nvme: add csi, ms and nuse to sysfs
v6 cover nvme: add csi, ms and nuse to sysfs
Rebased on the current nvme/nvme-6.8 branch and added the Review tags. Also addressed the printk format issue pointed out by Chaitanya.
About PI: I've added a note to the first commit pointing on this topic. I understood from the discussion, that yes it is an issue but it already exists and we will address this in future.
Thanks,
Daniel
libnvme changes:
https://github.com/igaw/libnvme/tree/tree-no-cmd
changes:
v6:
- rebased on nvme/nvme-6.8
- collected Reviewed tags
- updated printk format
v5:
- reverted trigger happy conversion to nvme_ns_head in nvme_zns_alloc_report_buffer
- removed debug output
- added refactoring patches
- ratelimit silence suppress messages
- added reviewed tags
- https://lore.kernel.org/linux-nvme/[email protected]/
v4:
- drop 'use nvme_ns_head instead nvme_ns' patches
- ratelimit nuse update per namespace not globally
- rename ns attribute group
- added non-multipath nuse update logic
- added cacheline optimization
- https://lore.kernel.org/linux-nvme/[email protected]/
v3:
- cut overlong lines shorter
- fixed disk (queuedata) initialization order
- more testing with blktest
- added nuse ratelimit
- added reviewed tags
- https://lore.kernel.org/linux-nvme/[email protected]/
v2:
- moved ns id data to nvme_ns_head
- dropped ds, nsze
- https://lore.kernel.org/linux-nvme/[email protected]/
v1:
- initial version
- https://lore.kernel.org/linux-nvme/[email protected]/
Daniel Wagner (6):
nvme: move ns id info to struct nvme_ns_head
nvme: refactor ns info helpers
nvme: refactor ns info setup function
nvme: rename ns attribute group
nvme: add csi, ms and nuse to sysfs
nvme: repack struct nvme_ns_head
drivers/nvme/host/core.c | 168 ++++++++++++++++++---------------- drivers/nvme/host/ioctl.c | 8 +- drivers/nvme/host/multipath.c | 2 +- drivers/nvme/host/nvme.h | 44 +++++---- drivers/nvme/host/rdma.c | 4 +- drivers/nvme/host/sysfs.c | 99 ++++++++++++++++++-- drivers/nvme/host/zns.c | 35 +++---- 7 files changed, 233 insertions(+), 127 deletions(-)
Signature · 1 line
2.43.0
PATCH v6 1/6 nvme: move ns id info to struct nvme_ns_head
Move the namesapce info to struct nvme_ns_head, because it's the same for all associated namespaces.
Note: with multipathing enabled the PI information is shared between all paths. If a path is using a different PI configuration it will overwrite the previous settings. This is obviously not correct and such configuration will be rejected in future. For the time being we expect a correctly configured storage.
--- drivers/nvme/host/core.c | 80 ++++++++++++++++++++------------------- drivers/nvme/host/ioctl.c | 8 ++-- drivers/nvme/host/nvme.h | 28 +++++++------- drivers/nvme/host/rdma.c | 2 +- drivers/nvme/host/zns.c | 17 +++++---- 5 files changed, 69 insertions(+), 66 deletions(-)
drivers/nvme/host/core.c +41 -39
@@ -312,12 +312,12 @@ static void nvme_log_error(struct request *req)struct nvme_request *nr =nvme_req (req);if (ns) { - pr_err_ratelimited("%s: %s(0x%x) @ LBA %llu, %llu blocks, %s (sct 0x%x / sc 0x%x) %s%s\n", + pr_err_ratelimited("%s: %s(0x%x) @ LBA %llu, %u blocks, %s (sct 0x%x / sc 0x%x) %s%s\n", ns->disk ? ns->disk->disk_name : "?", nvme_get_opcode_str(nr->cmd->common.opcode), nr->cmd->common.opcode, - (unsigned long long)nvme_sect_to_lba(ns, blk_rq_pos(req)), - (unsigned long long)blk_rq_bytes(req) >> ns->lba_shift, + nvme_sect_to_lba(ns, blk_rq_pos(req)), + blk_rq_bytes(req) >> ns->head->lba_shift, nvme_get_error_status_str(nr->status), nr->status >> 8 & 7, /* Status Code Type */ nr->status & 0xff, /* Status Code */
@@ -792,7 +792,7 @@ static blk_status_t nvme_setup_discard(struct nvme_ns *ns, struct request *req,if (queue_max_discard_segments (req->q) ==1 ) {u64 slba =nvme_sect_to_lba (ns,blk_rq_pos (req)); -u32 nlb =blk_rq_sectors (req) >> (ns->lba_shift -9 ); +u32 nlb =blk_rq_sectors (req) >> (ns->head->lba_shift -9 ); range[0 ].cattr =cpu_to_le32 (0 ); range[0 ].nlb =cpu_to_le32 (nlb);
@@ -801,7 +801,7 @@ static blk_status_t nvme_setup_discard(struct nvme_ns *ns, struct request *req, }else {__rq_for_each_bio (bio, req) {u64 slba =nvme_sect_to_lba (ns, bio->bi_iter.bi_sector); -u32 nlb = bio->bi_iter.bi_size >> ns->lba_shift; +u32 nlb = bio->bi_iter.bi_size >> ns->head->lba_shift;if (n < segments) { range[n].cattr =cpu_to_le32 (0 );
@@ -839,7 +839,7 @@ static void nvme_set_ref_tag(struct nvme_ns *ns, struct nvme_command *cmnd,u64 ref48;/* both rw and write zeroes share the same reftag format */ -switch (ns->guard_type) { +switch (ns->head->guard_type) {case NVME_NVM_NS_16B_GUARD : cmnd->rw.reftag =cpu_to_le32 (t10_pi_ref_tag (req));break ;
@@ -869,15 +869,16 @@ static inline blk_status_t nvme_setup_write_zeroes(struct nvme_ns *ns, cmnd->write_zeroes.slba =cpu_to_le64 (nvme_sect_to_lba (ns,blk_rq_pos (req))); cmnd->write_zeroes.length = -cpu_to_le16 ((blk_rq_bytes (req) >> ns->lba_shift) -1 ); +cpu_to_le16 ((blk_rq_bytes (req) >> ns->head->lba_shift) -1 ); -if (!(req->cmd_flags &REQ_NOUNMAP ) && (ns->features &NVME_NS_DEAC )) +if (!(req->cmd_flags &REQ_NOUNMAP ) && + (ns->head->features &NVME_NS_DEAC )) cmnd->write_zeroes.control |=cpu_to_le16 (NVME_WZ_DEAC );if (nvme_ns_has_pi (ns)) { cmnd->write_zeroes.control |=cpu_to_le16 (NVME_RW_PRINFO_PRACT ); -switch (ns->pi_type) { +switch (ns->head->pi_type) {case NVME_NS_DPS_PI_TYPE1 :case NVME_NS_DPS_PI_TYPE2 :nvme_set_ref_tag (ns, cmnd, req);
@@ -910,12 +911,13 @@ static inline blk_status_t nvme_setup_rw(struct nvme_ns *ns, cmnd->rw.cdw3 =0 ; cmnd->rw.metadata =0 ; cmnd->rw.slba =cpu_to_le64 (nvme_sect_to_lba (ns,blk_rq_pos (req))); - cmnd->rw.length =cpu_to_le16 ((blk_rq_bytes (req) >> ns->lba_shift) -1 ); + cmnd->rw.length = +cpu_to_le16 ((blk_rq_bytes (req) >> ns->head->lba_shift) -1 ); cmnd->rw.reftag =0 ; cmnd->rw.apptag =0 ; cmnd->rw.appmask =0 ; -if (ns->ms) { +if (ns->head->ms) {/* * If formated with metadata, the block layer always provides a * metadata buffer if CONFIG_BLK_DEV_INTEGRITY is enabled. Else
@@ -928,7 +930,7 @@ static inline blk_status_t nvme_setup_rw(struct nvme_ns *ns, control |=NVME_RW_PRINFO_PRACT ; } -switch (ns->pi_type) { +switch (ns->head->pi_type) {case NVME_NS_DPS_PI_TYPE3 : control |=NVME_RW_PRINFO_PRCHK_GUARD ;break ;
