ChanServ changed the topic of #armlinux to: ARM kernel talk [Upstream kernel, find your vendor forums for questions about their kernels] | https://libera.irclog.whitequark.org/armlinux
mraynal has quit [Remote host closed the connection]
mraynal has joined #armlinux
apritzel has quit [Ping timeout: 248 seconds]
SallyAhaj_ has joined #armlinux
SallyAhaj has quit [Ping timeout: 268 seconds]
amitk has joined #armlinux
luispm has joined #armlinux
luispm has quit [Ping timeout: 256 seconds]
luispm has joined #armlinux
macromorgan_ has joined #armlinux
macromorgan_ is now known as macromorgan
macromorgan is now known as Guest4316
Guest4316 has quit [Ping timeout: 255 seconds]
apritzel has joined #armlinux
frieder has joined #armlinux
Pali has joined #armlinux
apritzel has quit [Ping timeout: 256 seconds]
geertu has quit [Ping timeout: 255 seconds]
geertu has joined #armlinux
hanetzer has joined #armlinux
Pali has quit [Ping timeout: 256 seconds]
headless has joined #armlinux
alexels has joined #armlinux
prabhakarlad has joined #armlinux
dok has joined #armlinux
tre has joined #armlinux
SallyAhaj_ has quit [Remote host closed the connection]
SallyAhaj_ has joined #armlinux
SallyAhaj_ has quit [Remote host closed the connection]
apritzel__ has joined #armlinux
linusw_ has joined #armlinux
SallyAhaj has joined #armlinux
SallyAhaj has quit [Remote host closed the connection]
SallyAhaj has joined #armlinux
Misotauros has quit [Ping timeout: 268 seconds]
headless has quit [Quit: Konversation terminated!]
Misotauros has joined #armlinux
sudeepholla_ has quit [Ping timeout: 256 seconds]
sudeepholla_ has joined #armlinux
headless has joined #armlinux
sudeepholla_ has quit [Ping timeout: 252 seconds]
sudeepholla_ has joined #armlinux
headless has quit [Quit: Konversation terminated!]
<ardb>
sudeepholla: any clue why -rc1 on seattle gives me:
<ardb>
CPU features: No Cache Writeback Granule information, assuming 128
<ardb>
hmm actually i see the same in a vm on tx2
<broonie>
ardb: Did I mess that up with the conversion of CTR_EL0 to automatic generation?
<ardb>
perhaps - let me check
<broonie>
Ah, yeah. CTR_EL0_MASK was defined as 15 which is the *unshifted* value but our standard masks are the shifted values.
<broonie>
Same for all the other fields in CTR. Sorry about that :/
<broonie>
CTR_EL0_CWG_MASK rather
<ardb>
ah ok
<j`ey>
yeah, just noticed that. x >> y & z instead of x & z >> y
<broonie>
This sort of stuff is, of course, one of the reasons why generating this stuff is helpful :/
<ardb>
indeed
<ardb>
but grep is no longer my friend :-(
<ardb>
where is this stuff now?
<broonie>
The generated defines end up in arch/arm64/include/generated/asm/sysreg-defs.h
<broonie>
but won't show up in git grep because they're generated.
<ardb>
arch/arm64/tools/sysreg ?
<broonie>
Yeah, that's the source file.
<broonie>
Still slowly grinding my way through the registers (and trying to push anyone who touches sysreg.h to do any registers they touch).
cambrian_invader has quit [Ping timeout: 268 seconds]
<ardb>
what file format is that?
<ardb>
i'd imagine you can generate this from the ARM ARM xml
<broonie>
see gen-sysreg.awk, null defined this originally.
<ardb>
(and make the by-lines top 10 on lwn next cycle)
<broonie>
That is indeed the dream, at least as a starting point - we're going to want to diverge from the Arm in some places - there's some registers that are *weirdly* defined, and a lot of the enums don't have readily usable names.
<broonie>
but I want something that'll get you 90% of the way there automatically. The script and format were originally written by null.
<ardb>
yeah stuff like this needs to be machine generated, clearly
tre has quit [Remote host closed the connection]
<broonie>
It should be better going forward, most of the things that are problems are older registers. We probably will always want some manual review/override stage though.