| |
| |
| |
|
Comments:
<0> hi
<1> Hm, is this a good place to ask about tmpfs?
<0> no
<2> maybe
<0> sshot
<1> key2: ah, I got an answer in ##linux, anyway.
<1> So no worries.
<3> (Pardon me if I don't express this correctly.) Does the VFS layer do read-ahead with the more generic block drivers? I.e. How concerned do I need to be about doing (any of) my own read-ahead if I'm implementing a block driver?
<4> hiiii....
<4> simple question : may i get strange messages about soft lockup on SMP mashine just for uncleanly removed smbfs module? :/
<4> bcause now i have dead system ~480km away, and... it's just hard to go such a way just to "reboot" it ;-)
<4> smth like ... <1>Unable to handle kernel paging request at virtual address f8c9a408
<4> stack trace
<4> later just
<4> <3>BUG: soft lockup detected on CPU#0!
<4> again stack trace
<4> i can log in over ssh, but no one command works, reboot -f hung it all...
<4> so, a question - it may be hardware flaw, kernel bug (2.6.15.2) or simply - just need to reboot after rmmod -f smbfs
<4> ?
<5> ABLomas: simple, work out what causes it, It could simply be a piece of ****ty code
<4> maybe, maybe
<5> & unfortunately smb in itself is a crappy protocol too begin with
<4> but now i have only mostly dead system far far away, so i'm just asking for ideas - someone saw bugreport/etc about smth similar?
<5> so coders have too do some pretty ****ty programming too keep it functional
<4> i mean, removing smbfs module may hangup system, cause strange and incorrect kernel messages, cause system nor to reboot, but not to do something other task either?
<4> ;-)
<5> hence all the kernel patches & modules dedicated too smb
<4> huh
<4> i just tried to change it to CIFS... unsuccessfully :/
<5> ABLomas: does the same thing happen when you switch these other smb patches on?
<4> and, it seems, pretty simple task, doing it remote was failure
<4> hm? Which other patches?
<5> I know back in the time when I had too use smb, it was a nightmare, now I have no need of it at all thankfully :)
<5> nfs rocks :)
<4> i did not used any patches on this kernel. Just plain vanilla kernel, compiled by myself, mostly "all in", just some parts as modules
<4> yeah... but i can't select what to use
<4> main data-exchange dir is on w2k3 domain controler, so...
<4> again, back to question
<6> does nvidia provide any specs to developers for their mobo chipsets like intel does?
<4> how do u mean, it maybe caused _only_ by module removal?
<5> ABLomas: the most productive thing I can suggest is too start reading kernel mailing lists & perhaps using your favourite search engine in relation too the errors you get & the things you've tried
<4> again, i did it, as already said, it do not is directly related to smbfs
<4> but nvrmnd then
<5> magyar: they have no need too, according too them they pay sumone too write a module so you need too know nothing about the specs, etc. My choice is not too use nvidia.
<5> even if I was given the hardware with money too use it :)
<6> gb, so i need to stay away from any nforce based mother boards
<5> magyar: your choice, many people persist & are happy, I just dont believe in rewarding such behaviour by hardware manufacturers
<5> I prefer open everything, but thats my bigotted attitude :)
<6> i belive Intel does provide hardware specs , or am i wrong?
<5> I'm a firm believer of not only free speech, but also free beer & free pizza :)
<5> magyar: I believe they do, other hardware I avoid because it's **** are anything with rtl8139 chipsets
<5> & many mainboards include this onboard at no charge too you apparrently
<6> realteck chipset for sound or net?
<5> ethernet
<5> also checkout any onboard sound, some chipsets are a royal pita too get functional & cause performance drops also through downtime & other issues
<4> hm ;-)
<5> most items onboard are the cheapest possible available
<4> one more short question : if after msg's in syslog about "Unable to handle kernel paging request at virtual address"... i found this one :
<5> ie!==the best chipset too use for that particular function
<4> EIP is at sync_filesystems+0x3a/0x12a
<6> gb, i off the onboard sound, i stick to HW based cards from SB
<4> may i think, that kernel is hung trying to sync filesystems?
<5> magyar: same except I no longer use desktops & choice in notebooks is reduced even further :)
<4> and if so, does it have some sort of timeout? I got kernel.panic = 5 in sysctl, but it is not such a case..
<2> i have some onboard ketchup on my motheboard after i was eating sandwiches near it
<5> caffiene time
<5> balrog-kun: how bout some chocolate too speed up your cpu?
<5> oh its not april 1st yet
<2> nah, ketchup gets dry and sticky and looks much worse than chocolate
<7> hmm
<8> hmm
<8> not good
<5> I should script that into irssi
<9> i am running 2.6.15.5 on a latitude d610 and when i try to resume from suspend and hibernate the touchpad mouse disappears, and when i say disappears, i mean it is not even in /proc/bus/input/devices anymore. ... This is an alps glidepoint and it is usually detected as a ps2 mouse. any ideas?
<10> Can I run make clean and still build a module? (example: fuse)?
<11> I have a bus address on x86 architecture. I need to find out which virtual address it is ioremapped to. I am using 2.6.13. Which function / macro can let me do that ?
<5> manas: difficult too see, always in motion is it
<11> gb: Sorry, I do not understand.
<5> uhuh
<5> you never watched empire strikes back?
<12> is it possible to write a kernel module that writes over executable kernel memory..
<12> like would overwrite a builtin function to the kernel
<5> geckosenator: sure, anything is possible, not altogether wisem but possible :)
<5> araujo is a slaqa
<13> hy; I try to copy a file to NTFS; ls -l reports the same size: 1024 bytes, and cp complains: ntfs_setattr : changes in i_size not supported ; why would i_size change when both files are 1024 bytes ?
<2> try dd instead of cp
<2> with conv=notrunc (or whatever)
<13> ah ... since 2h I grab, at last an idea worth to try
<13> kernel stopped complaining, but command is not efficient
<14> hi
<14> doublehp: u have 2.6.15 kernel?
<15> 2.6.15.5.w00t.^_^.rofl
<14> how good the 2.6.15 ntfs r/w works?
<15> dunno dont have any ntfs formatted partitions
<13> paistis: I think so :)
<14> okey :)
<5> ntfs??Not that friggin ****e?
<15> heh
<14> :D
<15> i h8 joo ntfs
<16> i have a general question about designing routing protocols
<16> the one im implementing needs quite a bit of logic in the kernel, the only stuff im doing from userlevel at the moment is sending udp-packets (which i then p*** to the kernel via netlink, because i need to make decisions based on this information there) and creating ipip or gre tunnels
<16> now i'm pondering whether to move those two bits from userspace to kernel or just leave everything as is
<17> i need to know as to how i can set debug enable/disable in driver built-in kernel, so i can turn on/off debug globally/individually - ldd3 debug example is for building external modules
<18> hiall
<18> hi all
<2> hi metakiller
<18> is there french here?
<18> how i can disable support for ia64 processors
<18> ?
<19> I am using fedora.. my kernel does not allow me to read /proc/kcore.. though I have read permissions on the file.. how do I diable that ..
<20> hi all! would this be an appropriate place to ask about a lower-level networking problem with level 2, 3, and 4 packet reconstruction?
<21> is there a "right" way to find out exactly when a given line in a given source file [dis]appeared?
<21> i'm trying to apply a 2.6.16-rc5 to a 2.6.9 redhat kernel
<21> er, 2.6.16-rc5 *patch*
<21> and trying to reconcile a couple minor differences
<22> hi
<22> does anyone know about a bug whitin 2.6.12 and reiserfs ?
<22> permission denied while trying to access some files
<22> nobody at all ?
<22> ok
Return to
#kernel
or
Go to some related
logs:
#python
inspiron 9300 xorgconfig
one each of u types of candy
#kde
ubuntu xforward
litone tool
gxiso no module named bz2
uninstall fedora
#php
libwx_gtk2ud ubuntu
|
|