Use Intel NIC PXE image mod for 2.1 SLIC?

Discussion in 'MDL Projects and Applications' started by SirSilentBob, Aug 25, 2009.

  1. Lithium

    Lithium MDL Novice

    Sep 2, 2007
    18
    0
    0
    One more...

    Intel PRO/1000 MT Desktop Adapter, VendorID 8086, DeviceID 1076
    Flashed new WindSLIC ROM (Attachment 3717).

    Asus K8N, clean Win7 32bit & 64bit and again success.
    No problem at all...

    That thing with big intel chip on it was intel LAN card but named as Compaq N...something. NO GO, not allowed to flash built-in BIOS chip with ibautil.exe

    All best
     
  2. sebus

    sebus MDL Guru

    Jul 23, 2008
    6,384
    2,037
    210
    This mobo's BIOS can be modded with Andy's tool...
    So as such no need for NIC mod really

    sebus
     
  3. SB7

    SB7 MDL Novice

    Aug 24, 2008
    10
    2
    0
    sebus,

    Yes, I'm aware that I can and have modded my bios many times .
    This was a proof of concept to see that it worked. I was doing it for a couple of colleagues of mine that have intel MB's and cannot mod their BIOS. They are happily running Win7 ultimate right now.

    Personally I prefer to do it with a NIC as it's portable .. I've installed my NICs on 6 different PC's and had it activate Win7 on them without having to mod the original BIOS. IMHO it is far safer to plug in a NIC than to mod a BIOS and potentially brick a board. Yes, the chances of bricking a BIOS are quite low, but if I don't touch it the probability falls to near zero.
    It's just personal choice.

    I do appreciate you pointing out that my MB is modifiable, as I may not have already been aware of that.

    Cheers

    SB7
     
  4. sebus

    sebus MDL Guru

    Jul 23, 2008
    6,384
    2,037
    210
    I see, proof of concept, very good.
    And you are right in all you said!

    Thanks for comments

    sebus
     
  5. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    #205 Apokrif, Nov 9, 2009
    Last edited by a moderator: Apr 29, 2017
  6. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    Did a little troubleshooting, got (what’s left on the screen):
    Checking for new e820 High-Water-Mark...
    [some memory ranges]
    ^-Skipping e820 line for reason: [00000001]
    ...
    Done Getting H820 line...
    Skipping Replacement Int15H Handler...


    Tested on:
    Dell Optiplex 745 (with SLIC already in BIOS), supposed to get “Detected SLIC length is”…
    Dell Dimension 9100 (no SLIC at all) – very similar output…

    Don’t know what to do next… Yeah, I’m that stupid! :confused:
     
  7. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    Couple more things:
    ACPI ver. 2.0 on all systems (although, it doesn’t matter much)
    "Checking for new e820 High-Water-Mark..." All memory ranges have type 2, - no one with type 1… :confused:
     
  8. truthinjection

    truthinjection MDL Member

    Aug 27, 2009
    247
    46
    10
    Hrm. I'm out of town at the moment, and I haven't looked at the WindSLIC stuff for a while, but let's see..

    How much RAM is in each of those PCs? When last I looked at the WindSLIC stuff, I discovered that it wasn't set up to work if you had more than 4GB of memory (due to my overly simplistic logic at choosing a donor memory region).

    Does the PC with the SLIC already embedded in its BIOS show that in the RW-Everything report?

    When I get back from out of town I'll see if I can fix at least the 4GB problem.

    Thanks,
    -tij-
     
  9. Lithium

    Lithium MDL Novice

    Sep 2, 2007
    18
    0
    0
    4GB Problem

    Yes MR. truthinjection, fix the 4GB problem please.
     
  10. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    All 2GB.

    Yes. I put NIC into it just to see the difference in debug output.

    Ok. What do you think about: (As I wrote early)
    Debug show: All memory ranges have type 2, - no one with type 1…
    Still wondering if it's unique for Dell BIOS or same for all Phoenix BIOS?
    Yeah, I’m that stupid! :confused:
     
  11. truthinjection

    truthinjection MDL Member

    Aug 27, 2009
    247
    46
    10
    OK, so if they're all 2GB, then it's not the 4+GB problem at least.

    I don't have a Dell or Phoenix machine to test on, unfortunately. My code probably has a bug with the Debug output on the "all type 2" problem. I had been trying to develop a newer version, but it didn't go well, so I'm going to have to start over at the published version again and see if I can work out the issue.

    Thanks,
    -tij-
     
  12. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    #212 Apokrif, Nov 12, 2009
    Last edited by a moderator: Apr 29, 2017
  13. truthinjection

    truthinjection MDL Member

    Aug 27, 2009
    247
    46
    10
  14. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    Ok… I’ll be able to run it Monday only… Will PM you at once!

    Ok. I didn’t brick anything… yet! Beginner’s luck? :confused:

    Sounds scary… :eek: Don’t they have a simple way to test NIC PCI slot in - and if it’s in the wrong one (odd, even, whatever), bail out. If we flush a buggy code, we can move NIC into different slot and re-flush it?

