| |
| |
| |
|
Page: 1 2
Comments:
<0> when I try to make realtime-lsm I get this error
<0> http://pastebin.com/728843
<0> any ideas?
<1> Hi I'm needing a bit of help
<1> the problems is that I'm in a hurry
<1> I have a source of a kernel
<1> which it's exactly the same as one I've compiled and running
<1> I don't have the source of the last one
<1> but, as I said, It's exactly the same as the one I have
<1> but, when I used the source I have to build a module
<1> it says this when I'm loading it:
<1> version magic '2.6.15-1-486 486 gcc-3.3' should be '2.6.15-486 preempt 486 gcc-3.3'
<1> Question: Which file should I modify to say "this is kernel <2.6.15-486 preempt 486>" ???
<1> Please, anyone could help me? :\
<1> Ok :(
<1> Come on guys!
<1> HELP!!!!
<1> :(
<1> does anyone read me? :\
<2> Partial guess: /usr/src/linux/include/linux/version.h
<2> The UTS_RELEASE define in there may be exactly what you're looking for
<2> However, whether or not it makes *any* sense to wedge the phrase "preempt" into a given string has yet to be seen.
<1> Thanks
<2> Smarter it would be to simply get the source from the last kernel, or just *build a new one*
<2> Changing that string isn't going to let you load an explicitly versioned module
<2> ...and I'd have my doubts about lying to a module or the kernel about preemptive tasking
<1> I would do that If I were in a faster computer than a Pentium 1 166 now
<2> I think you're begging for kernel panics, personally.
<2> You don't sleep?
<1> the problem Is that I can't get the original source
<2> In any case, that sounds like good incentive for learning how to package a kernel into something you can transfer to other machines and have it *work*, provided you have something else faster than that to work with
<1> and I can't build a new one, besides of the time it would take, that the bootloader keeps failing writing on to MBR
<2> You "can't" or you don't feel like looking for it
<1> I can't
<1> I'm not at the place I did it
<2> Turn off BIOS protection of the boot sector if it can't write to the MBR.
<1> I'm about 1000 km away
<2> You're aware that you're on the internet, right?
<1> there's no such one enabled
<2> Pretty much every kernel from 1.0 on is readily available on this thing if you but look
<2> Shoot the machine then,.
<2> Writing to sector 0 on an IDE disk is trivial
<2> If the machine can't do it, it or at least the disk itself is ready for the boneyard.
<1> ok, as I said, It keeps failing
<1> well, It's all what I have here
<1> I can't afford a new one
<1> but, thanks for the good intentions :\
<1> Well, I see that the "UTS_RELEASE" was correct... :\
<2> I'd just like to point out that as sorry as you're feeling for yourself for having a P166, I was saddled with a P150 notebook about a year ago, and originally put up my first internet site on a 486 DX-33
<2> Things that are worthwhile sometimes take time.
<2> Just be thankful you don't have to hand-hold a kernel build all night long and that it can mostly manage to complete on it's own.
<2> Patience wins.
<2> I'm on an Athlon 3200 and I'm still about to go out for some coffee while a gargantuan run completes here. ;)
<1> yes and I agreed, in other case I would start compiling it again, but, as I said, I really tried many other ways
<1> I'm not a master with this
<1> I just do what I can and seems not to be enough
<1> sorry if you find me "stupid" with my arguments
<1> my apologies
<3> compile on a different computer, just copy the .config
<1> This is the only one I have
<1> and I CAN'T reach the original source
<1> If could copy the .config, I would use the whole thing
<3> ahm, then change the UTS_RELEASE (or whatever) manually..
<3> and compile on someone else's computer
<1> It was in the correct way
<1> I think I would need a Cyber coffee to do that, but, which HAS linux on it? :\
<3> i think it would take some 10 minutes on a computer like metawire.org (they give free accounts)
<3> but they're down currently :P
<1> GREAT! :\
<1> But It's a good Idea
<3> do you have access to the source tree in which the modules was built?
<1> Yes, I'm trying to load a module with a kernel I have that's exactly the same
<1> as the one it's loaded
<3> then change the string in the module, it should be somewhere in the .c file processes by modpost
<3> processed
<1> let me check
<2> If it were "exactly" the same, you wouldn't have to tamper with it
<3> can't you replace the running kernel with the one you're building and reboot?
<1> That's the problem, I'm not building it
<1> I just have the source
<3> how are you building the module then?
<1> the modules requieres kernel headers, right?
<1> well, I have them on the source
<1> do you understand?
<1> ok, gonna go to a cyber and compile the whole kernel again
<2> The question should probably be, "Are you afraid yet?", but that's just my cheery outlook. ;)
<1> Dagmar: that's for your sarcasm
<1> It's really a great help for my
<1> s/my/me/
<1> Thanks both!
<3> be sure to save the entire build tree somewhere :)
<2> heheh
<4> yay
<4> I compiled my first kernel happy hpapy
<4> *happy
<5> spanglesontoast: congrats
<4> ty
<4> now all I need to do is fix the scary errors at bootup
<5> heh
<4> I did find it hard to hunt people down and ask them how to setup the thing
<4> so I may make a guide
<6> spanglesontoast: most of us have no problem withe errors at boot
<6> ]the errors at boot are usually easily resolved by recompiling the kernel with different configuration or using a different kernel
<6> the biggest issue are the errors which occur when the machine is under load
<7> CONFIG_PACKET_MMAP - Will this help generic network speed? Is it worth enableing?
<6> whats the help say?
<7> If you are unsure say N. :-)
<6> then that's your answer :P
<6> if it was safe & would improve it, the default would be yes :P
<7> Packet protocol driver will use an IO mechanism that results in faster communication.
<7> So I know that but if it is possible to increase the performance of my system then it would be nice of course.
<7> I know what a packet is, and I know what a protocol is, but the description is just too generic for me really.
<7> Will it make a difference if I remove parallel port support if I do not plan on using my parallel port for anything?
<6> fholmes: removing modules generally will affect your system in no way
<6> if it's compiled into the kernel it will reduce the footprint in ram by a few bytes, but it's overall affect will probably be negligable
<6> but in the ideal kernel only the things you need too mount / will be in the kernel & everything else you might need should be compiled as modules
<6> the only advantage too not selecting the modules not needed by your system would be too reduce overall compile time & the disk footprint said modules would consume
<7> I have heard obviously that you should only compile in the things you need, but still. That makes more sense to me now, thank you.
<8> anyone here good with usb voodoo magic? :)
<8> I have a usb 2.0 card reader that has given me fits since i first bought it
<8> usb 1-9: device not accepting address 6, error -110
<8> get that for all the address of the card...
<8> but i thought ohci is usb 1.1 and ehci 2.0
<8> if i remove ohci module it doesnt see it at all on only ehci
<8> insert ohci and it at least finds something
<6> biohazrd: even if sumone was, what makes you think they'd admit it & then have you harr*** them with a vague problem for the next few weeks?
<8> lol
<6> no lamer hunting please
<8> I was just looking for help
<8> someone told me to try various kernel versions as sometimes usb works with some devices with one version and broken in another
<6> uhuh, funny how many think help & vague data go hand in hand
<8> can i pastebin the error from dmesg for you to look at?
<6> not for i, as I dont do usb
<6> or microsoft or anything related too proprietry interfaces
<8> i have searched google for weeks and havent found a solution
<6> it could be that there are none or that you are searching on the wrong criteria
<6> you found the problem in google at least?
<8> sort of, but not specific to this
<6> ok then your criteria is probably wrong, you need too generalise your search a tad more
<6> or look through the lkml's
<2> Sounds to me like the board's USB implementation is lame
<6> quite possibly
<2> I've never had a device *refuse* connectivity. They either work or screw up the bus because they're so noncompliant
<6> there are many devices out there that plain **** because the manufacturers did their implementation on the cheap
<2> I can think of some webcams like that
<2> "CyberGear" toy spycams, for instance
<6> & it aint restricted too usb either
<2> Che-eez cameras
<6> 8139 chipset for nics is a good example :)
Return to
#kernel
or
Go to some related
logs:
ubuntu java-package multiverse sources.list
#css
terminal service ubuntu clipboard
meaning of pollsa
aptitude ajunta linux
#perl
#ubuntu
#linux
#linux
#php
|
|