| |
| |
| |
|
Page: 1 2
Comments:
<0> hellooo, im just about to start compiling my own kernel, so i have copied the config for my current kernel to /usr/src/kernel-2.6.8/.conf will menuconfig use that as the starting point?
<1> smoo2: name it .config and then it will
<1> i don't know if .conf would work
<0> ah ok! thanks
<0> balrog-kun, is there a command to check which modules existin the current kernel (running debian)
<1> smoo2: i guess something like "grep =m .config" might work
<1> ah
<1> there's some file in /lib/modules containing the list of modules too
<0> ah ok
<0> as a general point, will having less in the kernel increase the speed or memory useage in any noticable way? eg loosing things like ntfs support which i never use
<1> smoo2: it might save some kilobytes of memory
<1> it won't increase the speed
<2> how can you change the kjournald commit interval?
<3> kjournald?
<4> gb: 2.6ism
<4> Renacor: no idea sorry.. nothing in sys that I can see..
<3> cef: I thought thecommit interval was adjustable filesystem by filesystem using tunefs
<4> gb: perhaps it is.. I've not really played with the fs side much
<3> at least thats how I set them in 2.4
<3> guess some people are just too lazy too rtfm in /usr/src/linux/Documentation/filesystems/
<4> nah, just too busy
<4> speaking of which, work to do
<3> hehe
<5> Hola.
<5> Just curious. When will 2.8 be released, approximately?
<5> I mean... '06, '07?
<3> indio: wigar
<5> gb: What?
<1> i think no one knows that
<1> it will be new version of 2.6 until someone introduces some really big changes
<1> and these changes are often unplanned
<6> like kernel panics? :-d
<5> wigar = when i get a release?
<5> hehe
<7> my xircom (xircom_cs) network card work fine under normal use, but hangs when tranfering large files??
<7> it's doing routing , nat, and ip masquerade also, would that contribute to the proble in any way
<7> .problem in any way
<6> indio: most of my crap is in $HOME :)
<6> choke: how well do you trust the pcmcia/cardbus in your laptop? that bus, afaik only handles 54mbits of throughoutput (eg, like usb 1.1, but only 4-5 times better)
<7> how do you know my bus, i only mentioned the card's kernel module
<5> Supaplex: After I wrote that, I realized I only freed 200MB. :)
<7> CardBus bridge: Texas Instruments PCI1131
<7> ok, i got a wifi card in it too, but it doesnt hang, also t unhang the card, i do a 'cardctl eject 1' and then a 'cardctl insert 1' and then it works fine
<5> Anyone using "The low performance USB storage driver (ub)."?
<7> is their a way to prevent the card from hanging, like by putting limits on the data throughput
<6> choke: the _cs gave it away
<7> i got a 108MB wifi card, hmm m:-/
<6> choke: I wish i knew. I have a p266mmx laptop that apparently suffers issues similar to what you've described. Only the simple 802.11b wireless may have worked (memory escapes me)
<6> yea, that's what marketing does for ya...
<6> humph, 88mb in Desktop/Trash. that had to be there for ages.
<7> i need the g+ to get a decent signal, but that card works fine, the xircom 10/100 card, will hang but the wifi card works fine
<6> my xircom seems to work "ok". I've never really put it in its paces to discover otherwise.
<7> im not even transfering at that fast of a rate, the internet is slower then 10MBit and i even tried limiting the connection along the way to 10Mbit 1/2 duplex, it bewonders me!
<7> kybe!
<8> This may be more of a question for #gentoo, but here goes: Can I 'emerge -C linux-headers' and symlink /usr/src/linux/include/linux to /usr/include/linux?
<8> The gentoo linux-headers package is for kernel 2.6.11 :-/
<1> why would you do that?
<1> besides, you usually need also asm and asm-generic in /usr/include/ in addition to linux
<9> what does it means :
<9> PREMPT_NONE
<9> PREMPT
<9> PREMPT_VOLUNTERY
<9> in .config of the kernel
<9> its sheduler? parameters ?
<8> balrog-kun: I'm just worried because the headers installed in /usr/include/linux are for kernel 2.6.11 and I'm running 2.6.16-rc2-mm1.
<10> hi :)
<10> are there any notable restrictions/caveats when invoking fsync regarding permissions? Over at GnomeVFS we have to flush all the ressources ***ociated with usually (p)mounted USB stick media before invoking (p)umount. Does the fsync fd have to be r/w?
<10> 11:11:15 <Quiznos> Manny do you mean to say "does the fd given to fsync() have to have rw perms"?
