Success with kernel-swsusp and Nvidia-closed-source

Success with kernel-swsusp and Nvidia-closed-source

[liste-9nAOAgdJVo4b1SvskN2V4Q]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 1228 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salut Folks,
just had a bit of spare time to re-iterate the test, if my Acer TM634 
(with it's Nvidia 2 Go) would now be ready for suspend-to-disk.

Since some time ago, it is no problem to suspend to disk, if using the 
open-source (i.e. nv) driver for X. As it has problems with perfomance, 
e.g. DVD-playback is so slow, that the audio stops playing... I need to 
use the closed source driver.
With plain Linux 2.6.9 it nearly worked. It reported:
Could not suspend device 0000:01:00.0: error -1

After a quick search on Nvidia's forums I steped over Haggis dirty hack 
for the NVIDIA-Linux-x86-1.0-6111-pkg1. I attached this. Now suspending to 
disk works relyably for me.

Cheers
   hartwig felger

Hartwig Felger informatics
P.S.: with mplayer I use the gl2 device with the nvidia-driver.
- -- 
1024D/339FD693 Hartwig Felger <hgfelger-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org>
Key fingerprint = FB2F 3EE9 345A D55B 6FF2  0EC1 F5B0 684F 339F D693
For the pulic keys, please visit my page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBiMBD9bBoTzOf1pMRAj1cAKDjGomgzQFyoxXBQ+lhQODItBRdTwCg0s6R
tQp8JIj4hrb+4Wn8Sr2HAjE=
=/olf
-----END PGP SIGNATURE-----

[-- Attachment #2: Type: TEXT/PLAIN, Size: 1678 bytes --]

--- NVIDIA-Linux-x86-1.0-6111-pkg1/usr/src/nv/nv.c.ori	2004-11-02 21:03:38.000000000 +0100
+++ NVIDIA-Linux-x86-1.0-6111-pkg1/usr/src/nv/nv.c	2004-11-02 21:03:38.000000000 +0100
@@ -19,6 +19,9 @@
 MODULE_ALIAS_CHARDEV_MAJOR(NV_MAJOR_DEVICE_NUMBER);
 #endif
 
+/* fix for 2.6.9-rc# kernels where __VMALLOC_RESERVE is not exported */
+#define __VMALLOC_RESERVE (128 << 20)
+
 /*
  * our global state; one per device
  */
@@ -3694,19 +3697,24 @@
 
     switch (state)
     {
-        case PM_SAVE_STATE:
-            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event\n");
+      //        case PM_SAVE_STATE:
+      //            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event\n");
+        case NV_PM_ACPI_STANDBY:
+            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event: %d\n", state);
             status = rm_power_management(nv, 0, NV_PM_ACPI_STANDBY);
             break;
 
-        case PM_RESUME:
-            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event\n");
+	    //        case PM_RESUME:
+	    //           nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event\n");
+
+     default:
+        nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event: %d\n", state);
             status = rm_power_management(nv, 0, NV_PM_ACPI_RESUME);
             break;
 
-        default:
-            nv_printf(NV_DBG_WARNINGS, "NVRM: ACPI: unsupported event: %d\n", state);
-            return -1;
+	    //        default:
+	    //            nv_printf(NV_DBG_WARNINGS, "NVRM: ACPI: unsupported event: %d\n", state);
+	    //            return -1;
     }
 
     if (status != RM_OK)

Re: Success with kernel-swsusp and Nvidia-closed-source

[Christian Hesse]

[-- Attachment #1: Type: text/plain, Size: 1215 bytes --]

On Wednesday 03 November 2004 12:25, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> Salut Folks,
> just had a bit of spare time to re-iterate the test, if my Acer TM634
> (with it's Nvidia 2 Go) would now be ready for suspend-to-disk.
>
> Since some time ago, it is no problem to suspend to disk, if using the
> open-source (i.e. nv) driver for X. As it has problems with perfomance,
> e.g. DVD-playback is so slow, that the audio stops playing... I need to
> use the closed source driver.
> With plain Linux 2.6.9 it nearly worked. It reported:
> Could not suspend device 0000:01:00.0: error -1
>
> After a quick search on Nvidia's forums I steped over Haggis dirty hack
> for the NVIDIA-Linux-x86-1.0-6111-pkg1. I attached this. Now suspending to
> disk works relyably for me.
>
> Cheers
>    hartwig felger
>
> Hartwig Felger informatics
> P.S.: with mplayer I use the gl2 device with the nvidia-driver.