@@ -1663,9 +1665,9 @@ static void nvme_init_integrity(struct gendisk *disk, struct nvme_ns *ns, {struct blk_integrity integrity = { }; -switch (ns->pi_type) { +switch (ns->head->pi_type) {case NVME_NS_DPS_PI_TYPE3 : -switch (ns->guard_type) { +switch (ns->head->guard_type) {case NVME_NVM_NS_16B_GUARD : integrity.profile = &t10_pi_type3_crc; integrity.tag_size =sizeof (u16) +sizeof (u32);
@@ -1683,7 +1685,7 @@ static void nvme_init_integrity(struct gendisk *disk, struct nvme_ns *ns,break ;case NVME_NS_DPS_PI_TYPE1 :case NVME_NS_DPS_PI_TYPE2 : -switch (ns->guard_type) { +switch (ns->head->guard_type) {case NVME_NVM_NS_16B_GUARD : integrity.profile = &t10_pi_type1_crc; integrity.tag_size =sizeof (u16);
@@ -1704,7 +1706,7 @@ static void nvme_init_integrity(struct gendisk *disk, struct nvme_ns *ns,break ; } - integrity.tuple_size = ns->ms; + integrity.tuple_size = ns->head->ms;blk_integrity_register (disk, &integrity);blk_queue_max_integrity_segments (disk->queue, max_integrity_segments); }
@@ -1763,11 +1765,11 @@ static int nvme_init_ms(struct nvme_ns *ns, struct nvme_id_ns *id)int ret =0 ;u32 elbaf; - ns->pi_size =0 ; - ns->ms =le16_to_cpu (id->lbaf[lbaf].ms); + ns->head->pi_size =0 ; + ns->head->ms =le16_to_cpu (id->lbaf[lbaf].ms);if (!(ctrl->ctratt &NVME_CTRL_ATTR_ELBAS )) { - ns->pi_size =sizeof (struct t10_pi_tuple ); - ns->guard_type =NVME_NVM_NS_16B_GUARD ; + ns->head->pi_size =sizeof (struct t10_pi_tuple ); + ns->head->guard_type =NVME_NVM_NS_16B_GUARD ; goto set_pi; }
@@ -1790,13 +1792,13 @@ static int nvme_init_ms(struct nvme_ns *ns, struct nvme_id_ns *id)if (nvme_elbaf_sts (elbaf)) goto free_data; - ns->guard_type =nvme_elbaf_guard_type (elbaf); -switch (ns->guard_type) { + ns->head->guard_type =nvme_elbaf_guard_type (elbaf); +switch (ns->head->guard_type) {case NVME_NVM_NS_64B_GUARD : - ns->pi_size =sizeof (struct crc64_pi_tuple ); + ns->head->pi_size =sizeof (struct crc64_pi_tuple );break ;case NVME_NVM_NS_16B_GUARD : - ns->pi_size =sizeof (struct t10_pi_tuple ); + ns->head->pi_size =sizeof (struct t10_pi_tuple );break ;default :break ;
@@ -1805,10 +1807,10 @@ static int nvme_init_ms(struct nvme_ns *ns, struct nvme_id_ns *id) free_data:kfree (nvm); set_pi: -if (ns->pi_size && (first || ns->ms == ns->pi_size)) - ns->pi_type = id->dps &NVME_NS_DPS_PI_MASK ; +if (ns->head->pi_size && (first || ns->head->ms == ns->head->pi_size)) + ns->head->pi_type = id->dps &NVME_NS_DPS_PI_MASK ;else - ns->pi_type =0 ; + ns->head->pi_type =0 ;return ret; }
@@ -1822,8 +1824,8 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)if (ret)return ret; - ns->features &= ~(NVME_NS_METADATA_SUPPORTED |NVME_NS_EXT_LBAS ); -if (!ns->ms || !(ctrl->ops->flags &NVME_F_METADATA_SUPPORTED )) + ns->head->features &= ~(NVME_NS_METADATA_SUPPORTED |NVME_NS_EXT_LBAS ); +if (!ns->head->ms || !(ctrl->ops->flags &NVME_F_METADATA_SUPPORTED ))return 0 ;if (ctrl->ops->flags &NVME_F_FABRICS ) {
@@ -1835,7 +1837,7 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)if (WARN_ON_ONCE (!(id->flbas &NVME_NS_FLBAS_META_EXT )))return 0 ; - ns->features |=NVME_NS_EXT_LBAS ; + ns->head->features |=NVME_NS_EXT_LBAS ;/* * The current fabrics transport drivers support namespace
@@ -1847,7 +1849,7 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)* gain the ability to use other metadata formats. */ if (ctrl->max_integrity_segments &&nvme_ns_has_pi (ns)) - ns->features |=NVME_NS_METADATA_SUPPORTED ; + ns->head->features |=NVME_NS_METADATA_SUPPORTED ; }else {/* * For PCIe controllers, we can't easily remap the separate
@@ -1856,9 +1858,9 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)* We allow extended LBAs for the passthrough interface, though. */ if (id->flbas &NVME_NS_FLBAS_META_EXT ) - ns->features |=NVME_NS_EXT_LBAS ; + ns->head->features |=NVME_NS_EXT_LBAS ;else - ns->features |=NVME_NS_METADATA_SUPPORTED ; + ns->head->features |=NVME_NS_METADATA_SUPPORTED ; }return 0 ; }
@@ -1885,14 +1887,14 @@ static void nvme_update_disk_info(struct gendisk *disk, struct nvme_ns *ns, struct nvme_id_ns *id) { sector_t capacity = nvme_lba_to_sect(ns, le64_to_cpu(id->nsze)); -u32 bs =1U << ns->lba_shift; +u32 bs =1U << ns->head->lba_shift;u32 atomic_bs, phys_bs, io_opt =0 ;/* * The block layer can't support LBA sizes larger than the page size * yet, so catch this early and don't allow block I/O. */ -if (ns->lba_shift >PAGE_SHIFT ) { +if (ns->head->lba_shift >PAGE_SHIFT ) { capacity =0 ; bs = (1 <<9 ); }
@@ -1935,9 +1937,9 @@ static void nvme_update_disk_info(struct gendisk *disk,* I/O to namespaces with metadata except when the namespace supports * PI, as it can strip/insert in that case. */ -if (ns->ms) { +if (ns->head->ms) {if (IS_ENABLED (CONFIG_BLK_DEV_INTEGRITY ) && - (ns->features &NVME_NS_METADATA_SUPPORTED )) + (ns->head->features &NVME_NS_METADATA_SUPPORTED ))nvme_init_integrity (disk, ns, ns->ctrl->max_integrity_segments);else if (!nvme_ns_has_pi (ns))
@@ -2031,7 +2033,7 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,blk_mq_freeze_queue (ns->disk->queue); lbaf =nvme_lbaf_index (id->flbas); - ns->lba_shift = id->lbaf[lbaf].ds; + ns->head->lba_shift = id->lbaf[lbaf].ds;nvme_set_queue_limits (ns->ctrl, ns->queue); ret =nvme_configure_metadata (ns, id);
@@ -2057,7 +2059,7 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,* do not return zeroes. */ if ((id->dlfeat &0x7 ) ==0x1 && (id->dlfeat & (1 <<3 ))) - ns->features |=NVME_NS_DEAC ; + ns->head->features |=NVME_NS_DEAC ;set_disk_ro (ns->disk,nvme_ns_is_readonly (ns, info));set_bit (NVME_NS_READY , &ns->flags);blk_mq_unfreeze_queue (ns->disk->queue);
drivers/nvme/host/ioctl.c +4 -4
@@ -224,10 +224,10 @@ static int nvme_submit_io(struct nvme_ns *ns, struct nvme_user_io __user *uio)return -EINVAL ; } - length = (io.nblocks +1 ) << ns->lba_shift; + length = (io.nblocks +1 ) << ns->head->lba_shift;if ((io.control &NVME_RW_PRINFO_PRACT ) && - ns->ms ==sizeof (struct t10_pi_tuple )) { + ns->head->ms ==sizeof (struct t10_pi_tuple )) {/* * Protection information is stripped/inserted by the * controller.
