karolherbst

karolherbst , vor 13 Stunden an Random Englisch

I'm currently looking into what's the best way to support SVM/USM in #rusticl and one thing I'm wondering about is, if there are any drawbacks doing:

mmap(some_chosen_address, ram_size, PROT_NONE, MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED_NOREPLACE, 0, 0);

and reserve a lot of virtual memory, and then suballocate SVM allocations out of this region with "PROT_READ | PROT_WRITE" and "MAP_FIXED"?

Alternatively, I was considering allocating smaller heaps on demand.

karolherbst OP , vor 13 Stunden

@jhwgh1968 nah, FIXED_NOREPLACE just means the mmap call won't replace existing mappings.

karolherbst OP , vor 13 Stunden

@jhwgh1968 mhhh.. I can't really sub-allocate, because after munmap the original mapping is just gone as well...

though maybe I can make it work without reserving such a huge range, but I'm kinda worried about fragmentation, but... maybe that's fine...

karolherbst OP , vor einer Stunde

@lina @jhwgh1968 yeah, it should.

Which just brings me back to the original question, if reserving that amount of VM space is a good idea and if there are other options.

I also wonder how well that would work with things like libasan, which just allocates 20TB of virtual memory here.

Anyway, it seems like Intel's CL stack just hopes nothing conflicts and maybe I just do the same, because it should be fine (tm).

karolherbst , vor 14 Stunden an Random Englisch

so.. which distribution is not going to patch out this secret chromium extension?

karolherbst OP , vor 13 Stunden

@jhwgh1968 apparently chromium has a private extension users can't disable, which grants *.google.com domains special privileges and access to private APIs.

Apparently the same extension is also available on edge.

See https://fedi.simonwillison.net/@simon/112757810519145581

karolherbst , vor 5 Tagen an Random Englisch

soooo.. in OpenCL C builtins like get_global_offset have to return defined values on out of bound accesses, but apparently Intel's driver also gets this wrong. Now I'm considering if I should keep it broken as well, or actually fix it as defined in the spec...

@bashbaug any thoughts on this? Maybe we should add OpenCL CTS tests for this and annoy everybody else with this?

karolherbst OP , vor einem Tag

@bashbaug we have some custom CL tests in piglit and this one seems to fail on my Intel one: https://gitlab.freedesktop.org/mesa/piglit/-/blob/main/tests/cl/program/execute/global-offset.cl?ref_type=heads#L97

It's the last "[test]" section in that file in case you want to adjust your local test program.

Here it returns "3" instead of "0" in the last line.

karolherbst OP , vor einem Tag

@bashbaug I think this issue should exist with all builtins, but I'm wondering why the SPIRV-Translator is the best place, because I was seeing this with Intel-compute-runtime using OpenCL C as well.

Not saying that it won't make sense to fix it in the translator, but I think it depends on what's meant with "The mapping from an OpenCL C built-in function to the SPIR-V BuiltIn is informational and non-normative." in the env spec.

karolherbst OP , vor einem Tag

@bashbaug oh, I wasn't aware that you are also using SPIR-V internally. Is that an old change or something more recent?

karolherbst , vor 2 Tagen an Random Englisch

what a relief to be honest

karolherbst OP , vor 2 Tagen

@jonkoops @karl I just hope that whatever government will be formed after that won't do the same mistakes as we e.g. see in Germany.

But it's going to be tough I bet.

18+ karolherbst , vor 4 Tagen an Random Englisch

the social democratic leadership sees austerity as the only option.

The same as 95 years ago.

It's really heartbreaking seeing political leaders doing the same mistakes over and over again.

karolherbst OP , vor 4 Tagen

@nCrazed I'm sure they do it deliberately in the meaning of "that's how the system works and that's what we have to do", but not deliberately in the sense of "haha, we want to ruin our system".

And people like the FDP finance minister just say, that they don't believe the studies claiming it does...

Maybe it's deliberate from the "liberals", but uhh.. I think they are just ideologically blinded.

karolherbst , vor 6 Tagen an Random Englisch

I think the funniest part of the new POSIX version is the removal of "ioctl"

karolherbst OP , vor 6 Tagen

@VegaHarmonia yeah, it was kinda weird having it in posix in the first place, because it was all implementation defined anyway. So they removed it, because POSIX really has no business there.

karolherbst OP , vor 6 Tagen

@VegaHarmonia it's mostly funny in regards to POSIX purists and them not wanting to use non POSIX stuff in software.

Granted, you almost never need to use ioctl directly in most cases, but still...

karolherbst , vor 6 Tagen an Random Englisch

I've finally finished the work to fix the spirv linker so it can link programs together originating from llvm-17 or newer.

Yes, this is a dirty workaround, but apparently we have no other options at this point.

https://github.com/KhronosGroup/SPIRV-Tools/pull/5534

karolherbst OP , vor 6 Tagen

@Man2Dev this PR only really matters if you link two spir-vs together, which I think besides mesa almost nobody is doing anyway, I think.

TheEvilSkeleton , vor 7 Tagen an Random Englisch

"Keep politics out of FOSS" is an interesting political statement

karolherbst , vor 7 Tagen

@TheEvilSkeleton what did I miss, or is it that new fancy web browser thing?

18+ karolherbst , vor 8 Tagen an Random Englisch

protect yourself from all openssh vulnerabilities with this little trick:

systemctl stop sshd; exit

18+ karolherbst OP , vor 8 Tagen

@brayd I've warned y'all

karolherbst , vor 8 Tagen an Random Englisch

anyway, some 🦀 content today:

I'm considering using serde in rusticl, however my use case is not to parse arbitrary data, but always structs with no custom data allowed. In short: the struct defines the data layout.

And we have some helpers in mesa dealing with reading/writing data to a blob.

So I'm wondering if a custom serde serialize/deserializer is what I want, or if I want something more special/focused as serde seems to be too generic for my use case here.

karolherbst OP , vor 8 Tagen

@matt it's C code and we use it for e.g. serializing GPU shaders so it can be stored in a shader cache. But it's all manual serialization and deserialization code, it's just mostly there to deal with alignment, padding and buffer overruns (like you can check if the buffer used for reading/writing wasn't big enough and handle it).

So, because writing the code yourself is annoying I was wondering if something can take a struct and generate the code for me in rust.