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
dormito has quit [Quit: WeeChat 3.6]
apritzel has quit [Ping timeout: 252 seconds]
Guest57 has joined #armlinux
heat has quit [Ping timeout: 248 seconds]
sally has quit [Quit: sally]
sally_ has joined #armlinux
sally_ is now known as sally
jclsn has quit [Ping timeout: 272 seconds]
Guest57 has quit [Ping timeout: 240 seconds]
jclsn has joined #armlinux
Peng_Fan has joined #armlinux
rvalue- has joined #armlinux
rvalue has quit [Ping timeout: 246 seconds]
rvalue- is now known as rvalue
monstr has joined #armlinux
apritzel has joined #armlinux
headless has joined #armlinux
cbeznea_ has joined #armlinux
apritzel has quit [Ping timeout: 244 seconds]
heat has joined #armlinux
dormito has joined #armlinux
prabhakalad has quit [Quit: Konversation terminated!]
prabhakalad has joined #armlinux
headless has quit [Quit: Konversation terminated!]
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
apritzel has joined #armlinux
apritzel has quit [Ping timeout: 248 seconds]
apritzel_ has joined #armlinux
sudeepholla_ has quit [Ping timeout: 252 seconds]
sszy has joined #armlinux
heat_ has joined #armlinux
heat has quit [Read error: Connection reset by peer]
apritzel has joined #armlinux
apritzel has quit [Ping timeout: 246 seconds]
luispm has quit [Ping timeout: 252 seconds]
dmart has joined #armlinux
luispm has joined #armlinux
<ajb-linaro>
should kernel_gpu_begin() be in the same headers for all arches for portability?
* ajb-linaro
can't build i915 for arm64 because it uses #include <asm/fpu/api.h> instead of include/asm/fpu.h for most other arches
<ajb-linaro>
kernel_fpu_begin!
<ajb-linaro>
although I guess the inline asm will be more of a challenge!
<broonie>
I was going to say...
<ajb-linaro>
broonie: so in theory an arm64 kernel under emulation can access a host i915 using virtio-gpu with native-context
<ajb-linaro>
although that might be an overly optimistic theory - its the graphics I have on my main x86 dev box
<broonie>
I was more thinking about what'd be inside the begin/end block than why you'd want to do it.
<ardb>
it's an interesting idea but i'm skeptical it will be worth the hassle
siak has joined #armlinux
System_Error has quit [Remote host closed the connection]
<ajb-linaro>
broonie: it seems to be assuming it can do an accelerated memcpy although it doesn't seem to fallback so I guess it expects all chips that support i915 have the movntdqa instructions
<broonie>
Given how those things are shipped that does seem like something they could do.
<ajb-linaro>
ardb: I thought it would at least be worth trying... the x86_64 guest kernel keeps crashing with ""error: kvm run failed Bad address" when I login to the WM. So far I'm failing at getting a decent backtrace via gdbstub
<ardb>
ajb-linaro: so the host is arm64 and the guest is x86_64?? so where is the i915 device?
<ajb-linaro>
ardb: no host is x86_64 - guest is x86_64 under kvm or arm64 under TCG
<ajb-linaro>
virtio-vulkan works quite nicely cross arch - I get a higher FPS running vkmark under an emulated neoverse system than I do on my AVA + Radeon...
<ajb-linaro>
although that probably says more about the broken PCI on AVA and the cost of the fixups
<ardb>
most definitely
<ajb-linaro>
but still smooth GUI animation under emulation is pretty sweet
cbeznea_ has quit [Ping timeout: 245 seconds]
System_Error has joined #armlinux
monstr has quit [Remote host closed the connection]