I'm using software suspend 2.1. For me suspend works fine, but the system 
freezes as soon as it is resumed, I only see a black screen then. Driver nv 
works without any problems. Any hints what could be the cause?

-- 
Christian Hesse

geek by nature
linux by choice

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]

Re: Success with kernel-swsusp and Nvidia-closed-source

[liste-9nAOAgdJVo4b1SvskN2V4Q]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salut Christian,
On Wed, 3 Nov 2004, Christian Hesse wrote:
> I'm using software suspend 2.1. For me suspend works fine, but the system
> freezes as soon as it is resumed, I only see a black screen then. Driver nv
> works without any problems. Any hints what could be the cause?
I did not patch in anything into the kernel. I used a vanilla 2.6.9 (only 
the Makefile was patched to show 2.6.9 instead of 8 ... maybe somebody 
watched a bit too much tv... :-)
So I did not use swsusp 2.1. And I could not manage to freeze on resume 
this time. So, why didn't you try the vanilla-kernel, with the 
Haggi' s patch for the nvidia driver?
When you test - be sure to sync before suspending. It's the second best, 
besides having a jung backup handy.

Cheers
   hartwig felger

Hartwig Felger informatics

- -- 
1024D/339FD693 Hartwig Felger <hgfelger-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org>
Key fingerprint = FB2F 3EE9 345A D55B 6FF2  0EC1 F5B0 684F 339F D693
For the pulic keys, please visit my page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBiVvr9bBoTzOf1pMRAlyYAKDMN861PRr34cB1fwhpq/l+bWDpkACeJ2yO
NJqtHOq5blb98z1bkTyD1x0=
=gBOM
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click

Re: Success with kernel-swsusp and Nvidia-closed-source

[Nigel Cunningham]

Hi.

On Thu, 2004-11-04 at 09:29, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
> 
> Salut Christian,
> On Wed, 3 Nov 2004, Christian Hesse wrote:
> > I'm using software suspend 2.1. For me suspend works fine, but the system
> > freezes as soon as it is resumed, I only see a black screen then. Driver nv
> > works without any problems. Any hints what could be the cause?
> I did not patch in anything into the kernel. I used a vanilla 2.6.9 (only 
> the Makefile was patched to show 2.6.9 instead of 8 ... maybe somebody 
> watched a bit too much tv... :-)
> So I did not use swsusp 2.1. And I could not manage to freeze on resume 
> this time. So, why didn't you try the vanilla-kernel, with the 
> Haggi' s patch for the nvidia driver?
> When you test - be sure to sync before suspending. It's the second best, 
> besides having a jung backup handy.

I guess I'm doing something driver related that Pavel isn't or vv :>.
Can I get a pointer to the driver source?

Regards,

Nigel
-- 
Nigel Cunningham
Pastoral Worker
Christian Reformed Church of Tuggeranong
PO Box 1004, Tuggeranong, ACT 2901

Everyone lives by faith. Some people just don't believe it.
Want proof? Try to prove that the theory of evolution is true.



