Summary: | ACPI backlight control is broken for Samsung R519 | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | PM <mitaspiotr> |
Component: | [OLD] Unspecified | Assignee: | Gentoo Kernel Bug Wranglers and Kernel Maintainers <kernel> |
Status: | RESOLVED UPSTREAM | ||
Severity: | normal | CC: | gregkh |
Priority: | High | ||
Version: | unspecified | ||
Hardware: | AMD64 | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Attachments: |
my kernel config
X log |
Description
PM
2011-01-10 18:59:05 UTC
Created attachment 259487 [details]
my kernel config
Created attachment 259490 [details]
X log
You're not the only one with backlight problems on Samsung laptops; apparently they didn't implement the normal ACPI controls thus requiring a custom driver -- which unfortunately doesn't seem to be in mainline kernels. Take a look at this: https://patchwork.kernel.org/patch/43284/ Not sure if the patch will still apply to your quite-recent kernel, but maybe without too much hacking... Thanks for the reply. However this patch is over a year old. If it worked, wouldn't it be in the kernel as of now? I tried applying it, but it failed and I'm not very good (or rather at all good) at kernel hacking... I also found a driver in the file drivers/staging/samsung-laptop/samsung-laptop.c which looks very similar to the one from your patch. I tried it before and it didn't work. On the other hand I found the way to set the brightness using setpci on some blog. So I'm kind of saved for now... This samsung R519 brightness problem is at the kernel level, not X, which relies on a working kernel driver for brightness control. The workaround pokes PCI registers manually, which is the sort of thing that kernel drivers normally handle. You're right that the samsung brightness driver did make it's way into staging, and the reason it's still in staging is that the author lacked the relevant hardware and didn't get enough feedback that it worked. In fact it _doesn't_ work yet for R519 model, so somebody who has one should give him that feedback and offer to test proposed patches -- are you game? You'll be asked to use one of the standard kernels (gentoo-sources or vanilla-sources) for testing any such patches. If not, this bug will have to be closed until a willing tester shows up. Sure, I can do that. How should I contact the author? Directly? Or post to some mailing list? Actually, the author happens to be a gentoo developer - Greg Kroah-Hartman. So I'll go ahead and assign this bug to him. Sorry, but I only had a N128 laptop, and got that to work, and wrote the driver that is in the kernel tree for it. Samsung only supports that single laptop for Linux, for any other, sorry, I really can't help as I don't have one, and the interface is different. Hm guess we have to give up then on getting a real driver until a kernel developer ends up with one of these laptops... sorry we couldn't be more help, but at least you've got a workaround for now I just saw Greg's post on Planet Gentoo: http://www.kroah.com/log/linux/samsung_laptop.html, so it seems the situation changed rather drastically from "Samsung only supports that single laptop for Linux" to "Linux can now properly support all known Samsung laptop devices" ;) However my R519 still doesn't work with that driver (module won't load with "no such device" error). So, if things really started moving, maybe I can help with adding support for this model? I'm reopening the bug, feel free to close it if the situation hasn't really changed. Why on Earth would anyone even contemplate reassigning a bug to bug-wranglers? It sounds like you have a request to support an unsupported laptop. You will need to take this upstream to bugzilla.kernel.org |