Driver complaint to SD Host Controller Specification 3.0

[Mudit Jain]

Hi all,

I have procured a RPi Model B and was testing the ported driver on the
same.
This is where it hangs.
--------------------------------------------------------------------------------------------------
*** LIBBSD SDHCI 01 TEST ***
nexus0: <RTEMS Nexus device>
bcm_dma0: <BCM2835 DMA Controller> on nexus0
sdhci_bcm0: <Broadcom 2708 SDHCI controller> on nexus0
<Hangs>
--------------------------------------------------------------------------------------------------
It hangs because it tries to look for a mbox device,
Ideally, mbox0 should be on nexus0.

I suspect the problem is with the bus handles and the IRQs

Presently, the resource structure defined in the nexus-devices.h code is as
follows :
------------------------------------------------------------------------------------------------------------------------
static const rtems_bsd_device_resource rpi_mbox_res[] = {
    {
        .type = RTEMS_BSD_RES_MEMORY,
        .start_request = 0,
        .start_actual = 0x2000B880
    }, {
        .type = RTEMS_BSD_RES_IRQ,
        .start_request = 0,
        .start_actual = BCM2835_IRQ_ID_MAILBOX_0
    }
};
------------------------------------------------------------------------------------------------------------
The value of BCM2835_IRQ_ID_MAILBOX_0 is 65 in irq.h

Is this resource structure appropriate for mbox ?

In addition to that, as Andre pointed out that from the bsd mailbox code
includes framebuffer operations, which may be redundant with the current
RPI framebuffer support.
The main benefit from the FreeBSD code may be the DMA access to the
mailboxes.

Github Link <https://github.com/spark1729/rtems-libbsd/commits/RPI_SD>

Thanks
Mudit

On Sat, Jun 4, 2016 at 9:51 AM, Chris Johns <chrisj at rtems.org> wrote:

> On 3/06/2016 10:20 PM, Sebastian Huber wrote:
>
>> On 03/06/16 13:20, Mudit Jain wrote:
>>
>>> In the latest commit I have followed the FreeBSD port guidelines given
>>> libbsd.txt using the tools to port the code.
>>>
>>
>> Sorry for being pedantic, but please follow the guidelines exactly:
>>
>>
> I will add we have to be very precise about doing this and patches can
> only be accepted if this is the case.
>
> Sebastian, I wonder if we can automate the testing for this?
>
> Chris
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.rtems.org/pipermail/devel/attachments/20160607/e769474c/attachment.html>