@@ -237,11 +237,11 @@ static int nvme_submit_io(struct nvme_ns *ns, struct nvme_user_io __user *uio) meta_len =0 ; metadata =NULL ; }else { - meta_len = (io.nblocks +1 ) * ns->ms; + meta_len = (io.nblocks +1 ) * ns->head->ms; metadata =nvme_to_user_ptr (io.metadata); } -if (ns->features &NVME_NS_EXT_LBAS ) { +if (ns->head->features &NVME_NS_EXT_LBAS ) { length += meta_len; meta_len =0 ; }else if (meta_len) {
drivers/nvme/host/nvme.h +14 -14
@@ -446,6 +446,17 @@ struct nvme_ns_head {bool shared;int instance;struct nvme_effects_log *effects; +int lba_shift; +u16 ms; +u16 pi_size; +u16 sgs; +u32 sws; +u8 pi_type; +u8 guard_type; +#ifdef CONFIG_BLK_DEV_ZONED +u64 zsze; +#endif +unsigned long features;struct cdev cdev;struct device cdev_device;
@@ -487,17 +498,6 @@ struct nvme_ns {struct kref kref;struct nvme_ns_head *head; -int lba_shift; -u16 ms; -u16 pi_size; -u16 sgs; -u32 sws; -u8 pi_type; -u8 guard_type; -#ifdef CONFIG_BLK_DEV_ZONED -u64 zsze; -#endif -unsigned long features;unsigned long flags;#define NVME_NS_REMOVING 0#define NVME_NS_ANA_PENDING 2
@@ -514,7 +514,7 @@ struct nvme_ns {/* NVMe ns supports metadata actions by the controller (generate/strip) */ static inline bool nvme_ns_has_pi (struct nvme_ns *ns) { -return ns->pi_type && ns->ms == ns->pi_size; +return ns->head->pi_type && ns->head->ms == ns->head->pi_size; }struct nvme_ctrl_ops {
@@ -648,7 +648,7 @@ static inline int nvme_reset_subsystem(struct nvme_ctrl *ctrl)*/ static inline u64 nvme_sect_to_lba (struct nvme_ns *ns,sector_t sector) { -return sector >> (ns->lba_shift -SECTOR_SHIFT ); +return sector >> (ns->head->lba_shift -SECTOR_SHIFT ); }/*
@@ -656,7 +656,7 @@ static inline u64 nvme_sect_to_lba(struct nvme_ns *ns, sector_t sector)*/ static inline sector_t nvme_lba_to_sect (struct nvme_ns *ns,u64 lba) { -return lba << (ns->lba_shift -SECTOR_SHIFT ); +return lba << (ns->head->lba_shift -SECTOR_SHIFT ); }/*
drivers/nvme/host/rdma.c +1 -1
@@ -1418,7 +1418,7 @@ static int nvme_rdma_map_sg_pi(struct nvme_rdma_queue *queue, goto mr_put;nvme_rdma_set_sig_attrs (blk_get_integrity (bio->bi_bdev->bd_disk), c, - req->mr->sig_attrs, ns->pi_type); + req->mr->sig_attrs, ns->head->pi_type);nvme_rdma_set_prot_checks (c, &req->mr->sig_attrs->check_mask);ib_update_fast_reg_key (req->mr,ib_inc_rkey (req->mr->rkey));
drivers/nvme/host/zns.c +9 -8
@@ -11,7 +11,7 @@ int nvme_revalidate_zones(struct nvme_ns *ns) {struct request_queue *q = ns->queue; -blk_queue_chunk_sectors (q, ns->zsze); +blk_queue_chunk_sectors (q, ns->head->zsze);blk_queue_max_zone_append_sectors (q, ns->ctrl->max_zone_append);return blk_revalidate_disk_zones (ns->disk,NULL );
@@ -99,11 +99,12 @@ int nvme_update_zone_info(struct nvme_ns *ns, unsigned lbaf) goto free_data; } - ns->zsze =nvme_lba_to_sect (ns,le64_to_cpu (id->lbafe[lbaf].zsze)); -if (!is_power_of_2 (ns->zsze)) { + ns->head->zsze = +nvme_lba_to_sect (ns,le64_to_cpu (id->lbafe[lbaf].zsze)); +if (!is_power_of_2 (ns->head->zsze)) {dev_warn (ns->ctrl->device,"invalid zone size:%llu for namespace:%u\n" , - ns->zsze, ns->head->ns_id); + ns->head->zsze, ns->head->ns_id); status = -ENODEV ; goto free_data; }
@@ -128,7 +129,7 @@ static void *nvme_zns_alloc_report_buffer(struct nvme_ns *ns,sizeof (struct nvme_zone_descriptor ); nr_zones = min_t(unsigned int, nr_zones, -get_capacity (ns->disk) >>ilog2 (ns->zsze)); +get_capacity (ns->disk) >>ilog2 (ns->head->zsze)); bufsize =sizeof (struct nvme_zone_report ) + nr_zones *sizeof (struct nvme_zone_descriptor );
@@ -162,7 +163,7 @@ static int nvme_zone_parse_entry(struct nvme_ns *ns, zone.type =BLK_ZONE_TYPE_SEQWRITE_REQ ; zone.cond = entry->zs >>4 ; - zone.len = ns->zsze; + zone.len = ns->head->zsze; zone.capacity =nvme_lba_to_sect (ns,le64_to_cpu (entry->zcap)); zone.start =nvme_lba_to_sect (ns,le64_to_cpu (entry->zslba));if (zone.cond ==BLK_ZONE_COND_FULL )
@@ -196,7 +197,7 @@ int nvme_ns_report_zones(struct nvme_ns *ns, sector_t sector, c.zmr.zrasf =NVME_ZRASF_ZONE_REPORT_ALL ; c.zmr.pr =NVME_REPORT_ZONE_PARTIAL ; - sector &= ~(ns->zsze -1 ); + sector &= ~(ns->head->zsze -1 );while (zone_idx < nr_zones && sector <get_capacity (ns->disk)) {memset (report,0 , buflen);
@@ -220,7 +221,7 @@ int nvme_ns_report_zones(struct nvme_ns *ns, sector_t sector, zone_idx++; } - sector += ns->zsze * nz; + sector += ns->head->zsze * nz; }if (zone_idx >0 )
Signature · 1 line
2.43.0
Re: [PATCH v6 1/6] nvme: move ns id info to struct nvme_ns_head
Quoted 11 lines · Daniel Wagner
Move the namesapce info to struct
nvme_ns_head, because it's the same for all associated namespaces.Note: with multipathing enabled the PI information is shared between all paths. If a path is using a different PI configuration it will overwrite the previous settings. This is obviously not correct and such configuration will be rejected in future. For the time being we expect a correctly configured storage.
Reviewed-by Christoph Hellwig <[email protected]>Signed-off-by Daniel Wagner <[email protected]>
Quoted 8 lines
--- drivers/nvme/host/core.c | 80 ++++++++++++++++++++------------------- drivers/nvme/host/ioctl.c | 8 ++-- drivers/nvme/host/nvme.h | 28 +++++++------- drivers/nvme/host/rdma.c | 2 +- drivers/nvme/host/zns.c | 17 +++++---- 5 files changed, 69 insertions(+), 66 deletions(-)
Cheers,
Hannes
Signature · 5 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald, Werner Knoblich
Re: [PATCH v6 1/6] nvme: move ns id info to struct nvme_ns_head
In reply to Daniel WagnerPATCH v6 2/6 nvme: refactor ns info helpers
In reply to Daniel WagnerPass in the nvme_ns_head pointer directly. This reduces the necessity on the caller side have the nvme_ns data structure present. Thus we can refactor the caller side in the next step as well.
