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
macromorgan has joined #armlinux
macromorgan has quit [Read error: Connection reset by peer]
macromorgan has joined #armlinux
mraynal has quit [Remote host closed the connection]
mraynal has joined #armlinux
apritzel has quit [Ping timeout: 276 seconds]
jclsn has quit [Ping timeout: 264 seconds]
jclsn has joined #armlinux
heat has quit [Ping timeout: 246 seconds]
mvaittin has joined #armlinux
cbeznea has joined #armlinux
amitk has joined #armlinux
hanetzer has quit [Ping timeout: 246 seconds]
hanetzer has joined #armlinux
luispm has joined #armlinux
monstr has joined #armlinux
luispm has quit [Ping timeout: 255 seconds]
luispm has joined #armlinux
sally has quit [Ping timeout: 260 seconds]
apritzel has joined #armlinux
apritzel has quit [Ping timeout: 272 seconds]
sally has joined #armlinux
gclement has joined #armlinux
<ukleinek>
geertu: if it only happens if eth was used in U-Boot it's something that isn't properly reset?!
<geertu>
ukleinek: I don't know if it only happens in that case. I always use TFTP for testing kernels.
prabhakalad has quit [Quit: Konversation terminated!]
headless has joined #armlinux
prabhakalad has joined #armlinux
luispm has quit [Ping timeout: 260 seconds]
sszy has joined #armlinux
luispm has joined #armlinux
heat has joined #armlinux
nsaenz has joined #armlinux
HerbY_NL_ has joined #armlinux
rgallaispou has joined #armlinux
HerbY_NL_ has quit [Client Quit]
nsaenz has quit [Remote host closed the connection]
<mvaittin>
My blindshooting guess would be that something hasn't finished reset/init. Unfortunatelty, guesses won't provide much value :) Maybe ping Tony? Perhaps he has an idea of what could've changed in v6.6? (more blindshooting)
<mvaittin>
geertu: ukleinek: --^
nsaenz has joined #armlinux
nsaenz has quit [Ping timeout: 252 seconds]
nsaenz has joined #armlinux
apritzel_ has joined #armlinux
HerbY_NL has joined #armlinux
nsaenz has quit [Remote host closed the connection]
headless has quit [Quit: Konversation terminated!]
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
lain6141_ has quit [Read error: Connection reset by peer]
lain6141_ has joined #armlinux
nsaenz has joined #armlinux
HerbY_NL has joined #armlinux
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
nsaenz has quit [Remote host closed the connection]
nsaenz has joined #armlinux
nsaenz has quit [Read error: Connection reset by peer]
dmart has joined #armlinux
HerbY_NL has joined #armlinux
nsaenz has joined #armlinux
nsaenz has quit [Ping timeout: 252 seconds]
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jelly has joined #armlinux
HerbY_NL has joined #armlinux
luispm has quit [Remote host closed the connection]
nsaenz has joined #armlinux
Amit_T has joined #armlinux
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
HerbY_NL has joined #armlinux
HerbY_NL has quit [Client Quit]
nsaenz has quit [Remote host closed the connection]
HerbY_NL has joined #armlinux
nsaenz has joined #armlinux
luispm has joined #armlinux
monstr has quit [Quit: Leaving]
monstr has joined #armlinux
HerbY_NL has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
nsaenz has quit [Remote host closed the connection]
nsaenz has joined #armlinux
nsaenz has quit [Remote host closed the connection]
nsaenz has joined #armlinux
nsaenz has quit [Remote host closed the connection]
nsaenz has joined #armlinux
headless has joined #armlinux
jbowen has quit [Ping timeout: 272 seconds]
jbowen has joined #armlinux
<ajb-linaro>
are aarch64 bigendian system calls the same numbers and reg layout as their little endian counterparts?
* ajb-linaro
assumes they are because that would be the sane way
<bjdooks>
is it not like the arm32 world where it just changes the load-store endian and the registers are always LE?
lain6141_ has quit [Read error: Connection reset by peer]
lain6141_ has joined #armlinux
nsaenz has quit [Remote host closed the connection]
nsaenz has joined #armlinux
nsaenz has quit [Remote host closed the connection]
<dmart>
ajb-linaro: Yes, BE/LE only changes the way the regs get stored to / loaded from memory.
<dmart>
ajb-linaro: Marshaling of __int128 *might* change for BE, but I don't think we have any of those in the syscall interface.
<dmart>
Encoding of extension regs (like SVE, SME etc.) in ptrace and the sigframe has some weird cases though.
macromorgan has quit [Read error: Connection reset by peer]
macromorgan has joined #armlinux
<ajb-linaro>
dmart: for QEMU we only care* about aarch64_be-linux-user - the system emulator relies on the guest switching modes
* ajb-linaro
notes * is relative, the fact the first ever aarch64_be bug was raised and it was asserts() on start-up indicates its not widely used.
<dmart>
ajb-linaro: switching modes> how do you mean?
nsaenz has joined #armlinux
nsaenz has quit [Remote host closed the connection]
<ajb-linaro>
arm_cpu_data_is_big_endian_a64() is driven by sctlr & (el ? SCTLR_EE : SCTLR_E0E)
<dmart>
ajb-linaro: Oh, so you mean there no distinct system emulator, but the guest can fun BE if it wants -- and the hardware remains just the same.
<ajb-linaro>
yes
<ajb-linaro>
which is nice, c.f. the myriad ppc binaries we build
gclement has quit [Ping timeout: 252 seconds]
<dmart>
ajb-linaro: I guess syscalls in general shouldn't be a huge problem then, what what about stuff that uses memory buffers like ioctl, perf and all that jazz?
<ajb-linaro>
dmart generally linux-user doesn't do device specific ioctls and translated perf is not supported (we don't emulate full PMUs in system mode either). My focus is just a simple hello world smoke test to catch any obvious breakage bobos
<arnd>
ajb-linaro: I think the preadv/pwritev/preadv2/preadv2 syscalls may hit an unexpected corner case on big-endian because they pass a 64-bit offset in a pair of 32-bit registers, with the lower half coming first
<arnd>
in general, system calls that split 64-bit arguments put the upper half first on big-endian ABIs
<arnd>
most syscalls pass 64-bit integers in a single register on a 64-bit kernel though
<arnd>
ajb-linaro: another issue you can hit a broken userspace that works by accident on little-endian
<arnd>
passing a little-endian 64-bit time_t into an interface that expects a little-endian 32-bit value does the right thing by accident, but breaks on big-endian
<arnd>
this mismatch is again much more common on 32-bit targets, e.g. a lot of accidental futex() calls that should have been futex_time64
<j`ey>
HdkR: off topic, where can I ask about fex?
<HdkR>
j`ey: We have an official Discord server at https://discord.gg/fexemu and technically an IRC channel #fex-emu@OFTC
<HdkR>
IRC channel is basically flatlined since the Discord server opened