Dell Inspiron 6000

[juniper]

man, STR is giving me some trouble...

I used mark lord's patchset on the 2.6.17.1 kernel and no go. i get the same input/output error i complained about above.

are people using any special boot parameters? all i am using is libata_enabled=1. do i need something else?

[amaroc]

man, STR is giving me some trouble...

I used mark lord's patchset on the 2.6.17.1 kernel and no go. i get the same input/output error i complained about above.

are people using any special boot parameters? all i am using is libata_enabled=1. do i need something else?

You don't need the libata_enabled=1 anymore (IIRC since 2.6.16). I don't believe that this is the reason. Anyway, this is how my boot-line looks like:

Code: Select all

kernel /kernel-2.6.17-gentoo-r4_20060726 root=/dev/sda6 resume=/dev/sda5 video=vesafb:1024x768-32@60

As you can see, I'm using latest stable gentoo-source now because
- no issues, STR+STD are stable

Code: Select all

# uptime
 00:10:30 up 7 days,  2:15,  1 user,  load average: 2.11, 1.68, 1.16

- vesa-tng patch included
- latest ACPI patches included
- I'm too lazy to patch any updated kernel by myself

However, I don't think your problem comes from either kernel sources or patches. Mark's patchset _should_ work - it covers almost every _known_ ACPI issue. I personally had this configuration running before switching to gentoo-sources.

Last question - have you tried the .config I've sent to you some weeks ago? Maybe it's worth to try (after a 'make oldconfig' .

[juniper]

@amaroc:

to be honest, amaroc, i got your email, but i can't for the life of me remember if i tried your kernel config with that 2.6.16 kernel. i have compiled probably 20 kernels in the last few weeks (not just on this laptop) so i simply can't remember.

I will give a try when i get home today. thanks.

[juniper]

@amaroc:

ok, i am about to compile your working patches against the 2.6.16.16 kernel (i think you use gentoo-sources, right?). before tweaking, i want to have everything EXACTLY like you do (just to see what is up).

out of curiosity, there is a suspend2 patch in the "others" tar file you sent me (if you recall you sent me a tar file of what patches you used and ones you didn't). i was under the impression that suspend 2 and kernel suspend were mutually exclusive. what does that suspend2 patch do?

[amaroc]

ok, i am about to compile your working patches against the 2.6.16.16 kernel (i think you use gentoo-sources, right?). before tweaking, i want to have everything EXACTLY like you do (just to see what is up).

OK, here's the history what should explain it.
I was using 2.6.16 vanilla sources together with Mark's patchset+vesa-tng patch since end of March - both STR and STD working. The only issues I had was with the builtin ipw2200 and wpa-supplicant that forced the dhcpd to time out after resume from STR. Therefore I had to use ipw2200 and ieee packages seperately.
After 2.6.17 with updated builtin ipw2200 became available I've run vanilla 2.6.17 with Marks _latest_ patchset.
After 2.6.17-gentoo-r4 became stable I've switched to gentoo-sources - what means a working kernel out-of-the-box now.

out of curiosity, there is a suspend2 patch in the "others" tar file you sent me (if you recall you sent me a tar file of what patches you used and ones you didn't). i was under the impression that suspend 2 and kernel suspend were mutually exclusive. what does that suspend2 patch do?

Yes, you have to decide either for bultin suspend or additional patches and suspend2. Mark preferred the suspend2 because of it's speed and beauty. But the builtin suspend works too and I peronally don't care about lower resume speed and nice progress information. Therefore I've skipped the suspend2 patch.

If 2.6.16 doesn't work for you and considering the trouble you had so far I would try the stable gentoo-source 2.6.17-gentoo-r4 together with the .config I've send to you. I have not changed the configuration so a make oldconfig should do the job. Maybe you have to enable builtin ipw2200 if you want WLAN functionality.

If there is still no go, let us know. Potential issues:
- hw-specific ACPI/SATA issues -> make use of latest patches from Mark or KML directly
- wrong ram.conf -> I've already posted my working config earlier
- issues with xorg or 855resolution

Good luck.

[juniper]

@amaroc:

BRILLIANT!!!!!!!!!!

yes, oddly, i don't really know why the other files did not work. i have not checked to see whether you have compiled support for the things I want but I can tell you that this post is being made after an STR cycle!!!!!!!!!

woooooooooooooooooooooooooooooooooo!!!!!!!!!!!!!!!!!!!!!

you don't know how many kernel i have compiled... i wonder what i have been missing this whole time. i have even used kernel configs that i friend of mine had for this exact same laptop that worked for him but not me.

thanks amaroc. thanks for the email, thanks for the help here and the time.

i also like suspend2, but for a few days i will bask in the glow of a working suspend1. i will see if the newer kernels work.

