Originally Posted by MALDATA
maintainers prefer that bug reports get filed upstream if possible
This is because the package maintainers are only responsible for packaging bugs; if the bug is in the code, then the package maintainers are not the best ones to fix the problem. It sounds like your problem is not a packaging issue, so filing upstream is preferable (and helps other distros too). Virtually all packages will have an upstream.
I recently got the upgrade from linux-firmware 20110227-1 to 20110512-2. On my Eee PC 901, this new package makes it hang when waking up from suspend. I rolled back the package to 20110227-1 and everything is fine again, so I figure I should file a bug report.
Looking at the linux-firmware page on the archlinux site
, it refers to the upstream repository at kernel.org. This makes sense, because linux-firmware contains the kernel drivers for devices that have proprietary code.
The kernel newbies site gives a quick guide
to reporting kernel bugs. More detail is in the REPORTING-BUGS file
in the kernel source. And another file contains the list of maintainers of the various parts
Now your problem is a bit complicated, because it involves the suspend/resume functionality, which will use a number of drivers and involves the ACPI. The first part is to figure out which one is causing the hang. It may take a bit of effort to localize the problem, but it will help to know who can best solve it (I'm assuming you've already googled to see if others have the same issue). Log files can be useful, configurations, etc.
One way of narrowing down the problem is to turn off various drivers (my first guess would be the wifi drivers), and see if you can suspend and resume when they are not loaded.
Once you have some information (the more the better, as far as you are able), then send it off to the appropriate maintainer. Good bug reports are appreciated.
Since it is a proprietary driver causing the problem, makes it harder to solve. But there might be a work around. An easy workaround for resume failures is to set up a script to shutdown the driver during the suspend, and then restore it during the resume; if it is something like a network driver, it is no big deal to lose the current state of the connection during a suspend.