Re: omapfb: help from userspace

[Tony Lindgren <tony@atomide.com>]

[-- Attachment #1: Type: text/plain, Size: 1409 bytes --]

* Nathan Monson <nmonson@gmail.com> [081016 15:15]:
> On Thu, Oct 16, 2008 at 3:10 PM, Tony Lindgren <tony@atomide.com> wrote:
> > No.. And I also got some -33 errors here too :( So back to the drawing
> > board.
> >
> > Could you post step by step instructions on how you can reproduce your
> > problem?
> 
> I'm running the DSP Bridge patches in my kernel.  I got them set up
> using this guide:
> http://elinux.org/BeagleBoard/DSP_Howto
> 
> That guide also points to some sample binaries, including 'ping.out',
> which is a simple round-trip test between the DSP and ARM.
> 
> To reproduce, I do this:
> 1) Boot kernel with mem=122m
> 2) modprobe bridgedriver phys_mempool_base=0x87a00000
> phys_mempool_size=0x600000
> base_img=/home/root/dsp/ddspbase_tiomap3430.dof64P
> 3) while true; do ./ping.out; done
> 
> This ALWAYS reproduces within seconds.
> 
> If the DSP patches are too much work to get set up, I've also heard
> that people can reproduce this issue by playing movie files with
> mplayer on Angstrom.

Thanks, I'll set that up when I get a chance.

Meanwhile, here's one more patch for you to test if you have a chance.
Maybe also do a power off for few seconds before booting with this
just in case.

If this still does not help, then we'll use the strongly ordered patch
for now, and tag v2.6.27-omap1 tomorrow. Then we can continue debugging
the issue until we fully understand it.

Tony

[-- Attachment #2: intc-write-readback.patch --]
[-- Type: text/x-diff, Size: 347 bytes --]

--- a/arch/arm/mach-omap2/irq.c
+++ b/arch/arm/mach-omap2/irq.c
@@ -53,6 +53,7 @@ static struct omap_irq_bank {
 static void intc_bank_write_reg(u32 val, struct omap_irq_bank *bank, u16 reg)
 {
 	__raw_writel(val, bank->base_reg + reg);
+	__raw_readl(bank->base_reg + reg);
 }
 
 static u32 intc_bank_read_reg(struct omap_irq_bank *bank, u16 reg)