Skip to contentSkip to conversation

scsi: ufs: add a quirk to disable FUA support

PATCH In review Jaegeuk Kim 5 participants 9 messages May 31, 2022 – Jun 7, 2022 SCSI

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 scsi: ufs: add a quirk to disable FUA support

UFS stack shows very low performance of FUA comparing to write and cache_flush. Let's add a quirk to adjust it.

E.g., average latency according to the chunk size of write

Write(us/KB) 4 64 256 1024 2048
FUA 873.792 754.604 995.624 1011.67 1067.99
CACHE_FLUSH 824.703 712.98 800.307 1019.5 1037.37

---
 drivers/scsi/ufs/ufshcd.c | 3 +++
 drivers/scsi/ufs/ufshcd.h | 5 +++++
 2 files changed, 8 insertions(+)
drivers/scsi/ufs/ufshcd.c +3 -0
index 3f9caafa91bf..811f3467879c 100644
@@ -5035,6 +5035,9 @@ static int ufshcd_slave_configure(struct scsi_device *sdev)
 	 */
 	sdev->silence_suspend = 1;
 
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_FUA)
+		sdev->broken_fua = 1;
+
 	ufshcd_crypto_register(hba, q);
 
 	return 0;
drivers/scsi/ufs/ufshcd.h +5 -0
index 94f545be183a..6c480c6741d6 100644
@@ -602,6 +602,11 @@ enum ufshcd_quirks {
 	 * support physical host configuration.
 	 */
 	UFSHCD_QUIRK_SKIP_PH_CONFIGURATION		= 1 << 16,
+
+	/*
+	 * This quirk disables FUA support.
+	 */
+	UFSHCD_QUIRK_BROKEN_FUA				= 1 << 17,
 };
 
 enum ufshcd_caps {
Signature · 1 line
2.36.1.255.ge46751e96f-goog
#1 lore raw
Quoted 34 lines · Jaegeuk Kim

UFS stack shows very low performance of FUA comparing to write and cache_flush. Let's add a quirk to adjust it.

E.g., average latency according to the chunk size of write

Write(us/KB) 4 64 256 1024 2048
FUA 873.792 754.604 995.624 1011.67 1067.99
CACHE_FLUSH 824.703 712.98 800.307 1019.5 1037.37

---
 drivers/scsi/ufs/ufshcd.c | 3 +++
 drivers/scsi/ufs/ufshcd.h | 5 +++++
 2 files changed, 8 insertions(+)
drivers/scsi/ufs/ufshcd.c +3 -0
index 3f9caafa91bf..811f3467879c 100644
@@ -5035,6 +5035,9 @@ static int ufshcd_slave_configure(struct scsi_device *sdev)
 	 */
 	sdev->silence_suspend = 1;
 
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_FUA)
+		sdev->broken_fua = 1;
+
 	ufshcd_crypto_register(hba, q);
 
 	return 0;
drivers/scsi/ufs/ufshcd.h +0 -0
index 94f545be183a..6c480c6741d6 100644

drivers/scsi/ufs/ufshcd.h enum ufshcd_quirks { · line 610

Quoted 9 lines
drivers/scsi/ufs/ufshcd.h +5 -0
@@ -602,6 +602,11 @@ enum ufshcd_quirks {
 	 * support physical host configuration.
 	 */
 	UFSHCD_QUIRK_SKIP_PH_CONFIGURATION		= 1 << 16,
+
+	/*
+	 * This quirk disables FUA support.
+	 */
+	UFSHCD_QUIRK_BROKEN_FUA				= 1 << 17,
 };

"Broken" is ambiguous. IIUC, the issue is that FUA performance is very bad, not that it doesn't work. Can you clarify the intent in the comment?

Also, this patch does nothing by itself. Which UFS host driver(s) need this quirk bit? Can you update them to use it? Or do they all need this, in which case a quirk bit would be unnecessary?

- Eric

#2 lore raw
Quoted 39 lines · Eric Biggers
Quoted 38 lines · Jaegeuk Kim

UFS stack shows very low performance of FUA comparing to write and cache_flush. Let's add a quirk to adjust it.

E.g., average latency according to the chunk size of write

Write(us/KB) 4 64 256 1024 2048
FUA 873.792 754.604 995.624 1011.67 1067.99
CACHE_FLUSH 824.703 712.98 800.307 1019.5 1037.37

---
 drivers/scsi/ufs/ufshcd.c | 3 +++
 drivers/scsi/ufs/ufshcd.h | 5 +++++
 2 files changed, 8 insertions(+)
drivers/scsi/ufs/ufshcd.c +3 -0
index 3f9caafa91bf..811f3467879c 100644
@@ -5035,6 +5035,9 @@ static int ufshcd_slave_configure(struct scsi_device *sdev)
 	 */
 	sdev->silence_suspend = 1;
 
