Mouse stuck after save/restore of guest.

PATCH In review Igor Mammedov 4 participants 11 messages Apr 14, 2011 – Apr 27, 2011

Subsystems (2)

Input drivers Xen hypervisor interface

lore

Share & tools

lore

Igor Mammedov

author

2011-04-14 15:45 UTC #0 lore raw

PATCH 1/1 Mouse stuck after save/restore of guest.

Mouse stuck after restore of PV guest but buttons are in working condition.
If driver has been configured for ABS coordinates at start it will get XENKBD_TYPE_POS events and then suddenly after restore it'll start getting XENKBD_TYPE_MOTION events, that will be dropped later and they won't get into user-space.

Regression was introduced by hunk 5 and 6 of 5ea5254 in upstream.

Driver on restore should ask xen for request-abs-pointer again if it's available. So restore parts that did it before 5ea5254.

---
 drivers/input/xen-kbdfront.c |   14 +++++++++++++-
 1 files changed, 13 insertions(+), 1 deletions(-)

drivers/input/xen-kbdfront.c +13 -1

index 53e6273..71e2fc5 100644
@@ -286,7 +286,7 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 				   enum xenbus_state backend_state)
 {
 	struct xenkbd_info *info = dev_get_drvdata(&dev->dev);
-	int val;
+	int ret, val;
 
 	switch (backend_state) {
 	case XenbusStateInitialising:

@@ -299,6 +299,18 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 
 	case XenbusStateInitWait:
 InitWait:
+		ret = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+                                   "feature-abs-pointer", "%d", &val);
+                if (ret < 0)
+                        val = 0;
+                if (val) {
+                        ret = xenbus_printf(XBT_NIL, info->xbdev->nodename,
+                                            "request-abs-pointer", "1");
+                        if (ret)
+                                printk(KERN_WARNING
+                                       "xenkbd: can't request abs-pointer");
+                }
+
 		xenbus_switch_state(dev, XenbusStateConnected);
 		break;

Signature · 1 line

1.7.1

Konrad Rzeszutek Wilk

2011-04-14 15:59 UTC #1 lore raw

Quoted 8 lines · Igor Mammedov

Mouse stuck after restore of PV guest but buttons are in working condition.
If driver has been configured for ABS coordinates at start it will get XENKBD_TYPE_POS events and then suddenly after restore it'll start getting XENKBD_TYPE_MOTION events, that will be dropped later and they won't get into user-space.

Quoted 6 lines

Regression was introduced by hunk 5 and 6 of 5ea5254 in upstream.
Driver on restore should ask xen for request-abs-pointer again if it's available. So restore parts that did it before 5ea5254.

Olaf?

Quoted 21 lines

---
 drivers/input/xen-kbdfront.c |   14 +++++++++++++-
 1 files changed, 13 insertions(+), 1 deletions(-)

drivers/input/xen-kbdfront.c +1 -1

index 53e6273..71e2fc5 100644
@@ -286,7 +286,7 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 				   enum xenbus_state backend_state)
 {
 	struct xenkbd_info *info = dev_get_drvdata(&dev->dev);
-	int val;
+	int ret, val;
 
 	switch (backend_state) {
 	case XenbusStateInitialising:

@@ -299,6 +299,18 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 
 	case XenbusStateInitWait:
 InitWait:

drivers/input/xen-kbdfront.c static void xenkbd_backend_changed(struct xenbus_device *dev, · line 311

Quoted 10 lines

drivers/input/xen-kbdfront.c +10 -0

+		ret = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+                                   "feature-abs-pointer", "%d", &val);
+                if (ret < 0)
+                        val = 0;
+                if (val) {
+                        ret = xenbus_printf(XBT_NIL, info->xbdev->nodename,
+                                            "request-abs-pointer", "1");
+                        if (ret)
+                                printk(KERN_WARNING
+                                       "xenkbd: can't request abs-pointer");

Any reason for not using the pr_warning as it was before?

Quoted 13 lines

drivers/input/xen-kbdfront.c +2 -0

+                }
+
 		xenbus_switch_state(dev, XenbusStateConnected);
 		break;

Signature · 7 lines

1.7.1


_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

Olaf Hering

2011-04-14 17:26 UTC #2 lore raw

Quoted 9 lines · Konrad Rzeszutek Wilk

Quoted 8 lines · Igor Mammedov
Mouse stuck after restore of PV guest but buttons are in working condition.
If driver has been configured for ABS coordinates at start it will get XENKBD_TYPE_POS events and then suddenly after restore it'll start getting XENKBD_TYPE_MOTION events, that will be dropped later and they won't get into user-space.

Quoted 8 lines

Quoted 6 lines
Regression was introduced by hunk 5 and 6 of 5ea5254 in upstream.
Driver on restore should ask xen for request-abs-pointer again if it's available. So restore parts that did it before 5ea5254.
Olaf?

This change is correct. Thanks for spotting, Igor.
I did not test suspend/resume, migration etc. Sorry for that. During resume the guest has to renegotiate with the new qemu-dm process.

Olaf

Konrad Rzeszutek Wilk

Acked-by Olaf Hering

2011-04-14 18:38 UTC #3 lore raw

Quoted 14 lines · Olaf Hering

Quoted 12 lines · Konrad Rzeszutek Wilk
Quoted 11 lines · Igor Mammedov
Mouse stuck after restore of PV guest but buttons are in working condition.
If driver has been configured for ABS coordinates at start it will get XENKBD_TYPE_POS events and then suddenly after restore it'll start getting XENKBD_TYPE_MOTION events, that will be dropped later and they won't get into user-space.
Regression was introduced by hunk 5 and 6 of 5ea5254 in upstream.

Quoted 7 lines

Quoted 5 lines
Quoted 3 lines
Driver on restore should ask xen for request-abs-pointer again if it's available. So restore parts that did it before 5ea5254.
Olaf?
This change is correct. Thanks for spotting, Igor.

Dmitry,

Was wondering if you are OK pushing this for 2.6.39-rc3 or whether you are OK with me doing. It fixes a regression introduced by the last hunk of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db

The patch rebase on top of 2.6.39-rc3 looks as so:

(and the patch is in git://git.kernel.org/pub/scm/linux/kernel/git/konrad/xen.git stable/bug-fixes-rc3) commit 876f162fe6831273fc5d678fa066f26f58be4a2c

Author: Igor Mammedov <[email protected]>
Date:   Thu Apr 14 17:45:07 2011 +0200

Input: xen-kbdfront: Mouse stuck after save/restore of guest.

Mouse stuck after restore of PV guest but buttons are in working condition.

If driver has been configured for ABS coordinates at start it will get XENKBD_TYPE_POS events and then suddenly after restore it'll start getting XENKBD_TYPE_MOTION events, that will be dropped later and they won't get into user-space.

Regression was introduced by hunk 5 and 6 of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db
("Input: xen-kbdfront - advertise either absolute or relative coordinates"

Driver on restore should ask xen for request-abs-pointer again if it's available. So restore parts that did it before 5ea5254.

CC Dmitry Torokhov <[email protected]>

Acked-by Olaf Hering <[email protected]>

Signed-off-by Igor Mammedov <[email protected]>

[v1: Expanded the commit description]

Signed-off-by Konrad Rzeszutek Wilk <[email protected]>

drivers/input/misc/xen-kbdfront.c +12 -1

index 7077f9b..6d119d5 100644
@@ -303,7 +303,7 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 				   enum xenbus_state backend_state)
 {
 	struct xenkbd_info *info = dev_get_drvdata(&dev->dev);
-	int val;
+	int ret, val;
 
 	switch (backend_state) {
 	case XenbusStateInitialising:

@@ -316,6 +316,17 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 
 	case XenbusStateInitWait:
 InitWait:
+		ret = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+                                   "feature-abs-pointer", "%d", &val);
+                if (ret < 0)
+                        val = 0;
+                if (val) {
+                        ret = xenbus_printf(XBT_NIL, info->xbdev->nodename,
+                                            "request-abs-pointer", "1");
+                        if (ret)
+                                pr_warning("xenkbd: can't request abs-pointer");
+                }
+
 		xenbus_switch_state(dev, XenbusStateConnected);
 		break;

Olaf Hering

2011-04-15 09:37 UTC #4 lore raw

Quoted 21 lines · Konrad Rzeszutek Wilk

Quoted 20 lines · Olaf Hering
Quoted 17 lines · Konrad Rzeszutek Wilk
Quoted 8 lines · Igor Mammedov
Mouse stuck after restore of PV guest but buttons are in working condition.
If driver has been configured for ABS coordinates at start it will get XENKBD_TYPE_POS events and then suddenly after restore it'll start getting XENKBD_TYPE_MOTION events, that will be dropped later and they won't get into user-space.
Quoted 6 lines
Regression was introduced by hunk 5 and 6 of 5ea5254 in upstream.
Driver on restore should ask xen for request-abs-pointer again if it's available. So restore parts that did it before 5ea5254.
Olaf?

Quoted 7 lines

Quoted 1 line
This change is correct. Thanks for spotting, Igor.
Dmitry,
Was wondering if you are OK pushing this for 2.6.39-rc3 or whether you are OK with me doing. It fixes a regression introduced by the last hunk of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db

Looking through my backlog, I noticed that the patch went also to [email protected] and is now in various trees already. So please forward this revert also to [email protected]

Olaf

Konrad Rzeszutek Wilk

2011-04-19 13:46 UTC #5 lore raw

Quoted 7 lines

Quoted 3 lines
Was wondering if you are OK pushing this for 2.6.39-rc3 or whether you are OK with me doing. It fixes a regression introduced by the last hunk of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db
Looking through my backlog, I noticed that the patch went also to [email protected] and is now in various trees already. So please forward this revert also to [email protected]

OK. The upstream git commit is c36b58e8a9112017c2bcc322cc98e71241814303.

Which stable tree did it go through? 2.6.38 only I presume?

Olaf Hering

2011-04-19 14:55 UTC #6 lore raw

Quoted 11 lines · Konrad Rzeszutek Wilk

Quoted 7 lines
Quoted 3 lines
Was wondering if you are OK pushing this for 2.6.39-rc3 or whether you are OK with me doing. It fixes a regression introduced by the last hunk of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db
Looking through my backlog, I noticed that the patch went also to [email protected] and is now in various trees already. So please forward this revert also to [email protected]
OK. The upstream git commit is c36b58e8a9112017c2bcc322cc98e71241814303.
Which stable tree did it go through? 2.6.38 only I presume?

All longterm versions, 2.6.32 and up.

Olaf

Igor Mammedov

author

2011-04-20 09:19 UTC #7 lore raw

Quoted 5 lines · Olaf Hering

Quoted 3 lines · Konrad Rzeszutek Wilk
Quoted 3 lines
Quoted 3 lines
Was wondering if you are OK pushing this for 2.6.39-rc3 or whether you are OK with me doing. It fixes a regression introduced by the last hunk of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db

Quoted 7 lines

Quoted 6 lines
Quoted 3 lines
Looking through my backlog, I noticed that the patch went also to [email protected] and is now in various trees already. So please forward this revert also to [email protected]
OK. The upstream git commit is c36b58e8a9112017c2bcc322cc98e71241814303.
Which stable tree did it go through? 2.6.38 only I presume?
All longterm versions, 2.6.32 and up.

In Fedora 13 it still works (2.6.34.*)

Quoted 6 lines

Olaf
_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

Konrad Rzeszutek Wilk

Acked-by Olaf Hering Acked-by Olaf Hering

2011-04-20 18:29 UTC #8 lore raw

Re: [Xen-devel] [PATCH] [PV Xen] Mouse stuck after save/restore of guest. - stable tree candidate.

In reply to Olaf Hering

Quoted 18 lines

Quoted 18 lines
Quoted 18 lines
Quoted 17 lines · Igor Mammedov
Quoted 8 lines
Mouse stuck after restore of PV guest but buttons are in working condition.
If driver has been configured for ABS coordinates at start it will get XENKBD_TYPE_POS events and then suddenly after restore it'll start getting XENKBD_TYPE_MOTION events, that will be dropped later and they won't get into user-space.
Quoted 6 lines
Regression was introduced by hunk 5 and 6 of 5ea5254 in upstream.
Driver on restore should ask xen for request-abs-pointer again if it's available. So restore parts that did it before 5ea5254.
Olaf?

Quoted 7 lines

Quoted 7 lines
Quoted 1 line
This change is correct. Thanks for spotting, Igor.
Dmitry,
Was wondering if you are OK pushing this for 2.6.39-rc3 or whether you are OK with me doing. It fixes a regression introduced by the last hunk of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db

Hello stable-tree maintainers.

Please apply the upstream git commit c36b58e8a9112017c2bcc322cc98e71241814303 (Input: xen-kbdfront - fix mouse getting stuck after save/restore) which fixes a regression introduced by 8c3c283e6bf463ab498d6e7823aff6c4762314b6 (Input: xen-kbdfront - advertise either absolute or relative coordinates)

The 2.6.38 and 2.6.37 tree both contain the regression.

For convience, attached and inline is the patch for stable tree submission.

commit c36b58e8a9112017c2bcc322cc98e71241814303

Author: Igor Mammedov <[email protected]>
Date:   Mon Apr 18 10:17:17 2011 -0700

Input: xen-kbdfront - fix mouse getting stuck after save/restore

Mouse gets "stuck" after restore of PV guest but buttons are in working condition.

Regression was introduced by hunk 5 and 6 of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db
("Input: xen-kbdfront - advertise either absolute or relative coordinates").

Driver on restore should ask xen for request-abs-pointer again if it is available. So restore parts that did it before 5ea5254.

Acked-by Olaf Hering <[email protected]>

Signed-off-by Igor Mammedov <[email protected]>

[v1: Expanded the commit description]

Signed-off-by Konrad Rzeszutek Wilk <[email protected]>

Signed-off-by Dmitry Torokhov <[email protected]>

drivers/input/misc/xen-kbdfront.c +12 -1

index 7077f9b..62bae99 100644
@@ -303,7 +303,7 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 				   enum xenbus_state backend_state)
 {
 	struct xenkbd_info *info = dev_get_drvdata(&dev->dev);
-	int val;
+	int ret, val;
 
 	switch (backend_state) {
 	case XenbusStateInitialising:

@@ -316,6 +316,17 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 
 	case XenbusStateInitWait:
 InitWait:
+		ret = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+				   "feature-abs-pointer", "%d", &val);
+		if (ret < 0)
+			val = 0;
+		if (val) {
+			ret = xenbus_printf(XBT_NIL, info->xbdev->nodename,
+					    "request-abs-pointer", "1");
+			if (ret)
+				pr_warning("xenkbd: can't request abs-pointer");
+		}
+
 		xenbus_switch_state(dev, XenbusStateConnected);
 		break;

commit c36b58e8a9112017c2bcc322cc98e71241814303

Author: Igor Mammedov <[email protected]>
Date:   Mon Apr 18 10:17:17 2011 -0700

Input: xen-kbdfront - fix mouse getting stuck after save/restore

Mouse gets "stuck" after restore of PV guest but buttons are in working condition.

Regression was introduced by hunk 5 and 6 of 5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db
("Input: xen-kbdfront - advertise either absolute or relative coordinates").

Driver on restore should ask xen for request-abs-pointer again if it is available. So restore parts that did it before 5ea5254.

Acked-by Olaf Hering <[email protected]>

Signed-off-by Igor Mammedov <[email protected]>

[v1: Expanded the commit description]

Signed-off-by Konrad Rzeszutek Wilk <[email protected]>

Signed-off-by Dmitry Torokhov <[email protected]>

drivers/input/misc/xen-kbdfront.c +12 -1

index 7077f9b..62bae99 100644
@@ -303,7 +303,7 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 				   enum xenbus_state backend_state)
 {
 	struct xenkbd_info *info = dev_get_drvdata(&dev->dev);
-	int val;
+	int ret, val;
 
 	switch (backend_state) {
 	case XenbusStateInitialising:

@@ -316,6 +316,17 @@ static void xenkbd_backend_changed(struct xenbus_device *dev,
 
 	case XenbusStateInitWait:
 InitWait:
+		ret = xenbus_scanf(XBT_NIL, info->xbdev->otherend,
+				   "feature-abs-pointer", "%d", &val);
+		if (ret < 0)
+			val = 0;
+		if (val) {
+			ret = xenbus_printf(XBT_NIL, info->xbdev->nodename,
+					    "request-abs-pointer", "1");
+			if (ret)
+				pr_warning("xenkbd: can't request abs-pointer");
+		}
+
 		xenbus_switch_state(dev, XenbusStateConnected);
 		break;

_______________________________________________
Xen-devel mailing list
[email protected]
http://lists.xensource.com/xen-devel

Olaf Hering

2011-04-27 16:31 UTC #9 lore raw

Re: [Xen-devel] [PATCH] [PV Xen] Mouse stuck after save/restore of guest. - stable tree candidate.