--- drivers/nvme/host/core.c | 36 +++++++++++++++++++++--------------- drivers/nvme/host/nvme.h | 12 ++++++------ drivers/nvme/host/rdma.c | 2 +- drivers/nvme/host/zns.c | 12 ++++++------ 4 files changed, 34 insertions(+), 28 deletions(-)
drivers/nvme/host/core.c +21 -15
@@ -316,7 +316,7 @@ static void nvme_log_error(struct request *req) ns->disk ? ns->disk->disk_name :"?" ,nvme_get_opcode_str (nr->cmd->common.opcode), nr->cmd->common.opcode, -nvme_sect_to_lba (ns,blk_rq_pos (req)), +nvme_sect_to_lba (ns->head,blk_rq_pos (req)),blk_rq_bytes (req) >> ns->head->lba_shift,nvme_get_error_status_str (nr->status), nr->status >> 8 & 7, /* Status Code Type */
@@ -372,9 +372,12 @@ static inline enum nvme_disposition nvme_decide_disposition(struct request *req)static inline void nvme_end_req_zoned (struct request *req) {if (IS_ENABLED (CONFIG_BLK_DEV_ZONED ) && -req_op (req) ==REQ_OP_ZONE_APPEND ) - req->__sector =nvme_lba_to_sect (req->q->queuedata, +req_op (req) ==REQ_OP_ZONE_APPEND ) { +struct nvme_ns *ns = req->q->queuedata; + + req->__sector =nvme_lba_to_sect (ns->head,le64_to_cpu (nvme_req (req)->result.u64)); + } }static inline void nvme_end_req (struct request *req)
@@ -791,7 +794,7 @@ static blk_status_t nvme_setup_discard(struct nvme_ns *ns, struct request *req, }if (queue_max_discard_segments (req->q) ==1 ) { -u64 slba =nvme_sect_to_lba (ns,blk_rq_pos (req)); +u64 slba =nvme_sect_to_lba (ns->head,blk_rq_pos (req));u32 nlb =blk_rq_sectors (req) >> (ns->head->lba_shift -9 ); range[0 ].cattr =cpu_to_le32 (0 );
@@ -800,7 +803,8 @@ static blk_status_t nvme_setup_discard(struct nvme_ns *ns, struct request *req, n =1 ; }else {__rq_for_each_bio (bio, req) { -u64 slba =nvme_sect_to_lba (ns, bio->bi_iter.bi_sector); +u64 slba =nvme_sect_to_lba (ns->head, + bio->bi_iter.bi_sector);u32 nlb = bio->bi_iter.bi_size >> ns->head->lba_shift;if (n < segments) {
@@ -867,7 +871,7 @@ static inline blk_status_t nvme_setup_write_zeroes(struct nvme_ns *ns, cmnd->write_zeroes.opcode = nvme_cmd_write_zeroes; cmnd->write_zeroes.nsid =cpu_to_le32 (ns->head->ns_id); cmnd->write_zeroes.slba = -cpu_to_le64 (nvme_sect_to_lba (ns,blk_rq_pos (req))); +cpu_to_le64 (nvme_sect_to_lba (ns->head,blk_rq_pos (req))); cmnd->write_zeroes.length =cpu_to_le16 ((blk_rq_bytes (req) >> ns->head->lba_shift) -1 );
@@ -875,7 +879,7 @@ static inline blk_status_t nvme_setup_write_zeroes(struct nvme_ns *ns, (ns->head->features & NVME_NS_DEAC)) cmnd->write_zeroes.control |=cpu_to_le16 (NVME_WZ_DEAC ); -if (nvme_ns_has_pi (ns)) { +if (nvme_ns_has_pi (ns->head)) { cmnd->write_zeroes.control |=cpu_to_le16 (NVME_RW_PRINFO_PRACT );switch (ns->head->pi_type) {
@@ -910,7 +914,8 @@ static inline blk_status_t nvme_setup_rw(struct nvme_ns *ns, cmnd->rw.cdw2 =0 ; cmnd->rw.cdw3 =0 ; cmnd->rw.metadata =0 ; - cmnd->rw.slba =cpu_to_le64 (nvme_sect_to_lba (ns,blk_rq_pos (req))); + cmnd->rw.slba = +cpu_to_le64 (nvme_sect_to_lba (ns->head,blk_rq_pos (req))); cmnd->rw.length =cpu_to_le16 ((blk_rq_bytes (req) >> ns->head->lba_shift) -1 ); cmnd->rw.reftag =0 ;
@@ -925,7 +930,7 @@ static inline blk_status_t nvme_setup_rw(struct nvme_ns *ns,* namespace capacity to zero to prevent any I/O. */ if (!blk_integrity_rq (req)) { -if (WARN_ON_ONCE (!nvme_ns_has_pi (ns))) +if (WARN_ON_ONCE (!nvme_ns_has_pi (ns->head)))return BLK_STS_NOTSUPP ; control |=NVME_RW_PRINFO_PRACT ; }
@@ -1723,8 +1728,9 @@ static void nvme_config_discard(struct gendisk *disk, struct nvme_ns *ns)struct request_queue *queue = disk->queue;u32 size =queue_logical_block_size (queue); -if (ctrl->dmrsl && ctrl->dmrsl <=nvme_sect_to_lba (ns,UINT_MAX )) - ctrl->max_discard_sectors =nvme_lba_to_sect (ns, ctrl->dmrsl); +if (ctrl->dmrsl && ctrl->dmrsl <=nvme_sect_to_lba (ns->head,UINT_MAX )) + ctrl->max_discard_sectors = +nvme_lba_to_sect (ns->head, ctrl->dmrsl);if (ctrl->max_discard_sectors ==0 ) {blk_queue_max_discard_sectors (queue,0 );
@@ -1848,7 +1854,7 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)* Note, this check will need to be modified if any drivers * gain the ability to use other metadata formats. */ -if (ctrl->max_integrity_segments &&nvme_ns_has_pi (ns)) +if (ctrl->max_integrity_segments &&nvme_ns_has_pi (ns->head)) ns->head->features |=NVME_NS_METADATA_SUPPORTED ; }else {/*
@@ -1886,7 +1892,7 @@ static void nvme_set_queue_limits(struct nvme_ctrl *ctrl,static void nvme_update_disk_info (struct gendisk *disk,struct nvme_ns *ns,struct nvme_id_ns *id) { -sector_t capacity =nvme_lba_to_sect (ns,le64_to_cpu (id->nsze)); +sector_t capacity =nvme_lba_to_sect (ns->head,le64_to_cpu (id->nsze));u32 bs =1U << ns->head->lba_shift;u32 atomic_bs, phys_bs, io_opt =0 ;
@@ -1942,7 +1948,7 @@ static void nvme_update_disk_info(struct gendisk *disk, (ns->head->features & NVME_NS_METADATA_SUPPORTED))nvme_init_integrity (disk, ns, ns->ctrl->max_integrity_segments); - else if (!nvme_ns_has_pi(ns)) + else if (!nvme_ns_has_pi(ns->head)) capacity =0 ; }
@@ -1973,7 +1979,7 @@ static void nvme_set_chunk_sectors(struct nvme_ns *ns, struct nvme_id_ns *id) is_power_of_2(ctrl->max_hw_sectors)) iob = ctrl->max_hw_sectors; else - iob =nvme_lba_to_sect (ns,le16_to_cpu (id->noiob)); + iob =nvme_lba_to_sect (ns->head,le16_to_cpu (id->noiob));if (!iob)return ;
drivers/nvme/host/nvme.h +6 -6
@@ -512,9 +512,9 @@ struct nvme_ns { };/* NVMe ns supports metadata actions by the controller (generate/strip) */ -static inline bool nvme_ns_has_pi (struct nvme_ns *ns) +static inline bool nvme_ns_has_pi (struct nvme_ns_head *head) { -return ns->head->pi_type && ns->head->ms == ns->head->pi_size; +return head->pi_type && head->ms == head->pi_size; }struct nvme_ctrl_ops {
@@ -646,17 +646,17 @@ static inline int nvme_reset_subsystem(struct nvme_ctrl *ctrl)/* * Convert a 512B sector number to a device logical block number. */ -static inline u64 nvme_sect_to_lba (struct nvme_ns *ns,sector_t sector) +static inline u64 nvme_sect_to_lba (struct nvme_ns_head *head,sector_t sector) { -return sector >> (ns->head->lba_shift -SECTOR_SHIFT ); +return sector >> (head->lba_shift -SECTOR_SHIFT ); }/* * Convert a device logical block number to a 512B sector number. */ -static inline sector_t nvme_lba_to_sect (struct nvme_ns *ns,u64 lba) +static inline sector_t nvme_lba_to_sect (struct nvme_ns_head *head,u64 lba) { -return lba << (ns->head->lba_shift -SECTOR_SHIFT ); +return lba << (head->lba_shift -SECTOR_SHIFT ); }/*
drivers/nvme/host/rdma.c +1 -1
@@ -2012,7 +2012,7 @@ static blk_status_t nvme_rdma_queue_rq(struct blk_mq_hw_ctx *hctx, queue->pi_support && (c->common.opcode == nvme_cmd_write || c->common.opcode == nvme_cmd_read) && - nvme_ns_has_pi(ns)) + nvme_ns_has_pi(ns->head)) req->use_sig_mr = true; else req->use_sig_mr = false;
drivers/nvme/host/zns.c +6 -6
@@ -100,7 +100,7 @@ int nvme_update_zone_info(struct nvme_ns *ns, unsigned lbaf) } ns->head->zsze = -nvme_lba_to_sect (ns,le64_to_cpu (id->lbafe[lbaf].zsze)); +nvme_lba_to_sect (ns->head,le64_to_cpu (id->lbafe[lbaf].zsze));if (!is_power_of_2 (ns->head->zsze)) { dev_warn(ns->ctrl->device, "invalid zone size:%llu for namespace:%u\n",
