Re: Dell Inspiron 2500
From: "Baptiste Malguy" <baptiste@malguy.net>
>
> 1) My Inspiron 2500:
> --------------------
>
> I have to say that even if now it works pretty well with most of the
> functionnalities ready, it has not always been true.
> Many trouble with ALSA, X, APM, ACPI, PCMCIA ...
I have no trouble with X. ALSA was simple once I finally realized that the
.deb package wasn't creating a required link from /etc/modutils/alsa ->
/etc/alsa/modutil/0.9. APM is definitely toast. PCMCIA works fine for me,
but that seems fairly dependent on the cards you're using.
> My advices:
>
> About ACPI: just forget 2.2.x kernels since it is now reliable (>= 2.4.16
I agree.
> But there is still one main problem: it seems ACPI really hates to share
> its IRQ (9) with any other devices. Unfortunately, the PCI management
> system also gives this IRQ to the PCMCIA/Cardbus driver :(
> Which makes the machine hanging. Not always crashing, but most of the
> time, it does crash :(
I'm not having this problem. I'm getting the kernel messages:
IRQ routing conflict for 01:03.n, have irq 9, want irq 5
but everything is working. You are supposed to be able to turn off the ACPI
handling of irq routing with a kernel parameter (I think 'pci=noacpi') at
boot time, but I haven't tried it. It might make a difference for you.
I'm really just getting into ACPI, and haven't tried the latest patch.
> So I traced a little, read the doc to find a way to say the kernel to
> provide it another IRQ as Windows does (it gives IRQ 5 to the PCMCIA
> driver). Nothing available it seems. So I wrote a very dirty patch, but
> which makes me happy. That's the first time I "read" some kernel source,
> so please don't shout on me :):
> http://babou.org/ls/sources/patch-2.4.16-inspiron2500
> Works at least for 2.4.16 and 2.4.17
> You have to say to apply the patch via make config/menuconfig/...
> About the sound: the OSS driver is obviously not efficient at all,
probably. I don't require sound on my laptop, so I'm happy enough that I
got OSS sound last week.
> About the extra keys: I use the package hotkeys. It can run some
> commands, and do some basis stuff. I personally use it to set the
> sound level, play/stop/prev/next with xmms
Thanks. I don't know that I'll use it the way you do, but it's handy.
> To use the switch off button, as a shutdown, I use the package acpid.
> It watchs /proc/acpi/event to discover what happends with ACPI,
> including the suspend and power buttons. So XDM is enough to power off
Where is the 'suspend' on a inspiron 2500? The blue suspend key (Fn-Esc)
doesn't appear to send any signal that's caught by ACPI (but I'm just using
the ACPI support in the kernel, not the latest patch - would the patch
help?).
>
> About the touchpad: my /etc/gpm.conf:
> device=/dev/psaux
> responsiveness=
> repeat_type=raw
> type=ps2
> append=""
> sample_rate=
>
> and XF86Config-4:
> ...
> Section "InputDevice"
> Identifier "Touchpad"
> Driver "mouse"
> Option "CorePointer"
> Option "Device" "/dev/gpmdata"
> Option "Protocol" "PS/2"
> ...
>
> I prefer to say it as I personaly spent sometimes on it.
>
> Also, I experienced NO problem about 16bit mode under XFree86.
> If you want my complete XF86Config-4 to test, just ask me.
>
> For additionnal information about it (french/english):
>
> http://babou.org/equipment/lebanon.fr.html
> http://babou.org/equipment/lebanon.en.html
>
> 2) My questions:
> ----------------
>
> About this famous suspend-to-disk partition, actually I directly put
> another hard disk in the laptop (from my previous one) exchanging the
> two hard disks. So I didn't even look what were on the original disk.
> Just formatted it for my older laptop.
There is definitely no s2d partition on the i2500 as delivered. I haven't
been able to figure out _what_ Dell does on its 8100s with RedHat - but I
imagine we can duplicate that with enough imagination :-)
> I created a IBM Thinkpad hibernation partition, FAT16/32 partitions,
> at the beginning and the end of the disk, as primary and logical
> partition. Nothing happened.
>
> However, dmesg tells me:
> ACPI: System firmware supports S0 S1 S4 S5
> So ! It is available, but ?
First, the 'firmware' supports those sleep levels, but afaik there's no
software behind it to actually force a suspend. In fact, on the acpi-devel
list in the last couple of days Pavel Machek (one of the developers) said:
"[If you say S4 works I'll not believe you, and if you say S3 works,
I'll be very pleasently surprised. S1 has some chance of working.]"
Just like apcid runs init when you press the power button, I think we'll
need to supply the functionality to do the suspend to disk. I haven't
finished reading the acpi docs either...
My experience was that S1 seems to work (echo 1 > /proc/acpi/sleep), but the
inspiron is incapable of resuming from it (the keyboard can't regain
control). Hitting the power button does put it through an orderly shutdown.
S4 is ugly. I haven't checked exactly what it did, I just know it wasn't
nice, and I couldn't power down properly :-) S5 looks like it wanted to do
a suspend to disk but all it really did was an abrupt power-off.
If you're interested in working with me on this, I'd suggest we keep trying
and report back to this list and acpi-devel with our results. We seem to be
at a pretty similar point.
> I gave a look at the suspend-to-disk patch, because I experienced how
> Windows 2000 does suspend-to-disk: it actually uses its own filesystem,
> creating a hiber....sys file. But I use ext3, and the patch does not
> handle the corresponding kernel threads (kjournald), so it does not work
>
> So, here I would be very happy to get any information about it.
as would I. I haven't tried it yet - in fact my next change to the kernel
was going to be to add ext3 because a number of testers of acpi have
enthused about the value of journaling filesystems when they're testing
power management.
> Also, I cannot switch off the display. I mean, it is always powered or I
> have to close the laptop. xset dpms force off just does nothing.
> Do you experience the same problem ? Does it work fine for someone else?
I've just tried it. Not having a dark room, I can't be certain it really
powered off, but it blanked the screen. I know when my screen blanks
normally (after 15 minutes of inactivity) it does _not_ power off.
> It also does not switch off on the text console.
>
> Last question: I saw that under windows, the touchpad driver can make the
> right side of the pad working as scrolling wheel. If anyone knows to do it
under
> Linux with a detailed explanation (programs, conf files ...) ?
I quit trying to make the TP do anything but ordinary mouse movements,
because I found it too flakey - and I hate how it works under Windows ME -
but if you want it to work as a wheel mouse, you'd definitely need it to be
an IMPS/2 not a PS/2. I guess I have it working as a wheel mouse in WinME
but I can't figure out how to control the wheel function. It's always
scrolling on me when I don't want it to. Perhaps if I had this
functionality in Linux I'd get use to it, but I only use Windows for Quicken
:-)
derek
Reply to: