Skip to contentSkip to conversation

UASP: Add MaxNumStreams module parameter

PATCH In review Luben Tuikov 4 participants 10 messages Dec 8, 2010 – Dec 10, 2010

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 2/4 UASP: Add MaxNumStreams module parameter

---
The long story is that we see some host controllers misreport their PSA as they solved 2^v = streams, instead of 2^(v+1) = streams. Thus They report that they support 32 streams when in fact they support 16. When the device attempts to return status for stream > 15, the host says ACK(NumP=0), the device goes in flow control, blah, blah, this module parameter allows you to set a max cap on the number of streams the driver will ask XHCI HCD to allocate.

 drivers/usb/storage/uasp.c |   21 ++++++++++++++++++++-
 1 files changed, 20 insertions(+), 1 deletions(-)
drivers/usb/storage/uasp.c +20 -1
index 3b10efd..e5a26e9 100644
@@ -30,6 +30,19 @@
 #include <scsi/scsi_host.h>
 #include <scsi/scsi_tcq.h>
 
+int MaxNumStreams = -1;
+
+module_param(MaxNumStreams, int, 0444);
+MODULE_PARM_DESC(MaxNumStreams, "\n"
+	"\tSome host controllers and/or devices report a larger number of\n"
+	"\tstreams that they in fact support. This parameter allows you\n"
+	"\tto limit the number of streams this driver will request the XHCI\n"
+	"\tHCD to allocate. If set to -1, the default value, then this driver\n"
+	"\twill use the value reported by the attached device. Else the\n"
+	"\tnumber of streams will be limited to the minimum reported by the\n"
+	"\tattached device and this value. Valid values are -1, default,\n"
+	"\tand 1 to 0xFFEF.");
+
 /* Information unit types
  */
 #define IU_CMD   1
@@ -938,6 +951,8 @@ static int uasp_ep_conf(struct uasp_tport_info *tpinfo)
 					 tpinfo->eps[3]->desc.bEndpointAddress);
 
 	if (udev->speed == USB_SPEED_SUPER) {
+		int max_streams;
+
 		for (i = 1; i < 4; i++) {
 			if (tpinfo->max_streams == 0)
 				tpinfo->max_streams = USB_SS_MAX_STREAMS(tpinfo->eps[i]->ss_ep_comp.bmAttributes);
@@ -952,9 +967,13 @@ static int uasp_ep_conf(struct uasp_tport_info *tpinfo)
 		}
 		
 		tpinfo->use_streams = 1;
+		if (1 <= MaxNumStreams && MaxNumStreams <= 0xFFEF)
+			max_streams = min(MaxNumStreams, tpinfo->max_streams);
+		else
+			max_streams = tpinfo->max_streams;
 		tpinfo->num_streams = usb_alloc_streams(iface,
 							&tpinfo->eps[1], 3,
-							tpinfo->max_streams,
+							max_streams,
 							GFP_KERNEL);
 		if (tpinfo->num_streams <= 0) {
 			dev_err(&udev->dev,
Signature · 1 line
1.7.0.1
#1 lore raw
Quoted 9 lines · Luben Tuikov

---
The long story is that we see some host controllers misreport their PSA as they solved 2^v = streams, instead of 2^(v+1) = streams. Thus They report that they support 32 streams when in fact they support 16. When the device attempts to return status for stream > 15, the host says ACK(NumP=0), the device goes in flow control, blah, blah, this module parameter allows you to set a max cap on the number of streams the driver will ask XHCI HCD to allocate.

If this is an issue with a host then the work-around should be in the xHCI driver instead. It should be based on the vendor/device ID of the offending host instead of being a module parameter in the UAS driver. The only people who benefit from this patch are the people "in the know" about which hosts are buggy, not normal Linux users.

Streams could be used for things other than UAS in the future, so the fix should really be in the xHCI host controller driver. And please put the information about the host bug in the commit message, so your private comments don't get lost in the bowels of the linux-usb mailing list.

In the future, please Cc me on any patches that involve xHCI, or bulk streams in general. I'm not an enemy; I frankly don't care which UAS driver gets in the kernel, I just want one good, well documented driver, and for the xHCI driver to be as good as possible.