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_FUA)
+		sdev->broken_fua = 1;
+
 	ufshcd_crypto_register(hba, q);
 
 	return 0;
drivers/scsi/ufs/ufshcd.h +1 -0
index 94f545be183a..6c480c6741d6 100644
@@ -602,6 +602,11 @@ enum ufshcd_quirks {
 	 * support physical host configuration.
 	 */
 	UFSHCD_QUIRK_SKIP_PH_CONFIGURATION		= 1 << 16,
+
Quoted 8 lines

drivers/scsi/ufs/ufshcd.h enum ufshcd_quirks { · line 610

Quoted 5 lines
drivers/scsi/ufs/ufshcd.h +4 -0
+	/*
+	 * This quirk disables FUA support.
+	 */
+	UFSHCD_QUIRK_BROKEN_FUA				= 1 << 17,
 };

"Broken" is ambiguous. IIUC, the issue is that FUA performance is very bad, not that it doesn't work. Can you clarify the intent in the comment?

My intent is FUA was supposed to be better than write+cache_flush.

Quoted 4 lines

Also, this patch does nothing by itself. Which UFS host driver(s) need this quirk bit? Can you update them to use it? Or do they all need this, in which case a quirk bit would be unnecessary?

Likewise other quick bits, using this is up to SoC or UFS vendors. I think that combination is up to OEMs who is building the product.

Quoted 2 lines

- Eric

#3 lore raw
Quoted 6 lines · Jaegeuk Kim
Quoted 3 lines

Also, this patch does nothing by itself. Which UFS host driver(s) need this quirk bit? Can you update them to use it? Or do they all need this, in which case a quirk bit would be unnecessary?

Likewise other quick bits, using this is up to SoC or UFS vendors. I think that combination is up to OEMs who is building the product.

Of the UFS host drivers in the upstream kernel, which ones actually need this?

- Eric

#4 lore raw
Quoted 10 lines · Eric Biggers
Quoted 6 lines · Jaegeuk Kim
Quoted 3 lines

Also, this patch does nothing by itself. Which UFS host driver(s) need this quirk bit? Can you update them to use it? Or do they all need this, in which case a quirk bit would be unnecessary?

Likewise other quick bits, using this is up to SoC or UFS vendors. I think that combination is up to OEMs who is building the product.

Of the UFS host drivers in the upstream kernel, which ones actually need this?

Qualcomm UFSHC would need this.

Quoted 1 line

- Eric

#5 lore raw
In reply to Jaegeuk Kim
Quoted 8 lines · Jaegeuk Kim

UFS stack shows very low performance of FUA comparing to write and cache_flush. Let's add a quirk to adjust it.

E.g., average latency according to the chunk size of write

Write(us/KB) 4 64 256 1024 2048
FUA 873.792 754.604 995.624 1011.67 1067.99
CACHE_FLUSH 824.703 712.98 800.307 1019.5 1037.37

Wouldn't it depend on how much data might be in the cache? Do you have real-world use-cases where the difference is measurable?

Quoted 26 lines
---
 drivers/scsi/ufs/ufshcd.c | 3 +++
 drivers/scsi/ufs/ufshcd.h | 5 +++++
 2 files changed, 8 insertions(+)
drivers/scsi/ufs/ufshcd.c +3 -0
index 3f9caafa91bf..811f3467879c 100644
@@ -5035,6 +5035,9 @@ static int ufshcd_slave_configure(struct scsi_device *sdev)
 	 */
 	sdev->silence_suspend = 1;
 
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_FUA)
+		sdev->broken_fua = 1;
+
 	ufshcd_crypto_register(hba, q);
 
 	return 0;
drivers/scsi/ufs/ufshcd.h +0 -0
index 94f545be183a..6c480c6741d6 100644

drivers/scsi/ufs/ufshcd.h enum ufshcd_quirks { · line 609

Quoted 8 lines
drivers/scsi/ufs/ufshcd.h +5 -0
@@ -602,6 +602,11 @@ enum ufshcd_quirks {
 	 * support physical host configuration.
 	 */
 	UFSHCD_QUIRK_SKIP_PH_CONFIGURATION		= 1 << 16,
+
+	/*
+	 * This quirk disables FUA support.
+	 */
+	UFSHCD_QUIRK_BROKEN_FUA				= 1 << 17,

Wouldn't it be more appropriate to make it a UFS_DEVICE_QUIRK_ since it presumably depends on the UFS device not the host controller?

Also, as already commented by others, there needs to be a user of the quirk

Quoted 3 lines
drivers/scsi/ufs/ufshcd.h +0 -0
 };
 
 enum ufshcd_caps {
#6 lore raw
Quoted 11 lines · Adrian Hunter

drivers/scsi/ufs/ufshcd.h enum ufshcd_quirks { · line 609

Quoted 4 lines · Jaegeuk Kim
drivers/scsi/ufs/ufshcd.h +4 -0
+	/*
+	 * This quirk disables FUA support.
+	 */
+	UFSHCD_QUIRK_BROKEN_FUA				= 1 << 17,

Wouldn't it be more appropriate to make it a UFS_DEVICE_QUIRK_ since it presumably depends on the UFS device not the host controller?

