| |
| |
| |
|
Page: 1 2 3
Comments:
<0> i will be soon probably forced to learn to use that too
<1> hi all
<1> how dangerous is changing network parameters on the fly?
<1> ipfrag_high_thresh and low_thresh
<1> im trying to tune nfs performance
<1> and where can i see what happens in terms of network io?
<2> i'm quite certain there are a number of net monitors
<2> check your distro's packages
<1> hrrrmmm...
<0> hmmm any kernel bigger than about 10 megs won't boot :/
<0> and the device has 32 MB of RAM
<3> i'm trying to compile an old kernel module, it's complaining it can't find video/fbcon.h and video/fbcon-8/16/24/32.h i ***ume these have moved around/changed. is it going to be a huge job to make this module use the new way of doing things?
<3> anyone?
<0> how old?
<3> like 2 yearsish
<3> those are the only errors i get from a make -k
<0> no undefined symbols?
<3> nope
<0> then i guess you only need to remove these two #include lines and it might start working
<3> balrog-kun: ohh sorry, no, i get undefined stuff that i ***ume is because it can't find those headers
<0> but nothing sure can be said before you try
<0> hm ok
<3> everything undefined seems to start with fb
<0> then it might take some work, but i don't think anyone can tell you anythin before trying to actually port it
<3> balrog-kun: okay, but where would i find info about what's changed in the framebuffer stuff since this was written?
<0> there are the Changelogs, but it's easier to look at other similar modules and see what api they use and do the same in your file
<3> okay, so check it out, the pm3fb.c driver (drivers/video/pm3fb.c) does an #include <video/fbcon-cfb8.h> (as well as all the others) but there is no such header anymore (at least according to find). what's up with that?
<0> bazz: try <linux/fb.h> instead
<0> what's the first undefined symbol that it reports?
<3> balrog-kun: field 'gen' has incomplete type, then error: unknown field 'fb_get_fix' specified in initializer
<0> hm, none of these names is in <linux/fb.h>
<3> yeah
<3> actually, the module code already includes fb.h
<3> so none of my undefs are going to be in it
<4> hello. is there any way I can get the device name (eg: eth2) of a NIC if I know it's ioaddr. I tried with ioctl(SIOCGIFMAP) but it's driver dependent
<5> Hi -- I'm having an issue with my usbhid driver not working, and I'm guessing usbhid gets loaded before uhci_hcd, because if I rmmod usbhid and modprobe it again afterwards, it picks up my mouse -- is there an easy way to change the loading order?
<6> what is the stack size limit in linux
<2> 8K i believe
<6> what happens if a program push's more data that on a stack?
<2> oh, in userspace you have unlimited space I think
<2> "unlimited"
<2> i believe -- and it's been a while -- that UNIX systems generally will expand their stack as needed
<6> so the 8kb limit is for modules?
<2> ok hold on
<2> are we asking about user space or kernel space here?
<2> modules are kernel space
<0> well, today i learned that i cannot have 8 MB on stack in userspace :)
<2> programs are user space
<6> Promit: yeah i know
<0> i could have 1 MB on stack though
<2> the kernel's limit on stack is 8K, and if you exceed that you'll probably just crash the system
<6> Promit: does 8kb limit for modules only?
<2> modules and the kernel itself
<6> ok i get it now
<2> i tend to use a heap-allocated stack for any considerable chunk of data, but that's probably somewhat a result of a Windows upbringing
<6> ahh windows ***embly ;)
<6> the stack has to be in continous memory address right?
<2> *shrug* with virtual memory I never know what the hell's going on. Ask me in May :P
<2> i've only ever bothered to do ***embly for small routines embedded in C++
<2> generally for SSE, although I'm moving to intrinsics for those
<2> ah, i had a question
<2> when the kernel says that it's sending something to the system log
<2> where can I read that stuff?
<2> is it talking about dmesg, or something else?
<6> how is pic code different than regular code
<5> Does anyone have a clue as to why usbhid doesn't pick up my mouse at boot, but if I manually reload the module, it works?
<7> SuN: I have the same issue with my pcmcia services, which is why I hacked ascript too reload once the system is up
<8> SuN: so you want to tell modprobe about dependence on usb drivers?
<5> mmc: actually, that was a lame attempt at a hack, but even loading usb-hid from startx doesn't solve it, it only works if I insert it manually, somehow.
<7> SuN: It's quite possibly a dependency issue which should be displayed in a log file somewhere, please dont ask me which logfile :)
<5> I'm not seeing anything out of the order in messages, syslog or dmesg, unfortunately.
<7> SuN: no you probably wont, it's usually something trivial & unnoticable :)
<5> Well whatever it is, it's not logging, at least not to anything in /var/log.
<7> SuN: no it probably is logging it, just as something you'd probably consider as a standard output & not a warning
<5> Like "Everything is working OK!" when it's really not? ;)
<7> SuN: In such cases running your log files through a coloriser helps
<5> There's really not much there:
<5> input: Cypress Sem PS2/USB Browser Combo Mouse as /cl***/input/input4
<5> input: USB HID v1.00 Mouse [Cypress Sem PS2/USB Browser Combo Mouse] on usb-0000:00:1d.1-1
<5> usbcore: registered new driver usbhid
<5> drivers/usb/input/hid-core.c: v2.6:USB HID core driver
<5> That's it.
<7> remember anything too do with the kernel is a working feature,not a bug :)
<5> I'd call a working mouse a feature :)
<7> is that the log entry for when the mouse loads properly?
<5> Nope
<5> See what I mean? :P
<7> What you need too do is find both
<7> or do a dump of the i/o, irq, etc for when it's working & make sure it loads everything needed with the correct parameters on startup
<7> nomean feat probably or it'd probably be working already
<5> But the weird thing is, even when I load it from xinitrc, which is right before X starts and way after all the other stuff is loaded, it won't pick it up.
<5> Whereas when rmmod usbhid; modprobe hid is the first thing I run in an xterm, it does work.
<4> the kernel doesen't have bugs, they are just undocumented features
<7> finger: au contraire some are documented :)
<5> Alright then, so how do I fix the features? ;)
<7> SuN: you dont fix features, they're there too be enjoyed &savoured
<7> SuN: if you can get it working after boot, then write a script too do that which you can disable between kernel versions
<7> it will no doubt be fixed if you are patient
<7> alternatively try an older kernel which might not have the current feature
<9> hey guys
<9> iptables is broken on 2.6.16-rc1-mm4
<9> how do I fix it?
<5> You should either now or not run rc's :)
<9> SuN: this problem is not exclusive of this particular release
<9> if you're interested.
<9> SuN: Do you know of a possible fix?
<5> know* for that matter
<5> No, you haven't even told me the problem.
<9> ok
<9> see this
<9> bash-3.00# iptables -F
<9> iptables: Unknown error 4294967295
<9> bash-3.00# iptables -L
<9> ERROR: 0 not a valid target)
<9> SuN: besides the module only compiling with manual hacking of the .config file
<9> SuN: so, have you seen a fix for those problems on the internet?
<5> Your googling is as good as mine, I guess.
<9> There's a way to reverse the modifications on iptables and put Linus patches, wich supposedly fixes the problem
<9> put I don't know which patch I should get to reverse what was added
<10> Anyone experimented with compiling the kernel across multiple machines? e.g. multi-machine make?
<10> I'm sitting here with 5 computers on my desk and watching a compile. I'd like to reduce it to < 5 minutes.
<11> Is it possible to tweak oomkiller via /proc?
<12> i have a 550 Mhz Processor can I make it work like a 233 Mhz ?
<12> anybody ?
<13> sure
<13> install linux on it
<13> :))
<12> anybody ?
<14> Anyone know why my command line is booting into 640x480? I thought I explicitly set it to 1280x1024 by a kernel parameter as follows:
<14> kernel=/kernel-2.6.15-gentoo-bs-r1 root=/dev/sda4 video=radeonfb:1280x1024-32@72,mtrr:2,ywrap splash=silent,fadein,theme:fingerprint-bs quiet CONSOLE=/dev/tty1
<14> initrd (hd0,0)/fbsplash-fingerprint-bs-1280x1024
<14> (That's from my grub.conf)
<15> can i use the .config of kernel 2.6.11.12 to compile a recent 2.6.16-* kernel?
<16> there is a ring in the double-linked table of my module, the kernel version:2.6.13, hot can i debug it ?
<7> blah
<17> gb, ") due
<17> dude
<17> Jack-Bauer, :) man
<17> TheBonsai, lo lo man
<7> _RADIOhead: gday
<17> gb, how things dude
<7> sore still
<7> made the mistake of ***embling three tyres a few days back
<17> i was so busy this days
<7> now I pay for it the next few weeks
<17> car tyres
<7> no 16" track tyres
<7> oops truck rather
<17> :)
<16> there is a ring in the double-linked table of my module, the kernel version:2.6.13, hot can i debug it ?
Return to
#kernel
or
Go to some related
logs:
#gentoo
#perl
aterm copy paste
kde kicker dual head
gentoo 3M microtouch howto
#linux
yum_updatesd memory
#fedora
#linux
#linux
|
|