-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click

Re: Success with kernel-swsusp and Nvidia-closed-source

[Christian Hesse]

[-- Attachment #1: Type: text/plain, Size: 1188 bytes --]

On Wednesday 03 November 2004 23:29, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> Salut Christian,
>
> On Wed, 3 Nov 2004, Christian Hesse wrote:
> > I'm using software suspend 2.1. For me suspend works fine, but the system
> > freezes as soon as it is resumed, I only see a black screen then. Driver
> > nv works without any problems. Any hints what could be the cause?
>
> I did not patch in anything into the kernel. I used a vanilla 2.6.9 (only
> the Makefile was patched to show 2.6.9 instead of 8 ... maybe somebody
> watched a bit too much tv... :-)
> So I did not use swsusp 2.1. And I could not manage to freeze on resume
> this time. So, why didn't you try the vanilla-kernel, with the
> Haggi' s patch for the nvidia driver?
> When you test - be sure to sync before suspending. It's the second best,
> besides having a jung backup handy.

Vanilla kernel's suspend to disk is too slow for me. I needs 90 seconds to 
suspend the system (vs. 10 seconds with swsusp-2.1). I think I'll continue 
using nv until Nvidia releases a version of the driver that works without any 
problems for me.

-- 
Christian Hesse

geek by nature
linux by choice

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]

Re: Success with kernel-swsusp and Nvidia-closed-source

[liste-9nAOAgdJVo4b1SvskN2V4Q]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 2074 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salut Nigel,
On Thu, 4 Nov 2004, Nigel Cunningham wrote:
> Hi.
>
> On Thu, 2004-11-04 at 09:29, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Salut Christian,
>> On Wed, 3 Nov 2004, Christian Hesse wrote:
>>> I'm using software suspend 2.1. For me suspend works fine, but the system
>>> freezes as soon as it is resumed, I only see a black screen then. Driver nv
>>> works without any problems. Any hints what could be the cause?
>> I did not patch in anything into the kernel. I used a vanilla 2.6.9 (only
>> the Makefile was patched to show 2.6.9 instead of 8 ... maybe somebody
>> watched a bit too much tv... :-)
>> So I did not use swsusp 2.1. And I could not manage to freeze on resume
>> this time. So, why didn't you try the vanilla-kernel, with the
>> Haggi' s patch for the nvidia driver?
>> When you test - be sure to sync before suspending. It's the second best,
>> besides having a jung backup handy.
>
> I guess I'm doing something driver related that Pavel isn't or vv :>.
> Can I get a pointer to the driver source?

The closed source driver from Nvidia is at
http://download.nvidia.com/XFree86/Linux-x86/1.0-6111/NVIDIA-Linux-x86-1.0-6111-pkg1.run
For me to make it work, you first needed to extract it with --extract-only
Then apply Haggi's patch (attached). Otherwise it refused to suspend:
  Could not suspend device 0000:01:00.0: error -1
But all this is not with Swsusp 2.1, but with the 2.6.9-kernel, and works 
as it should.
I think Christian is also using 6.111??? nvidia driver with Swsusp 2.1.

Cheers
   hartwig felger

Hartwig Felger informatics

- -- 
1024D/339FD693 Hartwig Felger <hgfelger-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org>
Key fingerprint = FB2F 3EE9 345A D55B 6FF2  0EC1 F5B0 684F 339F D693
For the pulic keys, please visit my page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBihmk9bBoTzOf1pMRAnM1AJ4rWSQfKGi9jWO9HMO7i53djIQKmACcCdiB
DzMLUtyQ4P0IBDxhoA8ICgM=
=ZSKy
-----END PGP SIGNATURE-----

[-- Attachment #2: Type: TEXT/PLAIN, Size: 1678 bytes --]

--- NVIDIA-Linux-x86-1.0-6111-pkg1/usr/src/nv/nv.c.ori	2004-11-02 21:03:38.000000000 +0100
+++ NVIDIA-Linux-x86-1.0-6111-pkg1/usr/src/nv/nv.c	2004-11-02 21:03:38.000000000 +0100
@@ -19,6 +19,9 @@
 MODULE_ALIAS_CHARDEV_MAJOR(NV_MAJOR_DEVICE_NUMBER);
 #endif
 
+/* fix for 2.6.9-rc# kernels where __VMALLOC_RESERVE is not exported */
+#define __VMALLOC_RESERVE (128 << 20)
+
 /*
  * our global state; one per device
  */
@@ -3694,19 +3697,24 @@
 
     switch (state)
     {
-        case PM_SAVE_STATE:
-            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event\n");
+      //        case PM_SAVE_STATE:
+      //            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event\n");
+        case NV_PM_ACPI_STANDBY:
+            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event: %d\n", state);
             status = rm_power_management(nv, 0, NV_PM_ACPI_STANDBY);
             break;
 
-        case PM_RESUME:
-            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event\n");
+	    //        case PM_RESUME:
+	    //           nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event\n");
+
+     default:
+        nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event: %d\n", state);
             status = rm_power_management(nv, 0, NV_PM_ACPI_RESUME);
             break;
 
-        default:
-            nv_printf(NV_DBG_WARNINGS, "NVRM: ACPI: unsupported event: %d\n", state);
-            return -1;
+	    //        default:
+	    //            nv_printf(NV_DBG_WARNINGS, "NVRM: ACPI: unsupported event: %d\n", state);
+	    //            return -1;
     }
 
     if (status != RM_OK)

Re: Success with kernel-swsusp and Nvidia-closed-source

[Christian Hesse]

[-- Attachment #1: Type: text/plain, Size: 1843 bytes --]

On Thursday 04 November 2004 12:59, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> Salut Nigel,
>
> On Thu, 4 Nov 2004, Nigel Cunningham wrote:
> > Hi.
> >
> > On Thu, 2004-11-04 at 09:29, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> >> -----BEGIN PGP SIGNED MESSAGE-----
> >> Hash: SHA1
> >>
> >> Salut Christian,
> >>
> >> On Wed, 3 Nov 2004, Christian Hesse wrote:
> >>> I'm using software suspend 2.1. For me suspend works fine, but the
> >>> system freezes as soon as it is resumed, I only see a black screen
> >>> then. Driver nv works without any problems. Any hints what could be the
> >>> cause?
> >>
> >> I did not patch in anything into the kernel. I used a vanilla 2.6.9
> >> (only the Makefile was patched to show 2.6.9 instead of 8 ... maybe
> >> somebody watched a bit too much tv... :-)
> >> So I did not use swsusp 2.1. And I could not manage to freeze on resume
> >> this time. So, why didn't you try the vanilla-kernel, with the
> >> Haggi' s patch for the nvidia driver?
> >> When you test - be sure to sync before suspending. It's the second best,
> >> besides having a jung backup handy.
> >
> > I guess I'm doing something driver related that Pavel isn't or vv :>.
> > Can I get a pointer to the driver source?
>
> The closed source driver from Nvidia is at
> http://download.nvidia.com/XFree86/Linux-x86/1.0-6111/NVIDIA-Linux-x86-1.0-
>6111-pkg1.run For me to make it work, you first needed to extract it with
> --extract-only Then apply Haggi's patch (attached). Otherwise it refused to
> suspend: Could not suspend device 0000:01:00.0: error -1
> But all this is not with Swsusp 2.1, but with the 2.6.9-kernel, and works
> as it should.
> I think Christian is also using 6.111??? nvidia driver with Swsusp 2.1.

Yes, I use 1.0-6111 as well.

-- 
Christian

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]

Re: Success with kernel-swsusp and Nvidia-closed-source

[liste-9nAOAgdJVo4b1SvskN2V4Q]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salut Christian,
On Thu, 4 Nov 2004 liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
>>> Salut Christian,
>>> On Wed, 3 Nov 2004, Christian Hesse wrote:
>>>> I'm using software suspend 2.1. For me suspend works fine, but the 
>>>> system
>>>> freezes as soon as it is resumed, I only see a black screen then. 
>>>> Driver nv
>>>> works without any problems. Any hints what could be the cause?
I played a bit with my X setup. When I force X to use the ali-agp instead 
of the "NVIDIA internal" I get a nearly lockedup machine on resume. This 
is obviously some bug in the linux-agp-driver-tree. Also if I change to a 
runlevel without X I may rmmod the nvidia module but not the agp-modules. 
So they are really buggy (at least the ali-agp, which I would have to 
use). The workaround, if I get such a lockedup machine is to login over 
the network and
  killall -9 X
because kill alone will not be enought. So you may want to try
      Option      "NvAGP" "1"
in your /etc/X11/XF86Config.

