Source Code available

That sounds great! However I don't have the knowledge to really understand how this benefits us :?: . Could you please explain?

Things I wonder:
- Can we extend and recompile the kernel for Linpus?
- Does the source include all the drivers?
- Can we use the drivers and kernel from Acer Linpus to recompile for other distros?
- Can we derive full hardware support for other distros / OS' from this?

Thanks!
 
Am I being blind? I don't see the kernel source anywhere. I've only had a quick glance as I should be working at the mo ;)
 
Has anyone compiled the kernel from this source for Ubuntu? I'm surprised to see that Acer used the madwifi driver.
Things I expect this kernel could fix: wifi switch, MMC support (there is a driver) and correctly mirrored (or flipped) webcam. BTW Is Linpus using OSS instead of ALSA?

edit

Sorry even though it's under source, it's just the compiled version
 
Heh, looks like I am blind (or at least a victim or over-eager browser caching) :oops:
 
i tried compiling this but the included configs look like they are for test environments and not the aspire one hardware
 
uberducky said:
i tried compiling this but the included configs look like they are for test environments and not the aspire one hardware

I compiled the kernel on a different machine with Fedora 8 and all build tools.

-The include/asm directory in the source is supposed to be a symlink. Rename the existing 'asm' directory or delete it. I think it should point to the asm-i386 directory (not sure).
- Run 'make clean'
- Run 'make mrproper' to clean up the source
- Use the config file found on your AA1 in the /boot directory and rename it to .config (note the '.') and copy it to the kernel source directory
- Run 'make oldconfig
- Run 'make menuconfig' if you want to change things
- Edit the Makefile and change the extraversion of the kernel (to make the name differ from the existing kernel) for instance to '.9-custom'
- Last run 'make'

So far I have successfully built the kernel, but the following steps would be:

- Copy the whole source tree into the AA1 and run 'make install' (now it is important that it was differently named in the Makefile to not overwrite the existing kernel). This could be done manually if maketools are not installed. I'll check this later.
- Edit the grub configuration
- Reboot
 
I looked at the patch between the stock 2.6.23.9 and acer source. A lot of odd changes.

edit, looks like they applied another drm (esp an updated i915 driver).
edit: there is an sdhci driver that could be interresting. (though we would have to strip all the windows line endings... ugh).
 
I've just had a quick look over the changes to the kernel. They started from a 2.6.23.9 kernel and major items relative to that upstream are as follows:

- Addition of TuxOnIce (http://www.tuxonice.net/) patches to give rapid suspend/resume
- Updating the sk98lin driver to syskonnect's latest version
- Various DRM updates (possibly a backport from 2.6.25.4 judging from backup files also left in the tree)
- Adding bootsplash support (http://www.bootsplash.org/)
- Memory stick/xD card support for the Ricoh R5C832 controller (drivers/mtd/devices) - guessing this is the memory card reader they're using
- Bunch of random patches to the ubi mtd drivers - possibly a backport, given the recent work there upstream
- Adding support for the Realtek r8101 ethernet driver (see http://en.alessiotreglia.com/articl...-r8101-modules-for-realtek-ethernet-adapters/)
- Adding the iwlwifi (http://www.intellinuxwireless.org/?p=iwlwifi) drivers, which are already in 2.6.24
- Added aufs (http://aufs.sourceforge.net/) guessing for doing union filesystems with extra sd cards
- Added squashfs (http://squashfs.sourceforge.net/) - probably for smaller initrd's
- Bunch of mac80211 stuff - probably another backport

There's also some seriously random changes:
- Commenting half the devices out of the piix driver (as well as adding a new one)
- Various little hacks scattered all over the place (setting random variables from 0 to 1; adding a workaround for a particular manufacturer's ropey sdhci implementation)
- Misc netfilter patches (backports? shrug knows)

Good news is that *most* of it looks like it's either upstream already (i.e. get a 2.6.26.x kernel and you're good) or was already out there, as opposed to having to integrate random undocumented changes. OTOH, they do do a fair amount of random shit in there and some of it needs more testing when/if I get myself an Aspire One.
 
palfrey said:
I've just had a quick look over the changes to the kernel.
I am already looking forward to when/if you get your AA1, nice detective work. There are a few other changes, which you can not know about, because they've been added as modules afterwards, and are not included in the source tree.
 
I looked at their kernel source. It's strange. ALSA is disabled. Nothing about xD-card and ms/mspro (but this modules are in linpus).
It looks like they build ALSA separately. But why they don't share source?
I found simple patch for sdhci.c. In ubuntu we use "setpci -d 197b:2381 AE=47" to enable card-reader. They added something like that in power function(don't remember that code). I think it'll solve problem whith card's autodetecting (not only at booting time).
I'll try it today or tomorrow.
Waiting for alsa sources!!! and may be madwifi patches, cause my wifi doesn't work stable.
 
Back
Top