www.sixfingeredman.net .................................................. ::. . . . . . . |
HOME readme brain ideas todo writing photos graphics projects quotes recipes books movies links old site |
linux on thinkpad 380edDSL 2.2b:
Kubuntu 5.10 installer:
Debian 3.1r1a:
Probably because you haven't compiled your own kernel. With many ThinkPads, you need to compile the kernel with CONFIG_APM_ALLOW_INTS=y; otherwise the machine will hang up when trying to suspend. thoodATmailexciteDOTcom The APM BIOS rejects suspend (or hibernate) requests under certain conditions: e.g., on my TP600, when my Xircom Ethernet/modem card is inserted and the machine is on AC power. If I do "apm --suspend" then the kernel apm driver hears about the request first and tells apmd which runs apmd_proxy which runs the scripts in /etc/apm/event.d which do things like eject the PCMCIA cards; only then does the apm driver pass the request on to the BIOS. (I think.) If I do "tpctl --suspend" on the other hand, the BIOS hears about the request first and, if the PCMCIA card is still inserted, it rejects the request with a couple of beeps (high---low). "apm --suspend" is therefore to be preferred to "tpctl --suspend". Unfortunately there is no alternative to "tpctl --hibernate": the apm program and the kernel apm driver don't know how to ask for hibernation. Solution: write a little script that does "cardctl eject ; tpctl --hib". Fn-F4 does the same as "tpctl --suspend" and Fn-F12 does the same as "tpctl --hibernate". There is a problem with the current (3.0final-1) Debian apmd_proxy script which makes the situation worse. Suppose I've ejected PCMCIA cards and requested suspend or hibernate using tpctl; the BIOS accepts the request and tells the apm driver; the apm driver tells apmd; apmd runs apmd_proxy; but apmd_proxy ignores the event if the machine is on AC power. I have no idea why this "feature" was added to the script. I submitted a bug report against it and the author responded by saying that he agreed that rejecting system suspends shouldn't be the default behavior. Since /etc/apm/apmd_proxy is a conf file, though, he says that it is normal that I should edit it to suit my system. I think that for something like this the "reject system suspend" behavior should be selectable by setting a variable in something like /etc/default/apmd_proxy. Prior to kernel 2.2.16preSomething, there was a problem with the apm driver such that it did not handle repeated suspend requests issued by the BIOS and the result was that hibernation would pause for several seconds before proceeding. This problem was not related to the above and it has been fixed in the latest stable kernels. Everyone seems to have a lot of trouble configuring the Crystal cs4236 sound card in some thinkpad 38x laptops. The given solution is to copy the proper settings from windows. However, I found the following settings to work flawlessly on my thinkpad 380ED with Mandrake (you can use the sndconfig utility to edit /etc/modules.conf - available in Red Hat or as separate download - or edit /etc/modules.conf directly): In the /usr/sbin/sndconfig utility, select the Crystal cs4232 sound card and enter the folowing settings when prompted: io=0x534 irq=5 dma=1 dma2=0 mpuio=0x330 mpuirq=7 For double checking, or if you want to edit yourself, here is my /etc/modules.conf: options op13 io=0x388 alias sound-slot-0 cs4232 options sound dmabuf=1 alias synth0 op13 options op13 io=0x388 options cs4232 io=0x534 irq=5 dma=1 dma2=0 mpuio=0x330 mpuirq=7 linuxATdaobDOTddsDOTnl |
"Oh, bother," said the Borg. "We've assimilated Pooh."