Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Luca Capello]

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

Hello,

on 09/08/04 07:38, Zhu, Yi wrote:
> Here comes the 3rd version of the suspend-resume patch.
it works perfectly on my ASUS M6N, 2.6.8.1 with no patch and S4: I can
suspend/resume directly from X (Debian XFree86-4.3.0.dfsg.1-7 & GNOME
2.7.92). All is working correctly, not only the ipw2100 driver but all
the other drivers, too. I mean, actually I loaded a lot of different
modules (see below), I just did 'echo 4 > /proc/acpi/sleep' different
times (this is the 3rd in sequence).
=====
luca@gismo:~$ lsmod
Module                  Size  Used by
ipw2100                98164  0
ieee80211              18564  1 ipw2100
ieee80211_crypt         6020  1 ieee80211
ds                     20744  4
lp                     10536  0
thermal                18576  0
fan                     5644  0
button                  8472  0
battery                12428  0
ac                      6668  0
asus_acpi              11544  0
rfcomm                 39324  0
l2cap                  26116  5 rfcomm
bluetooth              49668  4 rfcomm,l2cap
firmware_class         10496  1 ipw2100
ohci1394               36356  0
ieee1394              110772  1 ohci1394
yenta_socket           21888  0
pcmcia_core            72008  2 ds,yenta_socket
bcm5700               140844  0
snd_intel8x0m          21064  4
snd_intel8x0           36360  4
snd_ac97_codec         73348  2 snd_intel8x0m,snd_intel8x0
snd_pcm_oss            55080  0
snd_mixer_oss          20352  5 snd_pcm_oss
snd_pcm                98824  3 snd_intel8x0m,snd_intel8x0,snd_pcm_oss
snd_timer              26244  1 snd_pcm
snd_page_alloc         12040  3 snd_intel8x0m,snd_intel8x0,snd_pcm
snd_mpu401_uart         8448  1 snd_intel8x0
snd_rawmidi            25508  1 snd_mpu401_uart
snd_seq_device          8712  1 snd_rawmidi
snd                    58980  18
snd_intel8x0m,snd_intel8x0,snd_ac97_codec,snd_pcm_oss,snd_mixer_oss,snd_pcm,snd_timer,snd_mpu401_uart,snd_rawmidi,snd_seq_device
soundcore              10592  5 snd
ehci_hcd               43652  0
uhci_hcd               33804  0
intel_agp              22944  1
agpgart                34984  1 intel_agp
parport_pc             23616  1
parport                26944  2 lp,parport_pc
pcspkr                  3944  0
joydev                 10176  0
evdev                   9856  1
xfs                   518008  0
ufs                    70660  0
udf                    90628  0
romfs                   9092  0
ntfs                   98516  0
jfs                   177608  0
hfsplus                60420  0
hfs                    47620  0
cramfs                 42996  0
zlib_inflate           22784  1 cramfs
cdfs                   28200  0
speedstep_centrino      8148  0
processor              25776  2 thermal,speedstep_centrino
fbcon                  31748  72
font                    8576  1 fbcon
radeonfb               63016  1
cfbcopyarea             4096  1 radeonfb
cfbimgblt               3328  1 radeonfb
cfbfillrect             3968  1 radeonfb
luca@gismo:~$
=====

There's only one minor problem: sometimes during the suspend process the
laptop doesn't shut down, the screen remains black and I have to
manually press 4-sec the power button. Then resuming is always ok. But
this is a minor issue ;-)

BTW it's about 40min from the latest resume and the ipw2100 is working
perfectly: 75kB/s downloading a Debian CD images with BitTorrent.

S3 doesn't work neither without the ipw2100: I get a resume suddenly
after the suspend process is finished and without errors.

> Thank you for all your testing effort!
You're welcome :-)

Thx, bye,
Gismo / Luca

PS
- - ACPI folks: I cced to let you know this improvement in the M6N
suspend/resume process

- - Herman: I searched in http://mrhammy2.ath.cx:81/forum/index.php, but
it seems that S4 is never mentioned in the last days, so I cced you :-)
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQFBPuHpVAp7Xm10JmkRAj/PAJ905mm9l62rwUR53GZ95sDujRsbjwCfYSEt
+vJPnfh/5ewDZ4IpXo6KF98=
=iIhf
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click

Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Luca Capello]

