Skip to contentSkip to conversation

nvme: fixes for command retry

PATCH Superseded Minwoo Im 4 participants 19 messages Jan 8, 2021 – Jan 15, 2021 NVMe

v1 of 3 Superseded Latest is v3 →

lore
Share & tools
lore

Minwoo Im

author
#0 lore raw

cover nvme: fixes for command retry

Hello,

This patchset fixes command retry feature in NVMe core driver. The first one supports command retry delay for admin commands also. Earlier, we only supported I/O commands and this patch made it support for admin commands too. The second one fixed checks for whether to enable Advanced Command Retry Enable(ACRE) or not by checking all Command Retry Delay Time(CRDT)s from only a single one CRDT[0]. The last one added a module parameter 'command_retry' to enable the command retry feature in this driver.

Please review.

Thanks,

Minwoo Im (3):
nvme: support command retry delay for admin command
nvme: check all retry delay times in Identify Controller
nvme: add parameter command_retry to enable retry

 drivers/nvme/host/core.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)
Signature · 7 lines
2.17.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
author
#1 lore raw

PATCH 1/3 nvme: support command retry delay for admin command

Once driver decides to retry a command that failed, it can have some delay based on Command Retry Delay(CRD) in Completion Queue Entry. If CRD indicates Command Retry Delay Time 1, 2, or 3, driver calcualtes delay in Identify Controller data structure.

This feature can be applied not only I/O queue commands, but also admin queue commands. Admin command during the initialization may fail due to some un-ready state of a controller, so it would be better to have this feature for admin commands also.

This patch simply removed getting namespace instance(struct nvme_ns) from the struct request and get a controller instance(struct nvme_ctrl) by nvme_req() helper. Now, nvme_retry_req() calculates the delay regardless to namespace instance which meant this request is for I/O command.

---
 drivers/nvme/host/core.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)
drivers/nvme/host/core.c +2 -3
index ce1b61519441..d0c99a9400ea 100644
@@ -280,14 +280,13 @@ static blk_status_t nvme_error_status(u16 status)
 
 static void nvme_retry_req(struct request *req)
 {
-	struct nvme_ns *ns = req->q->queuedata;
 	unsigned long delay = 0;
 	u16 crd;
 
 	/* The mask and shift result must be <= 3 */
 	crd = (nvme_req(req)->status & NVME_SC_CRD) >> 11;
-	if (ns && crd)
-		delay = ns->ctrl->crdt[crd - 1] * 100;
+	if (crd)
+		delay = nvme_req(req)->ctrl->crdt[crd - 1] * 100;
 
 	nvme_req(req)->retries++;
 	blk_mq_requeue_request(req, false);
Signature · 7 lines
2.17.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
#2 lore raw

Re: [PATCH 1/3] nvme: support command retry delay for admin command

Quoted 4 lines · Minwoo Im

This feature can be applied not only I/O queue commands, but also admin queue commands. Admin command during the initialization may fail due to some un-ready state of a controller, so it would be better to have this feature for admin commands also.

What is some un-ready state ?

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#3 lore raw

Re: [PATCH 1/3] nvme: support command retry delay for admin command

Hello,

Quoted 6 lines · Chaitanya Kulkarni
Quoted 4 lines · Minwoo Im

This feature can be applied not only I/O queue commands, but also admin queue commands. Admin command during the initialization may fail due to some un-ready state of a controller, so it would be better to have this feature for admin commands also.

What is some un-ready state ?

State that controller is unable to process successfully command as "Command Interrupt" status code represents.

Thanks,

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

#4 lore raw

Re: [PATCH 1/3] nvme: support command retry delay for admin command

Quoted 4 lines · Minwoo Im

Hello,

Quoted 1 line · Chaitanya Kulkarni

On 1/8/21 06:54, Minwoo Im wrote:

Quoted 8 lines
Quoted 5 lines
Quoted 4 lines

This feature can be applied not only I/O queue commands, but also admin queue commands. Admin command during the initialization may fail due to some un-ready state of a controller, so it would be better to have this feature for admin commands also.

What is some un-ready state ?

