mtip32xx: Set block queue boundary variables

PATCH In review Asai Thambi S P 3 participants 8 messages May 30, 2012 – Jun 1, 2012 Block layer

lore

Share & tools

lore

Asai Thambi S P

author

2012-05-30 01:42 UTC #0 lore raw

PATCH 5/11 mtip32xx: Set block queue boundary variables

Set the following block queue boundary variables
* max_hw_sectors
* max_segment_size
* nr_requests

Signed-off-by Asai Thambi S P <[email protected]>

---
 drivers/block/mtip32xx/mtip32xx.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)

drivers/block/mtip32xx/mtip32xx.c +4 -0

index 9fe897d..801e70c 100644
@@ -3631,7 +3631,11 @@ skip_create_disk:
 	set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
 	blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
 	blk_queue_physical_block_size(dd->queue, 4096);
+	blk_queue_max_hw_sectors(dd->queue, 0xffff);
+	blk_queue_max_segment_size(dd->queue, 0x400000);
 	blk_queue_io_min(dd->queue, 4096);
+	dd->queue->nr_requests = 255;
+
 	/*
 	 * write back cache is not supported in the device. FUA depends on
 	 * write back cache support, hence setting flush support to zero.

Signature · 1 line

1.7.1

Jens Axboe

maintainer

2012-05-31 06:44 UTC #1 lore raw

Quoted 16 lines · Asai Thambi S P

Set the following block queue boundary variables
* max_hw_sectors
* max_segment_size
* nr_requests
Signed-off-by Asai Thambi S P <[email protected]>
---
 drivers/block/mtip32xx/mtip32xx.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
drivers/block/mtip32xx/mtip32xx.c +0 -0
index 9fe897d..801e70c 100644

drivers/block/mtip32xx/mtip32xx.c skip_create_disk: · line 3637

Quoted 7 lines

drivers/block/mtip32xx/mtip32xx.c +3 -0

@@ -3631,7 +3631,11 @@ skip_create_disk:
 	set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
 	blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
 	blk_queue_physical_block_size(dd->queue, 4096);
+	blk_queue_max_hw_sectors(dd->queue, 0xffff);
+	blk_queue_max_segment_size(dd->queue, 0x400000);
 	blk_queue_io_min(dd->queue, 4096);
+	dd->queue->nr_requests = 255;

->nr_requests isn't a boundary variable you set for the queue. It's set by the core bits, or by the user via the sysfs interface.

So you should not touch that from the driver.

Signature · 1 line

Jens Axboe

Asai Thambi S P

author

2012-05-31 15:56 UTC #2 lore raw

Quoted 28 lines · Jens Axboe

Quoted 16 lines · Asai Thambi S P
Set the following block queue boundary variables
* max_hw_sectors
* max_segment_size
* nr_requests
Signed-off-by Asai Thambi S P <[email protected]>
---
 drivers/block/mtip32xx/mtip32xx.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
drivers/block/mtip32xx/mtip32xx.c +0 -0
index 9fe897d..801e70c 100644
drivers/block/mtip32xx/mtip32xx.c skip_create_disk: · line 3637
Quoted 7 lines
drivers/block/mtip32xx/mtip32xx.c +3 -0
@@ -3631,7 +3631,11 @@ skip_create_disk:
 	set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
 	blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
 	blk_queue_physical_block_size(dd->queue, 4096);
+	blk_queue_max_hw_sectors(dd->queue, 0xffff);
+	blk_queue_max_segment_size(dd->queue, 0x400000);
 	blk_queue_io_min(dd->queue, 4096);
+	dd->queue->nr_requests = 255;
->nr_requests isn't a boundary variable you set for the queue. It's set by the core bits, or by the user via the sysfs interface.

Quoted 2 lines

So you should not touch that from the driver.

Ok.

I saw scsi lib module changing it, so thought of changing the value close to device queue depth.

--
Regards,
Asai Thambi

Jeff Moyer

2012-05-31 18:08 UTC #3 lore raw

Quoted 30 lines · Asai Thambi S P

Quoted 28 lines · Jens Axboe
Quoted 16 lines · Asai Thambi S P
Set the following block queue boundary variables
* max_hw_sectors
* max_segment_size
* nr_requests
Signed-off-by Asai Thambi S P <[email protected]>
---
 drivers/block/mtip32xx/mtip32xx.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
drivers/block/mtip32xx/mtip32xx.c +0 -0
index 9fe897d..801e70c 100644
drivers/block/mtip32xx/mtip32xx.c skip_create_disk: · line 3637
Quoted 7 lines
drivers/block/mtip32xx/mtip32xx.c +3 -0
@@ -3631,7 +3631,11 @@ skip_create_disk:
 	set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
 	blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
 	blk_queue_physical_block_size(dd->queue, 4096);
+	blk_queue_max_hw_sectors(dd->queue, 0xffff);
+	blk_queue_max_segment_size(dd->queue, 0x400000);
 	blk_queue_io_min(dd->queue, 4096);
+	dd->queue->nr_requests = 255;
->nr_requests isn't a boundary variable you set for the queue. It's set by the core bits, or by the user via the sysfs interface.

Quoted 7 lines

Quoted 2 lines
So you should not touch that from the driver.
Ok.
I saw scsi lib module changing it, so thought of changing the value close to device queue depth.

That's actually a fair point. What is the device queue depth for this card?