@@ -164,12 +164,12 @@ static int nvme_zone_parse_entry(struct nvme_ns *ns, zone.type =BLK_ZONE_TYPE_SEQWRITE_REQ ; zone.cond = entry->zs >>4 ; zone.len = ns->head->zsze; - zone.capacity =nvme_lba_to_sect (ns,le64_to_cpu (entry->zcap)); - zone.start =nvme_lba_to_sect (ns,le64_to_cpu (entry->zslba)); + zone.capacity =nvme_lba_to_sect (ns->head,le64_to_cpu (entry->zcap)); + zone.start =nvme_lba_to_sect (ns->head,le64_to_cpu (entry->zslba));if (zone.cond ==BLK_ZONE_COND_FULL ) zone.wp = zone.start + zone.len;else - zone.wp =nvme_lba_to_sect (ns,le64_to_cpu (entry->wp)); + zone.wp =nvme_lba_to_sect (ns->head,le64_to_cpu (entry->wp));return cb (&zone, idx, data); }
@@ -201,7 +201,7 @@ int nvme_ns_report_zones(struct nvme_ns *ns, sector_t sector,while (zone_idx < nr_zones && sector <get_capacity (ns->disk)) {memset (report,0 , buflen); - c.zmr.slba =cpu_to_le64 (nvme_sect_to_lba (ns, sector)); + c.zmr.slba =cpu_to_le64 (nvme_sect_to_lba (ns->head, sector)); ret =nvme_submit_sync_cmd (ns->queue, &c, report, buflen);if (ret) {if (ret >0 )
@@ -240,7 +240,7 @@ blk_status_t nvme_setup_zone_mgmt_send(struct nvme_ns *ns, struct request *req, c->zms.opcode = nvme_cmd_zone_mgmt_send; c->zms.nsid =cpu_to_le32 (ns->head->ns_id); - c->zms.slba =cpu_to_le64 (nvme_sect_to_lba (ns,blk_rq_pos (req))); + c->zms.slba =cpu_to_le64 (nvme_sect_to_lba (ns->head,blk_rq_pos (req))); c->zms.zsa = action;if (req_op (req) ==REQ_OP_ZONE_RESET_ALL )
Signature · 1 line
2.43.0
Re: [PATCH v6 2/6] nvme: refactor ns info helpers
Quoted 7 lines · Daniel Wagner
Pass in the
nvme_ns_headpointer directly. This reduces the necessity on the caller side have thenvme_nsdata structure present. Thus we can refactor the caller side in the next step as well.Reviewed-by Christoph Hellwig <[email protected]>Reviewed-by Sagi Grimberg <[email protected]>Signed-off-by Daniel Wagner <[email protected]>
Quoted 7 lines
--- drivers/nvme/host/core.c | 36 +++++++++++++++++++++--------------- drivers/nvme/host/nvme.h | 12 ++++++------ drivers/nvme/host/rdma.c | 2 +- drivers/nvme/host/zns.c | 12 ++++++------ 4 files changed, 34 insertions(+), 28 deletions(-)
Cheers,
Hannes
Signature · 5 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald, Werner Knoblich
PATCH v6 3/6 nvme: refactor ns info setup function
In reply to Daniel WagnerUse nvme_ns_head instead of nvme_ns where possible. This reduces the coupling between the different data structures.
--- drivers/nvme/host/core.c | 106 +++++++++++++++++++-------------------- drivers/nvme/host/zns.c | 16 +++--- 2 files changed, 62 insertions(+), 60 deletions(-)
drivers/nvme/host/core.c +53 -53
@@ -1665,14 +1665,14 @@ int nvme_getgeo(struct block_device *bdev, struct hd_geometry *geo) }#ifdef CONFIG_BLK_DEV_INTEGRITY -static void nvme_init_integrity (struct gendisk *disk,struct nvme_ns *ns, -u32 max_integrity_segments) +static void nvme_init_integrity (struct gendisk *disk, +struct nvme_ns_head *head,u32 max_integrity_segments) {struct blk_integrity integrity = { }; -switch (ns->head->pi_type) { +switch (head->pi_type) {case NVME_NS_DPS_PI_TYPE3 : -switch (ns->head->guard_type) { +switch (head->guard_type) {case NVME_NVM_NS_16B_GUARD : integrity.profile = &t10_pi_type3_crc; integrity.tag_size =sizeof (u16) +sizeof (u32);
@@ -1690,7 +1690,7 @@ static void nvme_init_integrity(struct gendisk *disk, struct nvme_ns *ns,break ;case NVME_NS_DPS_PI_TYPE1 :case NVME_NS_DPS_PI_TYPE2 : -switch (ns->head->guard_type) { +switch (head->guard_type) {case NVME_NVM_NS_16B_GUARD : integrity.profile = &t10_pi_type1_crc; integrity.tag_size =sizeof (u16);
@@ -1711,26 +1711,26 @@ static void nvme_init_integrity(struct gendisk *disk, struct nvme_ns *ns, break; } - integrity.tuple_size = ns->head->ms; + integrity.tuple_size = head->ms;blk_integrity_register (disk, &integrity);blk_queue_max_integrity_segments (disk->queue, max_integrity_segments); }#else -static void nvme_init_integrity (struct gendisk *disk,struct nvme_ns *ns, -u32 max_integrity_segments) +static void nvme_init_integrity (struct gendisk *disk, +struct nvme_ns_head *head,u32 max_integrity_segments) { }#endif /* CONFIG_BLK_DEV_INTEGRITY */ -static void nvme_config_discard (struct gendisk *disk,struct nvme_ns *ns) +static void nvme_config_discard (struct nvme_ctrl *ctrl,struct gendisk *disk, +struct nvme_ns_head *head) { -struct nvme_ctrl *ctrl = ns->ctrl;struct request_queue *queue = disk->queue;u32 size =queue_logical_block_size (queue); -if (ctrl->dmrsl && ctrl->dmrsl <=nvme_sect_to_lba (ns->head,UINT_MAX )) +if (ctrl->dmrsl && ctrl->dmrsl <=nvme_sect_to_lba (head,UINT_MAX )) ctrl->max_discard_sectors = -nvme_lba_to_sect (ns->head, ctrl->dmrsl); +nvme_lba_to_sect (head, ctrl->dmrsl);if (ctrl->max_discard_sectors ==0 ) {blk_queue_max_discard_sectors (queue,0 );
@@ -1761,21 +1761,21 @@ static bool nvme_ns_ids_equal(struct nvme_ns_ids *a, struct nvme_ns_ids *b) a->csi == b->csi; } -static int nvme_init_ms (struct nvme_ns *ns,struct nvme_id_ns *id) +static int nvme_init_ms (struct nvme_ctrl *ctrl,struct nvme_ns_head *head, +struct nvme_id_ns *id) {bool first = id->dps &NVME_NS_DPS_PI_FIRST ;unsigned lbaf =nvme_lbaf_index (id->flbas); -struct nvme_ctrl *ctrl = ns->ctrl;struct nvme_command c = { };struct nvme_id_ns_nvm *nvm;int ret =0 ;u32 elbaf; - ns->head->pi_size =0 ; - ns->head->ms =le16_to_cpu (id->lbaf[lbaf].ms); + head->pi_size =0 ; + head->ms =le16_to_cpu (id->lbaf[lbaf].ms);if (!(ctrl->ctratt &NVME_CTRL_ATTR_ELBAS )) { - ns->head->pi_size =sizeof (struct t10_pi_tuple ); - ns->head->guard_type =NVME_NVM_NS_16B_GUARD ; + head->pi_size =sizeof (struct t10_pi_tuple ); + head->guard_type =NVME_NVM_NS_16B_GUARD ; goto set_pi; }
@@ -1784,11 +1784,11 @@ static int nvme_init_ms(struct nvme_ns *ns, struct nvme_id_ns *id)return -ENOMEM ; c.identify.opcode = nvme_admin_identify; - c.identify.nsid =cpu_to_le32 (ns->head->ns_id); + c.identify.nsid =cpu_to_le32 (head->ns_id); c.identify.cns =NVME_ID_CNS_CS_NS ; c.identify.csi =NVME_CSI_NVM ; - ret =nvme_submit_sync_cmd (ns->ctrl->admin_q, &c, nvm,sizeof (*nvm)); + ret =nvme_submit_sync_cmd (ctrl->admin_q, &c, nvm,sizeof (*nvm));if (ret) goto free_data;
@@ -1798,13 +1798,13 @@ static int nvme_init_ms(struct nvme_ns *ns, struct nvme_id_ns *id)if (nvme_elbaf_sts (elbaf)) goto free_data; - ns->head->guard_type =nvme_elbaf_guard_type (elbaf); -switch (ns->head->guard_type) { + head->guard_type =nvme_elbaf_guard_type (elbaf); +switch (head->guard_type) {case NVME_NVM_NS_64B_GUARD : - ns->head->pi_size =sizeof (struct crc64_pi_tuple ); + head->pi_size =sizeof (struct crc64_pi_tuple );break ;case NVME_NVM_NS_16B_GUARD : - ns->head->pi_size =sizeof (struct t10_pi_tuple ); + head->pi_size =sizeof (struct t10_pi_tuple );break ;default :break ;
