Re: [iPAQ] 2.4.2 PCMCIA/CF IRQ problems

On Friday 30 March 2001 10:38, Nils Faerber wrote:
> Hi all!
> Using the current CVS kernel snapshot (03/29) I have problems with
> PCMCIA or compact flash cards. For example a CF flash card is recognized
> once, after a few seconds I get a kernel message that:
>
> irq_desc[irq].action = 0
> disabled irq 42

I just got my 2.4.2 kernel (from CVS as of last night) build working with
jffs2 and I'm seeing the same problems (I see irq 42 getting disabled with my
CF ethernet card). In fact I'm also seeing lots of:

         irq_desc[irq].action = 0
         disabled irq 17

as well as the ones above.

>
> and after that the card is dead. After removal and insertion again I get
> lots of
>
> ide0: ports already in use, skipping probe

Hmmm...with my CF memory card I see:

ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx
cardmgr[61]: get dev info on socket 0 failed: resource temporarily unavailable

>
> and finally failling initialization.
> A CF ethernet card (Kingston - which used to work with 2.4.0) works for
> a few seconds and I get the above mentioned "disabled irq..." messages
> and the card is dead again. Here a removal and re-insertion works; but
> just for few seconds.
> I have no real idea what the reason might be and hope that someone else
> does; especially the person that did the changes to the IRQ or PCMCIA
> stuff ;))

Actually here I'm using a Hawking Tech CF682T and about 20 seconds after
insertion I see:

         irq_desc[irq].action = 0
         disabled irq 42

But the card seems to be usable and I can ifconfig and ping other boxes on my
LAN. Interesting.

> PS: I also noticed that the power button now suspends and resumes,
> great! Just the X-Server does not seem to be very happy about this.
> After resume the screen is blanked (i.e. LCD off) for some reason and if
> the screen was on at suspend time it will not come back on after resume
> since the X-server still thinks it is already on ;) Maybe we should just
> keep the LCD on after resume... (on resume it is switched on for about a
> 10th of a second)

Indeed. It seems like the screensaver mechanism forgets which state is on or
off if the ipaq is suspended while the LCD is on.

The other strange behavior I'm having with my iPAQ is that the LCD backlight
control seems to be in a very strange state after a resume. If I activate the
screensaver first (to avoid the above mentioned bug) and then suspend, when I
resume the backlight will sometimes flicker as I move my stylus across the
screen. A more reliable way to get this behavior is to activate and then
deactivate the screensaver after an iteration of suspend/resume. In this case
the backlight will flicker very badly when touching the touchscreen.

--
Regards, Devin.