Let us know if this gets you any further. When I find time, I will give 
swsusp 2.1 another try myself!

Cheers
   hartwig felger

Hartwig Felger informatics

- -- 
1024D/339FD693 Hartwig Felger <hgfelger-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org>
Key fingerprint = FB2F 3EE9 345A D55B 6FF2  0EC1 F5B0 684F 339F D693
For the pulic keys, please visit my page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBilqb9bBoTzOf1pMRAsHwAKDoowrQMXCOB6trvObNNufIxwmdtgCgh9nj
O+UsJxDEuvujti5IfyWheuA=
=x2FD
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click

Re: Success with kernel-swsusp and Nvidia-closed-source

[Christian Hesse]

[-- Attachment #1: Type: text/plain, Size: 1558 bytes --]

On Thursday 04 November 2004 17:36, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> Salut Christian,
>
> On Thu, 4 Nov 2004 liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> >>> Salut Christian,
> >>>
> >>> On Wed, 3 Nov 2004, Christian Hesse wrote:
> >>>> I'm using software suspend 2.1. For me suspend works fine, but the
> >>>> system
> >>>> freezes as soon as it is resumed, I only see a black screen then.
> >>>> Driver nv
> >>>> works without any problems. Any hints what could be the cause?
>
> I played a bit with my X setup. When I force X to use the ali-agp instead
> of the "NVIDIA internal" I get a nearly lockedup machine on resume. This
> is obviously some bug in the linux-agp-driver-tree. Also if I change to a
> runlevel without X I may rmmod the nvidia module but not the agp-modules.
> So they are really buggy (at least the ali-agp, which I would have to
> use). The workaround, if I get such a lockedup machine is to login over
> the network and
>   killall -9 X
> because kill alone will not be enought. So you may want to try
>       Option      "NvAGP" "1"
> in your /etc/X11/XF86Config.
>
> Let us know if this gets you any further. When I find time, I will give
> swsusp 2.1 another try myself!

I disabled AGP in my kernel config, recompiled the module and added the line 
mentioned above in my xorg.conf - nothing changed. Screen is still black 
after resume. But I can confirm that I can log in via network and killall -9 
X.

-- 
Christian Hesse

geek by nature
linux by choice

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]

Re: Success with kernel-swsusp and Nvidia-closed-source

[liste-9nAOAgdJVo4b1SvskN2V4Q]

[-- Attachment #1: Type: TEXT/PLAIN, Size: 2092 bytes --]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salut Folks,
On Thu, 4 Nov 2004 liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> Let us know if this gets you any further. When I find time, I will give 
> swsusp 2.1 another try myself!
sadly I found out, that my suspend successes where with Linux 2.6.8, but 
none the less, it works the same with 2.6.9, as well as with 2.6.9 patched 
with swsusp-2.1.1
Regarding stability I had a bit more luck with the vanilla-kernels - so 
one time with swsusp 2.1.1 I could not revive the screen. finaly I 
logged-in through the network and issued a "shutdown -r now". As the X was 
going down, the virtual-console did show up again. So maybe I had the bad 
luck, that the keyboard did hang then. I sometimes needed to change to a 
virtual-console, or back to X to get a restored screen, but sometimes it 
worked without doing anything.
I tryed today Linux-2.6.10-rc1 (without Swsusp-2.1.1). To get the 
prorietary Nvidia-driver compiled, you may want to apply the attached 
patch nv-2.9.10-rc1-incl-haggis.diff. I did not try, to have only the 
changes needed for 2.6.9- 2.6.10-rc1 (which is the other attached patch). 
Don't even know, if the compiler likes it. Both patches are aginst the 
Nvidia-driver...
But for me, I could not get the machine going down to suspend. It shutdown 
the screen, and then sit there as long as you let it. By issuing some ACPI 
event, I always got it back to live, e.g. lid-switch, power-button ...