@@ -1813,25 +1813,25 @@ static int nvme_init_ms(struct nvme_ns *ns, struct nvme_id_ns *id) free_data:kfree (nvm); set_pi: -if (ns->head->pi_size && (first || ns->head->ms == ns->head->pi_size)) - ns->head->pi_type = id->dps &NVME_NS_DPS_PI_MASK ; +if (head->pi_size && (first || head->ms == head->pi_size)) + head->pi_type = id->dps &NVME_NS_DPS_PI_MASK ;else - ns->head->pi_type =0 ; + head->pi_type =0 ;return ret; } -static int nvme_configure_metadata (struct nvme_ns *ns,struct nvme_id_ns *id) +static int nvme_configure_metadata (struct nvme_ctrl *ctrl, +struct nvme_ns_head *head,struct nvme_id_ns *id) { -struct nvme_ctrl *ctrl = ns->ctrl;int ret; - ret =nvme_init_ms (ns, id); + ret =nvme_init_ms (ctrl, head, id);if (ret)return ret; - ns->head->features &= ~(NVME_NS_METADATA_SUPPORTED |NVME_NS_EXT_LBAS ); -if (!ns->head->ms || !(ctrl->ops->flags &NVME_F_METADATA_SUPPORTED )) + head->features &= ~(NVME_NS_METADATA_SUPPORTED |NVME_NS_EXT_LBAS ); +if (!head->ms || !(ctrl->ops->flags &NVME_F_METADATA_SUPPORTED ))return 0 ;if (ctrl->ops->flags &NVME_F_FABRICS ) {
@@ -1843,7 +1843,7 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)if (WARN_ON_ONCE (!(id->flbas &NVME_NS_FLBAS_META_EXT )))return 0 ; - ns->head->features |=NVME_NS_EXT_LBAS ; + head->features |=NVME_NS_EXT_LBAS ;/* * The current fabrics transport drivers support namespace
@@ -1854,8 +1854,8 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)* Note, this check will need to be modified if any drivers * gain the ability to use other metadata formats. */ -if (ctrl->max_integrity_segments &&nvme_ns_has_pi (ns->head)) - ns->head->features |=NVME_NS_METADATA_SUPPORTED ; +if (ctrl->max_integrity_segments &&nvme_ns_has_pi (head)) + head->features |=NVME_NS_METADATA_SUPPORTED ; }else {/* * For PCIe controllers, we can't easily remap the separate
@@ -1864,9 +1864,9 @@ static int nvme_configure_metadata(struct nvme_ns *ns, struct nvme_id_ns *id)* We allow extended LBAs for the passthrough interface, though. */ if (id->flbas &NVME_NS_FLBAS_META_EXT ) - ns->head->features |=NVME_NS_EXT_LBAS ; + head->features |=NVME_NS_EXT_LBAS ;else - ns->head->features |=NVME_NS_METADATA_SUPPORTED ; + head->features |=NVME_NS_METADATA_SUPPORTED ; }return 0 ; }
@@ -1889,18 +1889,18 @@ static void nvme_set_queue_limits(struct nvme_ctrl *ctrl,blk_queue_write_cache (q, vwc, vwc); } -static void nvme_update_disk_info (struct gendisk *disk, -struct nvme_ns *ns,struct nvme_id_ns *id) +static void nvme_update_disk_info (struct nvme_ctrl *ctrl,struct gendisk *disk, +struct nvme_ns_head *head,struct nvme_id_ns *id) { -sector_t capacity =nvme_lba_to_sect (ns->head,le64_to_cpu (id->nsze)); -u32 bs =1U << ns->head->lba_shift; +sector_t capacity =nvme_lba_to_sect (head,le64_to_cpu (id->nsze)); +u32 bs =1U << head->lba_shift;u32 atomic_bs, phys_bs, io_opt =0 ;/* * The block layer can't support LBA sizes larger than the page size * yet, so catch this early and don't allow block I/O. */ -if (ns->head->lba_shift >PAGE_SHIFT ) { +if (head->lba_shift >PAGE_SHIFT ) { capacity =0 ; bs = (1 <<9 ); }
@@ -1917,7 +1917,7 @@ static void nvme_update_disk_info(struct gendisk *disk,if (id->nsfeat &NVME_NS_FEAT_ATOMICS && id->nawupf) atomic_bs = (1 +le16_to_cpu (id->nawupf)) * bs;else - atomic_bs = (1 + ns->ctrl->subsys->awupf) * bs; + atomic_bs = (1 + ctrl->subsys->awupf) * bs; }if (id->nsfeat &NVME_NS_FEAT_IO_OPT ) {
@@ -1943,20 +1943,20 @@ static void nvme_update_disk_info(struct gendisk *disk,* I/O to namespaces with metadata except when the namespace supports * PI, as it can strip/insert in that case. */ -if (ns->head->ms) { +if (head->ms) {if (IS_ENABLED (CONFIG_BLK_DEV_INTEGRITY ) && - (ns->head->features &NVME_NS_METADATA_SUPPORTED )) -nvme_init_integrity (disk, ns, - ns->ctrl->max_integrity_segments); -else if (!nvme_ns_has_pi (ns->head)) + (head->features &NVME_NS_METADATA_SUPPORTED )) +nvme_init_integrity (disk, head, + ctrl->max_integrity_segments); +else if (!nvme_ns_has_pi (head)) capacity =0 ; }set_capacity_and_notify (disk, capacity); -nvme_config_discard (disk, ns); +nvme_config_discard (ctrl, disk, head);blk_queue_max_write_zeroes_sectors (disk->queue, - ns->ctrl->max_zeroes_sectors); + ctrl->max_zeroes_sectors); }static bool nvme_ns_is_readonly (struct nvme_ns *ns,struct nvme_ns_info *info)
@@ -2042,13 +2042,13 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns, ns->head->lba_shift = id->lbaf[lbaf].ds;nvme_set_queue_limits (ns->ctrl, ns->queue); - ret =nvme_configure_metadata (ns, id); + ret =nvme_configure_metadata (ns->ctrl, ns->head, id);if (ret <0 ) {blk_mq_unfreeze_queue (ns->disk->queue); goto out; }nvme_set_chunk_sectors (ns, id); -nvme_update_disk_info (ns->disk, ns, id); +nvme_update_disk_info (ns->ctrl, ns->disk, ns->head, id);if (ns->head->ids.csi ==NVME_CSI_ZNS ) { ret =nvme_update_zone_info (ns, lbaf);
@@ -2078,7 +2078,7 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,if (nvme_ns_head_multipath (ns->head)) {blk_mq_freeze_queue (ns->head->disk->queue); -nvme_update_disk_info (ns->head->disk, ns, id); +nvme_update_disk_info (ns->ctrl, ns->head->disk, ns->head, id);set_disk_ro (ns->head->disk,nvme_ns_is_readonly (ns, info));nvme_mpath_revalidate_paths (ns); blk_stack_limits(&ns->head->disk->queue->limits,
drivers/nvme/host/zns.c +9 -7
@@ -148,7 +148,8 @@ static void *nvme_zns_alloc_report_buffer(struct nvme_ns *ns,return NULL ; } -static int nvme_zone_parse_entry (struct nvme_ns *ns, +static int nvme_zone_parse_entry (struct nvme_ctrl *ctrl, +struct nvme_ns_head *head,struct nvme_zone_descriptor *entry,unsigned int idx,report_zones_cb cb,void *data)
@@ -156,20 +157,20 @@ static int nvme_zone_parse_entry(struct nvme_ns *ns,struct blk_zone zone = { };if ((entry->zt &0xf ) !=NVME_ZONE_TYPE_SEQWRITE_REQ ) { -dev_err (ns->ctrl->device,"invalid zone type %#x\n" , +dev_err (ctrl->device,"invalid zone type %#x\n" , entry->zt);return -EINVAL ; } zone.type =BLK_ZONE_TYPE_SEQWRITE_REQ ; zone.cond = entry->zs >>4 ; - zone.len = ns->head->zsze; - zone.capacity =nvme_lba_to_sect (ns->head,le64_to_cpu (entry->zcap)); - zone.start =nvme_lba_to_sect (ns->head,le64_to_cpu (entry->zslba)); + zone.len = head->zsze; + zone.capacity =nvme_lba_to_sect (head,le64_to_cpu (entry->zcap)); + zone.start =nvme_lba_to_sect (head,le64_to_cpu (entry->zslba));if (zone.cond ==BLK_ZONE_COND_FULL ) zone.wp = zone.start + zone.len;else - zone.wp =nvme_lba_to_sect (ns->head,le64_to_cpu (entry->wp)); + zone.wp =nvme_lba_to_sect (head,le64_to_cpu (entry->wp));return cb (&zone, idx, data); }
@@ -214,7 +215,8 @@ int nvme_ns_report_zones(struct nvme_ns *ns, sector_t sector,break ;for (i =0 ; i < nz && zone_idx < nr_zones; i++) { - ret =nvme_zone_parse_entry (ns, &report->entries[i], + ret =nvme_zone_parse_entry (ns->ctrl, ns->head, + &report->entries[i], zone_idx, cb, data);if (ret) goto out_free;
Signature · 1 line
2.43.0
Re: [PATCH v6 3/6] nvme: refactor ns info setup function
Quoted 11 lines · Daniel Wagner
Use
nvme_ns_headinstead ofnvme_nswhere possible. This reduces the coupling between the different data structures.Reviewed-by Christoph Hellwig <[email protected]>Reviewed-by Sagi Grimberg <[email protected]>Signed-off-by Daniel Wagner <[email protected]>--- drivers/nvme/host/core.c | 106 +++++++++++++++++++-------------------- drivers/nvme/host/zns.c | 16 +++--- 2 files changed, 62 insertions(+), 60 deletions(-)
Cheers
Hannes
Signature · 5 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald, Werner Knoblich
PATCH v6 4/6 nvme: rename ns attribute group
In reply to Daniel WagnerDrop the 'id' part of the attribute group name because we want to expose non 'id' related attributes via the ns attribute group.