By the way, just to let you know, i used a vanilla kernel 2.6.16.16. the gentoo sources had a whole bunch of the patches already applied and also the kernel would compile. so, i used the 2.6.16.16 kernel with EXACTLY your working patches and exactly your .config (except i changed the resume device to /dev/sda6, my swap).

[juniper]

one small problem i seem to be having.

sometimes, when my computer sleeps and i try to wake it up (by opening the lid for example) it goes right back to sleep. but then if i wake it up again it wakes up fine.

is this happening to anyone else? is there a fix?

[amaroc]

but I can tell you that this post is being made after an STR cycle!!!!!!!!!

congratulations

sometimes, when my computer sleeps and i try to wake it up (by opening the lid for example) it goes right back to sleep. but then if i wake it up again it wakes up fine.
is this happening to anyone else? is there a fix?

I don't have this issue here. Did you configure STR to be executed when you close the lid? If yes you have to handle "lid open" exactly.
Anyway, you should have a look to /etc/acpi/default.sh. This is how it would look like when you only want to switch on&off the backligth, like I have it configured:

Code: Select all

lid)
         	    grep -q open /proc/acpi/button/lid/LID/state && logger "ACPI action: LID OPEN" && vbetool dpms on
                    ;;

If there is something written with "hibernate-ram" you should change it to

Code: Select all

 grep -q close /proc/acpi/button/lid/LID/state && logger "ACPI action: LID CLOSE" && hibernate-ram &

[juniper]

@amaroc:

i deserve to be beaten with a stick; there were two

echo -n mem > /sys/power/state

lines in my sleep script (probably a cut and paste error). and, yes, removing one of the lines solved my problem.

[amaroc]

juniper,

now that you enjoy a working STR you know what you've missed before - right? And - it makes me happy as well if I could help you a little bit...

Last thing - I propose not to spend too much time for configuring suspend2. With STR the i6000 stays for almost one week from a fully loaded battery. So it might happen, that you will do STD only for testing after installing a new kernel.
Considering battery life time it is also much better to charge the battery when the notebook is STR and to work from battery rather then AC supply. This keeps the battery as cool as possible. My battery has already lost 10% of it's capacity after 10 month:

Code: Select all

cat /proc/acpi/battery/BAT0/info
present:                 yes
design capacity:         7200 mAh
last full capacity:      6489 mAh

[juniper]

i have had mine for 15 months and i am not doing so well either.

Code: Select all

cat /proc/acpi/battery/BAT0/info
present:                 yes
design capacity:         7200 mAh
last full capacity:      4880 mAh

but, yes, STR is rather nice and you very much did help setting it up.

[nybbles]

Would someone mind saying where these config files you guys are talking about can be found? Or did you just send them to each other through email?

I've got an i6k with the ATI 300X card and am not able to get Suspend-to-Ram working. I'd really like to see a working configuration so I can somehow get it working on my machine too!

Thanks..

[juniper]

ok. i put the patches and kernel config file up at

http://juniper1982.googlepages.com/home

if you have trouble, let us know. oh, btw, i have a intel 915 video card whereas you have an ati. i don't really know how that will play out, but give the kernel/patches a try.

[nybbles]

if you have trouble, let us know. oh, btw, i have a intel 915 video card whereas you have an ati. i don't really know how that will play out, but give the kernel/patches a try.

Thanks! I'll give it a try when I get home

--EDIT--

Argh, not having any luck with ACPI these days..

I have acpid running and the 'button' module loaded. But when I open/close my laptop lid, the event does not get registered. It does not appear when I run

Code: Select all

tail -f /var/log/acpi | grep "received event"

Regardless of whether the laptop lid is closed or open, I get the following:

Code: Select all

cat /proc/acpi/button/lid/LID/state
state:      closed

Any suggestions regarding what my problem might be? I have not seen any solutions to this problem in other posts, but it looks like people do have the problem..

Thanks for the config files and kernel patches. I see the kernel version is 2.6.16.16. Do these apply against gentoo-sources-2.6.16-r16 or something? I'm not sure how the version numbers work here.
Oh and where'd you get these patches from?

[Shucklak]

Just curious, I've been using Gentoo on my 6000 since last October, and I still have the lid problem: when I close my lid the screen will not come back on, even after using CTRL + SHFT +F* . The only way I can get the LCD to come back on is if I plug in an external monitor, use the FNC + LCD/CRT button to switch to the external monitor, and then switch back to the laptops LCD. Any one have a solution?

