_to_le32(bytes >> s->blkshift); > + buf->slba = cpu_to_le64(offset >> s->blkshift); > + buf->cattr = 0; > + > + qemu_iovec_init(&local_qiov, 1); > + qemu_iovec_add(&local_qiov, buf, 4096); > + > + req = nvme_get_free_req(ioq); > + assert(req); > + > + qemu_co_mutex_lock(&s->dma_map_lock); > + ret = nvme_cmd_map_qiov(bs, &cmd, req, &local_qiov); > + qemu_co_mutex_unlock(&s->dma_map_lock); > + > + if (ret) { > + req->busy = false; > + goto out; > + } > + > + trace_nvme_dsm(s, offset, bytes); > + > + nvme_submit_command(s, ioq, req, &cmd, nvme_rw_cb, &data); > + > + data.co = qemu_coroutine_self(); > + while (data.ret == -EINPROGRESS) { > + qemu_coroutine_yield(); > + } > + > + qemu_co_mutex_lock(&s->dma_map_lock); > + ret = nvme_cmd_unmap_qiov(bs, &local_qiov); > + qemu_co_mutex_unlock(&s->dma_map_lock); > + > + if (ret) { > + goto out; > + } > + > + ret = data.ret; > + trace_nvme_dsm_done(s, offset, bytes, ret); > +out: > + qemu_iovec_destroy(&local_qiov); > + qemu_vfree(buf); > + return ret; > + > +} > + > + > static int nvme_reopen_prepare(BDRVReopenState *reopen_state, > BlockReopenQueue *queue, Error **errp) > { > @@ -1259,6 +1341,7 @@ static BlockDriver bdrv_nvme = { > .bdrv_co_pwritev = nvme_co_pwritev, > > .bdrv_co_pwrite_zeroes = nvme_co_pwrite_zeroes, > + .bdrv_co_pdiscard = nvme_co_pdiscard, > > .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 651aa461d5..c61553b4b8 100644 > --- a/block/trace-events > +++ b/block/trace-events > @@ -153,6 +153,8 @@ nvme_write_zeroes(void *s, uint64_t offset, uint64_t bytes, int flags) "s %p off > 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" > +nvme_dsm(void *s, uint64_t offset, uint64_t bytes) "s %p offset %"PRId64" bytes %"PRId64"" > +nvme_dsm_done(void *s, uint64_t offset, uint64_t bytes, int ret) "s %p offset %"PRId64" bytes %"PRId64" ret %d" > nvme_dma_map_flush(void *s) "s %p" > nvme_free_req_queue_wait(void *q) "q %p" > nvme_cmd_map_qiov(void *s, void *cmd, void *req, void *qiov, int entries) "s %p cmd %p req %p qiov %p entries %d" > -- —js ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 0/2] block/nvme: add support for write zeros and discard 2019-09-13 13:36 [Qemu-devel] [PATCH v2 0/2] block/nvme: add support for write zeros and discard Maxim Levitsky 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 1/2] block/nvme: add support for write zeros Maxim Levitsky 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 2/2] block/nvme: add support for discard Maxim Levitsky @ 2019-10-28 10:35 ` Max Reitz 2019-10-29 13:33 ` John Snow 2 siblings, 1 reply; 8+ messages in thread From: Max Reitz @ 2019-10-28 10:35 UTC (permalink / raw) To: Maxim Levitsky, qemu-devel Cc: Fam Zheng, Kevin Wolf, qemu-block, Keith Busch, Paolo Bonzini, John Snow [-- Attachment #1.1: Type: text/plain, Size: 837 bytes --] On 13.09.19 15:36, Maxim Levitsky wrote: > This is the second part of the patches I prepared > for this driver back when I worked on mdev-nvme. > > V2: addressed review feedback, no major changes > > Best regards, > Maxim Levitsky > > Maxim Levitsky (2): > block/nvme: add support for write zeros > block/nvme: add support for discard > > block/nvme.c | 155 ++++++++++++++++++++++++++++++++++++++++++- > block/trace-events | 3 + > include/block/nvme.h | 19 +++++- > 3 files changed, 175 insertions(+), 2 deletions(-) Thanks, fixed the indentation in nvme.h in patch 1, and applied to my block branch: https://git.xanclic.moe/XanClic/qemu/commits/branch/block For the record, I don’t think !!x has benefits over x != 0 and I personally prefer bool y = x over any of it. O:-) Max [-- Attachment #2: OpenPGP digital signature --] [-- Type: application/pgp-signature, Size: 488 bytes --] ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 0/2] block/nvme: add support for write zeros and discard 2019-10-28 10:35 ` [PATCH v2 0/2] block/nvme: add support for write zeros and discard Max Reitz @ 2019-10-29 13:33 ` John Snow 2019-11-04 17:54 ` Maxim Levitsky 0 siblings, 1 reply; 8+ messages in thread From: John Snow @ 2019-10-29 13:33 UTC (permalink / raw) To: Max Reitz, Maxim Levitsky, qemu-devel Cc: Fam Zheng, Paolo Bonzini, qemu-block, Keith Busch, Kevin Wolf On 10/28/19 6:35 AM, Max Reitz wrote: > On 13.09.19 15:36, Maxim Levitsky wrote: >> This is the second part of the patches I prepared >> for this driver back when I worked on mdev-nvme. >> >> V2: addressed review feedback, no major changes >> >> Best regards, >> Maxim Levitsky >> >> Maxim Levitsky (2): >> block/nvme: add support for write zeros >> block/nvme: add support for discard >> >> block/nvme.c | 155 ++++++++++++++++++++++++++++++++++++++++++- >> block/trace-events | 3 + >> include/block/nvme.h | 19 +++++- >> 3 files changed, 175 insertions(+), 2 deletions(-) > Thanks, fixed the indentation in nvme.h in patch 1, and applied to my > block branch: > > https://git.xanclic.moe/XanClic/qemu/commits/branch/block > > For the record, I don’t think !!x has benefits over x != 0 and I > personally prefer bool y = x over any of it. O:-) > Well, that's even better :) For me, it's about making booleans obvious as booleans and that's all. --js ^ permalink raw reply [flat|nested] 8+ messages in thread
* Re: [PATCH v2 0/2] block/nvme: add support for write zeros and discard 2019-10-29 13:33 ` John Snow @ 2019-11-04 17:54 ` Maxim Levitsky 0 siblings, 0 replies; 8+ messages in thread From: Maxim Levitsky @ 2019-11-04 17:54 UTC (permalink / raw) To: John Snow, Max Reitz, qemu-devel Cc: Fam Zheng, Paolo Bonzini, qemu-block, Keith Busch, Kevin Wolf On Tue, 2019-10-29 at 09:33 -0400, John Snow wrote: > > On 10/28/19 6:35 AM, Max Reitz wrote: > > On 13.09.19 15:36, Maxim Levitsky wrote: > > > This is the second part of the patches I prepared > > > for this driver back when I worked on mdev-nvme. > > > > > > V2: addressed review feedback, no major changes > > > > > > Best regards, > > > Maxim Levitsky > > > > > > Maxim Levitsky (2): > > > block/nvme: add support for write zeros > > > block/nvme: add support for discard > > > > > > block/nvme.c | 155 ++++++++++++++++++++++++++++++++++++++++++- > > > block/trace-events | 3 + > > > include/block/nvme.h | 19 +++++- > > > 3 files changed, 175 insertions(+), 2 deletions(-) > > > > Thanks, fixed the indentation in nvme.h in patch 1, and applied to my > > block branch: > > > > https://git.xanclic.moe/XanClic/qemu/commits/branch/block > > > > For the record, I don’t think !!x has benefits over x != 0 and I > > personally prefer bool y = x over any of it. O:-) > > > > Well, that's even better :) For me, it's about making booleans obvious > as booleans and that's all. > > --js Thanks to all of you!! Best regards, Maxim Levitsky ^ permalink raw reply [flat|nested] 8+ messages in thread
end of thread, other threads:[~2019-11-04 17:57 UTC | newest] Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2019-09-13 13:36 [Qemu-devel] [PATCH v2 0/2] block/nvme: add support for write zeros and discard Maxim Levitsky 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 1/2] block/nvme: add support for write zeros Maxim Levitsky 2019-09-18 20:22 ` John Snow 2019-09-13 13:36 ` [Qemu-devel] [PATCH v2 2/2] block/nvme: add support for discard Maxim Levitsky 2019-09-18 20:24 ` John Snow 2019-10-28 10:35 ` [PATCH v2 0/2] block/nvme: add support for write zeros and discard Max Reitz 2019-10-29 13:33 ` John Snow 2019-11-04 17:54 ` Maxim Levitsky
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).