<qu1j0t3>
anyone here had experience with counterfeit linear chips?
<qu1j0t3>
e.g. Analog Devices. I've been, uh, burned once. Anyone know if cerdips are counterfeited as well as plastic dips?
<qu1j0t3>
also: is Hungary a known counterfeit source, or just that.. other place
<qu1j0t3>
i assume the Eastern European sellers are mostly waste reprocessors? i've never quite figured out why so many high brand value chips are sold from there
<qu1j0t3>
and environs
<qu1j0t3>
teach me irc
<prpplague>
qu1j0t3: less risk selling counterfit chips than selling fake cigarettes
jcarpenter2 has joined ##openfpga
unixb0y has quit [Ping timeout: 252 seconds]
unixb0y has joined ##openfpga
genii has joined ##openfpga
azonenberg_work has quit [Ping timeout: 264 seconds]
jcarpenter2 has quit [Remote host closed the connection]
jcarpenter2 has joined ##openfpga
azonenberg_work has joined ##openfpga
<whitequark>
awygle: poke?
jevinskie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
* prpplague
reports whitequark to HR for harrasment
<TD-Linux>
oh yeah I should finish testing my RevB
Bike has quit [Quit: Lost terminal]
jcarpenter2 has quit [Quit: Leaving]
s_frit has quit [Remote host closed the connection]
s_frit has joined ##openfpga
genii has quit [Remote host closed the connection]
<lain>
prpplague: thanks for fuelling my nightmares :P
<lain>
I think when we start sourcing parts for our ARM laptop (mostly for ourselves / internal use) we'll uhhhh x-ray all components and decap random samples D:
_whitelogger has joined ##openfpga
futarisIRCcloud has quit [Quit: Connection closed for inactivity]
<mrpackethead>
When someone produces an IP core "for sale" and you buy it.. is it encrypted or how do thye protect their IP?
<lain>
usually it is encrypted, but it has to be decrypted as part of the design process so it's like DRM... very easy but very illegal to bypass
GuzTech has joined ##openfpga
<azonenberg_work>
It depends on the specifics of the core
<azonenberg_work>
Sometimes you get a full source license
<azonenberg_work>
Sometimes you get a simulation model, plus a black box standard-cell "dummy" that contains pins for the IP core
<azonenberg_work>
and is replaced by the actual hard IP at the fab (you never see it)
<azonenberg_work>
(that's for asic)
<azonenberg_work>
For fpga, generally either source or encrypted netlist
<azonenberg_work>
Sometimes both are available depending on how much you pay
<azonenberg_work>
I mean, if you have enough arm will sell you a copy of the cortex-a53 source code that you can modify and turn into your own asic with additional propriretary cpu instructions or whatever
<azonenberg_work>
But i'm guessing arm source licneses start at the tens of millions for the lowest end cpus
<azonenberg_work>
an a53 source license probably costs you $50M or so
<azonenberg_work>
Plus royalties for each chip you make :p
<rqou>
really? tens of millions for an arm7tdmi?
<azonenberg_work>
i was thinking for like an m3
<azonenberg_work>
or something
<azonenberg_work>
but that was just a wild guess
<azonenberg_work>
i've never actually tried getting one :p
<rqou>
i was expecting "only" millions for an m3
<azonenberg_work>
either way, more than most companies can afford
<rqou>
and even less for an arm7tdmi/arm926ejs
<sorear>
they seem to not care very much about m3
<sorear>
https://github.com/cryptolu/maps is a functional simulator which was written by non-ARM employees while looking at the M3 Verilog code
<sorear>
it says something that that repo wasn't dmca'd in less than five minutes, I'm not quite sure what
<rqou>
i know there's no way i would pay tens of millions for an arm7tdmi
<rqou>
i'd rather just yolo a picorv32
<sorear>
i've also met someone IRL who claims that the ARM1176 has the same status as Night of the Living Dead due to rpi-related snafus; I trust this about 0% though
<sorear>
awfully specific thing to be making extremely specious claims about
<Bob_Dole>
talking about ASIC-making the other day, I was thinking about this guy, I'm sure you've all seen it before but: https://twitter.com/szeloof
<lain>
speaking of arm royalties, they recently released encrypted IP you can use in xilinx fpgas for free, it's a Cortex-M1 and I think an M3 is coming soon
<azonenberg_work>
at 0:06 i'd have yelled at him to get the muzzle on target NOW
<azonenberg_work>
and if it wasn't where it belonged immediately, off the range before it got any further
<azonenberg_work>
actually at 0:03 it looks like he sweeps way off to his right too
<rqou>
looks like the RSO was busy with another group
<rqou>
huh wow being an RSO is tough
<azonenberg_work>
That tells me they need more RSOs :p
<azonenberg_work>
At my club, the RSOs tend to watch people they dont recognize like a hawk
<rqou>
so no loaded gun selfies?
