dgilmore changed the topic of #fedora-riscv to: Fedora on RISC-V https://fedoraproject.org/wiki/Architectures/RISC-V || Logs: https://libera.irclog.whitequark.org/fedora-riscv || Alt Arch discussions are welcome in #fedora-alt-arches
esv__ has joined #fedora-riscv
davidlt has joined #fedora-riscv
zsun has joined #fedora-riscv
jcajka has joined #fedora-riscv
zsun has quit [Quit: Leaving.]
esv__ is now known as esv
<rwmjones_> davidlt: ^
<davidlt> rwmjones_, seems OK on a very fast glance
<davidlt> You could mention that C and some Zc* are also mandated by the profiles
<davidlt> Also objdump and "-M no-aliases" might be useful as that will prefix the compressed instructions if you want to "grep" them.
<rwmjones_> ah, good idea
<rwmjones_> clang/llvm configuration turns out to be much more difficult than I anticipated ... afaict it's not possible at all(?) to set the fine-grained arch string
<rwmjones_> if there are other languages I've forgotten about let me know
<davidlt> I think that could be the case for Rust too.
<davidlt> You can disable C on OpenJDK very easily IIRC. They have tons of options (again IIRC)
<davidlt> I think Zcb is mandatory in RVA23
<rwmjones_> oh yes java ..
* rwmjones_ does ocaml first
<rwmjones_> btw I'm using a new chat client (weechat) which means I may be able to add matrix support soon
<sorear> why isn't looking at e_flags an option for executables, instead of disassembling and grepping?
<sorear> rocket-chip supports disabling C at runtime if built with misaWritable; has anyone tested this on shipped hardware?
<rwmjones_> it's not reliable - I wrote a bit about it in that doco
<rwmjones_> hmm, interesting about rocket-chip, I wonder if the rocket-chip derived hardware like HiFive Unleashed suports that
<rwmjones_> I will test it at some point, but requires a bit of kernel hackery
<sorear> go may not support rvc at all; it was the last commit in my branch and never got merged
<davidlt> I keep forgetting why we should not trust elf flags
<davidlt> I haven't used weechat for quite some time
<rwmjones_> the elf RVC flag isn't set to anything useful by gas
<rwmjones_> it basically sets it to true if compressed instructions were ever possible at any stage
<rwmjones_> for example, if you use gas -march64=rv64gc (ie. compressed instructions possible)
<rwmjones_> but the input file has .option norvc
<rwmjones_> then the flag is still set
<rwmjones_> ideally it would only set the flag if any compressed instruction had been used, but it doesn't work like that at all right now
<rwmjones_> (I guess also it may need to decide on the flag for the header long before it gets round to emitted code)
sajcho has joined #fedora-riscv
sajcho has quit [Client Quit]
sajcho has joined #fedora-riscv
sajcho has quit [Client Quit]
iooi has quit [Quit: iooi]
sajcho has joined #fedora-riscv
sajcho has quit [Client Quit]
sajcho has joined #fedora-riscv
sajcho has quit [Client Quit]
jcajka has quit [Remote host closed the connection]
davidlt has quit [Ping timeout: 240 seconds]
fuwei has quit [Ping timeout: 264 seconds]
fuwei has joined #fedora-riscv