I just noticed the 4.9-20030914-PRERELEASE and made a set of installation boot floppies from it to see which of the devices in my new PC the PRERELEASE would support. It wedged hard during the boot autoconfiguration monologue, just after typing these lines on the console: atapci0: <Intel ICH5 ATA100 controller> port 0xf000-0xf00f,0-0x3,0-0x7,0-0x3,0-0x7 irq 0 at device 31.1 on pci0 ata0: at 0x1f0 irq 14 on atapci0 ata1: at 0x170 irq 15 on atapci0 atapci1: <Generic PCI ATA controller> port 0xd000-0xd00f,0xcc00-0xcc03,0xc800-0xc807,0xc400-0xc403,0xc000-0xc007 irq 5 at device 31.2 on pci 0 I had to free it with the system reset button. My motherboard is a Gigabyte 8KNXP which uses the Intel 875P/ICH5R support chipset. Device "atapci1" is the ICH5R serial ATA controller, configured to operate in "native" mode. When the 5.1-RELEASE generic kernel reaches this point, it continues with "ata2" and "ata3" and finishes booting. The 4.9-20030914-PRERELEASE generic kernel wedged solid. I tried to boot it twice. If I reconfigure the serial ATA controller to operate in "legacy" mode (in which it pretends to be a standard parallel ATA controller), the 4.9-20030914-PRERELEASE kernel boots correctly. Another "operating system" that wedges solid when it sees the serial ATA controller configured in native mode is Windows 98SE. I am a little shocked and disappointed that 4.9-20030914-PRERELEASE does no better than Windows since I was planning on running 4.9-RELEASE on my new PC. (5.1-release is a little too flaky for prime time.) It would be real nice if this could be fixed in the final release. I bet there are a bunch of other FreeBSD users out there with new motherboards using the 875P/ICH5R support chipset and some of them also have serial ATA drives and would rather not configure the serial ATA ports to operate in "legacy" mode (which renders one the parallel ATA ports unusable). Dan Strick strick@covad.net
On Sun, 14 Sep 2003, Dan Strick wrote:> I just noticed the 4.9-20030914-PRERELEASE and made a set of > installation boot floppies from it to see which of the devices > in my new PC the PRERELEASE would support.Try rolling back to a build previous to 9/10; theres a commit in ata that is causing long delays during probe in some systems and hangs in others. 9/9 would be a good pick. That or revert sys/dev/ata/ata-lowlevel.c to rev 1.10. -- Doug White | FreeBSD: The Power to Serve dwhite@gumbysoft.com | www.FreeBSD.org
>> > Try rolling back to a build previous to 9/10; theres a commit in ata that > is causing long delays during probe in some systems and hangs in others. > > 9/9 would be a good pick. > > That or revert sys/dev/ata/ata-lowlevel.c to rev 1.10. >>The only place I know to find builds in the stable branch is ftp://releng4.freebsd.org/pub/FreeBSD/snapshots/i386 and the next earlier build there is from 8/10. I am not set up to cvsup the stable branch and I don't know how to revert to a specific rev of the ata driver. (sorry) I can add just a little more diagnostic information. I tried booting the kernel with the -v option and got a couple of more lines out of the autoconfiguration monologue: atapci1: <Generic PCI ATA controller> port 0xd000-0xd00f,0xcc00-0xcc03,0xc800-0xc807,0xc400-0xc403,0xc000-0xc007 irq 5 at device 31.2 on pci 0 ata-: ata2 exists, using next available unit number ata2: iobase=0xc000 altiobase=0xc402 bmaddr=0xd000 At this time my goal is not to find a way to boot and use this particular release but to make sure the right people are aware of the problem and to offer any assistance I can in figuring it out. I don't know how to report 4.9-PRERELEASE problems. I am guessing that the people who are working on the release are reading freebsd-stable. (I gather that one of the purposes of this mailing list is communication between people running the STABLE branch, either via cvsup or snapshots.) Dan Strick strick@covad.net