[ECOS] Subject: Re: [ECOS] bug in mips arch code ????

[ECOS] Subject: Re: [ECOS] bug in mips arch code ????

[HG]

>The i8255x devices don't always obey the datasheets themselves. The
>i8255x driver has been worked on a lot to get it to work reliably.

on my platform the driver works very reliably for redboot. but when I run
applications like tftp_get it works reliably as long as the timer interrupt
does
not happen simultaneously to the 82559 interrupt.  (see my post of
yesterday)
any hints of what to look for???


thanks

Henri




-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] Subject: Re: [ECOS] bug in mips arch code ????

[Nick Garnett]

"HG" <henri@broadbandnetdevices.com> writes:

> >The i8255x devices don't always obey the datasheets themselves. The
> >i8255x driver has been worked on a lot to get it to work reliably.
> 
> on my platform the driver works very reliably for redboot. but when I run
> applications like tftp_get it works reliably as long as the timer interrupt
> does
> not happen simultaneously to the 82559 interrupt.  (see my post of
> yesterday)
> any hints of what to look for???
> 
> 
> thanks
> 
> Henri
> 

You need to find out exactly what happens when you get simultaneous
interrupts. It is difficult to tell at present whether this is a
hardware or a software problem.


-- 
Nick Garnett                    eCos Kernel Architect
http://www.ecoscentric.com/     The eCos and RedBoot experts


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: Re: [ECOS] bug in mips arch code ????

[HG]

> > >The i8255x devices don't always obey the datasheets themselves. The
> > >i8255x driver has been worked on a lot to get it to work reliably.
> >
> > on my platform the driver works very reliably for redboot. but when I
run
> > applications like tftp_get it works reliably as long as the timer
interrupt
> > does
> > not happen simultaneously to the 82559 interrupt.  (see my post of
> > yesterday)
> > any hints of what to look for???

>
> You need to find out exactly what happens when you get simultaneous
> interrupts. It is difficult to tell at present whether this is a
> hardware or a software problem.
>
Whatever is the process that acknowledges the interrupt (its not
HAL_INTERRUPT_ACKNOWLEDGE )
stops working (intreq signal stays low)
small file transfers are ok but long ones have more chances for this
event to happen . Apps that use the timer tick by themselves (no irq
produced by networking) are ok


thanks

Henri
> --
>
>


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss

Re: [ECOS] bug in mips arch code ????

[Nick Garnett]

"HG" <henri@broadbandnetdevices.com> writes:

> Whatever is the process that acknowledges the interrupt (its not
> HAL_INTERRUPT_ACKNOWLEDGE )
> stops working (intreq signal stays low)
> small file transfers are ok but long ones have more chances for this
> event to happen . Apps that use the timer tick by themselves (no irq
> produced by networking) are ok
> 

It looks like your platform has fallen into a gap between those that
acknowledge interrupts in the interrupt controller, and those that do
it in the device. Adding the device driver acknowledge macros may do
the trick.

-- 
Nick Garnett                    eCos Kernel Architect
http://www.ecoscentric.com/     The eCos and RedBoot experts


-- 
Before posting, please read the FAQ: http://sources.redhat.com/fom/ecos
and search the list archive: http://sources.redhat.com/ml/ecos-discuss