Cheers,
Jeff

Asai Thambi S P

author

2012-05-31 18:12 UTC #4 lore raw

Quoted 35 lines · Jeff Moyer

Quoted 33 lines · Asai Thambi S P
Quoted 30 lines · Jens Axboe
Quoted 16 lines · Asai Thambi S P
Set the following block queue boundary variables
* max_hw_sectors
* max_segment_size
* nr_requests
Signed-off-by Asai Thambi S P <[email protected]>
---
 drivers/block/mtip32xx/mtip32xx.c |    4 ++++
 1 files changed, 4 insertions(+), 0 deletions(-)
drivers/block/mtip32xx/mtip32xx.c +0 -0
index 9fe897d..801e70c 100644
drivers/block/mtip32xx/mtip32xx.c skip_create_disk: · line 3637
Quoted 7 lines
drivers/block/mtip32xx/mtip32xx.c +3 -0
@@ -3631,7 +3631,11 @@ skip_create_disk:
 	set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
 	blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
 	blk_queue_physical_block_size(dd->queue, 4096);
+	blk_queue_max_hw_sectors(dd->queue, 0xffff);
+	blk_queue_max_segment_size(dd->queue, 0x400000);
 	blk_queue_io_min(dd->queue, 4096);
+	dd->queue->nr_requests = 255;
->nr_requests isn't a boundary variable you set for the queue. It's set by the core bits, or by the user via the sysfs interface.
So you should not touch that from the driver.

Quoted 8 lines

Quoted 4 lines
Ok.
I saw scsi lib module changing it, so thought of changing the value close to device queue depth.
That's actually a fair point. What is the device queue depth for this card?

256

Jeff Moyer

2012-05-31 18:40 UTC #5 lore raw

Quoted 17 lines

Quoted 17 lines
Quoted 17 lines
Quoted 12 lines
drivers/block/mtip32xx/mtip32xx.c skip_create_disk: · line 3637
Quoted 8 lines
drivers/block/mtip32xx/mtip32xx.c +3 -0
@@ -3631,7 +3631,11 @@ skip_create_disk:
 	set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
 	blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
 	blk_queue_physical_block_size(dd->queue, 4096);
+	blk_queue_max_hw_sectors(dd->queue, 0xffff);
+	blk_queue_max_segment_size(dd->queue, 0x400000);
 	blk_queue_io_min(dd->queue, 4096);
+	dd->queue->nr_requests = 255;
->nr_requests isn't a boundary variable you set for the queue. It's set by the core bits, or by the user via the sysfs interface.
Quoted 2 lines
So you should not touch that from the driver.
Ok.

Quoted 7 lines

Quoted 6 lines
Quoted 2 lines
I saw scsi lib module changing it, so thought of changing the value close to device queue depth.
That's actually a fair point. What is the device queue depth for this card?
256

Jens, I actually think changing nr_requests makes sense; the only question in my mind is where to do it. (There's no point in artificially limiting the device by default.) IIRC, your general rule was to set this value to 2x the device's queue depth. So, given that this driver doesn't use the blk-tag interfaces, what do you think would be the cleanest way to bump nr_requests in this (and the more general) case?

Cheers,
Jeff

Jens Axboe

maintainer

2012-06-01 06:01 UTC #6 lore raw

Quoted 23 lines · Jeff Moyer

Quoted 23 lines
Quoted 23 lines
Quoted 12 lines
Quoted 12 lines
drivers/block/mtip32xx/mtip32xx.c skip_create_disk: · line 3637
Quoted 8 lines
drivers/block/mtip32xx/mtip32xx.c +3 -0
@@ -3631,7 +3631,11 @@ skip_create_disk:
 	set_bit(QUEUE_FLAG_NONROT, &dd->queue->queue_flags);
 	blk_queue_max_segments(dd->queue, MTIP_MAX_SG);
 	blk_queue_physical_block_size(dd->queue, 4096);
+	blk_queue_max_hw_sectors(dd->queue, 0xffff);
+	blk_queue_max_segment_size(dd->queue, 0x400000);
 	blk_queue_io_min(dd->queue, 4096);
+	dd->queue->nr_requests = 255;
->nr_requests isn't a boundary variable you set for the queue. It's set by the core bits, or by the user via the sysfs interface.
Quoted 7 lines
Quoted 2 lines
So you should not touch that from the driver.
Ok.
I saw scsi lib module changing it, so thought of changing the value close to device queue depth.
That's actually a fair point. What is the device queue depth for this card?

Quoted 8 lines

Quoted 1 line
256
Jens, I actually think changing nr_requests makes sense; the only question in my mind is where to do it. (There's no point in artificially limiting the device by default.) IIRC, your general rule was to set this value to 2x the device's queue depth. So, given that this driver doesn't use the blk-tag interfaces, what do you think would be the cleanest way to bump nr_requests in this (and the more general) case?

But it doesn't matter for this case, the driver isn't consuming request structures, hence the value set doesn't make any difference. So the only limiter of queue depth is the driver itself.

Signature · 1 line

Jens Axboe

Jeff Moyer

2012-06-01 13:11 UTC #7 lore raw

Quoted 10 lines · Jens Axboe

Quoted 6 lines
Jens, I actually think changing nr_requests makes sense; the only question in my mind is where to do it. (There's no point in artificially limiting the device by default.) IIRC, your general rule was to set this value to 2x the device's queue depth. So, given that this driver doesn't use the blk-tag interfaces, what do you think would be the cleanest way to bump nr_requests in this (and the more general) case?
But it doesn't matter for this case, the driver isn't consuming request structures, hence the value set doesn't make any difference. So the only limiter of queue depth is the driver itself.

Good point. Nevermind. ;-)

Cheers,
Jeff