big awoo notationI want to see if the thoughts of others align with mine.
An FPGA toolchain should optimise for:
An FPGA toolchain should optimise for:
highest worst-case Fmax
highest average Fmax
highest best-case Fmax
Poll ended at .
Luke Wren@lofty Depends how many parallel jobs I am allowed to submit to get that golden seed 😅
@wren6991 I'm computing distributions of 100 runs, so
@lofty If I'm comparing RTL changes for fmax then I'll usually do ~100 runs and take the median. Seems reasonable because 8 runs aiming for median frequency should have only a 0.4% chance of all failing. Most client machines have 8 parallel threads of execution available these days.
I don't particularly care about the mean because if I get a worst-case result then you can bet I'm re-rolling, so the tail values don't matter.
I don't think it's *good* that seed sweeping is the optimal way of using the tools, but the fact is it reduces variance as well as increasing expected fmax.
@lofty Also bear in mind my comparison point here is Vivado which has a >> 0.4% chance of just segfaulting for no apparent reason, so I find that kind of probability acceptable
lewurm
mei
Andrew Zonenberg
Aaron Sawdey, Ph.D.@azonenberg @mei @lofty @wren6991 It’s horrifying but only because they’re having to do this to released software. Otherwise, yeah I have done stuff like this, if rarely this elegantly. This same thing would be a tool for finding these errors if you add code to fill freed blocks with provocatively bad data. Dynamic linking games are always fun.
@acsawdey @azonenberg @mei @lofty > This same thing would be a tool for finding these errors if you add code to fill freed blocks with provocatively bad data
Isn't this what asan does with quarantining?
@wren6991 @azonenberg @mei @lofty yeah there are lots of things that do this, sometimes you have to roll your own .. like for instance if you can’t rebuild with asan.
I think it’s axiomatic that useful programming tools get reimplemented all the time.
@acsawdey @wren6991 @mei @lofty yeah it's more the horror of "let's just double all allocations and add a fudge factor in the expectation that the thing is full of buffer overflows" and "delay all frees in the expectation that it's full of UaFs".
Vivado leaks memory bad enough as is. I'd want another 128GB of RAM if I was using this...
@azonenberg @acsawdey @mei @lofty It's fine, it's just virtual memory that's being allocated. The padding has low cost if there are no overruns :) :) :)
ShutterBugged@wren6991 @azonenberg @acsawdey @mei @lofty I have in fact fed 128GB chunk after 128GB chunk of swap to a tool that ate all 64GB of ram and then lots more. (I soon had to switch to 256GB chunks) I don't remember how many I carved out of my SSD in the end, but I know it was only a 1TB SSD and I don't think I pulled any from spinning rust. It did eventually finish though. :P
And this wasn't even a Vendor tool, it was "research grade" tooling written in python.
Tim 🛥️🌈@azonenberg @mei @lofty @wren6991 I haven't seen this horrifying bodge in a long time, because it was deployed to catch a naughty Windows device driver which wrote back to user space memory after the driver handle was closed. The driver was eventually fixed.
the vessel of morganna@azonenberg @mei @lofty @wren6991 this sort of software bodge is literally built into the Windows compatibility shim and applied for a metric crapload of programs. we certainly live in times
@azonenberg @mei @lofty @wren6991 yep, oversized allocations and time-delayed free. the name "enterprise-malloc" is so incredibly accurate it hurts. the windows compat shim often gets tuned to limit the over-allocation size since it's a per-program profile, so it's not as noticeable
@astraleureka @mei @lofty @wren6991 the only thing worse than this shitpost is it not being a shitpost
F4GRX Sébastien
Daniel Carosone
kasia 
Wolf480pl@wolf480pl quite likely.
@nullenvk @lofty
when I was taking the Adanced Operating Systems class by @mwk, she once told a story how ISE had use-after-free bugs and in order to make it work, she had to use an LD_PRELOAD library that remembered what pointers ISE tried to free, and freed them a few calls later.
Which is part of what the code @mei posted does
Elias Mårtenson
zSchön 
@TheAlgorythm @mei C guarantees implicit zero initialisation of variables with static storage duration, though it's bad style to rely on it because the same does not apply to automatic storage duration. So the free of empty queue is avoided by the null check, and the position is initialised to zero.
AMS@TheAlgorythm @mei They're static so 0 on first call. Calling free on NULL is defined as no-op.
zSchön :disability_flag: (@[email protected])
@[email protected] @[email protected] Oh, ok. Didn’t know the exceptions of static.
Peter H. Fröhlich
Cassandrich
R
Steve Purcell
Chris L@mei @wren6991 @lofty impressive tho I fear for the quality of the code that this makes more reliable. LD_PRELOAD is a great power when you need it. I used it once to patch up code looking for 32 bit inodes on a system with 64 bit inodes, and to wrap ETL pipelines on one giant NFS so I could run them in a read only mode to build the run time call graph o allow splitting the giant NFS into one container per scheduled job.
Poul-Henning Kamp
Matteꙮ Italia@mei @lofty @wren6991 on one end I'm like "niiiice, but well, this can be improved, the queue can be made lock free, and we can drop the mutex from the initialization of the forwarded symbols through careful use of atomics; also, we may improve resilience by adjusting the return value to have some extra slack even at the start"; on the other hand I'm like NO THIS IS BAD LET IT BE SLOW
@cvtsi2sd serialising all the calls to free might also be helping with race conditions
Michał Trojnara 
Mae
zev@tubetime @wren6991 @lofty @mei Alternately (or perhaps as a second level of bumper bowling), use
__attribute__((constructor)) to install a SIGSEGV handler to just quietly "make the problem go away" a la https://people.csail.mit.edu/rinard/paper/osdi04.pdf (might also need to shim signal(2) and sigaction(2) to make sure it stays in place in case the Serious Enterprise Software tries to install its own crash handlers).
Rachel Barker@mei Disgusting, I love it!
Resuna
cliffordheath@cliffordheath @mei @lofty @wren6991
Yeah that's what I was thinking. Just loop over the area and free everything that isn't null.
I was just wondering if there was some interaction between valgrind and LD_PRELOAD that obviated it.
Gus
clang-enbyI like how this is called "enterprise malloc"
gudenau
endrift 🏳️⚧️
Vincent Sparks 🔜 Furlingame
Irenes (many)
bigblen
PsySal
Urix Turing@mei this is wild 😂
Bruno Levy