Yokemate of Keyboards
Posts: 12085 from 2003/5/22
From: Germany
> Our original plans to support 460 got massively delayed already and even there is quite
> some low level work still ahead. With all the additional issues showing up in public [...],
> it's not even sure if that will ever see the light of day anytime soon
I stumbled upon an interesting statement:
"
I found a guy from IBM which is a kind of ppc guru. He told me that the main cause of problems is the cpu itself (AMCC 460ex), as it lacks cache coherency."
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=37083&forum=33&start=20#718922
And I have found what reads like the original statements from the IBM guy:
"
for the 460EX, I am not surprised things aren't working with KMS & DRI2... the 460 is not cache coherent"
http://lists.freedesktop.org/archives/dri-devel/2012-February/019157.html
"
You will run into additional problems with 460 due to the fact that it's not cache coherent for DMA."
http://lists.ozlabs.org/pipermail/linuxppc-dev/2012-March/096888.html
"
DMA on 46x isn't cache coherent. The DRM plays interesting games with mapping/unmapping pages for DMA by the chip and I don't think we have the right hooks to do the appropriate cache flushing on these guys"
http://lists.ozlabs.org/pipermail/linuxppc-dev/2012-March/096925.html
Is this contributing to the porting difficulty? I remember that
Piru once said that MorphOS required cache coherency which he presented as reason that MorphOS couldn't be ported to MPC512x. If a MorphOS port to PPC460EX lacking cache coherency is possible, would that mean that a port to MPC512x is possible as well?