asciilifeform: i did notice.

asciilifeform: hey, futureproof!11111

asciilifeform: in all of the keys.

asciilifeform: mircea_popescu: i just verified, indeed, they left a 2048-bit long chunk of bits for the 65537 to live in !

asciilifeform: ... consider sharing ?

asciilifeform: kmalkki: how did you determine the fact about the sha256 at boot ?

asciilifeform: !!up kmalkki

asciilifeform: iirc cpu will run with a nulled smu rom.

asciilifeform: (smu is a southbridge animal)

asciilifeform: it was interesting, but does not help with the crippled psp.

asciilifeform: kmalkki: https://events.ccc.de/congress/2014/Fahrplan/system/attachments/2503/original/ccc-final.pdf << these ?

asciilifeform: perhaps it was simply to trace leaks.

asciilifeform: if it reduces to the same thing, given lack of revocation mechanism

asciilifeform: though it raises the question of why they would not simply share their master key with the OEMs then

asciilifeform: that would appear to be so

asciilifeform: any idea how amd revokes ?

asciilifeform: ah thx

asciilifeform: do you happen to know the format ?

asciilifeform: lolk

asciilifeform: do you know what an even rsa modulus implies ?

asciilifeform: e.g., 'avalon' PspSecureDebug.Key , ends in 0e a8 92 70 .

asciilifeform: kmalkki: something is strange. the format appears to be little-endian, but if we look at some of the pubkeys in http://wotpaste.cascadianhacker.com/pastes/k081w/?raw=true , they would appear to be EVEN

asciilifeform: !!up kmalkki

asciilifeform: Framedragger: did you ever post your converter-to-rfc4880 script ?

asciilifeform: any idea whether the latter are derived from the modulus ?

asciilifeform: interesting

asciilifeform: 0x01...0x03 seems to be a constant, 1. but what's 0x04..0x38.. ?

asciilifeform: kmalkki: http://phuctor.nosuchlabs.com .

asciilifeform: kmalkki: if it is an rsa key, i would like to get it into phuctor.

asciilifeform: ergo - more expensive.

asciilifeform: mircea_popescu: narrows the space, yes.

asciilifeform: mircea_popescu: pubkey-with-header, presumably, so i'd like to know 1) where is the modulus 2) is there checksum etc.

asciilifeform: (i.e. if the collision needs to be a turd in a similar format)

asciilifeform: mircea_popescu: can be pretty expensive if there is any structure checking

asciilifeform: of one single sha2 ?

asciilifeform: kmalkki: so all we need is a collision to break the thing ?

asciilifeform: you add short polymer chains, they 'drag', so to speak, the water.

asciilifeform: the sea, lol, but apparently it is possible to 'lubricate' tapwater, reduces vorticing and friction loss. discovered in '70s in su, iirc, never used for anything.

asciilifeform: https://github.com/coreboot/coreboot/blob/master/util/amdfwtool/amdfwtool.c << this appears to show where it ends up sitting down in the coreboot flash image, but not the format.

asciilifeform: !!up kmalkki

asciilifeform: kmalkki: do you perchance know the format of the amd public keys seen here : http://btcbase.org/log/2016-10-04#1552732 ( from coreboot dist ) ☝︎

asciilifeform: and never did.

asciilifeform: but no, i don't have access to any amd members-only whatevers.

asciilifeform: (or rather, that it was not in any of the public datashits.)

asciilifeform: kmalkki: i know that it is in the private one, yes.

asciilifeform: if you would like to post them publicly, send them pgp-encrypted to any of the folks here.

asciilifeform: kmalkki: i do not and will not sign any nda.

asciilifeform: kmalkki: but the other prong of this is your apparent discovery that new g-series boards disable hdt somehow

asciilifeform: so that folks do not need to rely on the extinct smartprobe and can make own debugger (the pinout is public)

asciilifeform: kmalkki: now more interestingly, hdt is simply a protocol on top of jtag, imho the main scientific interest in the smartprobe fw would be to extract the protocol.

asciilifeform: kmalkki: the stellaris arm also had jtag pins, handily brought out to pcb, as seen here, http://www.loper-os.org/pub/sage/test_points.jpg , i plugged it into a busblaster and eventually stepped through the execution from reset to where it checked the serial.

asciilifeform: kmalkki: ida happily eats the update payload.

asciilifeform: kmalkki: i received your find, it is very spiffy.

asciilifeform: mircea_popescu: kmalkki is an experimenter with sage/amd g-series/etc.

