Skip to contentSkip to conversation

nvme: some cleanup

PATCH Superseded Edmund Nadolski 5 participants 9 messages Apr 29, 2019 – Apr 30, 2019 NVMe

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

cover nvme: some cleanup

This series clarifies the code by adding some macros, comments, and fixing a few miscellaneous typos.

Edmund Nadolski (2):
nvme: nvme_set_queue_count should use descriptive macros
nvme: add clarifying comments and fix some typos

 drivers/nvme/host/core.c |  8 ++++++--
 drivers/nvme/host/nvme.h | 32 ++++++++++++++++----------------
 drivers/nvme/host/pci.c  |  2 +-
 3 files changed, 23 insertions(+), 19 deletions(-)
Signature · 1 line
2.20.1
#1 lore raw

PATCH 1/2 nvme: nvme_set_queue_count should use descriptive macros

Implement macros to set/get the number of submission and/or completion queues requested by the Set Features command. This replaces the bit masking/shifting code which is harder to read and more error prone to maintain.

Signed-off-by Edmund Nadolski <ednadols at linux.microsoft.com>
---
 drivers/nvme/host/core.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
drivers/nvme/host/core.c +6 -2
index 3dd043aa6d1f..b3804dbdcc30 100644
@@ -1132,9 +1132,13 @@ static int nvme_set_features(struct nvme_ctrl *dev, unsigned fid, unsigned dword
 	return ret;
 }
 
+#define SET_NUMQ(nsqr, ncqr)	(((nsqr) - 1) | (((ncqr) - 1) << 16))
+#define GET_NSQA(dw)		(((dw) & 0xffff) + 1)
+#define GET_NCQA(dw)		(((dw) >> 16) + 1)
+
 int nvme_set_queue_count(struct nvme_ctrl *ctrl, int *count)
 {
-	u32 q_count = (*count - 1) | ((*count - 1) << 16);
+	u32 q_count = SET_NUMQ(*count, *count);
 	u32 result;
 	int status, nr_io_queues;
 
@@ -1152,7 +1156,7 @@ int nvme_set_queue_count(struct nvme_ctrl *ctrl, int *count)
 		dev_err(ctrl->device, "Could not set queue count (%d)\n", status);
 		*count = 0;
 	} else {
-		nr_io_queues = min(result & 0xffff, result >> 16) + 1;
+		nr_io_queues = min(GET_NSQA(result), GET_NCQA(result));
 		*count = min(*count, nr_io_queues);
 	}
 
Signature · 1 line
2.20.1
maintainer
#2 lore raw

PATCH 1/2 nvme: nvme_set_queue_count should use descriptive macros

Looks fine,

Reviewed-by Sagi Grimberg <sagi at grimberg.me>
#3 lore raw

PATCH 1/2 nvme: nvme_set_queue_count should use descriptive macros

In reply to Edmund Nadolski
Quoted 6 lines · Edmund Nadolski

Implement macros to set/get the number of submission and/or completion queues requested by the Set Features command. This replaces the bit masking/shifting code which is harder to read and more error prone to maintain.

Signed-off-by Edmund Nadolski <ednadols at linux.microsoft.com>

I have to say I actually find this much harder to read than the old version.

#4 lore raw

PATCH 2/2 nvme: add clarifying comments and fix some typos

In reply to Edmund Nadolski

Add comment descriptions for struct fields per NVMe spec.

Signed-off-by Edmund Nadolski <ednadols at linux.microsoft.com>
---
 drivers/nvme/host/nvme.h | 32 ++++++++++++++++----------------
 drivers/nvme/host/pci.c  |  2 +-
 2 files changed, 17 insertions(+), 17 deletions(-)
drivers/nvme/host/nvme.h +16 -16
index 527d64545023..61736d1ef15e 100644
@@ -177,7 +177,7 @@ struct nvme_ctrl {
 	u16 cntlid;
 
 	u32 ctrl_config;
-	u16 mtfa;
+	u16 mtfa;		/* maximum time for firmware activation */
 	u32 queue_count;
 
 	u64 cap;
@@ -185,23 +185,23 @@ struct nvme_ctrl {
 	u32 max_hw_sectors;
 	u32 max_segments;
 	u16 crdt[3];
-	u16 oncs;
-	u16 oacs;
-	u16 nssa;
+	u16 oncs;		/* optional nvm command support */
+	u16 oacs;		/* optional admin command support */
+	u16 nssa;		/* nvme subsystem streams available */
 	u16 nr_streams;
 	u32 max_namespaces;
 	atomic_t abort_limit;
-	u8 vwc;
-	u32 vs;
-	u32 sgls;
-	u16 kas;
-	u8 npss;
-	u8 apsta;
-	u32 oaes;
+	u8 vwc;			/* volatile write cache */
+	u32 vs;			/* version */
+	u32 sgls;		/* scatter gather list support */
+	u16 kas;		/* keep alive support */
+	u8 npss;		/* number of power states supported */
+	u8 apsta;		/* autonomous power state transition */
+	u32 oaes;		/* optional asynchronous events supported */
 	u32 aen_result;
-	u32 ctratt;
+	u32 ctratt;		/* controller attribute */
 	unsigned int shutdown_timeout;
-	unsigned int kato;
+	unsigned int kato;	/* keep alive timeout */
 	bool subsystem;
 	unsigned long quirks;
 	struct nvme_id_power_state psd[32];
@@ -231,8 +231,8 @@ struct nvme_ctrl {
 	bool apst_enabled;
 
 	/* PCIe only: */
-	u32 hmpre;
-	u32 hmmin;
+	u32 hmpre;	/* host memory buffer preferred size */
+	u32 hmmin;	/* host memory buffer minimum size */
 	u32 hmminds;
 	u16 hmmaxd;
 
@@ -290,7 +290,7 @@ struct nvme_ns_ids {
 /*
  * Anchor structure for namespaces.  There is one for each namespace in a
  * NVMe subsystem that any of our controllers can see, and the namespace
- * structure for each controller is chained of it.  For private namespaces
+ * structure for each controller is chained off it.  For private namespaces
  * there is a 1:1 relation to our namespace structures, that is ->list
  * only ever has a single entry for private namespaces.
  */
drivers/nvme/host/pci.c +1 -1
index c1eecde6b853..368371c5cafc 100644
@@ -227,7 +227,7 @@ struct nvme_iod {
 };
 
 /*
- * Check we didin't inadvertently grow the command struct
+ * Check we didn't inadvertently grow the command struct
  */
 static inline void _nvme_check_size(void)
 {
Signature · 1 line
2.20.1
maintainer
#5 lore raw

PATCH 2/2 nvme: add clarifying comments and fix some typos

Quoted 2 lines

Add comment descriptions for struct fields per NVMe spec.

Lets split typos from this patch. Not sure that the comments are all that useful, as we usually have a spec/tp at hand...

#6 lore raw

PATCH 2/2 nvme: add clarifying comments and fix some typos

Quoted 5 lines · Sagi Grimberg
Quoted 2 lines

Add comment descriptions for struct fields per NVMe spec.

Lets split typos from this patch. Not sure that the comments are all that useful, as we usually have a spec/tp at hand...

They could be useful to save a lookup (or when a spec isn't handy). I had also thought of doing the nvm_id_ctrl and nvme_id_ns structs, but I'll hold off as it seems folks see this as extraneous.

I will split out the typos out and re-send.

Thanks,
Ed

#7 lore raw

PATCH 2/2 nvme: add clarifying comments and fix some typos

In reply to Edmund Nadolski
Quoted 10 lines

-----Original Message-----
From: Linux-nvme [mailto:linux-nvme-bounces at lists.infradead.org] On Behalf Of Edmund Nadolski
Sent: Tuesday, April 30, 2019 2:36 AM
To: linux-nvme at lists.infradead.org; ednadols at linux.microsoft.com
Subject: [PATCH 2/2] nvme: add clarifying comments and fix some typos

Add comment descriptions for struct fields per NVMe spec.

