[syslinux] EFI booting over network - can't then load anything

Spike White spikewhitetx at gmail.com
Wed Jun 11 17:20:52 PDT 2014


On 06/09/2014 12:09 AM, Andrew Rae wrote:
>
> Many thanks!
>
> I think we thought that it was more functional in EFI mode than it is -
apologies !
>
> I/We tried to do a RedHat 7 RC UEFI Network Boot - and that seemed to
work OK & as expected....
> If we could only get Centos 6.5 and some version of WindowsPE to boot! -
we could switch to using EFI booting on the new pc's/laptops to image them
up!
>
> Thanks - I'll keep my eye on the mailing list & try it on a few of the
newer releases as they come out
> Andrew

Andrew,

Yesterday I tested on OL 6.5 (Redhat-compatible kernel).  So like Centos
6.5,
Oracle Linux is merely a re-spin of RHEL6.5.  Tried on physical hardware.
 As
you say, the computer crashes and reboots.  This was syslinux 6.03-pre6.

But does crashes in an interesting place.  *After* it loads kernel and
seemingly the full initramfs.

Then I tested against the bootx64.efi NBP supplied with grub2. CentOS 6.5
will
EFI PXE boot with that NBP. bootx64.efi bootloader Gives an interesting
message:

   trying to allocate 1000 pages for vmlinuz

Also yesterday, Redhat released RHEL7 (GA version).  Today I tested that.

RHEL7 itself bundles kernel version 3.10, so I assume the boot media's
vmllinuz
is of comparable vintage.

As H. Peter predicted, syslinux.efi NBP boots the RHEL7 install media.
(It's a
new enough kernel).

So -- if we ignore this problem long enough (syslinux.efi can't EFI PXE boot
ancient kernels) -- the problem will go away.

In the meantime, the grub2 bootx64.efi bootloader will EFI PXE boot these
archaic kernels.

I'm curious how grub2's bootx64.efi bootstraps these older kernels.

Spike


More information about the Syslinux mailing list