State that controller is unable to process successfully command as "Command Interrupt" status code represents.

Given that controller ready has a specific meaning in NVMe I think a different wording here would be useful.

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

#7 lore raw

Re: [PATCH 1/3] nvme: support command retry delay for admin command

In reply to Christoph Hellwig
Quoted 2 lines · Christoph Hellwig

Except for the little wording issue this looks good to me. I'll pick it up for nvme-5.11 and will fix up the commit log.

Given that we never retry admin command right now I've moved this off nvme-5.11 into the pending queue.

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#8 lore raw

Re: [PATCH 1/3] nvme: support command retry delay for admin command

Quoted 6 lines · Christoph Hellwig
Quoted 2 lines · Christoph Hellwig

Except for the little wording issue this looks good to me. I'll pick it up for nvme-5.11 and will fix up the commit log.

Given that we never retry admin command right now I've moved this off nvme-5.11 into the pending queue.

Thanks for sharing!

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#9 lore raw

PATCH 2/3 nvme: check all retry delay times in Identify Controller

In reply to Minwoo Im

If none of Command Retry Delay Time(CRDT)s are set, then we can simply skip to configure Advanced Command Retry Enable(ACRE).

CRDT selection will be made in Completion Queue Entry that controller has responsibility to decide.

This patch checks all CRDT[0, 1, and 2] in Identify Controller data structrure instead of checking the first one: CRDT[0].

---
 drivers/nvme/host/core.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/nvme/host/core.c +1 -1
index d0c99a9400ea..6e428fdc25a8 100644
@@ -2501,7 +2501,7 @@ static int nvme_configure_acre(struct nvme_ctrl *ctrl)
 	int ret;
 
 	/* Don't bother enabling the feature if retry delay is not reported */
-	if (!ctrl->crdt[0])
+	if (!ctrl->crdt[0] && !ctrl->crdt[1] && !ctrl->crdt[2])
 		return 0;
 
 	host = kzalloc(sizeof(*host), GFP_KERNEL);
Signature · 7 lines
2.17.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
#10 lore raw

Re: [PATCH 2/3] nvme: check all retry delay times in Identify Controller

Quoted 8 lines · Minwoo Im

If none of Command Retry Delay Time(CRDT)s are set, then we can simply skip to configure Advanced Command Retry Enable(ACRE).

CRDT selection will be made in Completion Queue Entry that controller has responsibility to decide.

This patch checks all CRDT[0, 1, and 2] in Identify Controller data structrure instead of checking the first one: CRDT[0].

Why?

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#11 lore raw

Re: [PATCH 2/3] nvme: check all retry delay times in Identify Controller

Hello,

Quoted 11 lines · Christoph Hellwig
Quoted 8 lines · Minwoo Im

If none of Command Retry Delay Time(CRDT)s are set, then we can simply skip to configure Advanced Command Retry Enable(ACRE).

CRDT selection will be made in Completion Queue Entry that controller has responsibility to decide.

This patch checks all CRDT[0, 1, and 2] in Identify Controller data structrure instead of checking the first one: CRDT[0].

Why?

I thought we should check whether there's at least one CRDT fields are set or not regardless to CRDT number (0, 1, and 2). If CRDT[0] is not set, but CRDT[1] is set, then we can go ahead with CRDT[1] not [0].

Thanks!

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#12 lore raw

PATCH 3/3 nvme: add parameter command_retry to enable retry

In reply to Minwoo Im

nvme_init_request() has set REQ_FAILFAST_DRIVER to make requests non-retryable. This command flag value is checked in nvme_decide_disposition() to decide whether to RETRY or other operations. In that point, blk_noretry_request() macro will be used to check if command flags have one of REQ_FAILFAST_*. If so, it just decides to complete the request without retrying.

This patch added a module parameter named command_retry to turn on the command retry feature in this driver. If turning it on, REQ3_FAILFAST_DRIVER will not be set to requests so that retry can be reached out to nvme_retry_req() based on the module parameter.

