[Cbe-oss-dev] [RFC] Reading Local Store base address from SPU using stop-and-signal with kernel support

[D. Herrendoerfer]

Hi Cédric,

there is already an effort underway to make this happen. The
planned method, however is slightly different.
We plan to extend the SPE ABI to pass the Local Store base
address to the crt0. The startup code can then store it for 
later use.
It's good to see that others need this too.

Best regards,

D. Herrendoerfer

On Tue, 2007-05-22 at 23:56 +0200, Cédric Augonnet wrote:
> Hi all,
> 
> Sometimes, the SPU needs to know where its Local Store is mapped on
> main memory. The issue is that the user would have to somehow
> explicitly supply that information to the SPU as far as i know.
> Therefore a library needing that base address would hardly get that
> information.
> 
> I would have liked to have your opinion about using a stop-and-signal
> receiving some implicit support from the kernel in order to get it.
> 
> This is doable using the same approach as for system calls used from
> the SPUs : the argument supplied during the stop-and-signal call would
> be specific to that functionality (as it is already the case 0x2104
> for syscalls). The kernel would recognize such a specific case and
> handle it in consequence.
> 
> Here is a draft of the patch that would normally make it possible :
> http://perso.ens-lyon.fr/cedric.augonnet/Kernel/spu_get_ls.patch
> 
> Please note that this patch could *not* been tested yet since I don't
> have available hardware. But this shows the basic outlines of the
> proposition.
> 
> To make it clearer, I also wrote a sample of a SPU code intended to
> show how this functionnality could be used so that the SPU could get
> the LS base address transparently.
> http://perso.ens-lyon.fr/cedric.augonnet/Kernel/sample.c
> 
> So my questions are :
> 1/ Can such a stop-and-signal argument be reserved for so specific a
> purpose, as it was done with syscalls ?
> 
> 2/ Should we rather be able to retrieve the spu identifier using that
> same technique, and then make use of a couple of DMA transfer to
> explicitly fetch the base address from the PPU, provided the speid ?
> 
> I'm aware that providing such functionnalities through macros on the
> SPU might really hide the possibly huge cost of such a stop-and-signal
> call, but if the programmer is aware of such a cost, can't it make spu
> library simpler to write?
> 
> Perhaps this is silly proposition, I don't know, thus I just keep hearing
> for any comment about it.
> 
> Best regards,
> Cedric
> _______________________________________________
> cbe-oss-dev mailing list
> cbe-oss-dev at ozlabs.org
> https://ozlabs.org/mailman/listinfo/cbe-oss-dev