On Fri, 2003-02-21 at 11:25, James Sparenberg wrote: > On Fri, 2003-02-21 at 06:58, Praedor Atrebates wrote: > > -----BEGIN PGP SIGNED MESSAGE----- > > Hash: SHA1 > > > > On Thursday 20 February 2003 08:07 pm, Vincent Danen wrote: > > > On Thu Feb 20, 2003 at 10:58:03AM -0500, Praedor Atrebates wrote: > > > > > > [...] > > > > > > > Have people actually managed to BUILD 2.4.19-24mdk and use it? Or are > > > > you simply using the binary version. What did MandrakeSoft do to get it > > > > to build and actually work, if this is the case? To what god(s) did they > > > > offer sacrifice to so that it would work on any hardware at all? > > > > > > I built the 2.4.19-24mdk packages. They took a while, but they built (from > > > rpm). I rebuilt them once from source as a test to make sure they built, > > > but I don't recall if I booted from that compiled version or not. > > > > > > I have installed the binary package in vmware, on an "enterprise" machine, > > > on an "smp" machine, and on an "up" machine. In all cases, everything > > > worked just peachy. > > > > > > You seem to have a lot of problems with Linux that I just never come > > > across. I see quite often you having problems with stuff on the list. > > > Perhaps your luck just isn't very good. =) > > > > > > Seriously, a number of people tested this kernel before we put it out. We > > > can't test every piece of hardware out there, but anything major we can > > > usually spot quite quickly. I don't know why it doesn't work for you. > > > > > > The hardware is new but still old enough to have proven itself over time. I > > have tried completely removing the source package, kernel modules directory, > > source, etc, and then reinstalling the kernel-source rpm, and redoing the > > build. > > > > I have always ultimately rebuilt my own kernels, first from tarballs and more > > recently, from source rpms. It has only really been with 9.0 and the 2.4.19+ > > kernels that I have had problems. I had to dump my XFS filesystem to get a > > custom-built 2.4.19 9.0 kernel to work with my system (known issue - the > > kernel makefile indicates that for xfs support - and thus the whole kernel as > > a result, should be built with egcc?! Who uses that anymore? I thought that > > it and gcc had been merged over a year ago). > > > > In any case, the harddrive is new, the system memory is sound, the CPU is > > sound, the video card is sound (NVidia Ti4200). > > > > I was hoping someone would have run into this sort of kernel-uselessness > > situation and also found a way out of it. I am reluctant to simply install a > > binary kernel. They don't tend to work very well with my wireless devices. > > > > praedor > > > Praedor, > > Going back to your earlier fun with i82365.. I'm now having the same > trouble with all kernels since 2.4.19 Seems that something in i82365.c > doesn't compile right. If I kill pcmcia and then try to load the > modules directly > > # modprobe i82365 > /lib/modules/2.4.21pre4-9mdk/kernel/drivers/pcmcia/i82365.o.gz: > init_module: No such device > Hint: insmod errors can be caused by incorrect module parameters, > including invalid IO or IRQ parameters. > You may find more information in syslog or the output from dmesg > modprobe: insmod > /lib/modules/2.4.21pre4-9mdk/kernel/drivers/pcmcia/i82365.o.gz failed > modprobe: insmod i82365 failed > > if I try it directly > > Using /lib/modules/2.4.21pre4-9mdk/kernel/drivers/pcmcia/i82365.o.gz > /lib/modules/2.4.21pre4-9mdk/kernel/drivers/pcmcia/i82365.o.gz: > unresolved syl dead_socket_Rcf97f3bd > /lib/modules/2.4.21pre4-9mdk/kernel/drivers/pcmcia/i82365.o.gz: > unresolved syl pcmcia_get_card_services_info_R3c39c871 > /lib/modules/2.4.21pre4-9mdk/kernel/drivers/pcmcia/i82365.o.gz: > unresolved syl unregister_ss_entry_R0cdfb5e5 > /lib/modules/2.4.21pre4-9mdk/kernel/drivers/pcmcia/i82365.o.gz: > unresolved syl register_ss_entry_R705959cb > [EMAIL PROTECTED] linux-2.4.20]# /etc/init.d/pcmcia restart > Shutting down PCMCIA services: done. > > > The if I do /etc/init.d/pcmcia restart > > Shutting down PCMCIA services: done. > Starting PCMCIA services: using yenta_socket instead of i82365 > cardmgr[5411]: watching 2 sockets > done. > > > > My chipset needs to use i82365 not yenta_socket to work right just like > yours. Even removing the dang module it STILL loads it somehow > (yenta_socket) > > James
All, I remembered and checked one thing I've got a single kernel that i83265 does work correctly on. And it was compiled via 2.96 rather than 3.x....(older kernel so it could be the kernel) going to try this again and see what happens. James
Want to buy your Pack or Services from MandrakeSoft? Go to http://www.mandrakestore.com