<10> 11:11:20 <10> Quiznos, yes
<10> (from #linux)
<10> ok, it does have to be r/w, according to man fsync 2
<11> what is inter_module_get() being replaced with ?
<11> in the current linux kernel that is ?
<12> hiya... I need a little help with something. I've just done a fair bit of porting on a kernel for it to work on a certain arm-based embedded system I've got here. Now, I want to make a patch so that I can take a standard vanilla 2.4.27 kernel, apply this one patch, and come out the other side with everything sorted - all the various driver files etc I've edited, and the .config I've set up
<12> is it possible?
<3> willwork4foo: Anythings possible
<12> gb: do you know how to do it?
<3> willwork4foo: of course not :)
<12> outstanding.
<12> thanks for your help :)
<3> Would you rather I lie too you & give you ***istance that would waste your time?
<12> fair enough
<3> writing diffs is tedious
<12> yes. I have had a look already.
<12> seems fairly horrible.
<3> but there are tools out there too do it apparrently
<3> willwork4foo: sorry I cant offer you anything positive
<3> mind you bofhs are not suppossed too be helpful :)
<12> this is true.
<10> I fail to make int fd = open ("/media/myusbmount", O_SYNC | O_DSYNC); if (fd != -1) { fsync(fd); close (fd); } work as expected. The USB stick hasn't finished writing when fsync is left
<10> because sync()/sys_sync() seems to work (when running the "sync" program, everything is flushed), I suppose I'll have to flush the block device and not just it's mount point fd, right?
<1> i didn't know you can call open() on a mountpoint
<10> ouch
<10> ok, so I'll have to deal with the device node directly, right?
<1> i don't know, i think you cannot flush a single device on request
<1> you can flush all block devices by calling sync(), but i might be wrong
<10> I hoped to be able to directly poke around with the sync_blockdev API
<10> but these struct don't seem to be exported
<1> hmm, reading about fsync it seems you can do that
<1> you just need to call it on a descriptor for the device node
<1> i might be wrong again, though
<13> re
<13> ugh
<13> sorry, I didn't realize the split
<13> 15:32:59 <10> I just tried doing that
<13> was the last line I read
<14> hi
<14> is the 2.6.15 linux kernel so stable that i can put that in to server?
<14> and what are those big bugs in the 2.6 kernel?
<3> balrog-kun: no threats please
<3> paistis: roaches
<14> roaches?
<3> well roaches can be big bugs & it was the first thing too come too mind after reading what you typed
<14> okey
<15> bah. stock fc4 kernel has "agp aperture conflicts with pci mappings" or sth to that effect. roll-my own has the same complaint but hangs completely (no magic sysrq even)
<15> can't get a self-compiled kernel, with the fc4 stock kernel .config to get past AGP detection :o
<3> mord: What is fc4?
<15> gb: waitasec, booting.
<15> gb: 2.6.15-1.1831_FC4
<15> gb: or did you mean fedora core 4?
<3> mord: no I'm trying too guess stock what?
<3> redcarp, debian, suse, slackware, never heard of fc4 before tonight
<15> gb: fc == fedora core
<3> still meaningless too me
<15> gb: umm, you know. the free fork of redhat?
<3> tried a 2.4 kernel on same hardware?
<15> gb: no, and will not
<3> ahh ok so its a take on redcarp
<15> gb: it's the community version, forked when deadrat decided to only release the enterprise version
<3> mord: ok just trying too help you rule out a bios issue
<15> gb: tried shared vid mem sizes of 32, 64 and 128, to no avail
<3> occ***ionally you might have an irq conflict which will do similar things
<3> in which case it will do the saame regardless of kernel version
<15> gb: it does look like an irq conflict. the thing is i just trimmed the stock fc4 kernel .config and compiled a newer kernel and i get this
<15> gb: the fc kernels work just fine
<3> so previous 2.6 kernels function ok?
<15> gb: they just don't have the amd64 optimizations and are bloated
<3> ahuh
<15> gb: 2.6.15 works. running it right now
<15> gb: but it's the fc4 patch hell version
<3> good ok, just gathering info n building a piccy of a possible route too take
<15> gb: i'm going to try compiling 2.6.15.3 with the exact same .config as the fc4 kernel that works has
<3> seems like you might have tried everything I could suggest already though
Return to
#kernel
or
Go to some related
logs:
FATAL: Error inserting smbfs
portage remerges
ubuntu specview
pyparsing function call
xubunto
!!! This is a bug, please report it. (virtual/ghostscript-0)
#math
ubuntu reiserfs or ext3
#egghelp
westell 6100 401 authorization
|
|