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

Hello again,

on 09/08/04 12:41, Luca Capello wrote:
> it works perfectly on my ASUS M6N, 2.6.8.1 with no patch and S4: I can
> suspend/resume directly from X (Debian XFree86-4.3.0.dfsg.1-7 & GNOME
> 2.7.92). All is working correctly, not only the ipw2100 driver but all
> the other drivers, too. I mean, actually I loaded a lot of different
> modules (see below), I just did 'echo 4 > /proc/acpi/sleep' different
> times (this is the 3rd in sequence).
<cut>
> There's only one minor problem: sometimes during the suspend process the
> laptop doesn't shut down, the screen remains black and I have to
> manually press 4-sec the power button. Then resuming is always ok. But
> this is a minor issue ;-)
<cut>
> S3 doesn't work neither without the ipw2100: I get a resume suddenly
> after the suspend process is finished and without errors.
same /symptoms/ on 2.6.9-rc1 + ACPI 20040816: same success for
suspend/resume directly from XFree86 (with the included radeon driver)
and with the radeonfb loaded, same little problem with the black screen
in suspending and same problem for S3.

BTW is 'echo 4 > /proc/acpi/sleep' = 'echo disk > /sys/power/state'?

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQFBPwLmVAp7Xm10JmkRAr+FAJ9xkC+dGthAyfcMSa9WbUf9ra5kpQCfSiZA
+fRCZL6m5a0vBp/Lzz7WrOw=
=ISWc
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by BEA Weblogic Workshop
FREE Java Enterprise J2EE developer tools!
Get your free copy of BEA WebLogic Workshop 8.1 today.
http://ads.osdn.com/?ad_id=5047&alloc_id=10808&op=click

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Pavel Machek]

Hi!

> > it works perfectly on my ASUS M6N, 2.6.8.1 with no patch and S4: I can
> > suspend/resume directly from X (Debian XFree86-4.3.0.dfsg.1-7 & GNOME
> > 2.7.92). All is working correctly, not only the ipw2100 driver but all
> > the other drivers, too. I mean, actually I loaded a lot of different
> > modules (see below), I just did 'echo 4 > /proc/acpi/sleep' different
> > times (this is the 3rd in sequence).
> <cut>
> > There's only one minor problem: sometimes during the suspend process the
> > laptop doesn't shut down, the screen remains black and I have to
> > manually press 4-sec the power button. Then resuming is always ok. But
> > this is a minor issue ;-)
> <cut>
> > S3 doesn't work neither without the ipw2100: I get a resume suddenly
> > after the suspend process is finished and without errors.
> same /symptoms/ on 2.6.9-rc1 + ACPI 20040816: same success for
> suspend/resume directly from XFree86 (with the included radeon driver)
> and with the radeonfb loaded, same little problem with the black screen
> in suspending and same problem for S3.
> 
> BTW is 'echo 4 > /proc/acpi/sleep' = 'echo disk > /sys/power/state'?

Yes, it should be equivalent.
								Pavel

-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!


-------------------------------------------------------
This SF.Net email is sponsored by: thawte's Crypto Challenge Vl
Crack the code and win a Sony DCRHC40 MiniDV Digital Handycam
Camcorder. More prizes in the weekly Lunch Hour Challenge.
Sign up NOW http://ad.doubleclick.net/clk;10740251;10262165;m

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Luca Capello]

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

Hello Pavel,

on 09/15/04 00:29, Pavel Machek wrote:
>> BTW is 'echo 4 > /proc/acpi/sleep' = 'echo disk > /sys/power/state'?
> Yes, it should be equivalent.
the same I supposed. But in this case, there's an error somewhere: I can
never resume properly if I use 'echo disk > /sys/power/state' and I get
a kernel panic with these words:

Resuming from /dev/hda7
Resuming from unknown-block(3,7)
Kernel panic - not syncing: Resume Machine: Unable to find
suspending-data signature (pmdisk-swa - mispelled?

Usually, my lilo is configured to append 'resume=/dev/hda7' and I tried
even at boot with 'pmdisk=/dev/hda7', but it doesn't matter. So probably
the code is different in the 2 cases. Can someone check, please? I
suppose the /sys call is better, as all the /proc fs is being ported to
/sys, am I right?

On the other hand, using the /sys call the machine power off correctly
in suspending, while using the /proc call actually I should press 4-sec
the power button after 15-20sec the latest suspending screen.

BTW, I'm on 2.6.9-rc2 with no other patches.

