bazz

Arch codes iPXE

May 25th, 2015
390
0
Never
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 2.83 KB | None | 0 0
  1. 08:28 -!- Irssi: Join to #ipxe was synced in 42 secs
  2. ...
  3. 08:44 < bazz_> Page 2 of https://tools.ietf.org/html/rfc4578 presents something usable
  4. ...
  5. 08:49 < NiXZe> bazz_: the reason that the "spec" is not quoted is because, spec and what is seen is the real world does not match.
  6. 08:55 < bazz_> NiXZe: still, I've been wondering for days where this table would be.. It'd be nice to have a page on iPXE detailing them, or at least as a referral.. Feel free to mention in this same page "may not match real world"
  7. 08:55 < bazz_> :P
  8. 08:55 < NiXZe> Acording to the spec and 9 is for x86-64, but I have never seen 9 in the real world, instead it uses 7 for x86-64 machines.
  9. 08:55 < NiXZe> Here is the ref you probably want: https://git.ipxe.org/ipxe.git/blob/15759e539e2fd0d33ef4f68c92942807f2048a29:/src/include/ipxe/dhcp.h#l249
  10. 08:56 < bazz_> My Macbook Pro reports: PXEClient:Arch:00009:UNDI:003010
  11.  
  12. 08:56 < NiXZe> http://forum.ipxe.org/showthread.php?tid=7725
  13. 08:57 < NiXZe> Its just "a mess that works" ;)
  14. 08:58 < NiXZe> or maybe Apple is following spec, and all other EFI implementors have it wrong.
  15.  
  16. 08:59 < p_l> I think I've seen 9 from all of my hardware
  17. 09:00 < p_l> also, technically, 7 might be requested by x86-64 or any other host so long as the firmware has BC interpreter
  18. 09:01 < p_l> bazz_: according to the spec, you can report 7 so long as you have interpreter :)
  19. 09:01 < NiXZe> It is partly documented here: http://ipxe.org/cfg/platform (wrong place and wrong info)
  20. 09:01 < p_l> I guess so, I haven't ever touched it
  21.  
  22. 09:02 < NiXZe> but all our dhcp servers is checking for 7 and boots new PC motherboards and laptops correctly.
  23. 09:03 < p_l> the problem with 7 is that I'm not sure what should be the correct behaviour for firmware supporting BC and let's say IA-32 or BC and x86-64
  24.  
  25. 09:03 < p_l> because sending x86-64 efi to firmware sending a 7 for the arch *breaks* the spec
  26. 09:03 < NiXZe> p_l: I agree with that, but not the rest of the PC industy it seems.
  27. 09:04 < NiXZe> industry*
  28. 09:04 < p_l> NiXZe: so long as the firmwares include BC interpreter they fulfill the spec by sending 7
  29. 09:04 < p_l> the real problem here is I think intel's stranglehold on BC compiler for C
  30. 09:04 < p_l> and I lost my copy of the old one
  31.  
  32. 09:06 < NiXZe> ok, so it really is that the PCs wants BC, but we send them x86-64 and they are happy anyway because they support that to.
  33.  
  34. 09:07 < NiXZe> mcb30: we need to put up a note on the Captcha "known not to work in Chrome" ?
  35. 09:08 < bazz_> NiXZe: scratch that, am installing firefox
  36. End of Lastlog
  37. 09:17 * bazz_ recorded this IRC conversation, but must get breakfast
  38. 09:17 * bazz_ couldn't find anything on DokuWiki captchas failing in Google Chrome
  39. 09:17 < bazz_> TTYL
  40. 09:21 -!- hpa [[email protected]] has quit [Ping timeout: 265 seconds]
  41. 09:22 -!- hpa [[email protected]] has joined #ipxe
Advertisement
Add Comment
Please, Sign In to add comment