2013/8/5 H. Peter Anvin <hpa at zytor.com>> On 08/05/2013 05:37 AM, Jonas Keidel wrote: > > I have problems with the current syslinux (6.02-pre16) booting CentOS 6.4 > > 64bit via LOCALBOOT 0 with the Intel DX79TO. > > Did it work in previous versions? >As i have written, in every previous syslinux version it didn't work.> > Also in every previous version there is problem exiting the pxe-stack. It > > boots again from bios. > > What do you mean with "exiting the PXE stack" in this context? >So i mean looking for the next boot device in the bios boot order. The screen output says: Boot from local hdd ... Exiting PXE. And then it reboots.> > -hpa > >
H. Peter Anvin
2013-Aug-05 13:09 UTC
[syslinux] Intel DX79TO localboot problem with CentOS
On 08/05/2013 06:04 AM, Jonas Keidel wrote:> 2013/8/5 H. Peter Anvin <hpa at zytor.com> > >> On 08/05/2013 05:37 AM, Jonas Keidel wrote: >>> I have problems with the current syslinux (6.02-pre16) booting CentOS 6.4 >>> 64bit via LOCALBOOT 0 with the Intel DX79TO. >> >> Did it work in previous versions? >> > > As i have written, in every previous syslinux version it didn't work. > > >>> Also in every previous version there is problem exiting the pxe-stack. It >>> boots again from bios. >> >> What do you mean with "exiting the PXE stack" in this context? >> > > So i mean looking for the next boot device in the bios boot order. The > screen output says: > > Boot from local hdd ... > Exiting PXE. > > And then it reboots. >I don't know how much we can do about that. "localboot 0" and "localboot -1" are effectively different ways to request the same service from the BIOS, but it seems one is broken. Other than having a huge quirks table, of course, but that is problematic in itself. chain.c32 may be a better option all around... instead of exiting to the BIOS to let it pick up the next device, it loads the boot data from disk directly. -hpa
2013/8/5 H. Peter Anvin <hpa at zytor.com>> On 08/05/2013 06:04 AM, Jonas Keidel wrote: > > 2013/8/5 H. Peter Anvin <hpa at zytor.com> > > > >> On 08/05/2013 05:37 AM, Jonas Keidel wrote: > >>> I have problems with the current syslinux (6.02-pre16) booting CentOS > 6.4 > >>> 64bit via LOCALBOOT 0 with the Intel DX79TO. > >> > >> Did it work in previous versions? > >> > > > > As i have written, in every previous syslinux version it didn't work. > > > > > >>> Also in every previous version there is problem exiting the pxe-stack. > It > >>> boots again from bios. > >> > >> What do you mean with "exiting the PXE stack" in this context? > >> > > > > So i mean looking for the next boot device in the bios boot order. The > > screen output says: > > > > Boot from local hdd ... > > Exiting PXE. > > > > And then it reboots. > > > > I don't know how much we can do about that. "localboot 0" and > "localboot -1" are effectively different ways to request the same > service from the BIOS, but it seems one is broken. Other than having a > huge quirks table, of course, but that is problematic in itself. > > chain.c32 may be a better option all around... instead of exiting to the > BIOS to let it pick up the next device, it loads the boot data from disk > directly. > >So i've also tried to use chain.c32, but there are also some problems. For example FreeBSD doesn't boot with chain.c32, so we've decided to use localboot, because it works in 90%, except with that board. Maybe you have an idea how to fix that? Or what do you think is broken?> -hpa >
On Mon, Aug 5, 2013 at 9:04 AM, Jonas Keidel <jonas at jonas-keidel.de> wrote:> 2013/8/5 H. Peter Anvin <hpa at zytor.com> > >> On 08/05/2013 05:37 AM, Jonas Keidel wrote: >> > I have problems with the current syslinux (6.02-pre16) booting CentOS 6.4 >> > 64bit via LOCALBOOT 0 with the Intel DX79TO. >> >> Did it work in previous versions? > > As i have written, in every previous syslinux version it didn't work.Could you be specific on what versions you've tried? 3.86, 4.07 and 4.02 would be the first suggest.>> > Also in every previous version there is problem exiting the pxe-stack. It >> > boots again from bios. >> >> What do you mean with "exiting the PXE stack" in this context? > > So i mean looking for the next boot device in the bios boot order. The > screen output says: > > Boot from local hdd ... > Exiting PXE. > > And then it reboots.I think I had that issue with a first-generation uEFI-based system (Dell Latitude E6410) that did this on me. On Tue, Aug 6, 2013 at 2:17 AM, Jonas Keidel <jonas at jonas-keidel.de> wrote:> so further information towards the chain.c32 problem: > we can't boot from a software raid with this, because if the first hdd is > swapped (and empty), the system normally can boot from the secondary hdd. > But with chain.c32 that isn't possible. > So we need localboot working.How would your BIOS handle this if PXE were not involved? Are you using the driver-based RAID feature of your onboard SATA controller? I believe for pre-OS environments, the drives will act like one. For chain.c32, that'd be hd0. -- -Gene
2013/8/6 Gene Cumm <gene.cumm at gmail.com>> On Mon, Aug 5, 2013 at 9:04 AM, Jonas Keidel <jonas at jonas-keidel.de> > wrote: > > 2013/8/5 H. Peter Anvin <hpa at zytor.com> > > > >> On 08/05/2013 05:37 AM, Jonas Keidel wrote: > >> > I have problems with the current syslinux (6.02-pre16) booting CentOS > 6.4 > >> > 64bit via LOCALBOOT 0 with the Intel DX79TO. > >> > >> Did it work in previous versions? > > > > As i have written, in every previous syslinux version it didn't work. > > Could you be specific on what versions you've tried? 3.86, 4.07 and > 4.02 would be the first suggest. >I've tried 3.82 and 4.07.> > >> > Also in every previous version there is problem exiting the > pxe-stack. It > >> > boots again from bios. > >> > >> What do you mean with "exiting the PXE stack" in this context? > > > > So i mean looking for the next boot device in the bios boot order. The > > screen output says: > > > > Boot from local hdd ... > > Exiting PXE. > > > > And then it reboots. > > I think I had that issue with a first-generation uEFI-based system > (Dell Latitude E6410) that did this on me. >And did You find any solution towards this?> > On Tue, Aug 6, 2013 at 2:17 AM, Jonas Keidel <jonas at jonas-keidel.de> > wrote: > > so further information towards the chain.c32 problem: > > we can't boot from a software raid with this, because if the first hdd is > > swapped (and empty), the system normally can boot from the secondary hdd. > > But with chain.c32 that isn't possible. > > So we need localboot working. > > How would your BIOS handle this if PXE were not involved? Are you > using the driver-based RAID feature of your onboard SATA controller? > I believe for pre-OS environments, the drives will act like one. For > chain.c32, that'd be hd0. > >It is a software raid, so the bios don't know about that. And both disks are usually bootable (except after replacement). So, localboot selects the next boot device after network card, if this is not bootable it will got to the next. If the first hdd is empty (after replacement), it will automatically boot from the second hdd. chain.c32 doesn't do that.> -- > -Gene > _______________________________________________ > Syslinux mailing list > Submissions to Syslinux at zytor.com > Unsubscribe or set options at: > http://www.zytor.com/mailman/listinfo/syslinux > Please do not send private replies to mailing list traffic. > >