Re: [Qemu-devel] [PATCH 04/10] rocker: add register programming guide

[Eric Blake]

On 12/29/2014 10:14 PM, address@hidden wrote:
> From: Scott Feldman <address@hidden>
> 
> This is the register programming guide for the Rocker device.  It's intended
> for driver writers and device writers.  It covers the device's PCI space,
> the register set, DMA interface, and interrupts.
> 
> Signed-off-by: Scott Feldman <address@hidden>
> Signed-off-by: Jiri Pirko <address@hidden>
> ---
>  hw/net/rocker/reg_guide.txt |  957 
> +++++++++++++++++++++++++++++++++++++++++++
>  1 file changed, 957 insertions(+)
>  create mode 100644 hw/net/rocker/reg_guide.txt
> 
> diff --git a/hw/net/rocker/reg_guide.txt b/hw/net/rocker/reg_guide.txt
> new file mode 100644
> index 0000000..21ed9f7
> --- /dev/null
> +++ b/hw/net/rocker/reg_guide.txt
> @@ -0,0 +1,957 @@
> +Rocker Network Switch Register Programming Guide
> +Scott Feldman <address@hidden>
> +Neil Horman <address@hidden>
> +Version 0.11, 12/29/2014
> +

Missing a copyright and license clause.  Although this means that it
defaults to GPLv2+, per the top-level files, it's better to be explicit
(especially if you intended a different license).


> +
> +XXXXX this section needs work A generation bit in each descriptors comp_err

You'll want to fix this in your v2 patch.

> +field marks hardware's progression through the descriptor ring.  Each time 
> the
> +ring wraps, the generation bit is toggled.  The generation bit allows 
> software
> +to know which descriptors are DMA complete.  Software must also toggle its
> +sense of the generation bit each pass through the ring to remain synchronized
> +with hardware.  Hardware will write '1' to generation bit on first pass 
> through
> +ring, so software should zero-set descriptor ring memory to be syncronized 
> with

s/syncronized/synchronized/


> +
> +Interrupt credits
> +^^^^^^^^^^^^^^^^^
> +

> +masked but the device generates an interrupt, signally the driver that more

s/signally/signaling/

> +outstansing work is available.

s/outstansing/outstanding/

> +
> +Various buffer address and sizes should be tested to verify no address 
> boundary
> +issue exists.  In particular, buffers that start on odd-8-byte boudary and/or

s/boudary/boundary/


> +SECTION 8: CPU Packet Processing
> +================================
> +
> +For packets ingressing on switch ports that are not forwarded by the switch 
> but
> +rather directed to the host CPU for further processing are delievered in the

s/delievered/delivered/


> +Rx Packet Processing
> +--------------------
> +
> +For packets ingressing on switch ports that are not forwarded by the switch 
> but
> +rather directed to the host CPU for further processing are delievered in the

again

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature