<Larhzu>
I've skimmed riscv-cfi (Zicfisslp) draft for fun. I wonder if I spotted a mistake but being an outsider to RISC-V development I'm not very confident.
<Larhzu>
After the support for software guarded branches was committed, the asm pseudo-instruction "call" seems to no longer need a landing pad target.
<Larhzu>
But after the next two commits it gets simplified. Sorry if I should have just made an issue on GH instead of asking here, I just feel I might be missing something stupid. :)
<Larhzu>
s/stupid/simple/
<bjdooks>
conchuod: ta, have re-issed patch with correct fixes id
<conchuod>
I was so confused looked at it at first, because the diff looks like a hunk from what you had in the fixes tag
<conchuod>
+When i
SpaceCoaster has quit [Quit: Bye]
gianluca has quit [Server closed connection]
gianluca has joined #riscv
aparri has joined #riscv
<Larhzu>
Oh well, I'll make an issue on GH. :-)
<palmer>
Larhzu: I haven't looked at the CFI stuff that much, but it wouldn't be surprising to find an issue in the spec as it's pretty new
<Larhzu>
palmer: Thanks for encouragement. :-)
<palmer>
;)
<palmer>
no promisises the spec people will admit there's a bug, but might as well try
jobol has quit [Quit: Leaving]
<conchuod>
lmao
<palmer>
ya, I guess it's kind of funny ;)
<conchuod>
funny and sad
<palmer>
ya, I'm not sure which is more ;)
<palmer>
I guess I kind of bounce back and forth on that one
<Larhzu>
;-) I'm just passing by, reading specs for fun. I guess being an outsider helps spotting certain things but at the same time I probably can make really stupid questions.
<palmer>
I always err on the side of "if someone's confused, we should improve it"
<palmer>
it's kind of like code reviews and comments and such: sure maybe the review isn't correct, but if someone got tripped up it's probably worth making it more clear
aburgess_ has quit [Ping timeout: 258 seconds]
<Larhzu>
:-) The specs I have read have been quite readable. So when something doesn't feel clear it stands out.
<palmer>
Larhzu: you must be smarter than me, I generally find the specs completely unreadable ;)
<Larhzu>
Hah, I have mostly read the easy ones: unprivileged, Zba/Zbb/Zbc/Zbs, and such.
Amanieu has quit [Server closed connection]
Amanieu has joined #riscv
ashtin has quit [Quit: Leaving]
<Larhzu>
From my biased viewpoint, Zba feels kind of mandatory on RV64, I have so much uint32_t used for array indexing.
<Larhzu>
I hadn't even thought about it before as on x86 it doesn't matter much. :-|
<palmer>
ya, there's loads of important stuff missing from rv64gc
MaxGanzII has quit [Ping timeout: 240 seconds]
Xark has joined #riscv
raym has joined #riscv
<dh`>
which one is Zba? i can't keep them straight
Tenkawa has joined #riscv
danilogondolfo has quit [Quit: Leaving]
jacklsw has joined #riscv
vagrantc has joined #riscv
jacklsw has quit [Quit: Back to the real world]
heat_ is now known as neat
neat is now known as heat
EchelonX has joined #riscv
raym has quit [Ping timeout: 260 seconds]
<muurkha>
one of the bit-manipulation ones I guess
<muurkha>
I guess rotation could be useful if you were implementing division in software? and a couple of the rest are like "this is rarely used but for some software it's the hot inner loop and it's a lot faster in hardware than in software so it's better to include it"
<muurkha>
and then you have things like andn, max, maxu, min, and minu which are very common but were omitted from RV64I in the pursuit of an almost Oberon-like minimality
<dzaima[m]>
that's Zbb? Zba is the one with sh1add, sh2add, sh2add.uw, etc
<muurkha>
that's Zba
<muurkha>
oh, no
<muurkha>
you're right, I managed to read the wrong section
<muurkha>
doh
<muurkha>
now that I'm looking at the actual Zba, Larhzu's comment about array indexing makes a lot more sense :)
<muurkha>
thanks for the correction
pecastro has quit [Ping timeout: 258 seconds]
Gravis has quit [Ping timeout: 240 seconds]
Ohhtwo_ashtin has quit [Quit: Leaving]
raym has joined #riscv
Gravis has joined #riscv
aburgess has joined #riscv
Gravis_ has joined #riscv
prabhakarlad has quit [Quit: Client closed]
Gravis has quit [Ping timeout: 260 seconds]
SpaceCoaster has joined #riscv
prabhakarlad has joined #riscv
ssb has quit [Ping timeout: 240 seconds]
aparri has left #riscv [#riscv]
ssb has joined #riscv
SpaceCoaster has quit [Quit: Bye]
SpaceCoaster has joined #riscv
MaxGanzII has joined #riscv
pecastro has joined #riscv
BootLayer has quit [Quit: Leaving]
Noisytoot has quit [Server closed connection]
Noisytoot has joined #riscv
SpaceCoaster has quit [Quit: Bye]
SpaceCoaster has joined #riscv
MaxGanzII has quit [Ping timeout: 240 seconds]
heat has quit [Remote host closed the connection]
heat has joined #riscv
<conchuod>
bjdooks: is that pl2 driver's info printing even helpful?
<conchuod>
Once per cpu, for something you can get out of sysfs..
<bjdooks>
Hmm. Not sure if it is totally useful. I've not checked but I think some of the info should be available from common sysfs.
<bjdooks>
You'd have to check as I'm closer to a beer glass than an computer
<conchuod>
I think other than the ways, it's all info you can get from sysfs
___nick___ has quit [Ping timeout: 265 seconds]
MaxGanzII has joined #riscv
prabhakarlad has quit [Quit: Client closed]
zBeeble42 has quit [Quit: Leaving]
jacklsw has joined #riscv
jacklsw has quit [Client Quit]
prabhakarlad has joined #riscv
<bjdooks>
It certainly doesn't news 6 lines. Now imagine something like horsecreek1 which has 4 of those
pjw_ has quit [Server closed connection]
pjw_ has joined #riscv
zjason has quit [Read error: Connection reset by peer]
zjason has joined #riscv
MaxGanzII has quit [Quit: Leaving]
EchelonX has quit [Quit: Leaving]
MaxGanzII has joined #riscv
MaxGanzII has quit [Remote host closed the connection]