<azonenberg_work>
you'd be out on your ass long before it got to that point
<azonenberg_work>
If you're a regular they trust, they keep an eye out in case you slip up but arent super paranoid
<azonenberg_work>
If it's somebody new they look extra close
<azonenberg_work>
Now that I'm certified as one myself, they're a little less down my neck when I have a new shooter in my lane
<azonenberg_work>
But that's because they know me, they know i know what i'm doing, and i normally deal with issues long before they get out of hand
<azonenberg_work>
The only time i remember the on-duty RSO overruling me on a safety issue is when i was paying too much attention to muzzle control and missed that the guy's hand was too high up the grip
<azonenberg_work>
and he risked being pinched by the slide when it came back
<azonenberg_work>
But that's a relatively minor risk vs shooting somebody else there :p
<azonenberg_work>
If i'm watching somebody i normally complain if they're more than ~30 degrees off target because it's a bad habit
<azonenberg_work>
Even if they're nowhere near any of the other lanes
<azonenberg_work>
i want them thinking about it BEFORE it gets to the point i have to grab the barrel and point it in a safe direction
<rqou>
but have you ever seen an idiot trying to take a selfie?
<azonenberg_work>
No
<azonenberg_work>
Like i said, we deal with idiots before they get to that point
<azonenberg_work>
Minor muzzle control issues (~30 degrees off axis or so) get a polite correction, anything anywhere close to another shooter a ceasefire followed by a very stern warning
<azonenberg_work>
If i saw someone actually sweep another shooter with a loaded weapon they'd be off the range immediately
<rqou>
btw thoughts on those "gangsta" selfies with presumably unloaded guns?
<azonenberg_work>
I have zero respect for anyone who demonstrates poor weapons handling whether it's loaded or not
<azonenberg_work>
That's how people win darwin awards
<rqou>
does that include police officers? (btw we really should have a ##openfpga-offtopic to absorb the stuff that ##whitequark hasn't absorbed yet)
<azonenberg_work>
Or worse yet, shoot a stranger by accident
<azonenberg_work>
The police I know personally are very competent about such things
<azonenberg_work>
I cant comment on other departments
<azonenberg_work>
If they're anything like the military ranges i've heard stories of, their training does not mess around
<rqou>
and yet NYPD/LAPD/etc. keep "accidentally" shooting people
<azonenberg_work>
Like i said, i cant comment on other departments
<azonenberg_work>
What i do know is that NYPD issues Glocks, which have no manual safety catch and rely on you not pulling the trigger when you don't want to fire it, and as a bit of a crutch
<azonenberg_work>
they have a nonstandard excessively high trigger pull
<azonenberg_work>
Which does make an unintentional shot slightly harder to fire, but also reduces accuracy since you have to put a lot of force on the trigger
<azonenberg_work>
i've heard close to ten pounds
<azonenberg_work>
Which could very well lead to significant sideways motion as you fire if you're not experienced with it (and will reduce accuracy even for a skilled shooter)
<azonenberg_work>
Not saying that's the cause, but it certainly could contribute in some cases
<azonenberg_work>
to an intended shot going in an unintended direction
Laksen has joined ##openfpga
<q3k>
azonenberg_work: yeah, i'm not firearms expert, but when I shot a glock the trigger was somewhat heavy
<q3k>
azonenberg_work: took a lot of practice to pull it steadily without losing your aim sideways
<q3k>
azonenberg_work: well, a lot. after an hour i basically got it :)
_whitelogger has joined ##openfpga
s_frit has quit [Ping timeout: 252 seconds]
florolf has quit [Remote host closed the connection]
florolf has joined ##openfpga
rohitksingh_work has quit [Read error: Connection reset by peer]
scrts has quit [Ping timeout: 252 seconds]
Bike has joined ##openfpga
scrts has joined ##openfpga
rohitksingh has joined ##openfpga
Laksen has quit [Ping timeout: 252 seconds]
scrts has quit [Ping timeout: 268 seconds]
scrts has joined ##openfpga
Aleksandar-K has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
rohitksingh has joined ##openfpga
emeb has joined ##openfpga
Miyu has joined ##openfpga
genii has joined ##openfpga
m4ssi has quit [Remote host closed the connection]
jevinskie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jevinskie has joined ##openfpga
mumptai has joined ##openfpga
rohitksingh has quit [Quit: Leaving.]
jevinskie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jevinskie has joined ##openfpga
GuzTech has joined ##openfpga
<azonenberg_work>
q3k: most glocks are not that heavy
<azonenberg_work>
NYPD had theirs specifically modified
<azonenberg_work>
the theory was to reduce liability of negligent discharges
<azonenberg_work>
most folks who know anything about shooting think that's BS
<azonenberg_work>
might be a little harder to shoot by accident but if you hit the guy next to your perp and not him, that isnt much use
rohitksingh has joined ##openfpga
GuzTech has quit [Remote host closed the connection]
rohitksingh has quit [Quit: Leaving.]
azonenberg_work has quit [Ping timeout: 268 seconds]
azonenberg_work has joined ##openfpga
<prpplague>
how was orconf this year?