So besides the problem described in 
http://bugme.osdl.org/show_bug.cgi?id=3669, that the laptop does no long 
poweroff, since 2.6.9, S4 as such broke for me with 2.6.10-rc1.

Cheers
   hartwig felger

Hartwig Felger informatics

- -- 
1024D/339FD693 Hartwig Felger <hgfelger-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org>
Key fingerprint = FB2F 3EE9 345A D55B 6FF2  0EC1 F5B0 684F 339F D693
For the pulic keys, please visit my page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBjU159bBoTzOf1pMRAjx+AJ0WXZhZLsIJ+fxVeGYlwW4uT4oLNACfY9/q
oJ8dV4ZYKacSSZzOmRESBVc=
=1BqI
-----END PGP SIGNATURE-----

[-- Attachment #2: Type: TEXT/PLAIN, Size: 2584 bytes --]

--- usr/src/nv/nv.c.ori	2004-11-06 16:14:26.000000000 +0100
+++ usr/src/nv/nv.c	2004-11-06 16:14:26.000000000 +0100
@@ -19,6 +19,9 @@
 MODULE_ALIAS_CHARDEV_MAJOR(NV_MAJOR_DEVICE_NUMBER);
 #endif
 
+/* fix for 2.6.9-rc# kernels where __VMALLOC_RESERVE is not exported */
+#define __VMALLOC_RESERVE (128 << 20)
+
 /*
  * our global state; one per device
  */
@@ -639,7 +642,7 @@
     u8     cap_ptr;
     int    func, slot;
 
-    dev = pci_find_class(class << 8, NULL);
+    dev = pci_get_class(class << 8, NULL);
     do {
         for (func = 0; func < 8; func++) {
             slot = PCI_SLOT(dev->devfn);
@@ -650,7 +653,7 @@
             if (cap_ptr)
                 return fn;
         }
-        dev = pci_find_class(class << 8, dev);
+        dev = pci_get_class(class << 8, dev);
     } while (dev);
 
     return NULL;
@@ -3483,12 +3486,12 @@
     struct pci_dev *dev = NULL;
     int count = 0;
 
-    dev = pci_find_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
+    dev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
     while (dev)
     {
         if ((dev->vendor == 0x10de) && (dev->device >= 0x20))
             count++;
-        dev = pci_find_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
+        dev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
     }
     return count;
 }
@@ -3694,19 +3697,24 @@
 
     switch (state)
     {
-        case PM_SAVE_STATE:
-            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event\n");
+      //        case PM_SAVE_STATE:
+      //            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event\n");
+        case NV_PM_ACPI_STANDBY:
+            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received standby event: %d\n", state);
             status = rm_power_management(nv, 0, NV_PM_ACPI_STANDBY);
             break;
 
-        case PM_RESUME:
-            nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event\n");
+	    //        case PM_RESUME:
+	    //           nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event\n");
+
+     default:
+        nv_printf(NV_DBG_INFO, "NVRM: ACPI: received resume event: %d\n", state);
             status = rm_power_management(nv, 0, NV_PM_ACPI_RESUME);
             break;
 
-        default:
-            nv_printf(NV_DBG_WARNINGS, "NVRM: ACPI: unsupported event: %d\n", state);
-            return -1;
+	    //        default:
+	    //            nv_printf(NV_DBG_WARNINGS, "NVRM: ACPI: unsupported event: %d\n", state);
+	    //            return -1;
     }
 
     if (status != RM_OK)

[-- Attachment #3: Type: TEXT/PLAIN, Size: 1087 bytes --]