---
 drivers/nvme/host/core.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
drivers/nvme/host/core.c +6 -1
index 6e428fdc25a8..e1836ca9956f 100644
@@ -44,6 +44,10 @@ static unsigned char shutdown_timeout = 5;
 module_param(shutdown_timeout, byte, 0644);
 MODULE_PARM_DESC(shutdown_timeout, "timeout in seconds for controller shutdown");
 
+static bool command_retry;
+module_param(command_retry, bool, 0644);
+MODULE_PARM_DESC(command_retry, "retry commands up to nvme_max_retries");
+
 static u8 nvme_max_retries = 5;
 module_param_named(max_retries, nvme_max_retries, byte, 0644);
 MODULE_PARM_DESC(max_retries, "max number of retries a command may have");
@@ -560,7 +564,8 @@ static inline void nvme_init_request(struct request *req,
 	else /* no queuedata implies admin queue */
 		req->timeout = NVME_ADMIN_TIMEOUT;
 
-	req->cmd_flags |= REQ_FAILFAST_DRIVER;
+	if (!command_retry)
+		req->cmd_flags |= REQ_FAILFAST_DRIVER;
 	nvme_clear_nvme_request(req);
 	nvme_req(req)->cmd = cmd;
 }
Signature · 7 lines
2.17.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
#13 lore raw

Re: [PATCH 3/3] nvme: add parameter command_retry to enable retry

Quoted 33 lines · Minwoo Im

nvme_init_request() has set REQ_FAILFAST_DRIVER to make requests non-retryable. This command flag value is checked in nvme_decide_disposition() to decide whether to RETRY or other operations. In that point, blk_noretry_request() macro will be used to check if command flags have one of REQ_FAILFAST_*. If so, it just decides to complete the request without retrying.

This patch added a module parameter named command_retry to turn on the command retry feature in this driver. If turning it on, REQ3_FAILFAST_DRIVER will not be set to requests so that retry can be reached out to nvme_retry_req() based on the module parameter.

---
  drivers/nvme/host/core.c | 7 ++++++-
  1 file changed, 6 insertions(+), 1 deletion(-)
drivers/nvme/host/core.c +4 -0
index 6e428fdc25a8..e1836ca9956f 100644
@@ -44,6 +44,10 @@ static unsigned char shutdown_timeout = 5;
  module_param(shutdown_timeout, byte, 0644);
  MODULE_PARM_DESC(shutdown_timeout, "timeout in seconds for controller shutdown");
 
+static bool command_retry;
+module_param(command_retry, bool, 0644);
+MODULE_PARM_DESC(command_retry, "retry commands up to nvme_max_retries");
+
  static u8 nvme_max_retries = 5;
  module_param_named(max_retries, nvme_max_retries, byte, 0644);
  MODULE_PARM_DESC(max_retries, "max number of retries a command may have");

drivers/nvme/host/core.c static inline void nvme_init_request(struct request *req, · line 568

Quoted 6 lines
drivers/nvme/host/core.c +2 -1
@@ -560,7 +564,8 @@ static inline void nvme_init_request(struct request *req,
  	else /* no queuedata implies admin queue */
  		req->timeout = NVME_ADMIN_TIMEOUT;
 
-	req->cmd_flags |= REQ_FAILFAST_DRIVER;
+	if (!command_retry)
+		req->cmd_flags |= REQ_FAILFAST_DRIVER;

In abnormal scenarios, such as request time out, connection process may takes long time or the admin command waits for long time. Retry only for non-host errors may be a better choice. Maybe we can make some optimizations in nvme_decide_disposition.

drivers/nvme/host/core.c static inline void nvme_init_request(struct request *req, · line 571

Quoted 4 lines
drivers/nvme/host/core.c +0 -0
  	nvme_clear_nvme_request(req);
  	nvme_req(req)->cmd = cmd;
  }

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#14 lore raw

Re: [PATCH 3/3] nvme: add parameter command_retry to enable retry

Hello,

Quoted 34 lines · Chao Leng
Quoted 31 lines · Minwoo Im

nvme_init_request() has set REQ_FAILFAST_DRIVER to make requests non-retryable. This command flag value is checked in nvme_decide_disposition() to decide whether to RETRY or other operations. In that point, blk_noretry_request() macro will be used to check if command flags have one of REQ_FAILFAST_*. If so, it just decides to complete the request without retrying.