<prpplague>
the schedule looked fantastic
genii has quit [Read error: Connection reset by peer]
mumptai has quit [Quit: Verlassend]
Aleksandar-K has quit [Quit: Leaving]
genii has joined ##openfpga
scrts has quit [Ping timeout: 268 seconds]
kristianpaul has left ##openfpga [##openfpga]
scrts has joined ##openfpga
uovo has joined ##openfpga
oeuf has quit [Ping timeout: 252 seconds]
uovo has quit [Ping timeout: 246 seconds]
pie_ has joined ##openfpga
<kc8apf>
EAR requires a license for exports to designated terrorists. All requests for such licenses are summarily denied.
<whitequark>
ear?
<kc8apf>
US export restrictions
<kc8apf>
less restrictive than ITAR
<whitequark>
ah
<zkms>
why would they require a licence instead of just like...prohibiting altogether export to designated terrorists
<kc8apf>
zkms: bureaucrats love their jobs
<whitequark>
wait
<whitequark>
TERRORISTS
<kc8apf>
or maybe there's the occasional terrorist organization US gov't wants to sell weapons to
<whitequark>
not TERRITORIES
<whitequark>
wtf
<kc8apf>
yes, terrorists
<zkms>
another day at the EAR license office; everyone keeps asking me if they can get a license to export articles to designated terrorists, buddy, they won't even let *me* export them
<whitequark>
...
<whitequark>
tweet that
<TD-Linux>
ah, is this one of those cases where silicon valley innovation can ask for forgiveness later?
<zkms>
whitequark: i dont have a source for appropriate context and i dont want someone misunderstanding me lol
<whitequark>
oh
<whitequark>
right
<shapr>
Anyone have an ORConf writeup? I'm especially curious about the clash-lang presentation.
<Bob_Dole>
maybe exporting ears of corn as a peace-offering to halt terrorist activity is expected under EAR?
* shapr
plays the rimshot sound effect
<kc8apf>
TD-Linux: not usually. I mean, it regularly comes down to paying a fine and being restricted from doing those types of exports for a while
<openfpga-github>
[Glasgow] whitequark commented on issue #64: @kevinmehall Thanks, this worked quite well. Though I didn't anticipate that SET INTERFACE would reset the EP toggles and was really puzzled for a while there. https://github.com/whitequark/Glasgow/issues/64#issuecomment-428356957
<whitequark>
you can finally ^C applets and not have junk being read the next time you run it
<whitequark>
I really hated that and the only way to fix it would be to open the python console and do some bulk reads...
<travis-ci>
whitequark/Glasgow#57 (master - 5ac4c8d : whitequark): The build passed.
<whitequark>
it's really something amazing to fix a bug that caused incessant intermittent confusing failures and have everything work perfectly every time now
<felix_>
whitequark: did you have a look at the trace i linked you? if it looks sort-of useable, i'd try to get a trace at 24mhz with sigrok-cli on another computer running natively under linux. seems though that the frequency of the signal on the cc line is much higher that the 300khz
Bike has quit [Ping timeout: 256 seconds]
oeuf has joined ##openfpga
mumptai has joined ##openfpga
mumptai has quit [Client Quit]
s_frit has joined ##openfpga
<whitequark>
felix_: i haven't been able to install saleae sw for some reason, will try again today
<whitequark>
unrelated: anyone here understands sd/mmc?
<whitequark>
like there's spi mode, mmc mode, and... sd mode? do sd cards support mmc mode?
<whitequark>
rqou: i swear this used to not be a thing...
<whitequark>
a few years ago
<whitequark>
many years ago, ok
<rqou>
no it's definitely been around for a while
<whitequark>
you could do SPI but the entire SD thing was under NDA or something.
<whitequark>
hm ok
<whitequark>
weird.