Sarah Sharp

Quoted 59 lines
 drivers/usb/storage/uasp.c |   21 ++++++++++++++++++++-
 1 files changed, 20 insertions(+), 1 deletions(-)
drivers/usb/storage/uasp.c +20 -1
index 3b10efd..e5a26e9 100644
@@ -30,6 +30,19 @@
 #include <scsi/scsi_host.h>
 #include <scsi/scsi_tcq.h>
 
+int MaxNumStreams = -1;
+
+module_param(MaxNumStreams, int, 0444);
+MODULE_PARM_DESC(MaxNumStreams, "\n"
+	"\tSome host controllers and/or devices report a larger number of\n"
+	"\tstreams that they in fact support. This parameter allows you\n"
+	"\tto limit the number of streams this driver will request the XHCI\n"
+	"\tHCD to allocate. If set to -1, the default value, then this driver\n"
+	"\twill use the value reported by the attached device. Else the\n"
+	"\tnumber of streams will be limited to the minimum reported by the\n"
+	"\tattached device and this value. Valid values are -1, default,\n"
+	"\tand 1 to 0xFFEF.");
+
 /* Information unit types
  */
 #define IU_CMD   1
@@ -938,6 +951,8 @@ static int uasp_ep_conf(struct uasp_tport_info *tpinfo)
 					 tpinfo->eps[3]->desc.bEndpointAddress);
 
 	if (udev->speed == USB_SPEED_SUPER) {
+		int max_streams;
+
 		for (i = 1; i < 4; i++) {
 			if (tpinfo->max_streams == 0)
 				tpinfo->max_streams = USB_SS_MAX_STREAMS(tpinfo->eps[i]->ss_ep_comp.bmAttributes);
@@ -952,9 +967,13 @@ static int uasp_ep_conf(struct uasp_tport_info *tpinfo)
 		}
 
 		tpinfo->use_streams = 1;
+		if (1 <= MaxNumStreams && MaxNumStreams <= 0xFFEF)
+			max_streams = min(MaxNumStreams, tpinfo->max_streams);
+		else
+			max_streams = tpinfo->max_streams;
 		tpinfo->num_streams = usb_alloc_streams(iface,
 							&tpinfo->eps[1], 3,
-							tpinfo->max_streams,
+							max_streams,
 							GFP_KERNEL);
 		if (tpinfo->num_streams <= 0) {
 			dev_err(&udev->dev,
Signature · 7 lines
1.7.0.1


--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html
#2 lore raw
Quoted 17 lines · Sarah Sharp
Quoted 3 lines · Luben Tuikov

---
  The long story is that we see some host

controllers misreport their

Quoted 1 line

  PSA as they solved 2^v = streams, instead of

2^(v+1) = streams. Thus

Quoted 1 line

  They report that they support 32 streams when in

fact they support 16.

Quoted 1 line

  When the device attempts to return status for

stream > 15, the host

Quoted 1 line

  says ACK(NumP=0), the device goes in flow

control, blah, blah, this

Quoted 1 line

  module parameter allows you to set a max cap on

the number of streams

Quoted 1 line

  the driver will ask XHCI HCD to allocate.

Quoted 9 lines

If this is an issue with a host then the work-around should be in the
xHCI driver instead.  It should be based on the vendor/device ID of the
offending host instead of being a module parameter in the UAS driver.
The only people who benefit from this patch are the people "in the know"
about which hosts are buggy, not normal Linux users.

The OS might want to limit the number of streams for reasons other than a buggy HC. So this module parameter add flexibility other than making a buggy HC work.

(I didn't know you were reading my patches and yet to a "competing" driver of a driver where you're listed as a co-author...)

Quoted 9 lines

Streams could be used for things other than UAS in the future, so the
fix should really be in the xHCI host controller driver.  And please put
the information about the host bug in the commit message, so your
private comments don't get lost in the bowels of the linux-usb mailing
list.

I've an updated HC of this particular make sitting on my desk which I've not had time to test... The HC this was found in is an early version of that make.

Quoted 3 lines

In the future, please Cc me on any patches that involve xHCI, or bulk
streams in general.

In this case, you should probably not add 1 to the number of streams a protocol driver requests, if that number is already a power of 2. To extend this and make it foolproof, the XHCI should probably know if the number of streams came from a descriptor or the driver made it up (in which case you know more precisely what to do). Now, since you return the number of streams the driver can use, implementing the former is probably just enough.

Also please add a comment as to the return value therein, to the effect that the range of streams the caller can use is [1, the value returned], WLG.

Quoted 3 lines

I'm not an enemy; I frankly don't
care which UAS
driver gets in the kernel,

I see.