--- usr/src/nv/nv.c.ori2	2004-11-06 16:14:26.000000000 +0100
+++ usr/src/nv/nv.c	2004-11-06 16:14:26.000000000 +0100
@@ -642,7 +642,7 @@
     u8     cap_ptr;
     int    func, slot;
 
-    dev = pci_find_class(class << 8, NULL);
+    dev = pci_get_class(class << 8, NULL);
     do {
         for (func = 0; func < 8; func++) {
             slot = PCI_SLOT(dev->devfn);
@@ -653,7 +653,7 @@
             if (cap_ptr)
                 return fn;
         }
-        dev = pci_find_class(class << 8, dev);
+        dev = pci_get_class(class << 8, dev);
     } while (dev);
 
     return NULL;
@@ -3486,12 +3486,12 @@
     struct pci_dev *dev = NULL;
     int count = 0;
 
-    dev = pci_find_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
+    dev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
     while (dev)
     {
         if ((dev->vendor == 0x10de) && (dev->device >= 0x20))
             count++;
-        dev = pci_find_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
+        dev = pci_get_class(PCI_CLASS_DISPLAY_VGA << 8, dev);
     }
     return count;
 }

Re: Success with kernel-swsusp and Nvidia-closed-source

[Christian Hesse]

[-- Attachment #1: Type: text/plain, Size: 1754 bytes --]

On Thursday 04 November 2004 18:12, Christian Hesse wrote:
> On Thursday 04 November 2004 17:36, liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> > Salut Christian,
> >
> > On Thu, 4 Nov 2004 liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> > >>> Salut Christian,
> > >>>
> > >>> On Wed, 3 Nov 2004, Christian Hesse wrote:
> > >>>> I'm using software suspend 2.1. For me suspend works fine, but the
> > >>>> system
> > >>>> freezes as soon as it is resumed, I only see a black screen then.
> > >>>> Driver nv
> > >>>> works without any problems. Any hints what could be the cause?
> >
> > I played a bit with my X setup. When I force X to use the ali-agp instead
> > of the "NVIDIA internal" I get a nearly lockedup machine on resume. This
> > is obviously some bug in the linux-agp-driver-tree. Also if I change to a
> > runlevel without X I may rmmod the nvidia module but not the agp-modules.
> > So they are really buggy (at least the ali-agp, which I would have to
> > use). The workaround, if I get such a lockedup machine is to login over
> > the network and
> >   killall -9 X
> > because kill alone will not be enought. So you may want to try
> >       Option      "NvAGP" "1"
> > in your /etc/X11/XF86Config.
> >
> > Let us know if this gets you any further. When I find time, I will give
> > swsusp 2.1 another try myself!
>
> I disabled AGP in my kernel config, recompiled the module and added the
> line mentioned above in my xorg.conf - nothing changed. Screen is still
> black after resume. But I can confirm that I can log in via network and
> killall -9 X.

Tried with nvidias new driver 1.0-6629 today. I managed to make it work once, 
but I couldn't reproduce it... :(

-- 
Christian

[-- Attachment #2: Type: application/pgp-signature, Size: 190 bytes --]

Re: Success with kernel-swsusp and Nvidia-closed-source

[liste-9nAOAgdJVo4b1SvskN2V4Q]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salut Christian,
On Sun, 7 Nov 2004, Christian Hesse wrote:
>>> Let us know if this gets you any further. When I find time, I will give
>>> swsusp 2.1 another try myself!
>>
>> I disabled AGP in my kernel config, recompiled the module and added the
>> line mentioned above in my xorg.conf - nothing changed. Screen is still
>> black after resume. But I can confirm that I can log in via network and
>> killall -9 X.
>
> Tried with nvidias new driver 1.0-6629 today. I managed to make it work once,
> but I couldn't reproduce it... :(
I am verry impressed. I had a try myself today and did not succeed. Then I 
took a look at the driver (6629 -> nv.c) and discovered, that they removed 
all ACPI support, which was in 6111. So your once success with 6629 is a 
little miracle. But maybe ACPI is not allways needed :-))))
So, from time to time, I have the trouble, that my keyboard is not 
behaving well after suspend (which comes from X). When I login over the 
network, and issue a
  chvt 6
  chvt 7
sequence, all will be back to normal. I start now including a change to 
vt-6 before suspending into my script. I will report back, if this is a 
better solution.
So for the record: working for me: 2.6.9-vanilla+nv-6111+haggis
                                    2.6.8-vanilla+nv-6111+haggis
                                    2.6.9+sw2.1.1+nv-6111+haggis