This patch added a module parameter named command_retry to turn on the command retry feature in this driver. If turning it on, REQ3_FAILFAST_DRIVER will not be set to requests so that retry can be reached out to nvme_retry_req() based on the module parameter.

---
  drivers/nvme/host/core.c | 7 ++++++-
  1 file changed, 6 insertions(+), 1 deletion(-)
drivers/nvme/host/core.c +4 -0
index 6e428fdc25a8..e1836ca9956f 100644
@@ -44,6 +44,10 @@ static unsigned char shutdown_timeout = 5;
  module_param(shutdown_timeout, byte, 0644);
  MODULE_PARM_DESC(shutdown_timeout, "timeout in seconds for controller shutdown");
+static bool command_retry;
+module_param(command_retry, bool, 0644);
+MODULE_PARM_DESC(command_retry, "retry commands up to nvme_max_retries");
+
  static u8 nvme_max_retries = 5;
  module_param_named(max_retries, nvme_max_retries, byte, 0644);
  MODULE_PARM_DESC(max_retries, "max number of retries a command may have");
Quoted 8 lines

drivers/nvme/host/core.c static inline void nvme_init_request(struct request *req, · line 567

Quoted 6 lines
drivers/nvme/host/core.c +2 -1
@@ -560,7 +564,8 @@ static inline void nvme_init_request(struct request *req,
  	else /* no queuedata implies admin queue */
  		req->timeout = NVME_ADMIN_TIMEOUT;
-	req->cmd_flags |= REQ_FAILFAST_DRIVER;
+	if (!command_retry)
+		req->cmd_flags |= REQ_FAILFAST_DRIVER;

In abnormal scenarios, such as request time out, connection process may takes long time or the admin command waits for long time. Retry only for non-host errors may be a better choice. Maybe we can make some optimizations in nvme_decide_disposition.

Thanks for your review!

Oh, I agreed that it might wait for so long time in connecting process. Restricting some of commands that should be retried would be better as you mentiond.

Do you mean that maybe we can check this module parameter in nvme_decide_disposition()? Like, even if blk_noretry_request(req) says that it's non-retriable, if this module parameter is enabled, then we can retry rather than failfast?

Thanks,

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

#15 lore raw

Re: [PATCH 3/3] nvme: add parameter command_retry to enable retry

Quoted 48 lines · Minwoo Im

Hello,

Quoted 34 lines · Chao Leng
Quoted 31 lines · Minwoo Im

nvme_init_request() has set REQ_FAILFAST_DRIVER to make requests non-retryable. This command flag value is checked in nvme_decide_disposition() to decide whether to RETRY or other operations. In that point, blk_noretry_request() macro will be used to check if command flags have one of REQ_FAILFAST_*. If so, it just decides to complete the request without retrying.

This patch added a module parameter named command_retry to turn on the command retry feature in this driver. If turning it on, REQ3_FAILFAST_DRIVER will not be set to requests so that retry can be reached out to nvme_retry_req() based on the module parameter.

---
   drivers/nvme/host/core.c | 7 ++++++-
   1 file changed, 6 insertions(+), 1 deletion(-)
drivers/nvme/host/core.c +4 -0
index 6e428fdc25a8..e1836ca9956f 100644
@@ -44,6 +44,10 @@ static unsigned char shutdown_timeout = 5;
   module_param(shutdown_timeout, byte, 0644);
   MODULE_PARM_DESC(shutdown_timeout, "timeout in seconds for controller shutdown");
+static bool command_retry;
+module_param(command_retry, bool, 0644);
+MODULE_PARM_DESC(command_retry, "retry commands up to nvme_max_retries");
+
   static u8 nvme_max_retries = 5;
   module_param_named(max_retries, nvme_max_retries, byte, 0644);
   MODULE_PARM_DESC(max_retries, "max number of retries a command may have");
Quoted 8 lines

drivers/nvme/host/core.c static inline void nvme_init_request(struct request *req, · line 567