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQFBR4cMVAp7Xm10JmkRAsDAAJ9dq5WlwwGRc1kFGD5TgjWTQ4FvHACeNoqj
o8CPipEfsMNMNa2NRul4W/A=
=LjO0
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by: thawte's Crypto Challenge Vl
Crack the code and win a Sony DCRHC40 MiniDV Digital Handycam
Camcorder. More prizes in the weekly Lunch Hour Challenge.
Sign up NOW http://ad.doubleclick.net/clk;10740251;10262165;m

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Karol Kozimor]

Thus wrote Luca Capello:
> >> BTW is 'echo 4 > /proc/acpi/sleep' = 'echo disk > /sys/power/state'?
> > Yes, it should be equivalent.
> the same I supposed. But in this case, there's an error somewhere: I can
> never resume properly if I use 'echo disk > /sys/power/state' and I get
> a kernel panic with these words:
> 
> Resuming from /dev/hda7
> Resuming from unknown-block(3,7)
> Kernel panic - not syncing: Resume Machine: Unable to find
> suspending-data signature (pmdisk-swa - mispelled?

Append noresume. swsusp1 kicks in earlier than pmdisk.
Best regards,

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


-------------------------------------------------------
This SF.Net email is sponsored by: thawte's Crypto Challenge Vl
Crack the code and win a Sony DCRHC40 MiniDV Digital Handycam
Camcorder. More prizes in the weekly Lunch Hour Challenge.
Sign up NOW http://ad.doubleclick.net/clk;10740251;10262165;m

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Luca Capello]

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

Hello Karol,

on 09/15/04 10:37, Karol Kozimor wrote:
> Append noresume. swsusp1 kicks in earlier than pmdisk.
this works indeed.

Now, my situation:
- - ASUS M6N with BIOS 0209A
- - kernel 2.6.9-rc2 with no other patches
- - quite all the possible modules for my machine loaded
- - Debian GNU/Linux experimental
- - Debian XFree86-4.3.0.dfsg.1-7 (using the native radeon driver)
- - Debian GNOME 2.7.92-1

I can *successfully* suspend/resume with the 2 calls (/proc or /sys),
directly from XFree86 and with all the modules loaded. As already
reported, the only difference is that with the /sys call, in suspending
I've always the machine which powers off, while with the /proc call I
should power off the machine by pressing 4-sec the power button.

Simple question: which should be the best method to use? /sys or /proc?

Another thing: why for the /sys call I should use 'append=noresume',
while I want to *resume*? I mean, it's not so clear...

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQFBSBgFVAp7Xm10JmkRAi8CAJ9mL0FRj3oGsvJRQ2hxaZ4ZFXSyiQCaA3Vo
ljpK7b0WuGf9VMa5Z2SjiD0=
=ONtH
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by: thawte's Crypto Challenge Vl
Crack the code and win a Sony DCRHC40 MiniDV Digital Handycam
Camcorder. More prizes in the weekly Lunch Hour Challenge.
Sign up NOW http://ad.doubleclick.net/clk;10740251;10262165;m

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Karol Kozimor]

Thus wrote Luca Capello:
> I can *successfully* suspend/resume with the 2 calls (/proc or /sys),
> directly from XFree86 and with all the modules loaded. As already
> reported, the only difference is that with the /sys call, in suspending
> I've always the machine which powers off, while with the /proc call I
> should power off the machine by pressing 4-sec the power button.
> 
> Simple question: which should be the best method to use? /sys or /proc?

Use /sys if it works. OTOH, /proc is more tested and actively maintained.
ACPI S4 triggered by /sys is slightly better in handling wake-up events
(think Wake On LAN, RTC) and other things like, say AC status updates.

> Another thing: why for the /sys call I should use 'append=noresume',
> while I want to *resume*? I mean, it's not so clear...

Because /sys is the pmdisk code, /proc is swsusp1 code (at least until the
merge in -mm kernels goes into mainline). swsusp1 uses resume and noresume,
pmdisk uses pmdisk (=/dev/hdXY) and, well... AFAIR pmdisk (="").
Best regards,

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


-------------------------------------------------------
This SF.Net email is sponsored by: thawte's Crypto Challenge Vl
Crack the code and win a Sony DCRHC40 MiniDV Digital Handycam
Camcorder. More prizes in the weekly Lunch Hour Challenge.
Sign up NOW http://ad.doubleclick.net/clk;10740251;10262165;m

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Pavel Machek]

Hi!

> On the other hand, using the /sys call the machine power off correctly
> in suspending, while using the /proc call actually I should press 4-sec
> the power button after 15-20sec the latest suspending screen.
> 
> BTW, I'm on 2.6.9-rc2 with no other patches.