--- drivers/nvme/host/core.c | 2 +- drivers/nvme/host/multipath.c | 2 +- drivers/nvme/host/nvme.h | 2 +- drivers/nvme/host/sysfs.c | 14 +++++++------- 4 files changed, 10 insertions(+), 10 deletions(-)
drivers/nvme/host/core.c +1 -1
@@ -3671,7 +3671,7 @@ static void nvme_alloc_ns(struct nvme_ctrl *ctrl, struct nvme_ns_info *info)up_write (&ctrl->namespaces_rwsem);nvme_get_ctrl (ctrl); -if (device_add_disk (ctrl->device, ns->disk, nvme_ns_id_attr_groups)) +if (device_add_disk (ctrl->device, ns->disk, nvme_ns_attr_groups)) goto out_cleanup_ns_from_list;if (!nvme_ns_head_multipath (ns->head))
drivers/nvme/host/multipath.c +1 -1
@@ -579,7 +579,7 @@ static void nvme_mpath_set_live(struct nvme_ns *ns)*/ if (!test_and_set_bit (NVME_NSHEAD_DISK_LIVE , &head->flags)) { rc =device_add_disk (&head->subsys->dev, head->disk, - nvme_ns_id_attr_groups); + nvme_ns_attr_groups);if (rc) {clear_bit (NVME_NSHEAD_DISK_LIVE , &ns->flags);return ;
drivers/nvme/host/nvme.h +1 -1
@@ -865,7 +865,7 @@ int nvme_ns_head_chr_uring_cmd(struct io_uring_cmd *ioucmd,int nvme_getgeo (struct block_device *bdev,struct hd_geometry *geo);int nvme_dev_uring_cmd (struct io_uring_cmd *ioucmd,unsigned int issue_flags); -extern const struct attribute_group *nvme_ns_id_attr_groups[]; +extern const struct attribute_group *nvme_ns_attr_groups[];extern const struct pr_ops nvme_pr_ops;extern const struct block_device_operations nvme_ns_head_ops;extern const struct attribute_group nvme_dev_attrs_group;
drivers/nvme/host/sysfs.c +7 -7
@@ -114,7 +114,7 @@ static ssize_t nsid_show(struct device *dev, struct device_attribute *attr, }static DEVICE_ATTR_RO (nsid); -static struct attribute *nvme_ns_id_attrs[] = { +static struct attribute *nvme_ns_attrs[] = { &dev_attr_wwid.attr, &dev_attr_uuid.attr, &dev_attr_nguid.attr,
@@ -127,7 +127,7 @@ static struct attribute *nvme_ns_id_attrs[] = { NULL, }; -static umode_t nvme_ns_id_attrs_are_visible (struct kobject *kobj, +static umode_t nvme_ns_attrs_are_visible (struct kobject *kobj,struct attribute *a,int n) { struct device *dev = container_of(kobj, struct device, kobj);
@@ -157,13 +157,13 @@ static umode_t nvme_ns_id_attrs_are_visible(struct kobject *kobj,return a->mode; } -static const struct attribute_group nvme_ns_id_attr_group = { - .attrs = nvme_ns_id_attrs, - .is_visible = nvme_ns_id_attrs_are_visible, +static const struct attribute_group nvme_ns_attr_group = { + .attrs = nvme_ns_attrs, + .is_visible = nvme_ns_attrs_are_visible, }; -const struct attribute_group *nvme_ns_id_attr_groups[] = { - &nvme_ns_id_attr_group, +const struct attribute_group *nvme_ns_attr_groups[] = { + &nvme_ns_attr_group,NULL , };
Signature · 1 line
2.43.0
Re: [PATCH v6 4/6] nvme: rename ns attribute group
Quoted 6 lines · Daniel Wagner
Drop the 'id' part of the attribute group name because we want to expose non 'id' related attributes via the ns attribute group.
Reviewed-by Christoph Hellwig <[email protected]>Reviewed-by Sagi Grimberg <[email protected]>Signed-off-by Daniel Wagner <[email protected]>
Quoted 7 lines
--- drivers/nvme/host/core.c | 2 +- drivers/nvme/host/multipath.c | 2 +- drivers/nvme/host/nvme.h | 2 +- drivers/nvme/host/sysfs.c | 14 +++++++------- 4 files changed, 10 insertions(+), 10 deletions(-)
Cheers,
Hannes
Signature · 5 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald, Werner Knoblich
PATCH v6 5/6 nvme: add csi, ms and nuse to sysfs
In reply to Daniel Wagnerlibnvme is using the sysfs for enumarating the nvme resources. Though there are few missing attritbutes in the sysfs. For these libnvme issues commands during discovering.
As the kernel already knows all these attributes and we would like to avoid libnvme to issue commands all the time, expose these missing attributes.
The nuse value is updated on request because the nuse is a volatile value. Since any user can read the sysfs attribute, a very simple rate limit is added (update once every 5 seconds). A more sophisticated update strategy can be added later if there is actually a need for it.