asciilifeform: kmalkki: http://btcbase.org/log/2016-10-04#1552690 << here's what you missed re thread. ☝︎

asciilifeform: hi Kyösti

asciilifeform: !!up kmalkki

asciilifeform: that would be hilarious

asciilifeform: (what prevents the substitution of another pubkey ?)

asciilifeform: i would think it would be in mask rom.

asciilifeform: mircea_popescu: is it just me or is it a strange thing that the pubkey is in there at all.

asciilifeform: ^ from coreboot src.

asciilifeform: default "3rdparty/blobs/southbridge/amd/kern/PSP/AmdPubKeyCZ.bin" if CPU_AMD_PI_00660F01

asciilifeform: default "3rdparty/blobs/southbridge/amd/avalon/PSP/AmdPubKey.bin" if CPU_AMD_PI_00730F01

asciilifeform: http://wotpaste.cascadianhacker.com/pastes/k081w/?raw=true http://wotpaste.cascadianhacker.com/pastes/wv3x3/?raw=true << have we heroes ?? ☟︎

asciilifeform: it is the key format that interested me.

asciilifeform: nobody is that stupid.

asciilifeform: but i cannot presently believe that the 'd' (private exponent) is actually in there.

asciilifeform: in a few seconds.

asciilifeform: realize, if we know the private exponent, we can demolish the N.

asciilifeform: ' |--key metadata--|--public modulus--|--public exponent--|--private exponent--|'

asciilifeform: ok am i dreaming, or did i just read...

asciilifeform: (unless they made an entirely other one for x86)

asciilifeform: https://android.googlesource.com/platform/hardware/samsung_slsi/exynos5/+/master/libkeymaster/tlcTeeKeymaster_if.c#831 << possibly the key format.

asciilifeform: will offload part of AGESA to the PSP, making memory init even part of the PSP :(' -- https://www.coreboot.org/Binary_situation .

asciilifeform: 'PSP: The Cortex A8 running trustzone firmware, implemented from family 15h model 60h (Carrizo) and family 16h model 30h (Mullins) and up. It is running Trustonic TEE OS licensed by AMD. PSP Boot ROM runs *before* x86 core. Then non-bootrom PSP parts are stored (zlib-compressed) in the main flash. Bypass mechanism available via strap pin, but dummy and AMD signed bypass binaries needs to be always run. It is expected that newer CPUs

asciilifeform: https://archive.is/biNPy << related lulz.

asciilifeform: it gets rolled into coreboot, or post-2013 chip doesn't release from reset on warmup.

asciilifeform: amd's shitball, as-released

asciilifeform: i'd like to get their pubkey into phuctor, but it is in some peculiar format. (if somebody can discern a modulus in there, please write in.)

asciilifeform: so these are what run on the apu2.

asciilifeform: mircea_popescu: it is what gets rolled into coreboot

asciilifeform: no need to extract, it is right there !

asciilifeform: only signed

asciilifeform: the payloads aren't, interestingly, encrypted

asciilifeform: https://github.com/coreboot/blobs/tree/master/southbridge/amd/kern/PSP << lulzy, on subj.

asciilifeform: 'AMD’s PSP is based around a single 32-bit ARM Cortex-A5, with its own isolated ROM and SRAM but has access to system memory and resources. It contains logic to deal with the x86 POST process but also features a cryptographic co-processor.' << they stuffed, finally, their 'fritz chip', into the g-series.

asciilifeform: trinque, shinohai , mircea_popescu , et al ^

asciilifeform: ACHTUNG, PANZERS! pc engines 'apu2' (the board with the intel nics - vs. 'apu1', with realtek) , turns out, is crippled, hdt probe barfs with it, the cpu is reputed to have a drm fuse set. ☟︎☟︎☟︎☟︎☟︎☟︎☟︎☟︎☟︎

asciilifeform: PeterL: what is the logic in ' if hexmessage[-1] == 'L': ' ?

asciilifeform: aha.

asciilifeform: notice, with my original, you can substitute in your favourite hasher without changing the code.

asciilifeform: (in actual practice, i would probably use sha512 or keccak; sha256 was for pedagogic example only)

asciilifeform: PeterL: neato

asciilifeform: 'The theory is that while a court can compel someone to not speak (a gag order), it cannot compel someone to lie.' << orly.

asciilifeform: to know which way the elevator is going.

asciilifeform: nah, a stalingrad, definitionally, suffices for the folks on the bus.

asciilifeform: mtgox, evidently, was not a sufficient stalingrad.

asciilifeform: sorta what it's for, neh

asciilifeform: the greatest american who ever lived, p. t. barnum, is not forgotten in his motherland. his work lives.