Re: [RFC PATCH v0.001] PCI: Add support for tango PCIe controller

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

 



On 21/03/2017 15:23, Liviu Dudau wrote:

> On Tue, Mar 21, 2017 at 02:01:57PM +0100, Mason wrote:
>
>> My PCIe controller is b/d/f 0/0/0.
>> It ignores all PCI addresses that do not fall within the range
>> defined in BAR0 of b/d/f 0/0/0.
>>
>> BAR0 has a configurable width of 2^(23+i) bytes, i < 8
>> It is implicitly split in 8 regions, which can map to
>> anywhere in CPU space. However, my understanding is that,
>> by default, the DMA framework expects PCI address X to
>> map to CPU address X...
>> (My understanding of that part is still a bit hazy.)
> 
> It looks very much like your PCIe controller (host bridge) has the configuration
> registers on the wrong side of the bridge (PCIe one vs the more normal host side).
> But otherwise it does look very much like an ATS system, where you program how the
> PCIe bus addresses map into the system (you call them CPU) addresses. Do you also
> have a way of controlling the direction of the mapping? (in other words, those 8
> regions are only for translating request coming out of the PCIe bus into system
> addresses, or can you also set the direct mapping of system address to PCIe address?).

There are two (asymmetric) features for these mappings.

For bus-to-system accesses, the scheme described above
of "bus addresses within BAR0 fall into 1 of 8 regions,
and 8 registers define 8 target system addresses".

For system-to-bus accesses, there is a single "offset"
register which gets added to the system address, to form
the bus address. I just use a 0 offset.

(This is made slightly more complex on later chips because
someone in marketing thought it would look good on the
data sheet to support 4 GB of RAM in a SoC using 32-bit
processors. There is an additional cpu-to-foo mapping.)

> As for DMA framework expectations, I'm not sure how that plays a role, unless you
> have some DMA engines on the PCIe bus.

I think the USB3 PCIe card I'm using is a bus master, so
it's able to push data to RAM without any help from the
ARM core.

Regards.

P.S. something in your message is confusing my MUA, likely
your Mail-Followup-To field. My MUA thinks it should not
send you my reply. Is that really what you want?




[Index of Archives]     [DMA Engine]     [Linux Coverity]     [Linux USB]     [Video for Linux]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]     [Greybus]

  Powered by Linux