    Oh, boy, don’t die, please! I mean, not yet! ;)
    BTW: Hackers don’t write code, bro! They can only hack somebody else programs. Am I mistaken? :cool:
     
  15. ll25031968

    ll25031968 MDL Novice

    Aug 24, 2009
    9
    0
    0
    #215 ll25031968, Nov 14, 2009
    Last edited: Nov 14, 2009
    Apokrif , i,m your same case:
    i,ve a primergy rx300 s5 with 12gb ram and when deactivate onboard nics and plug the intel1000ct with the new soft loaded, i get a blue screen.
    i have unplugged all my ram but 2gb and i get the same results.
    i hope truthinjection solve this problem

    thanks

    pd: the bios is Phoenix
     
  16. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    #216 Apokrif, Nov 14, 2009
    Last edited by a moderator: Apr 29, 2017
  17. ll25031968

    ll25031968 MDL Novice

    Aug 24, 2009
    9
    0
    0
    here you are

    with slic nic:

    BIOS-provided physical RAM map:
    BIOS-e820: 0000000000000000 - 000000000009CC00 (usable)
    BIOS-e820: 000000000009CC00 - 00000000000A0000 (reserved)
    BIOS-e820: 00000000000E4000 - 0000000000100000 (reserved)
    BIOS-e820: 0000000000100000 - 00000000BF7BB800 (usable)
    BIOS-e820: 00000000BF7BB800 - 00000000BF7C0000 (ACPI data)
    BIOS-e820: 00000000BF7C0000 - 00000000BF7CB000 (ACPI data)
    BIOS-e820: 00000000BF7CB000 - 00000000BF7CC000 (ACPI NVS)
    BIOS-e820: 00000000BF7CC000 - 00000000BF7FF000 (reserved)
    BIOS-e820: 00000000BF7FF000 - 00000000BF800000 (reserved)
    BIOS-e820: 00000000BF800000 - 00000000C0000000 (reserved)
    BIOS-e820: 00000000E0000000 - 00000000F0000000 (reserved)
    BIOS-e820: 00000000FEC00000 - 00000000FED00000 (reserved)
    BIOS-e820: 00000000FEE00000 - 00000000FEE01000 (reserved)
    BIOS-e820: 00000000FFA80000 - 0000000100000000 (reserved)
    BIOS-e820: 0000000100000000 - 0000000340000000 (usable)

    and without slic nic

    BIOS-provided physical RAM map:
    BIOS-e820: 0000000000000000 - 000000000009CC00 (usable)
    BIOS-e820: 000000000009CC00 - 00000000000A0000 (reserved)
    BIOS-e820: 00000000000CE000 - 00000000000D0000 (reserved)
    BIOS-e820: 00000000000E4000 - 0000000000100000 (reserved)
    BIOS-e820: 0000000000100000 - 00000000BF7C0000 (usable)
    BIOS-e820: 00000000BF7C0000 - 00000000BF7CB000 (ACPI data)
    BIOS-e820: 00000000BF7CB000 - 00000000BF7CC000 (ACPI NVS)
    BIOS-e820: 00000000BF7CC000 - 00000000BF7FF000 (reserved)
    BIOS-e820: 00000000BF7FF000 - 00000000BF800000 (reserved)
    BIOS-e820: 00000000BF800000 - 00000000C0000000 (reserved)
    BIOS-e820: 00000000E0000000 - 00000000F0000000 (reserved)
    BIOS-e820: 00000000FEC00000 - 00000000FED00000 (reserved)
    BIOS-e820: 00000000FEE00000 - 00000000FEE01000 (reserved)
    BIOS-e820: 00000000FFA80000 - 0000000100000000 (reserved)
    BIOS-e820: 0000000100000000 - 0000000340000000 (usable)
     
  18. truthinjection

    truthinjection MDL Member

    Aug 27, 2009
    247
    46
    10
  19. Apokrif

    Apokrif MDL Addicted

    Dec 7, 2008
    542
    35
    30
    Truthinjection,
    We really need this piece for our safety ;)
    Namely, I would suggest to add small piece of code:
    1. On boot (very first phase) NIC should report:
    A. Its PCI slot.
    B. Configured PCI slot (i.e. compiled in)
    C. "Bail out" status.

    I.e. if (NIC PCI slot) != (Configured PCI slot) no attempt to do any memory mapping/injection/whatever should be done.
    This way if there is a mistake in above code, we can move NIC to a different PCI slot and re-flush it.
    So we don’t have to do that: :eek:
    IMO, almost all code are there already. All we need is a little on-screen reporting.

    Again, since we are going to shaky debugging ground, extra safety pin won’t hurt. ;)
    I hope Truthinjection will take over flu real soon!

    And I’m anxious to get my dumps for examination!

    Truthinjection, could you share/PM dos_e820.com src, so I can look further in interesting regions, plz?
    dos_e820 output is nice, I might just add a little more details for ACPI data regions. Not sure how useful it’ll be tho.
     
  20. truthinjection

    truthinjection MDL Member

    Aug 27, 2009
    247
    46
    10
    #220 truthinjection, Nov 16, 2009
    Last edited: Nov 16, 2009
    That code is actually already in there, (buried in the Advanced part and disabled by default) except for the reporting part. Displaying anything onscreen seems to be the easiest way to cause a hang in the first place. Plus, the slot numbering isn't really human-readable, since it could be wired to be any bus, any slot, even if you only had one physical PCI slot.

    My first bailout code was based on PCI slot "even/odd" status, but then I realized that lots of motherboards only have *one* PCI slot these days, so I went towards the "Press F5" method, which makes it a lot simpler to fix (no physical unscrew/rescrew needed) and reflash if the screwed up code is in the BEV portion.

    On-screen reporting also makes the ROM a lot bigger than anything else. The Intel cards have plenty of space (128KB, usually), but most BIOSes only give between 32KB-96KB room, before they either don't load the ROM at all, or hang, which would be bad.

    Eventually I plan to release all the code I've got. I want to try and incorporate some of the ACPI stuff into a C-based program (currently the dos_e820 is turbo-c based, with some NASM code parts) instead of an ASM-only program (as my ACPI dumper is, since it's based on FDAPM code). The C-based stuff, once you get the ASM-bits working, is much cleaner to read/understand.

    -tij-