Karmic Ubuntu upup-450 full drivers SCSI 2.6.30.5 smp

[James C]

Switching to Flash 9 seems to have solved all of the browser crashing problems I mentioned earlier.Thanks to playdayz for pointing the fix out.

[Sit Heel Speak]

I am having good luck with a 2.6.31.6 kernel, compiled using my patched gcc. I gave it, besides the regular 4 patches which Barry uses, one additional aufs fix patch, plus the Ingo Molnar real-time preemption patchset and the complete Ubuntu Karmic kernel patchset. I guess it makes sense that if we are going to use application packages from Karmic, then probably we'd better have a Karmic kernel too. Seamonkey is much more stable ***edited: I retract this statement; Seamonkey 1.1.8 is not more stable.*** ***Unless you recompile it and XOrg correctly, using my gcc***

, and my Dell Mini-9 runs about 20 degrees Fahrenheit (11 C) cooler with this kernel, than with Puppies 4.21, Lighthouse, or regular Karmic UPup's 432 and 450 2.6.30.5 kernels.

If anyone wants my source (i.e., the subdir /usr/src/linux-2.6.31.6, tarr'ed up, ready to make menuconfig), and/or the vmlinuz and the /lib/modules/2.6.31.6-rt19 subdir (is there anything else needed, to plunk-in a kernel?) then I will post it.

But most people I presume will want to build it yourself, so here are the patches. Apply them in the order given. Warning: this is not for novices; you will see failed hunks. You must know what you're doing well enough, to go in and fix the failed hunks by hand. During the compile, make will choke on a few sourcefiles where silly mistakes have been made, e.g. symbol exported before it is defined. If you know what you're doing it is trivial to go in and fix these and resume the make.

The alsa patch is an updated one, courtesy of rerwin. The aufs patches are from latest aufs git source. The rt patch is from kernel.org. The Ubuntu patch comes from the Oregon State University mirror, but it is a quarter-million lines, so I have broken it out into 12 parts so you can easily see which hunks fail. Most of the failed hunks fail simply because they have already been incorporated into the mainline vanilla 2.6.31.6 source code.

After applying the Ubuntu patchset, I copied the subdir ubuntu/aufs into fs along with the subdirs for all the other filesystem types, and copied the header file aufs_type.h from /ubuntu/include/linux up to include/linux, then commented out the line in ubuntu/Makefile which builds the Ubuntu aufs. Make will automatically find and build the aufs that you put under fs, though you get warnings about choice value being used outside its choice group. Big deal.

ndiswrapper fails to fully build in 2.6.31.6 --look in my dotconfig and you will see where I disabled it-- owing to the fact that semaphore.h in 2.6.31.6 has had the part that declares init_MUTEX taken out. If this is a problem for you, then I suggest you grab the patch-2.6.31.5-rt18 patch from the rt subdir at kernel.org and use the 2.6.31.5 source with these same other patches applied; I compiled that as well, and it works great. Or, you could try taking the declaration and definition of init_MUTEX out of semaphore.h from 2.6.31.5 and sticking it back into semaphore.h in 2.6.31.6.

2.6.31.6 breaks the XOrg in Upup-450, so you will have to use XVesa...until I post the correctly-compiled XOrg, in a day or two.

2.6.31.5 with all these patches does not break XOrg, but you have to go into rc.sysinit and add a line to make the remount rw line unconditional (I put it immediately underneath the original, which I did not comment out). Note, I tested (and am working) only on a full install.

The patches:

http://www.puppylinux.asia/tpp/SitHeelS ... 6.31.patch
http://www.puppylinux.asia/tpp/SitHeelS ... base.patch
http://www.puppylinux.asia/tpp/SitHeelS ... uild.patch
http://www.puppylinux.asia/tpp/SitHeelS ... lone.patch
http://www.puppylinux.asia/tpp/SitHeelS ... .31.6-rt19
http://www.puppylinux.asia/tpp/SitHeelS ... 8--01.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--02.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--03.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--04.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--05.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--06.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--07.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--08.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--09.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--10.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--11.diff
http://www.puppylinux.asia/tpp/SitHeelS ... 8--12.diff
http://www.puppylinux.asia/tpp/SitHeelS ... .6.31.diff

Here's the original Ubuntu patch which I broke out, in case you want to check my work:

http://www.puppylinux.asia/tpp/SitHeelS ... 14.48.diff

The DOTconfig I used (latest iteration, 22 Nov 2009):
(this is for the Atom cpu, you will want to change it to your own)
http://www.puppylinux.asia/tpp/SitHeelS ... 6.31.6-SHS

[playdayz]

Hey SitHeelSpeak, your kernel 2.6.31 could be made into a pet package for inclusion in Woof, couldn't it? It would be just looking at the other kernel pets in Woof and then imitating them (at least that's how I usually proceed I would be happy to test it by building a Woof with it.

and/or the vmlinuz and the /lib/modules/2.6.31.6-rt19 subdir (is there anything else needed, to plunk-in a kernel?) then I will post it.

Or if you wanted to post the "stuff needed to plunk in a kernel" I would like to try that too. Or to test a uPuplet that you built with that kernel and Woof. I just want to get my hands on that kernel, I don't know the answer to the question there though.

[Sit Heel Speak]

Hey SitHeelSpeak, your kernel 2.6.31 could be made into a pet package for inclusion in Woof, couldn't it?...

Let's see if it's ready-for-prime-time first. I think /boot/vmlinuz and /lib/modules are all you need, right? Here it is:
http://www.murga-linux.com/puppy/viewtopic.php?t=49086

Back up your existing /lib/modules/firmware first, of course.