Quoted 6 lines
drivers/nvme/host/core.c +2 -1
@@ -560,7 +564,8 @@ static inline void nvme_init_request(struct request *req,
   	else /* no queuedata implies admin queue */
   		req->timeout = NVME_ADMIN_TIMEOUT;
-	req->cmd_flags |= REQ_FAILFAST_DRIVER;
+	if (!command_retry)
+		req->cmd_flags |= REQ_FAILFAST_DRIVER;

In abnormal scenarios, such as request time out, connection process may takes long time or the admin command waits for long time. Retry only for non-host errors may be a better choice. Maybe we can make some optimizations in nvme_decide_disposition.

Thanks for your review!

Quoted 8 lines

Oh, I agreed that it might wait for so long time in connecting process. Restricting some of commands that should be retried would be better as you mentiond.

Do you mean that maybe we can check this module parameter in nvme_decide_disposition()? Like, even if blk_noretry_request(req) says that it's non-retriable, if this module parameter is enabled, then we can retry rather than failfast?

No, I mean that add the local preferential retry which defined in the NVMe protocol.

---
 drivers/nvme/host/core.c | 18 +++++++++++++-----
 drivers/nvme/host/nvme.h |  1 +
 2 files changed, 14 insertions(+), 5 deletions(-)
drivers/nvme/host/core.c +13 -5
index 99f91efe3824..a25e9b4956b9 100644
@@ -316,15 +316,20 @@ static inline enum nvme_disposition nvme_decide_disposition(struct request *req)
 	if (likely(nvme_req(req)->status == 0))
 		return COMPLETE;
 
-	if (blk_noretry_request(req) ||
-	    (nvme_req(req)->status & NVME_SC_DNR) ||
+	if ((nvme_req(req)->status & NVME_SC_DNR) ||
 	    nvme_req(req)->retries >= nvme_max_retries)
 		return COMPLETE;
 
+	if (nvme_req(req)->ctrl->acre &&
+	    !nvme_is_path_error(nvme_req(req)->status) &&
+	    !blk_queue_dying(req->q))
+		return RETRY;
+
+	if (blk_noretry_request(req))
+		return COMPLETE;
+
 	if (req->cmd_flags & REQ_NVME_MPATH) {
-		if (nvme_is_path_error(nvme_req(req)->status) ||
-		    blk_queue_dying(req->q))
-			return FAILOVER;
+		return FAILOVER;
 	} else {
 		if (blk_queue_dying(req->q))
 			return COMPLETE;
@@ -2513,6 +2518,7 @@ static int nvme_configure_acre(struct nvme_ctrl *ctrl)
 	struct nvme_feat_host_behavior *host;
 	int ret;
 
+	ctrl->acre = false;
 	/* Don't bother enabling the feature if retry delay is not reported */
 	if (!ctrl->crdt[0])
 		return 0;
@@ -2524,6 +2530,8 @@ static int nvme_configure_acre(struct nvme_ctrl *ctrl)
 	host->acre = NVME_ENABLE_ACRE;
 	ret = nvme_set_features(ctrl, NVME_FEAT_HOST_BEHAVIOR, 0,
 				host, sizeof(*host), NULL);
+	if (!ret)
+		ctrl->acre = true;
 	kfree(host);
 	return ret;
 }
drivers/nvme/host/nvme.h +1 -0
index bfcedfa4b057..fd914b0dec88 100644
@@ -276,6 +276,7 @@ struct nvme_ctrl {
 #ifdef CONFIG_BLK_DEV_ZONED
 	u32 max_zone_append;
 #endif
+	bool acre;
 	u16 crdt[3];
 	u16 oncs;
 	u16 oacs;
Quoted 4 lines

Thanks,
.

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#16 lore raw

Re: [PATCH 3/3] nvme: add parameter command_retry to enable retry

Hello,

Quoted 55 lines · Chao Leng
Quoted 52 lines · Minwoo Im

Hello,

Quoted 34 lines · Chao Leng
Quoted 31 lines · Minwoo Im

nvme_init_request() has set REQ_FAILFAST_DRIVER to make requests non-retryable. This command flag value is checked in nvme_decide_disposition() to decide whether to RETRY or other operations. In that point, blk_noretry_request() macro will be used to check if command flags have one of REQ_FAILFAST_*. If so, it just decides to complete the request without retrying.