<rqou>
i was reading it back in the NDS homebrew era to try to debug a "DLDI" (cursed)
<whitequark>
I'm not even going to ask what that is
<rqou>
iirc I fixed it but never published the fix anywhere
<whitequark>
hm
<whitequark>
the URLs are predictable so you can technically download them all without ever accepting EULA
<whitequark>
I wonder if the courts would agree with me
<rqou>
the linux kernel also has a driver that knows how to send all the right commands
<whitequark>
rqou: wtf, sd and mmc are basically the same except different
<whitequark>
why does this exist
<rqou>
good question
<rqou>
afaik drm
<rqou>
mmc doesn't have that
<whitequark>
it does
<whitequark>
different kind of drm
<rqou>
wat
<whitequark>
also not public
<rqou>
wtf
<whitequark>
seriously
<rqou>
wtfwtfwtf
* rqou
shoots self
* ZipCPU
calls ambulance for rqou
<q3k>
>sd and mmc are basically the same except different
<q3k>
this is annoyingly accurate
<q3k>
god I hate those standards so much
<whitequark>
extremely cursed
<q3k>
seeing the specs actually makes it more difficult to understand them
<q3k>
linux drivers are more readable then the fucking specs
<q3k>
everything about SD/MMC is utter trash
<q3k>
the wire format
<q3k>
the documentation for the wire format
<q3k>
the 'standardized' host controller api (forgot what it's called)
<whitequark>
the wire format just seems kind of "meh" to me, but the command set is some real atrocity
<q3k>
and the documentation for that in turn
<whitequark>
SDHC
<Bob_Dole>
so uhhh, SD/MMC- is using just the spi interface thing viable for stuff that doesn't need a lot of speed?
<q3k>
nope
<zkms>
also IME most SD cards i've used have been the worst binned flash ever, just the slowest least enduring stuff there is
<q3k>
the SPI is totally separate
<q3k>
*SPI mode
<Bob_Dole>
goal of just getting storage onto a 65816
<q3k>
and the decision tree to figure out whether your card actually supports SPI mode is like a choose your own adventure book
<whitequark>
Bob_Dole: I *think* the SD SPI mode and MMC SPI mode are *also* separate things
<q3k>
twenty different logos
<zkms>
apparently they make SLC "industrial" SD cards and hopefully those aren't the worst-binned shit
<Bob_Dole>
Oh god
<q3k>
fifty different standards
<whitequark>
but I can't confirm it because the docs are awful
<ZipCPU>
whitequark: It's SPI both ways, but the messages are subtly different.
<q3k>
whitequark: seriously, what the _fuck_ where they smoking when they just numbered the commands
<kc8apf>
zkms: they're only slightly better
<whitequark>
ZipCPU: right, this is what I assumed after looking at the specs
<rqou>
uh, my experience is that SD card flash is ok as long as you blacklist any vendor who doesn't actually own the flash fab
<whitequark>
thanks
<whitequark>
I hate it
<q3k>
whitequark: instead of giving them names/identifiers
<rqou>
e.g. samsung sd cards are ok
<kc8apf>
very little in the way of health statistics available for SD. Hard to know when a card is starting to fail
<whitequark>
q3k: they give them names in the mmc spec
<whitequark>
but also still use numbers at random
<whitequark>
because fuck you, that's why
<q3k>
reading the spec from page 1 is like being dropped into the middle of a B movie
<q3k>
what the fuck is CMD1
<q3k>
what the fuck is CMD3
<whitequark>
oh yes.
<q3k>
what does any of this mean
<rqou>
oh lol yeah
<q3k>
why does this state diagram look like the kaballah
<q3k>
god damn JEDEC you had one job
unixb0y has quit [Ping timeout: 246 seconds]
<whitequark>
q3k: are you on ##whitequark yet btw
<whitequark>
the cursed tech channel
<oeuf>
whitequark: is it really about tech?
<oeuf>
I thought it was about cats and then went off topic
<whitequark>
oeuf: well it is about cursed tech *now* apparently
<whitequark>
it was about medicine.
<whitequark>
but then twitter happened to it
<rqou>
there's still medicine and lewd and cats and stuff
<oeuf>
rqou: are whitequark's cats lewd?
<rqou>
well, one of them had kittens, so yes?
<whitequark>
in a way, our bodies are the most cursed tech of it all
Miyu has quit [Ping timeout: 252 seconds]
<q3k>
whitequark: I thought ##openfpga was ##whitequark
unixb0y has joined ##openfpga
<whitequark>
q3k: eh it felt like i was going offtopic way too much on here
<whitequark>
plus on ##whitequark i don't have a filter at all :p
* zkms
gives whitequark a root raised cosine
<whitequark>
um,
* ZipCPU
suggests a Root Nyquist filter instead of the root-raised cosine, it has lower sidelobes
jevinskie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<sorear>
having read ieee 1364-2008, ZipCPU's blog, and the yosys manual (in that order) i'm now thinking about what reasonable practices for organizing projects / package management are
<ZipCPU>
sorear: You mean like svn vs git? File layout/structure?
<sorear>
file layout, handling dependencies
<sorear>
build/test/verification scripts
<ZipCPU>
I try to follow a document that was once posted on OpenCores for that purpose
<ZipCPU>
I think I've modified it somewhat in practice, but ... I've only documented it for my patreon sponsor's
<sorear>
i'm guessing we're not quite at the npm level of "easy to integrate third-party stuff"