This kernel breaks the XOrg which is supplied with Karmic Woof. I will release a correctly-compiled XOrg which it does not break, hopefully within the next week.

[playdayz]

Good news about Flash. i built a uPup with Woof 2009-1120 with all of xorg, devx, and SHS's gcc 4.4.1 and Flash 10 seems to be working. It is not crashing everything immediately anyway.

<Add> It even plays full screen in hulu.com. All right! I am using Namoroka/Firefox 3.6b5pre.

I put a note here. http://www.murga-linux.com/puppy/viewtopic.php?t=48984

Flash 10 is also working on a build of 4.32 uPup with all of xorg, in seamonkey 1.1.18,

[Jim1911]

ttuuxxx,

Hope that you are still working with upup. The main issue I still have is being unable to get my HP 5280 all-in-one printer to work. Still a problem with CUPS.

I believe that upup will make a good platform for the update of Living Water since it is already pretty stable and is able to run the latest Bibletime 2.2, if printing issues are resolved. However, I do like dpup, but have been unable to get any Bibletime later than 1.6.5 working on it.

Thanks,
Jim

[gposil]

Jim1911 said:

I do like dpup, but have been unable to get any Bibletime later than 1.6.5 working on it.

Because you asked....now in Dpup repo on ibiblio - Bibletime-2.4.1

http://distro.ibiblio.org/pub/linux/dis ... 6-dpup.pet

Cheers
Guy

[ttuuxxx]

ttuuxxx,

Hope that you are still working with upup. The main issue I still have is being unable to get my HP 5280 all-in-one printer to work. Still a problem with CUPS.

I believe that upup will make a good platform for the update of Living Water since it is already pretty stable and is able to run the latest Bibletime 2.2, if printing issues are resolved. However, I do like dpup, but have been unable to get any Bibletime later than 1.6.5 working on it.

Thanks,
Jim

Hi Jim In the early part of the new year(Jan,Feb) I'll be building the next Living Water, Living waters primary goals is Size, Quality, Outreach for older Pc's, I've had emails that some Christian schools are actually using it day to day.
I like to start with a small base and move forward, I did have the go-ahead from Barry to revive 2 series, 2 Series is small and well what we've done with 2.14X is outstanding compared to what we started with. So possibly 2 series, I also could make 2 versions a retro kernel and a newer kernel. As for Bibletime 2, I was happy with the last package I made up, It was stable, and opened/printed bibles etc. Bibletime is a reader of bibles and maps, Kind of like adobe acrobart, the first release was 1MB the last release is like 40MB, but all it does is read pdfs and prints, lol anyways I'll look at 2 bibletime, an see where we're at
ttuuxxx

[playdayz]

There is an experimental version of uPup made with Woof 20091120, with extra xservers (video drivers) plus the video enhancements GLX and DRI.

uPup 20091120 -> http://www.murga-linux.com/puppy/viewtopic.php?t=49665

With SitHeelspeak's gcc 4.4.1 devx supplement I have compiled the latest ATI Catalyst driver including Catalyst Control Center. It also runs SRWare Iron, Seamonkey 2, and Firefox 3.6b5pre nicely, as well as Flash 10.

[aarf]

As noted by playdayz epdf viewer fails in upup450. Installed a foxit pet and it also failed. Will dig details if required.

[aarf]

as with dpupbeta5,when using gprs connect in upup450 the required link fails to go into the network menu requiring #pgprs-connect in a console to complete the connection. 431 is ok in this respect.

edit: also needed is a wifi down script for when the wifi doesn't perform and you need to switch to gprs.
if the wifi signal is too weak or doesnt perform for some other reason but still connected, it stops the gprs connection from functioning.
wifidown is here.

if you are aiming for the third world then getting gprs connection correct is important because gprs connection is more commonly used there.

[aarf]

SM and PuppyBrowser crashing with .swf files:
attached is the .dat file run action in dpupbeta5 compare that with the run action for /root/.mozilla/pluginreg.dat

edit:run action wasnt the problem because playdayz upup452 has the same run action as upup450 but swf files dont crash the browsers so now at
upup452 instead

[aarf]

as with dpupbeta5,when using gprs connect in upup450 the required link fails to go into the network menu requiring #pgprs-connect in a console to complete the connection. 431 is ok in this respect.

edit: also needed is a wifi down script for when the wifi doesn't perform and you need to switch to gprs.
if the wifi signal is too weak or doesnt perform for some other reason but still connected, it stops the gprs connection from functioning.
wifidown is here.

if you are aiming for the third world then getting gprs connection correct is important because gprs connection is more commonly used there.

also note that the gprs links are now in the menu after how many reboots? this is behavior of earlier puppy not the current 431. in dpupbeta5 and the gprs links are still not there.

[playdayz]

hey aarf, epdf is working with uPup 4.52 http://www.murga-linux.com/puppy/viewtopic.php?t=49665

Also, the Foxit pet that is listed there towards the middle of the first message in the thread is working. Personally I suspect that including glx/dri might be what makes the pdf viewers work, and also makes Flash more stable (no proof, just an impression, well, no proof other than that they work ).

BarryK has made a number of changes to Woof since the version that produced ttuuxxx's 4.50. He has fed back several of the problems and fixes we have found with uPup 4.32 and 4.51 into the latest Woof. I would certainly appreciate your help in testing and improving uPup 4.52, with the goal of finding more problems and fixes to improve Woof and to make a uPup that is useful and stable right from the Woof version.

[aarf]

have been meaning to do a gprs test of 452 but my gprs connection has caught an extreme case of gposils internet malady. perhaps it will not show "packet data connection not available" on my phone come the new year.