Summary: | making pcmcia rc script probe yenta/i82365 automatically | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Alastair Tse (RETIRED) <liquidx> |
Component: | [OLD] Core system | Assignee: | Chad Huneycutt (RETIRED) <chadh> |
Status: | RESOLVED FIXED | ||
Severity: | enhancement | CC: | csimonut, pardsbane |
Priority: | High | ||
Version: | 1.3 | ||
Hardware: | x86 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
/etc/init.d/pcmcia update
/etc/conf.d/pcmcia patch to enable PCIC |
Description
Alastair Tse (RETIRED)
2002-08-21 11:09:16 UTC
Created attachment 3262 [details, diff]
/etc/init.d/pcmcia update
Created attachment 3263 [details, diff]
/etc/conf.d/pcmcia patch to enable PCIC
Hi I'd like to change patch for the init.d script from: + modprobe ${PCIC} + modprobe ds to: + modprobe pcmcia_core ${CORE_OPTS} + modprobe ${PCIC} ${PCIC_OPTS} + modprobe ds which would force those options (if needed!) for pcmcia_core and ${PCIC} to be appended to the modules. This saves us from hacking into modules.conf/modules.d manually --kirill Hi I'd like to change patch for the init.d script from: + modprobe ${PCIC} + modprobe ds to: + modprobe pcmcia_core ${CORE_OPTS} + modprobe ${PCIC} ${PCIC_OPTS} + modprobe ds which would force those options (if needed!) for pcmcia_core and ${PCIC} to be appended to the modules. This saves us from hacking into modules.conf/modules.d manually --kirill Here was my problem with this approach: How can you tell if the user has built the modules into the kernel, left them out, or built them as modules. Your test finds out if they are loaded *if they were built as modules*. We made the decision to make the user explicitly load his basic pcmcia modules (pcmcia-core,yenta_socket,ds, for example). That is not to say that we won't change the scripts. We also didn't want init scripts to be loading modules, although now that we have alsa, firewall, etc., this is not an issue. I guess I didn't explain why this would be helpful :) I'm on a laptop that needs the PCMCIA PCIC and associated modules to be re-loaded (rmmod/insmod) everytime I suspend/resume via apm. So adding the loading and unloading of modules in the pcmcia init script would help cleanly unload these modules and then reload them again. If they didn't have those compiled as modules, then the modprobes would just fail, we could obviously just redirect the output elsewhere and make it silent. If they didn't have the pcmcia support compiled anyway, then why would they start the pcmcia init scripts in the first place. If they have the modules compiled, then all is well and good. As for the unloading, if they compiled it in to the kernel, they can't unload it anyway, so that shouldn't conflict with the them not being able to do rmmods. I just figured this it is more logical to make the pcmcia script smarter and log the modules for the user rather than having them needing to add it tie modules.autoload. Not that it is very difficult to add it in there. *** Bug 12177 has been marked as a duplicate of this bug. *** Okay, I am convinced about loading the modules in the init scripts. I'll see what I can do about getting this init script in the latest pcmcia version soon. Okay, I made the changes to /etc/init.d/pcmcia and /etc/conf.d/pcmcia. Please remerge pcmcia-cs-3.2.3 and make sure the new init script works for you. I haven't heard anything on this bug, so I am going to close it out. There is a new bug about some errors with the script, though, so we'll continue there. Its a bit late, but I can confirm this works. |