Skip to contentSkip to conversation

nvmet: replace sysfs_emit() with sprintf() in configfs show attrs

PATCH In review Guixin Liu 3 participants 4 messages Jun 22 – Jun 26

Subsystems (2)
lore
Share & tools
lore

Guixin Liu

author
#0 lore raw

PATCH 1/1 nvmet: replace sysfs_emit() with sprintf() in configfs show attrs

sysfs_emit() is designed for sysfs attributes and internally checks that the buffer is page-aligned. Although configfs currently happens to use a page-aligned buffer allocated via kzalloc(PAGE_SIZE), using sysfs_emit() in configfs show callbacks is a semantic misuse of the API. Replace with sprintf() to match the convention used by all other configfs show functions in these files.

---
 drivers/nvme/target/configfs.c | 6 +++---
 drivers/nvme/target/pci-epf.c  | 6 +++---
 2 files changed, 6 insertions(+), 6 deletions(-)
drivers/nvme/target/configfs.c +3 -3
index b88f897f06e2..f4c77dd1f160 100644
@@ -798,7 +798,7 @@ CONFIGFS_ATTR_WO(nvmet_ns_, revalidate_size);
 
 static ssize_t nvmet_ns_resv_enable_show(struct config_item *item, char *page)
 {
-	return sysfs_emit(page, "%d\n", to_nvmet_ns(item)->pr.enable);
+	return sprintf(page, "%d\n", to_nvmet_ns(item)->pr.enable);
 }
 
 static ssize_t nvmet_ns_resv_enable_store(struct config_item *item,
@@ -1534,7 +1534,7 @@ static ssize_t nvmet_subsys_attr_ieee_oui_show(struct config_item *item,
 {
 	struct nvmet_subsys *subsys = to_subsys(item);
 
-	return sysfs_emit(page, "0x%06x\n", subsys->ieee_oui);
+	return sprintf(page, "0x%06x\n", subsys->ieee_oui);
 }
 
 static ssize_t nvmet_subsys_attr_ieee_oui_store_locked(struct nvmet_subsys *subsys,
@@ -1582,7 +1582,7 @@ static ssize_t nvmet_subsys_attr_firmware_show(struct config_item *item,
 {
 	struct nvmet_subsys *subsys = to_subsys(item);
 
-	return sysfs_emit(page, "%s\n", subsys->firmware_rev);
+	return sprintf(page, "%s\n", subsys->firmware_rev);
 }
 
 static ssize_t nvmet_subsys_attr_firmware_store_locked(struct nvmet_subsys *subsys,
drivers/nvme/target/pci-epf.c +3 -3
index 4e9db96ebfec..547eeb0cae85 100644
@@ -2482,7 +2482,7 @@ static ssize_t nvmet_pci_epf_portid_show(struct config_item *item, char *page)
 	struct config_group *group = to_config_group(item);
 	struct nvmet_pci_epf *nvme_epf = to_nvme_epf(group);
 
-	return sysfs_emit(page, "%u\n", le16_to_cpu(nvme_epf->portid));
+	return sprintf(page, "%u\n", le16_to_cpu(nvme_epf->portid));
 }
 
 static ssize_t nvmet_pci_epf_portid_store(struct config_item *item,
@@ -2515,7 +2515,7 @@ static ssize_t nvmet_pci_epf_subsysnqn_show(struct config_item *item,
 	struct config_group *group = to_config_group(item);
 	struct nvmet_pci_epf *nvme_epf = to_nvme_epf(group);
 
-	return sysfs_emit(page, "%s\n", nvme_epf->subsysnqn);
+	return sprintf(page, "%s\n", nvme_epf->subsysnqn);
 }
 
 static ssize_t nvmet_pci_epf_subsysnqn_store(struct config_item *item,
@@ -2543,7 +2543,7 @@ static ssize_t nvmet_pci_epf_mdts_kb_show(struct config_item *item, char *page)
 	struct config_group *group = to_config_group(item);
 	struct nvmet_pci_epf *nvme_epf = to_nvme_epf(group);
 
-	return sysfs_emit(page, "%u\n", nvme_epf->mdts_kb);
+	return sprintf(page, "%u\n", nvme_epf->mdts_kb);
 }
 
 static ssize_t nvmet_pci_epf_mdts_kb_store(struct config_item *item,
Signature · 1 line
2.43.7
#1 lore raw

Adding Greg to see if he agrees.

But if we add churn to change this, we should probably switch to using seq_file for memory safety, and have some generic configfs-level infrastruture for that.

#2 lore raw
Quoted 2 lines · Christoph Hellwig

Adding Greg to see if he agrees.

Ick, no, sysfs_emit() should be just fine, and safe, for configfs to use, please don't do conversions away from it.

Quoted 3 lines

But if we add churn to change this, we should probably switch to using seq_file for memory safety, and have some generic configfs-level infrastruture for that.

seq_file() would also be great, but really, sysfs_emit() should be all that you need to use here as you should only have one-value-per-file with no overflows possible anyway.

thanks,

greg k-h

author
#3 lore raw
Quoted 8 lines · Greg Kroah-Hartman
Quoted 1 line · Christoph Hellwig

Adding Greg to see if he agrees.

Ick, no, sysfs_emit() should be just fine, and safe, for configfs to use, please don't do conversions away from it.

Quoted 3 lines

But if we add churn to change this, we should probably switch to using seq_file for memory safety, and have some generic configfs-level infrastruture for that.

Quoted 7 lines

seq_file() would also be great, but really, sysfs_emit() should be all that you need to use here as you should only have one-value-per-file with no overflows possible anyway.

thanks,

greg k-h

Got it.

Best Regards,
Guixin Liu