I have just pushed out SYSLINUX 3.10-pre9. Please test it out, especially everyone who has been working with me on various systems trying to get them to work, or have had recent breakage. -hpa
>I have just pushed out SYSLINUX 3.10-pre9. Please test it out, >especially everyone who has been working with me on various systems >trying to get them to work, or have had recent breakage. > > -hpaIt seems to work now, but it takes about 10 seconds before it displays "SYSLINUX CBIOS etc.".
> I have just pushed out SYSLINUX 3.10-pre9. Please test it out, > especially everyone who has been working with me on various systems > trying to get them to work, or have had recent breakage.3.10pre9 fixes the issues with booting via USB. Thanks! Alex
>robotti at godmail.com wrote: >>>I have just pushed out SYSLINUX 3.10-pre9. Please test it out, >>>especially everyone who has been working with me on various systems >>>trying to get them to work, or have had recent breakage. >>> >>> -hpa >> >> >> It seems to work now, but it takes about 10 seconds before it >> displays "SYSLINUX CBIOS etc.". >> > >What that means is that it keeps trying and trying to access your broken >EBIOS, and eventually gives up and falls back to CBIOS (with the >1024-cylinder limit.) This is a good thing, because that's what the >code is supposed to do. There really doesn't seem to be any better >option on your particular system, and at least it works now. > >(syslinux -s would make it faster, and I'm considering an option to >force CBIOS, but as people already know, my general attitude is that it >should work automatically on as many systems as possible, even if it's >slow.)The EBIOS problem doesn't seem to exist when using syslinux to boot from my SATA hard drive (past 1024 cylinders), but just the USB drive. The USB drive is only 128MB so it doesn't need EBIOS, but it stumbles over it and falls back to CBIOS.
Hi!> -----Original Message----- > From: H. Peter Anvin [mailto:hpa at zytor.com] > Sent: Tuesday, August 16, 2005 8:47 PM > To: SYSLINUX at zytor.com > Subject: [syslinux] SYSLINUX 3.10-pre9 > > > I have just pushed out SYSLINUX 3.10-pre9. Please test it out, > especially everyone who has been working with me on various systems > trying to get them to work, or have had recent breakage. > > -hpaWorks like a charm. It boots with the regular speed, and uses EBIOS. So, the recent breakege was solved for me, thanks! Luis Correia
H. Peter Anvin wrote:> I have just pushed out SYSLINUX 3.10-pre9. Please test it out, > especially everyone who has been working with me on various systems > trying to get them to work, or have had recent breakage.The DISPLAY bug in pxelinux still exists. The one I reported in June where a <CAN>filename<newline> gets a bunch of jibberish appended to it so the tftp request for the graphic is corrupted. Both with the current pre-release with max_cmd_len equ 511 and the prior 1023. TFTP Read Request, File: /pxe/pxelinux.0, Transfer type: octet TFTP Read Request, File: /pxe/pxelinux.cfg/01-00-e0-81-ff-ff-ff, Transfer type: octet TFTP Read Request, File: /pxe/menu/benhead.msg, Transfer type: octet TFTP Read Request, File: /pxe/menu/benhead.rle??????'`?, Transfer type: octet However, a test revert it back to the old max_cmd_len equ 255 and the symptom vanished. TFTP Read Request, File: /pxe/pxelinux.0, Transfer type: octet TFTP Read Request, File: /pxe/pxelinux.cfg/01-00-e0-81-ff-ff-ff, Transfer type: octet TFTP Read Request, File: /pxe/menu/benhead.msg, Transfer type: octet TFTP Read Request, File: /pxe/menu/benhead.rle, Transfer type: octet The test file benhead.msg is exactly 18 bytes: <CAN>menu/benhead.rle<newline> I hope that is a clear explanation. The above excerpts are from a network trace of only the client request side of the pxe conversation. This is on an Tyan/Opteron system which uses AMIBIOS. ../C
Curtis Doty wrote:> H. Peter Anvin wrote: > >> I have just pushed out SYSLINUX 3.10-pre9. Please test it out, >> especially everyone who has been working with me on various systems >> trying to get them to work, or have had recent breakage. > > > > The DISPLAY bug in pxelinux still exists. The one I reported in June > where a <CAN>filename<newline> gets a bunch of jibberish appended to it > so the tftp request for the graphic is corrupted. Both with the current > pre-release with max_cmd_len equ 511 and the prior 1023. >What happens if you drop it to 255? -hpa
H. Peter Anvin wrote:> Curtis Doty wrote: > >> H. Peter Anvin wrote: >> >>> I have just pushed out SYSLINUX 3.10-pre9. Please test it out, >>> especially everyone who has been working with me on various systems >>> trying to get them to work, or have had recent breakage. >> >> >> The DISPLAY bug in pxelinux still exists. The one I reported in June >> where a <CAN>filename<newline> gets a bunch of jibberish appended to >> it so the tftp request for the graphic is corrupted. Both with the >> current pre-release with max_cmd_len equ 511 and the prior 1023. >> > > What happens if you drop it to 255?The bug symptom vanishes and my graphic is displayed as expected. ../C
Curtis Doty wrote:> > However, a test revert it back to the old max_cmd_len equ 255 and the > symptom vanished. >Just re-read your message, and my previous question is obviously in there. Could you try changing "9000h" in the file "layout.inc" to "0A000h"? -hpa