Quoted 3 lines

I just want one good, well
documented driver,
and for the xHCI driver to be as good as possible.

Uh-huh.

Quoted 104 lines

Sarah Sharp

Quoted 1 line

  drivers/usb/storage/uasp.c |   21

++++++++++++++++++++-

Quoted 1 line
  1 files changed, 20 insertions(+), 1

deletions(-)

Review

Quoted 2 lines
(quoted diff) +0 -0

b/drivers/usb/storage/uasp.c

drivers/usb/storage/uasp.c

Quoted 12 lines

index 3b10efd..e5a26e9 100644

drivers/usb/storage/uasp.c +0 -0
@@ -30,6 +30,19 @@
   #include <scsi/scsi_host.h>
   #include <scsi/scsi_tcq.h>

drivers/usb/storage/uasp.c +5 -0
+int MaxNumStreams = -1;
+
+module_param(MaxNumStreams, int, 0444);
+MODULE_PARM_DESC(MaxNumStreams, "\n"
+    "\tSome host controllers and/or

devices report a larger number of\n"

Quoted 1 line

+    "\tstreams that they in fact

support. This parameter allows you\n"

Quoted 1 line

+    "\tto limit the number of streams

this driver will request the XHCI\n"

Quoted 1 line

+    "\tHCD to allocate. If set to -1,

the default value, then this driver\n"

Quoted 1 line

+    "\twill use the value reported by

the attached device. Else the\n"

Quoted 1 line

+    "\tnumber of streams will be

limited to the minimum reported by the\n"

Quoted 1 line

+    "\tattached device and this value.

Valid values are -1, default,\n"

drivers/usb/storage/uasp.c static int uasp_ep_conf(struct

Quoted 6 lines

+    "\tand 1 to 0xFFEF.");
+
  /* Information unit types
   */
  #define IU_CMD   1

drivers/usb/storage/uasp.c +0 -0
@@ -938,6 +951,8 @@ static int uasp_ep_conf(struct

uasp_tport_info *tpinfo)

Quoted 1 line


     tpinfo->eps[3]->desc.bEndpointAddress);

Quoted 2 lines


      if (udev->speed ==

USB_SPEED_SUPER) {

Quoted 1 line

+        int

max_streams;

Quoted 2 lines

+
          for (i =

1; i < 4; i++) {

Quoted 1 line

    if (tpinfo->max_streams == 0)

Quoted 1 line

        tpinfo->max_streams
= USB_SS_MAX_STREAMS(tpinfo->eps[i]->ss_ep_comp.bmAttributes);

drivers/usb/storage/uasp.c static int uasp_ep_conf(struct

Quoted 1 line
drivers/usb/storage/uasp.c +0 -0
@@ -952,9 +967,13 @@ static int uasp_ep_conf(struct

uasp_tport_info *tpinfo)

Quoted 3 lines
drivers/usb/storage/uasp.c +0 -0
           }


tpinfo->use_streams = 1;

Quoted 1 line

+        if (1 <=

MaxNumStreams && MaxNumStreams <= 0xFFEF)

Quoted 1 line

+

    max_streams = min(MaxNumStreams,
tpinfo->max_streams);

Quoted 2 lines

+        else
+

    max_streams = tpinfo->max_streams;

Quoted 1 line

tpinfo->num_streams = usb_alloc_streams(iface,

Quoted 1 line



&tpinfo->eps[1], 3,

Quoted 1 line

-



tpinfo->max_streams,

Quoted 1 line

+


        max_streams,

Quoted 1 line


        GFP_KERNEL);

Quoted 1 line

          if

(tpinfo->num_streams <= 0) {

Quoted 1 line

    dev_err(&udev->dev,

Quoted 6 lines
Signature · 5 lines
1.7.0.1


--
To unsubscribe from this list: send the line

"unsubscribe linux-usb" in

Quoted 2 lines

the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

#3 lore raw
Quoted 28 lines · Luben Tuikov
Quoted 17 lines · Sarah Sharp
Quoted 3 lines · Luben Tuikov

---
  The long story is that we see some host

controllers misreport their

Quoted 1 line

  PSA as they solved 2^v = streams, instead of

2^(v+1) = streams. Thus

Quoted 1 line

  They report that they support 32 streams when in

fact they support 16.

Quoted 1 line

  When the device attempts to return status for

stream > 15, the host

Quoted 1 line

  says ACK(NumP=0), the device goes in flow

control, blah, blah, this

Quoted 1 line

  module parameter allows you to set a max cap on

the number of streams

Quoted 1 line

  the driver will ask XHCI HCD to allocate.

