[PATCH 1/2] PRUSS UIO driver support

[TK, Pratheesh Gangadhar]

> -----Original Message-----
> From: Hans J. Koch [mailto:hjk at hansjkoch.de]
> Sent: Friday, February 18, 2011 10:02 PM
> To: TK, Pratheesh Gangadhar
> Cc: davinci-linux-open-source at linux.davincidsp.com; hjk at hansjkoch.de;
> gregkh at suse.de; Chatterjee, Amit; linux-kernel at vger.kernel.org; linux-arm-
> kernel at lists.infradead.org
> Subject: Re: [PATCH 1/2] PRUSS UIO driver support
> 
> On Fri, Feb 18, 2011 at 08:35:29PM +0530, Pratheesh Gangadhar wrote:
> > Signed-off-by: Pratheesh Gangadhar <pratheesh at ti.com>
> >
> > This patch implements PRUSS (Programmable Real-time Unit Sub System)
> > UIO driver which exports SOC resources associated with PRUSS like
> > I/O, memories and IRQs to user space. PRUSS is dual 32-bit RISC
> > processors which is efficient in performing embedded tasks that
> > require manipulation of packed memory mapped data structures and
> > efficient in handling system events that have tight real time
> > constraints. This driver is currently supported on Texas Instruments
> > DA850, AM18xx and OMAPL1-38 devices.
> > For example, PRUSS runs firmware for real-time critical industrial
> > communication data link layer and communicates with application stack
> > running in user space via shared memory and IRQs.
> 
> I see a few issues, comments below.
> 
> Thanks,
> Hans
> 
> > ---
> >  drivers/uio/Kconfig     |   10 ++
> >  drivers/uio/Makefile    |    1 +
> >  drivers/uio/uio_pruss.c |  250
> +++++++++++++++++++++++++++++++++++++++++++++++
> >  3 files changed, 261 insertions(+), 0 deletions(-)
> >  create mode 100644 drivers/uio/uio_pruss.c
> >
> > diff --git a/drivers/uio/Kconfig b/drivers/uio/Kconfig
> > index bb44079..631ffe3 100644
> > --- a/drivers/uio/Kconfig
> > +++ b/drivers/uio/Kconfig
> > @@ -94,4 +94,14 @@ config UIO_NETX
> >  	  To compile this driver as a module, choose M here; the module
> >  	  will be called uio_netx.
> >
> > +config UIO_PRUSS
> > +	tristate "Texas Instruments PRUSS driver"
> > +	depends on ARCH_DAVINCI_DA850
> > +	default n
> 
> That line is unneccessary, "n" is already the default.
> 
Ok, will fix in the next version.
> > +	help
> > +	  PRUSS driver for OMAPL138/DA850/AM18XX devices
> > +	  PRUSS driver requires user space components
> > +	  To compile this driver as a module, choose M here: the module
> > +	  will be called uio_pruss.
> > +
> >  endif
> > +#define MAX_PRUSS_EVTOUT_INSTANCE	(8)
> 
> The brackets are not needed.
> 
Will fix in the next version.
> > +
> > +static struct clk *pruss_clk;
> > +static struct uio_info *info[MAX_PRUSS_EVTOUT_INSTANCE];
> 
> is it really neccessary to allocate that statically?
> 
> > +static void *ddr_virt_addr;
> > +static dma_addr_t ddr_phy_addr;
> > +
Agree - not necessary - will fix.

> > +static irqreturn_t pruss_handler(int irq, struct uio_info *dev_info)
> > +{
> > +	return IRQ_HANDLED;
> > +}
> 
> ROTFL. That reminds me of an old story. The last time I wrote this, and
> Greg dared to post it, we received this reply:
> 
> http://marc.info/?l=linux-kernel&m=116604101232144&w=2
> 
> So, if you really have a _very_ good reason why this _always_ works on
> _any_ DA850 board, add a comment that explains why. Otherwise the whole
> patch set will be doomed.
> 
It always worked for me during the tests on the h/w. So did not bother to dig into the details then. From AM1808/AM1810 ARM Microprocessor System Reference Guide (http://focus.ti.com/general/docs/lit/getliterature.tsp?literatureNumber=sprugm9a&fileType=pdf), section 11.3.1 Interrupt processing
The interrupt processing block does the following tasks:
	-Synchronization of slower and asynchronous interrupts
	- Conversion of polarity to active high
	- Conversion of interrupt type to pulse interrupts
After the processing block, all interrupts will be active-high pulses

Interrupt processing is the first step in INTC h/w which maps system interrupts to ARM (host) interrupts (FIQ, IRQ). 

However I am willing to clean this up to meet the kernel guidelines and good practices... 
> > +
> > +static int __devinit pruss_probe(struct platform_device *dev)
> > +{
> > +	int ret = -ENODEV;
> > +	int count = 0;
> > +	struct resource *regs_pruram, *regs_l3ram, *regs_ddr;
> > +	char *string;
> > +	for (count = 0; count < MAX_PRUSS_EVTOUT_INSTANCE; count++) {
> > +		info[count]->mem[0].addr = regs_pruram->start;
> > +		info[count]->mem[0].size =
> > +		    regs_pruram->end - regs_pruram->start + 1;
> > +		if (!info[count]->mem[0].addr
> > +		    || !(info[count]->mem[0].size - 1)) {
> 
> That size check looks fishy. If somebody forgot to set the size it's OK ?
> 
size is set just previous line, right? If regs_prum_ram->end== regs_prum_ram->start - then there is something wrong... 

> > +			dev_err(&dev->dev, "Invalid memory resource\n");
> > +			break;
> > +		}
> > +		info[count]->mem[0].internal_addr =
> > +		    ioremap(regs_pruram->start, info[count]->mem[0].size);
> > +		if (!info[count]->mem[0].internal_addr) {
> > +			dev_err(&dev->dev,
> > +				"Can't remap memory address range\n");
> > +			break;
> > +		}
> > +		/* Register PRUSS IRQ lines */
> > +		info[count]->irq = IRQ_DA8XX_EVTOUT0 + count;
> > +
> > +		info[count]->irq_flags = IRQF_SHARED;
> 
> How do you handle shared interrupts with the handler above?
> 
> > +		info[count]->handler = pruss_handler;
> 
> And how do you make sure your interrupts are not level triggered? The
> handler above will hang for level triggered interrupts.

As mentioned above interrupt controller (ARM INTC) converts interrupt type to pulse. After required processing is complete - user space handler clears the interrupt from PRUSS.

Thanks,
Pratheesh