<Bob_Dole>
Raptor's talking about a new lower-cost POWER board
<cr1901_modern>
This is mild off-topic, tbh. ##openfpga often evolves (devolves?) into a room about anime
<cr1901_modern>
Pretty sure azonenberg enforced the "no off topic" rule for a week at most when the channel was new :D.
<qu1j0t3>
Bob_Dole: interesting
mumptai has joined ##openfpga
<awygle>
that's cool
<shapr>
cr1901_modern: on the good side, I think we're get back on topic when something on-topic arrives
<balrog>
regarding that -- has anyone had to deal with lattice xp2 series?
<balrog>
Apple uses one as the GPU mux in the 2011 macbook pro, and a crazy idea is to reprogram it to never enable the discrete GPU
<balrog>
(since discrete GPU failure is a given in those machines)
<qu1j0t3>
balrog: i have a 2012 with gpu problems. if i don't disable acceleration in chrome, constant freezes for seconds. And Photoshop complains on launch and disables GPU... which means i can't acces the preference to turn it off pre-emptively *facepalm*
<qu1j0t3>
balrog: Machine otherwise is fine though :D
<balrog>
qu1j0t3: which 2012?
<balrog>
15" retina or non-retina?
<qu1j0t3>
Retina
<balrog>
that's not the GPU
<qu1j0t3>
it's already had afew logic board replacements
<balrog>
that's soldering under a power chip
<qu1j0t3>
why are all the problems gpu related ?
<balrog>
because the power chip is part of the GPU power circuit
<balrog>
so the GPU no longer gets clean power
<qu1j0t3>
(also the previous owner said that video card connections were a culprit iirc)
<qu1j0t3>
ah ok.
<balrog>
yeah, because that was what was believed
<balrog>
the chip/area in question is U8900
<balrog>
find the schematic and boardview for the model and poke around :)
<qu1j0t3>
balrog: let's take this to a more off topic channel
<balrog>
sure
m4ssi has quit [Remote host closed the connection]
<kc8apf>
Bob_Dole: i'm glad to see Raptor is still going. IBM kinda-sorta supports them but they've had a lot of challenges getting parts to market
keesj has quit [Quit: Changing server]
<azonenberg_work>
cr1901_modern: my rule is, when people want to talk on topic don't derail it
<azonenberg_work>
But if nothing interesting is going on i dont care
<cr1901_modern>
azonenberg_work: Are you perpetually at work lately :P?
<qu1j0t3>
who isn't
rohitksingh has quit [Quit: Leaving.]
<azonenberg_work>
cr1901_modern: actually today is my day off in theory, although i do have one call with a customer
<azonenberg_work>
working odd hours for this project since work is being done outside business hours
<cr1901_modern>
Lol, I just meant according to my client, you haven't used the handle "azonenberg" in 63 days
<qu1j0t3>
stalking
<balrog>
anyway back to the point: has anyone had to deal with lattice xp2 series?
<azonenberg_work>
cr1901_modern: oh, that
<azonenberg_work>
That's because the computer that has the "azonenberg" nickserv password on it
<azonenberg_work>
is in a cardboard box
<azonenberg_work>
and has been there for 2 months
keesj has joined ##openfpga
<cr1901_modern>
azonenberg_work: Ooooooooh, right makes sense
<awygle>
azonenberg has a bandana tied into a sack on the end of a stick
<awygle>
and roams the PNW by day while working in a labor camp at night
<azonenberg_work>
lol
<awygle>
some say he's still wandering to this day, bemoaning his fate and warning people against buying more house than they can afford :p
<azonenberg_work>
awygle: hey, i could afford the house just fine
<azonenberg_work>
It's making it habitable that is problematic :p
<awygle>
to quote myself from 30 seconds ago in another channel - you can't see me, but i'm rolling my eyes :p
<Ultrasauce>
adding hardware seems pretty hamfisted when there are uefi rootkits in the wild already
<qu1j0t3>
not if people are already alert to them (betting apple/amazon dc's are). That said, i take the whole article with a big grain of salt.
<zkms>
you can read-back flash and analyse it pretty easily, the modchip idea is neat because it lets you inject tainted stuff upon boot but software that tries to read it back later will just get the clean version
<zkms>
i dont know anything about the specific supermicro case but i do know that putting a modchip that touches flash SPI lines and the ethernet SMBus lines (the ones that Intel ME uses to talk to the network) has a lot of advantages over just UEFI fuckery
<Ultrasauce>
...I didn't want to know that they bridge ethernet over i2c
<rqou>
if you repeatedly inject ascii SCEA into your UEFI, you pwn the machine /s
<fseidel>
but it has to be 1 bit at a time
<fseidel>
and it's INTELINSIDE :-)
<rqou>
if you patch in AUTHENTICAMD instead, one of the dragons in spyro will yell at you :P
<rqou>
(idk how this particular antipiracy actually worked)
ayjay_t has quit [Read error: Connection reset by peer]
ayjay_t has joined ##openfpga
m_t has quit [Remote host closed the connection]
<zino>
iirc one of the bios calls has a undocumented argument that lets you know how many "SCEx" you got since... some point. Using it on known reads will of course return 1/3rd of the "SCEx" of you have the standard injection method of inserting all 3 regions one after the other in your modchip.
<kc8apf>
Apple/Amazon/FB/Google/MS all have process in place to not trust flash contents on new machines. First thing everyone does is reflash everything
<emeb>
anyone tried loading up that ARM cortex M1 design example in Vivado?
<cr1901_modern>
It's encrypted RTL :(
<emeb>
I understand that. I'm just talking about opening it up in the Xilinx tools to get an idea of how it sits in a design, what the tools let you do, etc.
<emeb>
I downloaded it and followed the (long) instructions on setting up all the prerequisites and it still bombed out when I tried to open up the design.
<whitequark>
you can also decrypt it
<cr1901_modern>
Oh, the keys that were already leaked will work?