--- drivers/nvme/host/core.c | 6 ++- drivers/nvme/host/nvme.h | 6 +++ drivers/nvme/host/sysfs.c | 85 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 96 insertions(+), 1 deletion(-)
drivers/nvme/host/core.c +5 -1
@@ -20,6 +20,7 @@#include <linux/ptrace.h> #include <linux/nvme_ioctl.h> #include <linux/pm_qos.h> +#include <linux/ratelimit.h> #include <asm/unaligned.h> #include "nvme.h"
@@ -1449,7 +1450,7 @@ static int nvme_identify_ns_descs(struct nvme_ctrl *ctrl,return status; } -static int nvme_identify_ns (struct nvme_ctrl *ctrl,unsigned nsid, +int nvme_identify_ns (struct nvme_ctrl *ctrl,unsigned nsid,struct nvme_id_ns **id) {struct nvme_command c = { };
@@ -2040,6 +2041,7 @@ static int nvme_update_ns_info_block(struct nvme_ns *ns,blk_mq_freeze_queue (ns->disk->queue); lbaf =nvme_lbaf_index (id->flbas); ns->head->lba_shift = id->lbaf[lbaf].ds; + ns->head->nuse =le64_to_cpu (id->nuse);nvme_set_queue_limits (ns->ctrl, ns->queue); ret =nvme_configure_metadata (ns->ctrl, ns->head, id);
@@ -3420,6 +3422,8 @@ static struct nvme_ns_head *nvme_alloc_ns_head(struct nvme_ctrl *ctrl, head->ns_id = info->nsid; head->ids = info->ids; head->shared = info->is_shared; +ratelimit_state_init (&head->rs_nuse,5 *HZ ,1 ); +ratelimit_set_flags (&head->rs_nuse,RATELIMIT_MSG_ON_RELEASE );kref_init (&head->ref);if (head->ids.csi) {
drivers/nvme/host/nvme.h +6 -0
@@ -16,6 +16,7 @@#include <linux/rcupdate.h> #include <linux/wait.h> #include <linux/t10-pi.h> +#include <linux/ratelimit_types.h> #include <trace/events/block.h>
@@ -451,6 +452,7 @@ struct nvme_ns_head {u16 pi_size;u16 sgs;u32 sws; +u64 nuse;u8 pi_type;u8 guard_type;#ifdef CONFIG_BLK_DEV_ZONED
@@ -458,6 +460,8 @@ struct nvme_ns_head {#endif unsigned long features; +struct ratelimit_state rs_nuse; +struct cdev cdev;struct device cdev_device;
@@ -862,6 +866,8 @@ int nvme_ns_chr_uring_cmd(struct io_uring_cmd *ioucmd, unsigned int issue_flags);int nvme_ns_head_chr_uring_cmd (struct io_uring_cmd *ioucmd,unsigned int issue_flags); +int nvme_identify_ns (struct nvme_ctrl *ctrl,unsigned nsid, +struct nvme_id_ns **id);int nvme_getgeo (struct block_device *bdev,struct hd_geometry *geo);int nvme_dev_uring_cmd (struct io_uring_cmd *ioucmd,unsigned int issue_flags);
drivers/nvme/host/sysfs.c +85 -0
97 lines not on this page · raw message
Signature · 1 line
2.43.0
Quoted 13 lines · Daniel Wagner
libnvme is using the sysfs for enumarating the nvme resources. Though there are few missing attritbutes in the sysfs. For these libnvme issues commands during discovering.
As the kernel already knows all these attributes and we would like to avoid libnvme to issue commands all the time, expose these missing attributes.
The nuse value is updated on request because the nuse is a volatile value. Since any user can read the sysfs attribute, a very simple rate limit is added (update once every 5 seconds). A more sophisticated update strategy can be added later if there is actually a need for it.
Quoted 9 lines
Reviewed-by Christoph Hellwig <[email protected]>Reviewed-by Sagi Grimberg <[email protected]>Signed-off-by Daniel Wagner <[email protected]>--- drivers/nvme/host/core.c | 6 ++- drivers/nvme/host/nvme.h | 6 +++ drivers/nvme/host/sysfs.c | 85 +++++++++++++++++++++++++++++++++++++++ 3 files changed, 96 insertions(+), 1 deletion(-)
Cheers
Hannes
Signature · 5 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald, Werner Knoblich
PATCH v6 6/6 nvme: repack struct nvme_ns_head
In reply to Daniel Wagnerns_id, lba_shift and ms are always accessed for every read/write I/O in nvme_setup_rw. By grouping these variables into one cacheline we can safe some cycles.
4k sequential reads:
baseline patched Bandwidth: 1620 1634 IOPs 66345579 66910939
--- drivers/nvme/host/nvme.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/nvme/host/nvme.h +4 -4
@@ -440,21 +440,21 @@ struct nvme_ns_head {struct list_head list;struct srcu_struct srcu;struct nvme_subsystem *subsys; -unsigned ns_id;struct nvme_ns_ids ids;struct list_head entry;struct kref ref;bool shared;int instance;struct nvme_effects_log *effects; +u64 nuse; +unsigned ns_id;int lba_shift;u16 ms;u16 pi_size; -u16 sgs; -u32 sws; -u64 nuse;u8 pi_type;u8 guard_type; +u16 sgs; +u32 sws;#ifdef CONFIG_BLK_DEV_ZONED u64 zsze;#endif
Signature · 1 line
2.43.0
Re: [PATCH v6 6/6] nvme: repack struct nvme_ns_head
Quoted 10 lines · Daniel Wagner
ns_id,lba_shiftand ms are always accessed for every read/write I/O innvme_setup_rw. By grouping these variables into one cacheline we can safe some cycles.4k sequential reads:
baseline patched
Bandwidth: 1620 1634
IOPs 66345579 66910939
Quoted 7 lines
Reviewed-by Christoph Hellwig <[email protected]>Reviewed-by Sagi Grimberg <[email protected]>Signed-off-by Daniel Wagner <[email protected]>--- drivers/nvme/host/nvme.h | 8 ++++---- 1 file changed, 4 insertions(+), 4 deletions(-)
Cheers,
Hannes
Signature · 5 lines
Dr. Hannes Reinecke Kernel Storage Architect [email protected] +49 911 74053 688 SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg HRB 36809 (AG Nürnberg), GF: Ivo Totev, Andrew McDonald, Werner Knoblich
Keith Busch
maintainerQuoted 2 lines · Daniel Wagner
Rebased on the current nvme/nvme-6.8 branch and added the Review tags. Also addressed the printk format issue pointed out by Chaitanya.
Thanks, added to nvme-6.8.
Hi Keith,
Quoted 5 lines · Keith Busch
Quoted 2 lines · Daniel Wagner
Rebased on the current nvme/nvme-6.8 branch and added the Review tags. Also addressed the printk format issue pointed out by Chaitanya.
Thanks, added to nvme-6.8.
I'm rebasing my fixes to the PI bug (which are not trivial) on top of nvme-6.8 and Daniel's series. I've found that this branch is missing some fixes from the master that I've added manually to local nvme-6.8 branch. When do you plan to rebase the nvme-6.8 on top of master ?
It will be helpful, otherwise my patches will not apply cleanly on your branch. :)
-Max.
Keith Busch
maintainerQuoted 9 lines · Max Gurtovoy
Hi Keith,
Quoted 5 lines · Keith Busch
Quoted 2 lines · Daniel Wagner
Rebased on the current nvme/nvme-6.8 branch and added the Review tags. Also addressed the printk format issue pointed out by Chaitanya.
Thanks, added to nvme-6.8.
Quoted 7 lines
I'm rebasing my fixes to the PI bug (which are not trivial) on top of nvme-6.8 and Daniel's series.
I've found that this branch is missing some fixes from the master that I've added manually to local nvme-6.8 branch.
When do you plan to rebase the nvme-6.8 on top of master ?
It will be helpful, otherwise my patches will not apply cleanly on your branch. :)
linux-nvme pulls go through Jens' linux-block repo, so we have to be based on the appropriate branch from there. We already have some late fixes in 6.7 that create minor merge conflicts for our existing 6.8 branch, so we may provide a reference resolution branch with the pull request if conflicts are non-trivial. That might be an option for your fixes if you can make them work atop existing nvme-6.8 without the conflicting 6.7 changes.
Sometime after the next merge window opens, Jens will rebase his 6.8 block tree to Linus', and we usually get a 2nd feature pull request within the merge window from that new baseline. That might be an option if you want to wait a couple more weeks.
Quoted 26 lines · Keith Busch
Quoted 9 lines · Max Gurtovoy
Hi Keith,
Quoted 5 lines · Keith Busch
Quoted 2 lines · Daniel Wagner
Rebased on the current nvme/nvme-6.8 branch and added the Review tags. Also addressed the printk format issue pointed out by Chaitanya.
Thanks, added to nvme-6.8.
Quoted 7 lines
I'm rebasing my fixes to the PI bug (which are not trivial) on top of nvme-6.8 and Daniel's series.
I've found that this branch is missing some fixes from the master that I've added manually to local nvme-6.8 branch.
When do you plan to rebase the nvme-6.8 on top of master ?
It will be helpful, otherwise my patches will not apply cleanly on your branch. :)linux-nvme pulls go through Jens' linux-block repo, so we have to be based on the appropriate branch from there. We already have some late fixes in 6.7 that create minor merge conflicts for our existing 6.8 branch, so we may provide a reference resolution branch with the pull request if conflicts are non-trivial. That might be an option for your fixes if you can make them work atop existing nvme-6.8 without the conflicting 6.7 changes.
Quoted 4 lines
Sometime after the next merge window opens, Jens will rebase his 6.8 block tree to Linus', and we usually get a 2nd feature pull request within the merge window from that new baseline. That might be an option if you want to wait a couple more weeks.
I see. I think that after finishing the latest testing I'll send the version on top of nvme-6.8 + my local branch patches I took from 6.7. We will start the reviewing process and probably will converge in a week or two. Then we will see what is the status of the linux-block repo and decide how to continue.