Quoted 9 lines

If this is an issue with a host then the work-around should be in the
xHCI driver instead.  It should be based on the vendor/device ID of the
offending host instead of being a module parameter in the UAS driver.
The only people who benefit from this patch are the people "in the know"
about which hosts are buggy, not normal Linux users.

Quoted 3 lines

The OS might want to limit the number of streams for reasons other than a buggy HC. So this module parameter add flexibility other than making a buggy HC work.

Ok, that's fine, but my comments still stand about fixing this particular bug in the xHCI driver, rather than in the UAS driver.

Quoted 2 lines

(I didn't know you were reading my patches and yet to a "competing" driver of a driver where you're listed as a co-author...)

I always read emails that seem like they might be xHCI related. Especially when they're a patch.

Quoted 9 lines
Quoted 5 lines

Streams could be used for things other than UAS in the future, so the fix should really be in the xHCI host controller driver.  And please put the information about the host bug in the commit message, so your private comments don't get lost in the bowels of the linux-usb mailing list.

I've an updated HC of this particular make sitting on my desk which I've not had time to test... The HC this was found in is an early version of that make.

If the earlier version of the HC is already out in the market, then the fix needs to get into the xHCI driver, regardless of whether the new version works. Otherwise some poor user is going to plug in their off-the-shelf UAS device into their off-the-shelf host controller and wonder why it doesn't work.

Quoted 5 lines
Quoted 2 lines

In the future, please Cc me on any patches that involve xHCI, or bulk streams in general.

In this case, you should probably not add 1 to the number of streams a protocol driver requests, if that number is already a power of 2.

Why? What if the host and endpoints can handle 32 streams, but the driver requests 16? Is there a reason the xHCI driver shouldn't allocate transfer rings for streams 1 through 16? Something special to do with powers of two?

There's a deeper design decision here that I never really resolved. Say a driver asks for 16 streams. Stream 0 is reserved, and can't be used after streams are enabled. The xHCI host controller spec forbids allocating a transfer ring for stream 0.

Now, what should the contract between the host controller and the driver be? Should the xHCI driver allocate stream rings for stream IDs 1 through 15, thereby guaranteeing 15 streams can be used? Or should it allocate streams 1 through 16, giving the driver the 16 active streams it actually wanted? I chose the latter, which is why I add 1 to the number of streams the driver requests.

Quoted 5 lines

To extend this and make it foolproof, the XHCI should probably know if the number of streams came from a descriptor or the driver made it up (in which case you know more precisely what to do). Now, since you return the number of streams the driver can use, implementing the former is probably just enough.

In either case, I look at the endpoint descriptors that the driver passes me, and never allocate more streams than those endpoints can handle. The driver also never allocates more streams than the host controller can handle. All that work is done in xhci_calculate_streams_and_bitmask().

Quoted 3 lines

Also please add a comment as to the return value therein, to the effect that the range of streams the caller can use is [1, the value returned], WLG.

It's documented in Documentation/usb/bulk-streams.txt under the "Picking new Stream IDs to use" section. But yes, it should probably be documented in the function comments above usb_alloc_streams().

Sarah Sharp

#4 lore raw
Quoted 29 lines · Sarah Sharp
Quoted 28 lines · Luben Tuikov
Quoted 26 lines · Sarah Sharp
Quoted 3 lines · Luben Tuikov

---
  The long story is that we see some host

controllers misreport their

Quoted 1 line

  PSA as they solved 2^v = streams, instead of

2^(v+1) = streams. Thus

Quoted 1 line

  They report that they support 32 streams when in

fact they support 16.

Quoted 1 line

  When the device attempts to return status for

stream > 15, the host

Quoted 1 line

  says ACK(NumP=0), the device goes in flow

control, blah, blah, this

Quoted 1 line

  module parameter allows you to set a max cap on

the number of streams

Quoted 1 line

  the driver will ask XHCI HCD to allocate.

If this is an issue with a host then the work-around should be in the
xHCI driver instead.  It should be based on the vendor/device ID of the
offending host instead of being a module parameter in the UAS driver.
The only people who benefit from this patch are the people "in the know"
about which hosts are buggy, not normal Linux users.

Quoted 6 lines
Quoted 3 lines

The OS might want to limit the number of streams for reasons other than a buggy HC. So this module parameter add flexibility other than making a buggy HC work.

