> Just one idle thought on disabling highmem on 32-bit : I used to have > a box with 4GB and radeon integrated-graphics - that took up some of > the RAM, so I assumed that in 32-bit I would not need highmem: wrong, > I got much less RAM and I think I had to go for the "big" kernel > congig option (maybe it was HIGHMEM_64G ?) to get the full 3.75 GB or > whatever it was. Just an idle thought because that box died, so I > can't go back and check it.
That's an intriguing idea! Mine's a invidious card. It should have its own RAM, but it has to map it somewhere. Boxes seem to die, that's why I like to build identical twins if possible. > I don't have any ideas why that would happen if you do not have a > radeon video chip. Have you looked at all the details about the > drivers in the X log ? From memory, several drivers may be listed, in > order, and then some will be discarded because either the driver does > not exist (that's in my systems - I don't build all of them) or > because of other errors (i.e. its the wrong driver). That's why I borrowed the one from the X86-64 system, mostly to get the right drivers. I disabled the ati glamor stanza mentioned in the book, and the radeon went away. But then, strangely, "X -configure" failed to find any devices. I didn't want to seed xorg.conf.d, hoping to maximize adaptability, but there were a few files there in the 64-bitter, so I brought those here and X came up--with the same kernel, of course. This isn't resolved yet. CTL-ALT-+/- doesn't change screen sizes, and "Don't Zoom" is off. So I've still got unresolved issues! > > You might also find it useful to compare the Xorg log in the working > x86_64, and confirm that you do indeed have the nouveau driver built > in i686, and also that whichever kernel settings are needed for > nouveau on x86_64 are also enabled in your 32-bit kernel config. Yes, it is/was. I checked /lib/modules/... > My memory says that 32-bit kernel configs were VERY different from 64- > bit, I will not be surprised if you have to make other changes from > those you had to make on x86_64. Yes, I did. But I did diff the configs and looked for differences NOT related to address lengths, to make sure the other stuff was the same. > After that, if I google for 'linux CPU or task stall' I get a lot of > results, and there are also some related searches listed there which > might be more like what you are seeing. No idea if any of them are There was one from the kernel docs that talks about it in some depth, but I didn't get any reccomendations for kernel config setting from it. Also unresolved. > ĸen Thanks for your thoughts, Ken. The "fat lady" hasn't appeared on stage yet! > -- > Brave Sir Nigel ran away! When reality reared its ugly head, Sir > Nigel turned his tail and fled. Brave brave brave Sir Nigel. Mom said, "Don't count your chickens before they hatch." You may find "buyer's remorse" abroad in the land before 50 goes down. -- Paul Rogers paulgrog...@fastmail.fm Rogers' Second Law: "Everything you do communicates." (I do not personally endorse any additions after this line. TANSTAAFL :-) -- http://www.fastmail.com - mmm... Fastmail... -- http://lists.linuxfromscratch.org/listinfo/blfs-support FAQ: http://www.linuxfromscratch.org/blfs/faq.html Unsubscribe: See the above information page
