答复: FILO 0.4 [PMX:#]

[Steven James]

Greetings,

You can use bootselect in utils/baremetal to choose between multiple
payloads.

G'day,
sjames


-------------------------steven james, director of research, linux labs
... ........ ..... ....                    230 peachtree st nw ste 2701
the original linux labs                             atlanta.ga.us 30303
      -since 1995                              http://www.linuxlabs.com
                                   office 404.577.7747 fax 404.577.7743
-----------------------------------------------------------------------


On Tue, 6 Apr 2004, YhLu wrote:

> So you must use filo as normal payload and Etherboot as fallback payload.
>
> Then if I want to boot from network in normal mode, how to do?
>
> Why just merge filo into Etherboot?
>
> So we can use tg3--filo.zelf instead of tg3--ide_disk.zelf.
>
> Regards
>
> YH
>
> -----邮件原件-----
> 发件人: ron minnich [mailto:rminnich at lanl.gov]
> 发送时间: 2004年4月6日 10:28
> 收件人: Eric W. Biederman
> 抄送: Takeshi Sone; linuxbios at clustermatic.org
> 主题: Re: FILO 0.4 [PMX:#]
>
> On 5 Apr 2004, Eric W. Biederman wrote:
>
> > I am looking at using FILO as a disked based bootloader.  So I might
> > just write that code and send you a patch.
>
> We want FILO in FLASH, not on disk. We are using FLASH-based FILO to
> replace FLASH-based Etherboot on Pink and will be using FILO to replace
> Etherboot on Lightning and Blue Steel to load images from IDE-FLASH. We
> are very happy with its capability in that mode. It's just a ton easier
> for a sysadmin to mount the IDE-FLASH as ext2 and change kernel file names
> instead of dd-ing raw kernel images onto an IDE-FLASH. FILO lets us treat
> IDE-FLASH as a file system, and Etherboot had us treating it as a raw
> disk. We prefer FILO at this point, and I think we will not be the only
> ones. Also we have found the build procedure for FILO to be less prone to
> error and problems than for Etherboot.
>
> That's not to say Etherboot does not have *many* applications: it is
> terrific for lots of things, just not for our clusters right now.
>
> > Right now I think a bootloader of this type with a config file is
> > not terribly useful.
>
> It is all context. I think we need to turn our thinking a bit. The BIOS
> you use is going to become more application-specific. What you have in an
> embedded "black box" will not need to be identical to what you have on
> your desk or your cluster. I think this is all to the good.
>
> ron
>
> _______________________________________________
> Linuxbios mailing list
> Linuxbios at clustermatic.org
> http://www.clustermatic.org/mailman/listinfo/linuxbios
> _______________________________________________
> Linuxbios mailing list
> Linuxbios at clustermatic.org
> http://www.clustermatic.org/mailman/listinfo/linuxbios
>