Signed-off-by Edmund Nadolski <ednadols at linux.microsoft.com>

IMHO, if you really want to provide description for structure attributes, why don't we have all of them instead describing few fields only.

Also It would be better if typo-related commit is separted from it.

Thanks,

#8 lore raw

PATCH 2/2 nvme: add clarifying comments and fix some typos

In reply to Edmund Nadolski

These fields are taken from the spec, so I don't see the need for adding comments as most of the fields are combination of the bits and for that you have to refer to the spec anyways.

Quoted 91 lines · Edmund Nadolski

Add comment descriptions for struct fields per NVMe spec.

Signed-off-by Edmund Nadolski <ednadols at linux.microsoft.com>
---
  drivers/nvme/host/nvme.h | 32 ++++++++++++++++----------------
  drivers/nvme/host/pci.c  |  2 +-
  2 files changed, 17 insertions(+), 17 deletions(-)
drivers/nvme/host/nvme.h +16 -16
index 527d64545023..61736d1ef15e 100644
@@ -177,7 +177,7 @@ struct nvme_ctrl {
  	u16 cntlid;
 
  	u32 ctrl_config;
-	u16 mtfa;
+	u16 mtfa;		/* maximum time for firmware activation */
  	u32 queue_count;
 
  	u64 cap;
@@ -185,23 +185,23 @@ struct nvme_ctrl {
  	u32 max_hw_sectors;
  	u32 max_segments;
  	u16 crdt[3];
-	u16 oncs;
-	u16 oacs;
-	u16 nssa;
+	u16 oncs;		/* optional nvm command support */
+	u16 oacs;		/* optional admin command support */
+	u16 nssa;		/* nvme subsystem streams available */
  	u16 nr_streams;
  	u32 max_namespaces;
  	atomic_t abort_limit;
-	u8 vwc;
-	u32 vs;
-	u32 sgls;
-	u16 kas;
-	u8 npss;
-	u8 apsta;
-	u32 oaes;
+	u8 vwc;			/* volatile write cache */
+	u32 vs;			/* version */
+	u32 sgls;		/* scatter gather list support */
+	u16 kas;		/* keep alive support */
+	u8 npss;		/* number of power states supported */
+	u8 apsta;		/* autonomous power state transition */
+	u32 oaes;		/* optional asynchronous events supported */
  	u32 aen_result;
-	u32 ctratt;
+	u32 ctratt;		/* controller attribute */
  	unsigned int shutdown_timeout;
-	unsigned int kato;
+	unsigned int kato;	/* keep alive timeout */
  	bool subsystem;
  	unsigned long quirks;
  	struct nvme_id_power_state psd[32];
@@ -231,8 +231,8 @@ struct nvme_ctrl {
  	bool apst_enabled;
 
  	/* PCIe only: */
-	u32 hmpre;
-	u32 hmmin;
+	u32 hmpre;	/* host memory buffer preferred size */
+	u32 hmmin;	/* host memory buffer minimum size */
  	u32 hmminds;
  	u16 hmmaxd;
 
@@ -290,7 +290,7 @@ struct nvme_ns_ids {
  /*
   * Anchor structure for namespaces.  There is one for each namespace in a
   * NVMe subsystem that any of our controllers can see, and the namespace
- * structure for each controller is chained of it.  For private namespaces
+ * structure for each controller is chained off it.  For private namespaces
   * there is a 1:1 relation to our namespace structures, that is ->list
   * only ever has a single entry for private namespaces.
   */
drivers/nvme/host/pci.c +1 -1
index c1eecde6b853..368371c5cafc 100644
@@ -227,7 +227,7 @@ struct nvme_iod {
  };
 
  /*
- * Check we didin't inadvertently grow the command struct
+ * Check we didn't inadvertently grow the command struct
   */
  static inline void _nvme_check_size(void)
  {