[amaroc]

Just curious, I've been using Gentoo on my 6000 since last October, and I still have the lid problem: when I close my lid the screen will not come back on, even after using CTRL + SHFT +F* . The only way I can get the LCD to come back on is if I plug in an external monitor, use the FNC + LCD/CRT button to switch to the external monitor, and then switch back to the laptops LCD. Any one have a solution?

You need vbetool in order to switch the backlight on. Simply unmask and emerge it.
As written earlier you could then add the following line to your /etc/acpi/default.sh

Code: Select all

			lid)	
			    grep -q open /proc/acpi/button/lid/LID/state && logger "ACPI action: LID OPEN" && vbetool dpms on 

It works here with the i915.

Argh, not having any luck with ACPI these days..

How did you configued your acpi?
http://www.gentoo.org/doc/en/power-management-guide.xml
http://gentoo-wiki.com/TIP_ACPI_basic_configuration
However, fo STR & STD I use the hibernate-script from suspend2 - it works with built-in suspend as well and all actions/options are handled within 3 files only:
actions: /etc/acpi/default.sh
STR: /etc/hibernate/ram.conf
STD: /etc/hibernate/hibernate.conf
All easy and self-explaining.

I see the kernel version is 2.6.16.16. Do these apply against gentoo-sources-2.6.16-r16 or something? I'm not sure how the version numbers work here.
Oh and where'd you get these patches from?

2.6.16.16 is related to vanilla sources - that was the actual kernel when those patches have been made available by Mark Lord http://rtr.ca/dell_i9300/.
This is a known to work kernel but - I still suggest to use the latest stable gentoo-kernel.
- no additional patches needed
- (should) work "out-of-the-box"
Note: You can still use the linked config-file after doing

Code: Select all

make oldconfig

If there are still problems let us know.

[Shucklak]

I emerged vbetool and put the following into my /etc/acpi/default.sh

Code: Select all

 lid)   
             grep -q open /proc/acpi/button/lid/LID/state && logger "ACPI action: LID OPEN" && vbetool dpms on

I then tried closing and opening the lid and nothing. Rebooted and tried again, still nothing. Then I noticed, the grep is calling for /proc/acpi/button/lid/LID/STATE and I don't have that in my /proc. I can go as far as /proc/acpi and that is it. There is not a button directory.

[amaroc]

Then I noticed, the grep is calling for /proc/acpi/button/lid/LID/STATE and I don't have that in my /proc. I can go as far as /proc/acpi and that is it. There is not a button directory.

I suppose you have either not enabled acpi within your kernel at all, missed some options like CONFIG_ACPI_BUTTON or didn't load some acpi-related modules. Please check:

Code: Select all

CONFIG_ACPI=y
CONFIG_ACPI_SLEEP=y
CONFIG_ACPI_SLEEP_PROC_FS=y
CONFIG_ACPI_SLEEP_PROC_SLEEP=y
CONFIG_ACPI_AC=y
CONFIG_ACPI_BATTERY=y
CONFIG_ACPI_BUTTON=y
CONFIG_ACPI_VIDEO=y
CONFIG_ACPI_HOTKEY=m
CONFIG_ACPI_FAN=y
CONFIG_ACPI_PROCESSOR=y
CONFIG_ACPI_THERMAL=y
# CONFIG_ACPI_ASUS is not set
# CONFIG_ACPI_IBM is not set
# CONFIG_ACPI_TOSHIBA is not set
CONFIG_ACPI_BLACKLIST_YEAR=0
# CONFIG_ACPI_DEBUG is not set
CONFIG_ACPI_EC=y
CONFIG_ACPI_POWER=y
CONFIG_ACPI_SYSTEM=y
CONFIG_X86_PM_TIMER=y
CONFIG_ACPI_CONTAINER=m

For you reference pls. check .config from http://juniper1982.googlepages.com/home

[juniper]

@ nybbles:

hmm. I don't know what is wrong with your lid. you appear to have acpi working (as it says the lid is closed) but not working well.

@ Shucklak: I concur with amaroc, I essentially have what he has but, I believe, I have it in my lm_lid.sh file in /etc/acpi/actions. but, since your /proc/acpi/LIDSTUFF is missing you may not have support in the kernel.

@ anyone: i am not using the ram.conf or hibernate.conf script. I am using the acpi_sleep script found at

http://socrates.homelinux.org/index.php ... spiron6000

if you notice though, that script has a cut and paste error (there is a whole repeated chunk, just get rid of the second instance of it).