Quoted 8 lines · Konrad Rzeszutek Wilk

Hello stable-tree maintainers.
Please apply the upstream git commit c36b58e8a9112017c2bcc322cc98e71241814303 (Input: xen-kbdfront - fix mouse getting stuck after save/restore) which fixes a regression introduced by 8c3c283e6bf463ab498d6e7823aff6c4762314b6 (Input: xen-kbdfront - advertise either absolute or relative coordinates)
The 2.6.38 and 2.6.37 tree both contain the regression.

I checked which trees contain the broken commit from mainline:

broken stable trees:
2.6.38
http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.38.y.git;a=commit;h=5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db
2.6.37
http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.37.y.git;a=commit;h=723517a796adf59f9ae628629fcb606c4cec8715
2.6.35
http://git.kernel.org/?p=linux/kernel/git/longterm/linux-2.6.35.y.git;a=commit;h=b88505e58ce6d36732cd15841cc821f97efca9bb
2.6.33
http://git.kernel.org/?p=linux/kernel/git/longterm/linux-2.6.33.y.git;a=commit;h=6997348a861d3540085bf9adf4166b86ac7a96e5
2.6.32
http://git.kernel.org/?p=linux/kernel/git/longterm/linux-2.6.32.y.git;a=commit;h=95e7148d08d0578687ea26cd3a8925aa5f7bc8fa
2.6.27
Does not have a variant of 8c3c283e6bf463ab498d6e7823aff6c4762314b6, yet.

So please apply upstream commit c36b58e8a9112017c2bcc322cc98e71241814303 to 2.6.38, 2.6.37, 2.6.35, 2.6.33 and 2.6.32 to fix the introduced regression.

Thanks.

Olaf

Greg KH

2011-04-27 23:32 UTC #10 lore raw

Re: [stable] [Xen-devel] [PATCH] [PV Xen] Mouse stuck after save/restore of guest. - stable tree candidate.

Quoted 27 lines · Olaf Hering

Quoted 8 lines · Konrad Rzeszutek Wilk
Hello stable-tree maintainers.
Please apply the upstream git commit c36b58e8a9112017c2bcc322cc98e71241814303 (Input: xen-kbdfront - fix mouse getting stuck after save/restore) which fixes a regression introduced by 8c3c283e6bf463ab498d6e7823aff6c4762314b6 (Input: xen-kbdfront - advertise either absolute or relative coordinates)
The 2.6.38 and 2.6.37 tree both contain the regression.
I checked which trees contain the broken commit from mainline:
broken stable trees:
2.6.38
http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.38.y.git;a=commit;h=5ea5254aa0ad269cfbd2875c973ef25ab5b5e9db
2.6.37
http://git.kernel.org/?p=linux/kernel/git/stable/linux-2.6.37.y.git;a=commit;h=723517a796adf59f9ae628629fcb606c4cec8715
2.6.35
http://git.kernel.org/?p=linux/kernel/git/longterm/linux-2.6.35.y.git;a=commit;h=b88505e58ce6d36732cd15841cc821f97efca9bb
2.6.33
http://git.kernel.org/?p=linux/kernel/git/longterm/linux-2.6.33.y.git;a=commit;h=6997348a861d3540085bf9adf4166b86ac7a96e5
2.6.32
http://git.kernel.org/?p=linux/kernel/git/longterm/linux-2.6.32.y.git;a=commit;h=95e7148d08d0578687ea26cd3a8925aa5f7bc8fa
2.6.27
Does not have a variant of 8c3c283e6bf463ab498d6e7823aff6c4762314b6, yet.

Quoted 3 lines

So please apply upstream commit c36b58e8a9112017c2bcc322cc98e71241814303 to 2.6.38, 2.6.37, 2.6.35, 2.6.33 and 2.6.32 to fix the introduced regression.

That commit fails to apply on the .38-stable tree, or anything else I tried. Please provide a backported version if you wish to have it applied.

Note that the .37-stable tree is now closed, so that doesn't matter anymore.

thanks,

greg k-h