nvmet: bump NVMET_NR_QUEUES to 128
Subsystems (2)
PATCH 1/1 nvmet: bump NVMET_NR_QUEUES to 128
Raise the max number of IO queues to 128. There are several hosts with more than 64 cpus/threads.
--- drivers/nvme/target/nvmet.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/nvme/target/nvmet.h +1 -1
@@ -314,7 +314,7 @@ u16 nvmet_copy_from_sgl(struct nvmet_req *req, off_t off, void *buf,u32 nvmet_get_log_page_len (struct nvme_command *cmd);#define NVMET_QUEUE_SIZE 1024 -#define NVMET_NR_QUEUES 64 +#define NVMET_NR_QUEUES 128#define NVMET_MAX_CMD NVMET_QUEUE_SIZE#define NVMET_KAS 10#define NVMET_DISC_KATO 120
Signature · 1 line
2.13.1
PATCH 1/1 nvmet: bump NVMET_NR_QUEUES to 128
Quoted 8 lines · James Smart
Raise the max number of IO queues to 128. There are several hosts with more than 64 cpus/threads.
Signed-off-by James Smart <james.smart at broadcom.com>--- drivers/nvme/target/nvmet.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
I guess it's about time as more and more systems use > 64 cpus,
PATCH 1/1 nvmet: bump NVMET_NR_QUEUES to 128
In reply to James SmartThis looks ok as a quick check, but any chance you could send a patch to make this configurable through configfs? Preferably also for the max queue size.
PATCH 1/1 nvmet: bump NVMET_NR_QUEUES to 128
Quoted 3 lines · Christoph Hellwig
This looks ok as a quick check, but any chance you could send a patch to make this configurable through configfs? Preferably also for the max queue size.
Yep - I'll put one together
-- james
PATCH 1/1 nvmet: bump NVMET_NR_QUEUES to 128
Quoted 6 lines · James Smart
Quoted 3 lines · Christoph Hellwig
This looks ok as a quick check, but any chance you could send a patch to make this configurable through configfs? Preferably also for the max queue size.
Yep - I'll put one together
Did you get a chance to look at that?
Christoph Hellwig
AppliedPATCH 1/1 nvmet: bump NVMET_NR_QUEUES to 128
In reply to James SmartThanks, applied to nvme-4.15.