Also, as already commented by others, there needs to be a user of the quirk

Another possibility is to use the generic SCSI blacklist mechanism. See also the scsi_static_device_list array. See also /proc/scsi/device_info. From scsi_devinfo.c:

/*
* proc_scsi_dev_info_write - allow additions to scsi_dev_info_list via
* /proc.
*
* Description: Adds a black/white list entry for vendor and model with
* an integer value of flag to the scsi device info list.
* To use, echo "vendor:model:flag" > /proc/scsi/device_info */

Thanks,

Bart.

#7 lore raw
Quoted 17 lines · Bart Van Assche
Quoted 11 lines · Adrian Hunter

drivers/scsi/ufs/ufshcd.h enum ufshcd_quirks { · line 609

Quoted 4 lines · Jaegeuk Kim
drivers/scsi/ufs/ufshcd.h +4 -0
+	/*
+	 * This quirk disables FUA support.
+	 */
+	UFSHCD_QUIRK_BROKEN_FUA				= 1 << 17,

Wouldn't it be more appropriate to make it a UFS_DEVICE_QUIRK_ since it presumably depends on the UFS device not the host controller?

Also, as already commented by others, there needs to be a user of the quirk

Another possibility is to use the generic SCSI blacklist mechanism. See also the scsi_static_device_list array. See also /proc/scsi/device_info. From scsi_devinfo.c:

Quoted 8 lines

/*
* proc_scsi_dev_info_write - allow additions to scsi_dev_info_list via
* /proc.
*
* Description: Adds a black/white list entry for vendor and model with
* an integer value of flag to the scsi device info list.
* To use, echo "vendor:model:flag" > /proc/scsi/device_info */

Good to know. Thank you, Bart.

Quoted 4 lines

Thanks,

Bart.

#8 lore raw
In reply to Adrian Hunter
Quoted 11 lines · Adrian Hunter
Quoted 8 lines · Jaegeuk Kim

UFS stack shows very low performance of FUA comparing to write and cache_flush. Let's add a quirk to adjust it.

E.g., average latency according to the chunk size of write

Write(us/KB) 4 64 256 1024 2048
FUA 873.792 754.604 995.624 1011.67 1067.99
CACHE_FLUSH 824.703 712.98 800.307 1019.5 1037.37

Wouldn't it depend on how much data might be in the cache?

I've got this average latency from 100 commands of write+cache_flush vs. write(FUA). I think the cached data should be the same as this chunk size.

Quoted 1 line

Do you have real-world use-cases where the difference is measurable?

I'm approaching this based on 1) f2fs uses FUA for checkpoint and fsync, and 2) iomap uses FUA for O_DIRECT|O_DSYNC case [1].

[1] https://lore.kernel.org/lkml/[email protected]/

Quoted 39 lines
Quoted 27 lines
---
 drivers/scsi/ufs/ufshcd.c | 3 +++
 drivers/scsi/ufs/ufshcd.h | 5 +++++
 2 files changed, 8 insertions(+)
drivers/scsi/ufs/ufshcd.c +3 -0
index 3f9caafa91bf..811f3467879c 100644
@@ -5035,6 +5035,9 @@ static int ufshcd_slave_configure(struct scsi_device *sdev)
 	 */
 	sdev->silence_suspend = 1;
 
+	if (hba->quirks & UFSHCD_QUIRK_BROKEN_FUA)
+		sdev->broken_fua = 1;
+
 	ufshcd_crypto_register(hba, q);
 
 	return 0;
drivers/scsi/ufs/ufshcd.h +0 -0
index 94f545be183a..6c480c6741d6 100644

drivers/scsi/ufs/ufshcd.h enum ufshcd_quirks { · line 609

Quoted 8 lines
drivers/scsi/ufs/ufshcd.h +5 -0
@@ -602,6 +602,11 @@ enum ufshcd_quirks {
 	 * support physical host configuration.
 	 */
 	UFSHCD_QUIRK_SKIP_PH_CONFIGURATION		= 1 << 16,
+
+	/*
+	 * This quirk disables FUA support.
+	 */
+	UFSHCD_QUIRK_BROKEN_FUA				= 1 << 17,

Wouldn't it be more appropriate to make it a UFS_DEVICE_QUIRK_ since it presumably depends on the UFS device not the host controller?

Quoted 2 lines

Also, as already commented by others, there needs to be a user of the quirk

Since I asked SoC vendors can verify the performance with this quirk, I need to wait for their reply. Meanwhile, I'm willing to disable FUA in Pixel devices, which I cannot post any patch directly to LKML.

Agreed that, if there's no other user in upstream, I'm okay to drop this.

Quoted 4 lines
Quoted 4 lines
drivers/scsi/ufs/ufshcd.h +0 -0
 };
 
 enum ufshcd_caps {