[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Who is maintaining the powerpc ep1a BSP ?



Till,

The ep1a is in use using a modified 4.6 tree.  I made the
appropriate modifications and rolled it into the 4.7 tree
and did a short test to verify that it was working at that
point.  Any mods made after that point have not been tested
on the board.

The board does use interrupts, but irq.c was modified and a
unique version of it is in the bsp.  The rsPMCQ1 does not have
a set interrupt, but must be read from the board at boot.
The base board may have multiple rsPMCQ1 boards attached to it
and unique interrupt numbers for each board is determined in
the rsPMCQ1 code.

I don't have access to the tree that is being used with the
ep1a at this point in time, so I can't do a comparison to
verify that something didn't get added to its unique 4.6 tree
after the merger into the 4.7 tree, and then didn't get merged
back into the trunk.

Hope this helps and thanks for the heads up on a possible problem.

Jennifer

> I'm trying to remove conditional compilation
> constructs from libbsp/powerpc/shared/openpic
> which still contains
>
> #if defined(mpc8240) || defined(mpc8245)
> ...
>
> I scanned for BSPs that use such a CPU and it
> seems that there are only mvme2100 and ep1a.
>
> However, ep1a (make/custom/ep1a.cfg) does
> currently *not* -Dmpc8240 (mvme2100 does)
> which means that the 8240-specific differences
> are not enabled on the ep1a BSP.
>
> In particular, it seems to me (from looking at the code)
> that the PCI interrupts (IRQs that use the openpic
> rather than the i8259 PIC) cannot work properly
> because the openpic register layout of the EPIC
> implementation (used by the mpc8240/mpc8245)
> is (slightly) different from the standard openpic.
> However, the modification to the layout invoked
> by the aforementioned cpp conditionals are not
> effective because the BSP does not -Dmpc8245.
>
> And yet, it seems that the ep1a BSP uses PCI
> interrupts (rsPMCQ1 driver). Has anyone used/tested
> that driver lately?
>
> -- Till
>
> PS (also to the ep1a maintainer). On the mvme2100
> we had issues with the EOI race-condition mentioned
> in the 8245 manual and there is a work-around that
> could be of interest for the ep1a, too. YMMV
>
> _______________________________________________
> rtems-users mailing list
> rtems-users at rtems.com
> http://rtems.rtems.org/mailman/listinfo/rtems-users
>