Go to -mm kernel. Vanilla has *very* old version of the patch.
									Pavel
-- 
People were complaining that M$ turns users into beta-testers...
...jr ghea gurz vagb qrirybcref, naq gurl frrz gb yvxr vg gung jnl!


-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Luca Capello]

[-- Attachment #1: Type: text/plain, Size: 781 bytes --]

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

Hello,

on 09/17/04 10:51, Pavel Machek wrote:
>>On the other hand, using the /sys call the machine power off correctly
>>in suspending, while using the /proc call actually I should press 4-sec
>>the power button after 15-20sec the latest suspending screen.
>>
>>BTW, I'm on 2.6.9-rc2 with no other patches.
> Go to -mm kernel. Vanilla has *very* old version of the patch.
it seems that the mail I sent last Friday never reached the
mailing-list. Here it is.

Thx, bye,
Gismo / Luca
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQFBV/9nVAp7Xm10JmkRAi93AKCEOn7+S31Lsu+l9iCpEUsSuKVUJACgjunD
2SnhVjwvXbyTej1m9loRFsg=
=4MSb
-----END PGP SIGNATURE-----

[-- Attachment #2: Attached Message --]
[-- Type: message/rfc822, Size: 2908 bytes --]

[-- Attachment #2.1.1: Type: text/plain, Size: 1523 bytes --]

Hello,

On Fri, Sep 17, 2004 at 10:51:10AM +0200, Pavel Machek wrote:
> > On the other hand, using the /sys call the machine power off correctly
> > in suspending, while using the /proc call actually I should press 4-sec
> > the power button after 15-20sec the latest suspending screen.
> > 
> > BTW, I'm on 2.6.9-rc2 with no other patches.
> Go to -mm kernel. Vanilla has *very* old version of the patch.
I tried 2.6.9-rc2-mm2 and here the results:

- echo 4 > /proc/acpi/sleep
after 10m the suspend process is still going on and the data written
to the swap partition are only 1% (and so I stopped the process ;-) )

- echo disk > /sys/power/state
the suspend process seems ok, but the resume arrives suddenly after the
suspend is finished (I mean, the computer doesn't power off)

Anyway, I saw that in the kernel there's no more the choice for pmdisk:
is this normal?

BTW, as I counted the time for -mm2, I did the same for a vanilla 2.6.9
+ latest ACPI patch:

- echo 4 > /proc/acpi/sleep
40s and now the computer automatically power off (AFAIR the only
driver upgrade is the 'ipw2100', where I used the latest patch posted in
this list, too)

- echo disk > /sys/power/state
30s to the power off

Now, Pavel, I ask you the same question Karol already answered: as I've
the choice between swsusp and pmdisk, Karol suggested to use pmdisk. And
you?

Thx, bye,
Gismo / Luca

PS all the test were performed as root and from GNOME/XFree86-4.3.0
(Debian, 'radeon' driver included)

[-- Attachment #2.1.2: Digital signature --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: Re: [Ipw2100-devel] [susp-resume] ipw2100 suspend patch for 0.54 (version 3)

[Luca Capello]

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

Hello all,

on 09/27/04 13:54, Luca Capello wrote:
> Now, Pavel, I ask you the same question Karol already answered: as I've
> the choice between swsusp and pmdisk, Karol suggested to use pmdisk. And
> you?
I can answer by myself after having read [1] (sorry, this is the only
info I found about the merging process for 'swsusp' and 'pmdisk'): I'm
trying to test -mm4 and have it working, so maybe I'll ask again for
help ;-)

Thx, bye,
Gismo / Luca

[1] http://www.kerneltraffic.org/kernel-traffic/latest.html#1
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)
Comment: Using GnuPG with Debian - http://enigmail.mozdev.org

iD8DBQFBWWdLVAp7Xm10JmkRAnMLAJ4+v2QNrfib/A5o3CQFht1eZq11lACfXQl+
CXltos4jCoCWqekkjYvPDi4=
=Uqai
-----END PGP SIGNATURE-----


-------------------------------------------------------
This SF.Net email is sponsored by: YOU BE THE JUDGE. Be one of 170
Project Admins to receive an Apple iPod Mini FREE for your judgement on
who ports your project to Linux PPC the best. Sponsored by IBM.
Deadline: Sept. 24. Go here: http://sf.net/ppc_contest.php