In message <200610302227.52575.nb_root_at_videotron.ca>, Nicolas Blais writes: >I saw that pci.c was updated today (to version 1.318) and I decided to give it >a try. As expected, it did not work, so I kept my /usr/src in sync, but reverted pci.c to 1.292.2.9 and my system booted fine! > >By diff'ing with 1.315 >(http://www.freebsd.org/cgi/cvsweb.cgi/src/sys/dev/pci/pci.c.diff?f=H&r1=text&tr1=1.292.2.9&r2=text&tr2=1.315) >we can find there's obviously something in the vpd code that my system >doesn't like, or that the vpd code doesn't like something about my system ;). FYI, jmg_at_ sent me the following patch that fixed a similar issue. The thing in common between systems seeing this hang appears to be that they have a Brooktree card installed. Seemingly, some Brooktree cards have a PCI capability claiming VPD support, but the VPD data is missing, which confuses the parser. One of the pci_read_vpd_reg() calls then gets stuck when it tries to read from an invalid address. Ian Index: pci.c =================================================================== RCS file: /dump/FreeBSD-CVS/src/sys/dev/pci/pci.c,v retrieving revision 1.316 diff -u -r1.316 pci.c --- pci.c 20 Oct 2006 21:28:11 -0000 1.316 +++ pci.c 31 Oct 2006 18:58:02 -0000 _at__at_ -630,7 +630,9 _at__at_ state = 5; break; default: /* XXX - unimplemented */ - state = 4; + end = 1; + cksumvalid = 0; + continue; break; } break;Received on Tue Oct 31 2006 - 18:08:03 UTC
This archive was generated by hypermail 2.4.0 : Wed May 19 2021 - 11:39:02 UTC