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
apritzel_ has quit [Ping timeout: 276 seconds]
tlwoerner_ has joined #armlinux
IgorPec_ has joined #armlinux
derRicha1d has joined #armlinux
geertu_ has joined #armlinux
bamse_ has joined #armlinux
tlwoerner has quit [*.net *.split]
Stary has quit [*.net *.split]
a3f has quit [*.net *.split]
Emantor has quit [*.net *.split]
IgorPec has quit [*.net *.split]
tom5760 has quit [*.net *.split]
tleb has quit [*.net *.split]
geertu has quit [*.net *.split]
vingu has quit [*.net *.split]
Stalevar has quit [*.net *.split]
derRichard has quit [*.net *.split]
mrutland has quit [*.net *.split]
bamse has quit [*.net *.split]
[florian] has quit [*.net *.split]
Stalevar has joined #armlinux
npcomp has quit [Ping timeout: 248 seconds]
jclsn has quit [Ping timeout: 272 seconds]
jclsn has joined #armlinux
tom5760 has joined #armlinux
a3f has joined #armlinux
Emantor has joined #armlinux
Stary has joined #armlinux
vingu has joined #armlinux
[florian] has joined #armlinux
mrutland has joined #armlinux
tleb has joined #armlinux
prabhakalad has quit [Ping timeout: 252 seconds]
prabhakalad has joined #armlinux
<wens>
found out that dma_alloc_coherent() will happily use an assigned reserved memory region, even if that region's address is beyond the device's dma address range (dma mask)
haritz has quit [Quit: ZNC 1.8.2+deb3.1+deb12u1 - https://znc.in]
monstr has joined #armlinux
monstr has quit [Ping timeout: 252 seconds]
apritzel_ has joined #armlinux
gclement has joined #armlinux
apritzel_ has quit [Ping timeout: 268 seconds]
oupton has joined #armlinux
System_Error has quit [Ping timeout: 264 seconds]
oupton has quit [Quit: WeeChat 3.8]
oupton has joined #armlinux
oupton has quit [Changing host]
oupton has joined #armlinux
headless has joined #armlinux
Poltawer has joined #armlinux
System_Error has joined #armlinux
krzk has quit [Quit: leaving]
krzk has joined #armlinux
<a3f>
wens, was the mask set by dma_set_coherent_mask?
<wens>
a3f: it's using the default, 32 bits
<wens>
whereas the reserved memory is allocated from the end of memory, which combined with the memory base address offset, is higher than 32 bits
<wens>
I was sort of expecting it could be an issue, but I think erroring out would be better than having the device blindly doing DMA to/from a truncated address
<wens>
in my case I was testing audio playback
headless has quit [Quit: Konversation terminated!]
sszy has joined #armlinux
<robmur01>
wens: if the DT specifies a carveout for a device with addressing limitations but doesn't provide a "reg" or "alloc-ranges" appropriate to those limitations then the DT is just wrong
<robmur01>
we don't generally go out of our way to try to defend against the DT being wrong
apritzel has joined #armlinux
Perflosopher038 has joined #armlinux
Perflosopher03 has quit [Ping timeout: 272 seconds]
Perflosopher038 is now known as Perflosopher03
<wens>
robmur01: in my case, it turns out the hardware actually does support >32 bits addressing, but the driver had not supported it
<wens>
in which case, I think an error or warning is valid?
<jn>
a warning, if it's easily detected, sounds reasonable
haritz has joined #armlinux
haritz has quit [Changing host]
haritz has joined #armlinux
d4ve has quit [Ping timeout: 248 seconds]
sfo has quit [Read error: Connection reset by peer]
d4ve has joined #armlinux
bryanb has quit [Write error: Connection reset by peer]
lane has quit [Write error: Connection reset by peer]
lane has joined #armlinux
bryanb has joined #armlinux
sfo has joined #armlinux
oupton has quit [Ping timeout: 265 seconds]
frieder has joined #armlinux
<robmur01>
perhaps, although I bet then we find out how many drivers are in the opposite situation of depending on carveouts working despite not bothering to set a proper DMA mask :)
bjoto has quit [Remote host closed the connection]