Ok, that's fine, but my comments still stand about fixing this particular bug in the xHCI driver, rather than in the UAS driver.

I agree.

Also, a module parameter will not work when your driver is connecting to different xHCI devices in the same system. Please never add module parameters anymore, there is usually never a need for them at all.

thanks,

greg k-h

#5 lore raw
Quoted 55 lines · Greg KH
Quoted 1 line · Sarah Sharp

On Thu, Dec 09, 2010 at 02:54:19PM -0800, Luben Tuikov

Quoted 1 line · wrote:
Quoted 1 line

--- On Thu, 12/9/10, Sarah Sharp <[email protected]>

Quoted 4 lines · wrote:
Quoted 4 lines
Quoted 4 lines
Quoted 3 lines · Luben Tuikov

---
  The long story is that we see some

host

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

controllers misreport their

Quoted 1 line

  PSA as they solved 2^v = streams,

instead of

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

2^(v+1) = streams. Thus

Quoted 1 line

  They report that they support 32

streams when in

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

fact they support 16.

Quoted 1 line

  When the device attempts to return

status for

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

stream > 15, the host

Quoted 1 line

  says ACK(NumP=0), the device goes in

flow

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

control, blah, blah, this

Quoted 1 line

  module parameter allows you to set a

max cap on

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

the number of streams

Quoted 1 line

  the driver will ask XHCI HCD to

allocate.

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

If this is an issue with a host then the

work-around should

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

be in the
xHCI driver instead.  It should be based on

the

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

vendor/device ID of the
offending host instead of being a module

parameter in the

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

UAS driver.
The only people who benefit from this patch

are the people

Quoted 2 lines
Quoted 2 lines
Quoted 2 lines

"in the know"
about which hosts are buggy, not normal

Linux users.

Quoted 2 lines
Quoted 2 lines

The OS might want to limit the number of streams

for reasons other

Quoted 1 line
Quoted 1 line

than a buggy HC. So this module parameter add

flexibility other than

Quoted 3 lines
Quoted 1 line

making a buggy HC work.

Ok, that's fine, but my comments still stand about

fixing this

Quoted 1 line

particular bug in the xHCI driver, rather than in the

UAS driver.

Quoted 8 lines

I agree.

Also, a module parameter will not work when your driver is connecting to
different xHCI devices in the same system.  Please never add module
parameters anymore, there is usually never a need for them at all.

I hesitated a lot before adding this module parameter. I did not want to add it since of course the obvious solution is to have a "black list" in the XHCI HCD. I think everyone knows this.

But at the time and still in some ways, it seemed like a valuable addition. Of course, when not set (default behavior) it doesn't impact the operation of the driver.

Luben

#6 lore raw
In reply to Luben Tuikov
Quoted 5 lines
Quoted 3 lines

The only people who benefit from this patch are the people "in the know"
about which hosts are buggy, not normal Linux users.

The OS might want to limit the number of streams for reasons other than a buggy HC. So this module parameter add flexibility other than making a buggy HC work.

Not of any use to anyone. Any stream limit stuff neees to b done automatically otherwise it'll be of value to maybe 1 person if that. It'll also never get tested so will in time rot, produce weird bugs or show up locking horrors.

Alan

#7 lore raw
In reply to Sarah Sharp
Quoted 7 lines · Sarah Sharp

In the future, please Cc me on any patches that involve xHCI, or bulk
streams in general.  I'm not an enemy; I frankly don't care which UAS
driver gets in the kernel, I just want one good, well documented driver,
and for the xHCI driver to be as good as possible.

Then do the noble thing and retract the still-born uas.c where you're listed as a co-author.

Luben

#8 lore raw
Quoted 11 lines · Luben Tuikov
Quoted 7 lines · Sarah Sharp

In the future, please Cc me on any patches that involve xHCI, or bulk
streams in general.  I'm not an enemy; I frankly don't care which UAS
driver gets in the kernel, I just want one good, well documented driver,
and for the xHCI driver to be as good as possible.

Then do the noble thing and retract the still-born uas.c where you're listed as a co-author.

That's just plain rude and obnoxious.

You just earned a place in my kill-file.

*plonk*

#9 lore raw
Quoted 6 lines · Greg KH

That's just plain rude and obnoxious.

You just earned a place in my kill-file.

*plonk*

But, Greg, it's what I would've done. If there is a better solution out there, let it be, instead of leading on this charade of your protecting someone else's interests.