Cheers
   hartwig felger

Hartwig Felger informatics

- -- 
1024D/339FD693 Hartwig Felger <hgfelger-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org>
Key fingerprint = FB2F 3EE9 345A D55B 6FF2  0EC1 F5B0 684F 339F D693
For the pulic keys, please visit my page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBj5T+9bBoTzOf1pMRAinaAKDk2Oi3ZpjR3Gw3saqgpR9nncyU3wCfXpZQ
gGblWm6+Um3MX66fuTdjP84=
=/onx
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click

Re: Success with kernel-swsusp and Nvidia-closed-source

[Karol Kozimor]

Thus wrote liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org:
> I am verry impressed. I had a try myself today and did not succeed. Then I 
> took a look at the driver (6629 -> nv.c) and discovered, that they removed 
> all ACPI support, which was in 6111. So your once success with 6629 is a 

Which was outright bogus...

> little miracle. But maybe ACPI is not allways needed :-))))

... because there's only one PM interface in 2.6 and it's not really
ACPI-specific. The driver might even work as expected now.
Note: I haven't seen the new code yet.
Best regards,

-- 
Karol 'sziwan' Kozimor
sziwan-DETuoxkZsSqrDJvtcaxF/A@public.gmane.org


-------------------------------------------------------
This SF.Net email is sponsored by:
Sybase ASE Linux Express Edition - download now for FREE
LinuxWorld Reader's Choice Award Winner for best database on Linux.
http://ads.osdn.com/?ad_id=5588&alloc_id=12065&op=click

Re: Success with kernel-swsusp and Nvidia-closed-source

[liste-9nAOAgdJVo4b1SvskN2V4Q]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Salut Folks,
On Sat, 6 Nov 2004 liste-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org wrote:
> I tryed today Linux-2.6.10-rc1 (without Swsusp-2.1.1). To get the prorietary 
> Nvidia-driver compiled, you may want to apply the attached patch 
> nv-2.9.10-rc1-incl-haggis.diff. I did not try, to have only the changes 
> needed for 2.6.9- 2.6.10-rc1 (which is the other attached patch). Don't even 
> know, if the compiler likes it. Both patches are aginst the Nvidia-driver...
> But for me, I could not get the machine going down to suspend. It shutdown 
> the screen, and then sit there as long as you let it. By issuing some ACPI 
> event, I always got it back to live, e.g. lid-switch, power-button ...
So with 2.6.10-rc2 this regression is gone for me. We have another working 
combination: vanilla 2.6.10-rc2 and nvidia-6111+haggis+my-patch.

Cheers
   hartwig felger

Hartwig Felger informatics

- -- 
1024D/339FD693 Hartwig Felger <hgfelger-9nAOAgdJVo4b1SvskN2V4Q@public.gmane.org>
Key fingerprint = FB2F 3EE9 345A D55B 6FF2  0EC1 F5B0 684F 339F D693
For the pulic keys, please visit my page.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.2 (GNU/Linux)

iD8DBQFBmSlF9bBoTzOf1pMRAlu7AKDeIiDSIRxxVa2qGa7JGtAixJAtPgCfTqGr
vrWbRPeq2UBuY4WaLZRUEsM=
=LZ+X
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by: InterSystems CACHE
FREE OODBMS DOWNLOAD - A multidimensional database that combines
robust object and relational technologies, making it a perfect match
for Java, C++,COM, XML, ODBC and JDBC. www.intersystems.com/match8