This patch added a module parameter named command_retry to turn on the command retry feature in this driver. If turning it on, REQ3_FAILFAST_DRIVER will not be set to requests so that retry can be reached out to nvme_retry_req() based on the module parameter.

---
   drivers/nvme/host/core.c | 7 ++++++-
   1 file changed, 6 insertions(+), 1 deletion(-)
drivers/nvme/host/core.c +4 -0
index 6e428fdc25a8..e1836ca9956f 100644
@@ -44,6 +44,10 @@ static unsigned char shutdown_timeout = 5;
   module_param(shutdown_timeout, byte, 0644);
   MODULE_PARM_DESC(shutdown_timeout, "timeout in seconds for controller shutdown");
+static bool command_retry;
+module_param(command_retry, bool, 0644);
+MODULE_PARM_DESC(command_retry, "retry commands up to nvme_max_retries");
+
   static u8 nvme_max_retries = 5;
   module_param_named(max_retries, nvme_max_retries, byte, 0644);
   MODULE_PARM_DESC(max_retries, "max number of retries a command may have");
Quoted 8 lines

drivers/nvme/host/core.c static inline void nvme_init_request(struct request *req, · line 567

Quoted 6 lines
drivers/nvme/host/core.c +2 -1
@@ -560,7 +564,8 @@ static inline void nvme_init_request(struct request *req,
   	else /* no queuedata implies admin queue */
   		req->timeout = NVME_ADMIN_TIMEOUT;
-	req->cmd_flags |= REQ_FAILFAST_DRIVER;
+	if (!command_retry)
+		req->cmd_flags |= REQ_FAILFAST_DRIVER;

In abnormal scenarios, such as request time out, connection process may takes long time or the admin command waits for long time. Retry only for non-host errors may be a better choice. Maybe we can make some optimizations in nvme_decide_disposition.

Thanks for your review!

Oh, I agreed that it might wait for so long time in connecting process. Restricting some of commands that should be retried would be better as you mentiond.

Quoted 5 lines
Quoted 4 lines

Do you mean that maybe we can check this module parameter in nvme_decide_disposition()? Like, even if blk_noretry_request(req) says that it's non-retriable, if this module parameter is enabled, then we can retry rather than failfast?

No, I mean that add the local preferential retry which defined in the NVMe protocol.

Thanks for your suggestion, Chao. I think this is much better idea than module parameter. Fixing the nvme_decide_disposition() would be better! I think it's going to be also an answer for Christoph's question on this patch. Please let me prepare next pach series.

Thanks,

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

#17 lore raw

Re: [PATCH 3/3] nvme: add parameter command_retry to enable retry

In reply to Minwoo Im
Quoted 4 lines · Minwoo Im

This patch added a module parameter named command_retry to turn on the command retry feature in this driver. If turning it on, REQ3_FAILFAST_DRIVER will not be set to requests so that retry can be reached out to nvme_retry_req() based on the module parameter.

Why would we want to do that, and if so why only conditional on a module parameter?

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme

author
#18 lore raw

Re: [PATCH 3/3] nvme: add parameter command_retry to enable retry

Hello,

Quoted 8 lines · Christoph Hellwig
Quoted 4 lines · Minwoo Im

This patch added a module parameter named command_retry to turn on the command retry feature in this driver. If turning it on, REQ3_FAILFAST_DRIVER will not be set to requests so that retry can be reached out to nvme_retry_req() based on the module parameter.

Why would we want to do that, and if so why only conditional on a module parameter?

If reqeust->cmd_flags are set with the REQ_FAILFAST_DRIVER in nvme_init_request(), then it will always go to COMPLETE case rather than RETRY in nvme_decide_disposition():

	if (blk_noretry_request(req) ||
		...

And I thought that to make command delay feature enabled, module parameter was a good start to toggle this feature.

Thanks!

_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme