It was OK on gentoo kernel 2.6.13. Yesterady I switched to 2.6.15 and emerge ltmodem 8.31_alpha10-r1 quits complaining about an undeclared PCI_DEVICE_ID_ATT_L56XMF. Indeed, on 2.6.13 there was a #define of it in the file: /usr/src/linux/include/linux/pci_ids.h but on 2.6.15 it disappeared! Workaround; just add the line: #define PCI_DEVICE_ID_ATT_L56XMF 0x0440 in file pci_ids.h Emerge works and the modem works, too! Why so many defines have been removed from pci_ids.h on 2.6.15 (including this one)?
The ebuild needs a bump to use the ltmodem-2.6-alk-8.tar.bz2 tarball instead of ltmodem-2.6-alk-7.tar.bz2 > Why so many defines have been removed from pci_ids.h on 2.6.15 (including this one)? Read all about it here -> http://www.linux-usenet.com/threadedpost733852.html#post733852 No real valid reason as far as I can see, just that the deciding dev(s) don't happen to currently be using the hardware for which the IDs were removed. Perhaps the thought is that these IDs should be defined from within the source of the programs that intend to use them. At any rate, this seems to be the result as the applications & drivers that continue to use the removed PCI_IDs are swallowing it & adapting ok, as can be seen with the latest tarball of ltmodem. The same issue also breaks the compilation of svgalib's kernel module, it's latest version (1.9.24 not yet in portage) has also addressed the problem, but that's another bug.
*** Bug 118149 has been marked as a duplicate of this bug. ***
(In reply to comment #1) ... > Read all about it here -> > http://www.linux-usenet.com/threadedpost733852.html#post733852 > Good catch, tx! El Fabre
fixed in ltmodem-8.31_alpha10-r2. it will be marked as stable after the probation period.
I've marked as stable, for help making kernel 2.6.15 stable. release team needs it for 2006.0.