[juniper]

i know it is a big no no to cross post, but i thought that many people here are probably having the same problem.

i have an i915 card and it has to do with xorg (7 in my case, but this was happening with 6.x), mplayer and STR/STD. basically, mplayer/xine/totem would lock X after a sleep cycle.

see

http://forums.gentoo.org/viewtopic-t-48 ... ight-.html

for full post. has anyone encountered this?

@ amaroc (or anyone else):

your STR kernel/config works great but another problem I am having is that when I wake up from STR my cdrom can no longer be read. my /dev/sr0 is missing, this is where it is usually detected. i tried restarting udevd, coldplug, hald and dbus to no avail. in dmesg it does not seem to have detected it.

EDIT: i checked cdrecord -scanbus and it found my cdrom, but still sr0 was not there. i put a cd in and after a while, it simply showed up and dmesg says

Code: Select all

sd 0:0:0:0: Attached scsi generic sg0 type 0
 1:0:0:0: Attached scsi generic sg1 type 5
sr0: scsi3-mmc drive: 24x/24x writer cd/rw xa/form2 cdda tray
Uniform CD-ROM driver Revision: 3.20
sr 1:0:0:0: Attached scsi CD-ROM sr0

so all is well, except it took forever.

[amaroc]

i have an i915 card and it has to do with xorg (7 in my case, but this was happening with 6.x), mplayer and STR/STD. basically, mplayer/xine/totem would lock X after a sleep cycle.
see
http://forums.gentoo.org/viewtopic-t-48 ... ight-.html

First, I'm not using xine or others very often - but I've seen something like this as well. However, I don't know whether it is the same problem or something different. This is how it happens:
1) boot -> STR -> Resume -> STR -> etc. : OK
2) boot -> [STR] -> xine (w/ exit) -> STR -> Resume: X-server does frequent restarts, have to issue halt() via acpi+power-button for a graceful shutdown

By chance I've found a workaround. When I've run xine and would like to do STR aftwerwards I have to leave one xine instance up. Because this is not really convenient after watching a DVD I simply use krusader's view option to instantiate a xine instance on some .mpg-file and press the pause button immediately. This window can be minimized but must not be closed completely. Now STR/Resume works as before.

I've not investigated the xorg-log and not googled much for this. I've read somewhere that it might come from the vesa-tng or drm-module or what ever. I'm still using stable Xorg 7.0 and will check again when 7.1 reaches stable and the recently anounced Intel-drivers will be available.

I've posted this answer here because I 'm too lazy for two posts - sorry.

your STR kernel/config works great but another problem I am having is that when I wake up from STR my cdrom can no longer be read. my /dev/sr0 is missing, this is where it is usually detected. i tried restarting udevd, coldplug, hald and dbus to no avail. in dmesg it does not seem to have detected it.

EDIT: i checked cdrecord -scanbus and it found my cdrom, but still sr0 was not there. i put a cd in and after a while, it simply showed up and dmesg says

I don't have this issue at all because /dev/sr0 is still there after a STR. But my hibernate-ram script does unloading of

Code: Select all

Unloading module snd_seq_dummy...
Unloading module snd_seq_oss...
Unloading module snd_pcm_oss...
Unloading module fuse...
Unloading module uhci_hcd...
Unloading module sr_mod...
Unloading module sg...
Unloading module evdev...
Unloading module ipw2200...
Unloading module ieee80211_crypt_tkip...
Unloading module snd_seq_midi_event...
Unloading module cdrom...
Unloading module ieee80211...
Unloading module firmware_class...
Unloading module snd_seq...
Unloading module ieee80211_crypt...
Unloading module snd_seq_device...

automatically by

Code: Select all

UnloadAllModules yes

within my ram.conf.
Maybe it's worth to check for unload/load sg or evdev module from your suspend script.

[juniper]

Code: Select all

Unloading module snd_seq_dummy...
Unloading module snd_seq_oss...
Unloading module snd_pcm_oss...
Unloading module fuse...
Unloading module uhci_hcd...
Unloading module sr_mod...
Unloading module sg...
Unloading module evdev...
Unloading module ipw2200...
Unloading module ieee80211_crypt_tkip...
Unloading module snd_seq_midi_event...
Unloading module cdrom...
Unloading module ieee80211...
Unloading module firmware_class...
Unloading module snd_seq...
Unloading module ieee80211_crypt...
Unloading module snd_seq_device...

automatically by

Code: Select all

UnloadAllModules yes

within my ram.conf.
Maybe it's worth to check for unload/load sg or evdev module from your suspend script.

ok, that gives me some info. I didn't change my modules.autoload file when i used your kernel. my sleep script unloads all modules listed in my modules.autoload file and only loads the ones in modules.autoload on wake up. i did not have sr_mod or sg in my autoload file. modprobing sr_mod did the trick. I just checked this works after and STR cycle, so great. by the way, if that last list is not a complete set of modules you load, can you post your autoload file?

as for the video card issue. hmmmm, your solution is not the most convenient. i am actually surprised that this isn't a hot topic given how inconvenient it is. I guess i will do a little more poking around.

thanks again, amaroc.

[amaroc]

by the way, if that last list is not a complete set of modules you load, can you post your autoload file?

Autoload is even shorter

Code: Select all

ieee80211
ieee80211_crypt
ieee80211_crypt_tkip
ipw2200
evdev
sg
cdrom
sr_mod
drm
i915
uhci_hcd
fuse

so, the hibernate-script is probably doing lsmod somehow.
I've also tried to unload drm and i915 explicitely - but as expected it didn't work even after chvt (video issue).

as for the video card issue. hmmmm, your solution is not the most convenient. i am actually surprised that this isn't a hot topic given how inconvenient it is. I guess i will do a little more poking around.

You're right, it's an issue and I should have complained about it earlier. But reading all the reports in relation to X-resolution, DRM and video-playback I was more than happy to have
- native X resolution 1920x1400
- boot+text console with about 130x40 chars
- mc and unicode in text and X-console
- DVD playback (used really rarely)

btw - does the solution with a suspended xine-instance work for you? I've verified it here because I had to reboot anyway (new baselayout).

If you got some further information or want me to do some verification let me know.

[juniper]

btw - does the solution with a suspended xine-instance work for you? I've verified it here because I had to reboot anyway (new baselayout).

If you got some further information or want me to do some verification let me know.

i don't really use xine, do you know if this trick works with mplayer? another problem with this solution, is a play most video files over nfs, so i don't know if this trick would work.

[amaroc]

i don't really use xine, do you know if this trick works with mplayer? another problem with this solution, is a play most video files over nfs, so i don't know if this trick would work.

I've done some testing with mplayer, both native and with KDE-frontend KMPlayer. I could do STR cycles w/ and w/o mplayer running - it didn't crashed X at all. If I leave an mplayer instance during STR active I could also start xine after a STR. However, if I used xine after a STR w/o having an instance of either mplayer or xine running it crashed immediately. The X.log shows very likely the same as you have seen:

Code: Select all

Error in I830WaitLpRing(), now is 266949712, start is 266947711
pgetbl_ctl: 0x7ffc0001 pgetbl_err: 0x3
ipeir: 0 iphdr: 1810000
LP ring tail: 580 head: 0 len: 1f801 start 0
eir: 0 esr: 10 emr: ffff
instdone: ffc0 instpm: 0
memmode: 108 instps: f0000
hwstam: fffe ier: 82 imr: 8 iir: 220
space: 129656 wanted 131064
(II) I810(0): [drm] removed 1 reserved context for kernel
(II) I810(0): [drm] unmapping 8192 bytes of SAREA 0xf8849000 at 0xa7847000

Fatal server error:
lockup

Error in I830WaitLpRing(), now is 266951725, start is 266949724
pgetbl_ctl: 0x7ffc0001 pgetbl_err: 0x3
ipeir: 0 iphdr: 1810000
LP ring tail: 588 head: 0 len: 1f801 start 0
eir: 0 esr: 10 emr: ffff
instdone: ffc0 instpm: 0
memmode: 108 instps: f0000
hwstam: ffff ier: 0 imr: ffff iir: 0
space: 129648 wanted 131064

FatalError re-entered, aborting
lockup

So, I seem to have the same issue - but not with mplayer. Anyway, leaving one instance of either mplayer or xine running during STR solves it - at least it doesn't hurt.
I've tested it with a vob from a music DVD.

Regarding network:
My homenetwork has two shares - a buffalo-link-station and a Hifidelio-audio-player. I've tried both smbfs and cifs and couldn't get it working with STR/Resume. I had to end applications (e.g. audio-playback) in order to unmount shares and unload the modules.
Therefore I've switched to ftp via fuse and curlftpfs. A mounted "ftp-share" reconnects automatically after a STR even an open file (RO) doesn't harm an application.

I don't know how to handle this with nfs - but it depends whether you want an application to be alive with open files from the network share. You can check with lsof.
If it's only the video-issue than copy a small vob-file to your i6000 and start mplayer once X is up. I'm pretty shure there are some parameter for mplayer to hide the window - but there are so much...