-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 I tried the kcs_detect() hack, and the driver hung at insertion time.
I also tried forcing type kcs, bt,smic combinating with different I/O addresses(default 0xca2, and 0x4e from DMI) still no success so at last my bet is that there is no BMC in my machine. There is no strong evidence, but the confusing specifications, and all the stuff we tried makes me think this. many thanks for your help and your time, Gabor Corey Minyard wrote: > Unfortunately, this is a difficult issue and very hard to resolve. > From what I can tell, it is one of the following problems: > > 1. The machine does not actually have a BMC. A lot of these machines > have optional add-in cards, and maybe there is an error in the > BIOS that always puts the DMI information in. > 2. The machine has a BMC, but the DMI information is wrong. Since > the slave address is bogus, this is probably the most likely. The > default address for IPMI is 0xca2 in the I/O space. Maybe that > would work? Maybe it's not KCS? But it's only a guess. > 3. The machine has a BMC at the given address, but there is something > wrong with the BMC or it's buggy and doesn't implement the state > machine properly. You should be getting more information out of > the kernel if this is the case, though. > 4. There is a bug in the IPMI driver. Since the driver works with a > very large number of machines, this is unlikely. > > Looking at the error you are getting, that means the driver read a > 0xff from the address given for the IPMI control port. The control > port cannot ever be 0xff, so that generally means there is nothing > there. You might having the kcs_detect() function in > drivers/char/ipmi/ipmi_kcs_sm.c always return 0, but I doubt that > will make a difference. > > -Corey > > Ivánszky Gábor wrote: >> -----BEGIN PGP SIGNED MESSAGE----- >> Hash: SHA1 >> >> Thanks for the hint. >> >> Since I couldn't find anything about slave address in my vendor's >> docs, >> I tried with 0x20 address, but it seems that ipmi_si doesn't care >> about >> slave_addrs as a loading time parameter :(. I doublechecked openipmi >> documentation(IPMI.txt), but unfortunately I was not able to find the >> root of this issue. >> >> # modprobe ipmi_si slave_addrs=0x20 >> FATAL: Error inserting ipmi_si >> (/lib/modules/2.6.20/kernel/drivers/char/ipmi/ipmi_si.ko): No such >> device >> >> IPMI System Interface driver. >> ipmi_si: Trying SMBIOS-specified kcs state machine at i/o address >> 0x4e, >> slave address 0x0, irq 0 >> ipmi_si: Interface detection failed >> ipmi_si: Unable to find any System Interface(s) >> > -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.6 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iD8DBQFF7ZgfDiqVI6Xm21MRAtY2AJ9HbTa2SAsKo1qgH+xVpvv37KuErACffe83 Jtf/3LVgQi8aQoyQwBXsoiw= =G33N -----END PGP SIGNATURE----- ------------------------------------------------------------------------- Take Surveys. Earn Cash. Influence the Future of IT Join SourceForge.net's Techsay panel and you'll get the chance to share your opinions on IT & business topics through brief surveys-and earn cash http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV _______________________________________________ Openipmi-developer mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/openipmi-developer
