Re: [PATCH 1/5] dt-bindings: dma: Document RZ/G2L bindings

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

 



Hi Biju,

On Mon, Jun 14, 2021 at 04:33:03PM +0000, Biju Das wrote:
> > On Mon, Jun 14, 2021 at 04:24:38PM +0000, Biju Das wrote:
> > > > On Mon, Jun 14, 2021 at 04:09:04PM +0000, Biju Das wrote:
> > > > > > On Mon, Jun 14, 2021 at 12:54:02PM +0000, Biju Das wrote:
> > > > > > > > On Fri, Jun 11, 2021 at 1:36 PM Biju Das wrote:
> > > > > > > > > Document RZ/G2L DMAC bindings.
> > > > > > > > >
> > > > > > > > > Signed-off-by: Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> > > > > > > > > Reviewed-by: Lad Prabhakar <prabhakar.mahadev-lad.rj@xxxxxxxxxxxxxx>
> > > > > > > >
> > > > > > > > Thanks for your patch!
> > > > > > > >
> > > > > > > > > --- /dev/null
> > > > > > > > > +++ b/Documentation/devicetree/bindings/dma/renesas,rz-dma
> > > > > > > > > +++ c.ya
> > > > > > > > > +++ ml
> > > > > > > > > @@ -0,0 +1,132 @@
> > > > > > > > > +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> > > > > > > > > +%YAML
> > > > > > > > > +1.2
> > > > > > > > > +---
> > > > > > > > > +$id:
> > > > > > > > > +https://jpn01.safelinks.protection.outlook.com/?url=http%
> > > > > > > > > +3A%2
> > > > > > > > > +F%2F
> > > > > > > > > +devi
> > > > > > > > > +cetree.org%2Fschemas%2Fdma%2Frenesas%2Crz-dmac.yaml%23&am
> > > > > > > > > +p;da
> > > > > > > > > +ta=0
> > > > > > > > > +4%7C
> > > > > > > > > +01%7Cbiju.das.jz%40bp.renesas.com%7C4b547e10cbe64b6f4d850
> > > > > > > > > +8d92
> > > > > > > > > +f2da
> > > > > > > > > +0c0%
> > > > > > > > > +7C53d82571da1947e49cb4625a166a4a2a%7C0%7C0%7C637592695286
> > > > > > > > > +8468
> > > > > > > > > +09%7
> > > > > > > > > +CUnk
> > > > > > > > > +nown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLC
> > > > > > > > > +JBTi
> > > > > > > > > +I6Ik
> > > > > > > > > +1haW
> > > > > > > > > +wiLCJXVCI6Mn0%3D%7C1000&amp;sdata=5Jh%2FxPaia5ZOY0CrViQCc
> > > > > > > > > +rNtz
> > > > > > > > > +uDej
> > > > > > > > > +p8wo
> > > > > > > > > +Nrx9iO0ht8%3D&amp;reserved=0
> > > > > > > > > +$schema:
> > > > > > > > > +https://jpn01.safelinks.protection.outlook.com/?url=http%
> > > > > > > > > +3A%2
> > > > > > > > > +F%2F
> > > > > > > > > +devi
> > > > > > > > > +cetree.org%2Fmeta-
> > > > > > schemas%2Fcore.yaml%23&amp;data=04%7C01%7Cbiju.das.
> > > > > > > > > +jz%40bp.renesas.com%7C4b547e10cbe64b6f4d8508d92f2da0c0%7C
> > > > > > > > > +53d8
> > > > > > > > > +2571
> > > > > > > > > +da19
> > > > > > > > > +47e49cb4625a166a4a2a%7C0%7C0%7C637592695286846809%7CUnkno
> > > > > > > > > +wn%7
> > > > > > > > > +CTWF
> > > > > > > > > +pbGZ
> > > > > > > > > +sb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwi
> > > > > > > > > +LCJX
> > > > > > > > > +VCI6
> > > > > > > > > +Mn0%
> > > > > > > > > +3D%7C1000&amp;sdata=5qQ1PljM3e4Bn4%2FjdldYUHRBQL3jArJgRIA
> > > > > > > > > +dLnh
> > > > > > > > > +Jraw
> > > > > > > > > +%3D&
> > > > > > > > > +amp;reserved=0
> > > > > >
> > > > > > *sigh*
> > > > > >
> > > > > > > > > +
> > > > > > > > > +title: Renesas RZ/G2L DMA Controller
> > > > > > > > > +
> > > > > > > > > +maintainers:
> > > > > > > > > +  - Biju Das <biju.das.jz@xxxxxxxxxxxxxx>
> > > > > > > > > +
> > > > > > > > > +allOf:
> > > > > > > > > +  - $ref: "dma-controller.yaml#"
> > > > > > > > > +
> > > > > > > > > +properties:
> > > > > > > > > +  compatible:
> > > > > > > > > +    items:
> > > > > > > > > +      - enum:
> > > > > > > > > +          - renesas,dmac-r9a07g044  # RZ/G2{L,LC}
> > > > > > > >
> > > > > > > > Please use "renesas,r9a07g044-dmac".
> > > > > > >
> > > > > > > OK. Will change.
> > > > > > >
> > > > > > > > > +      - const: renesas,rz-dmac
> > > > > > > >
> > > > > > > > Does this need many changes for RZ/A1H and RZ/A2M?
> > > > > > >
> > > > > > > It will work on both RZ/A1H and RZ/A2M. I have n't tested since I don't have the board.
> > > > > > > There is some difference in MID bit size. Other wise both identical.
> > > > > > >
> > > > > > > > > +  renesas,rz-dmac-slavecfg:
> > > > > > > > > +    $ref: /schemas/types.yaml#/definitions/uint32-array
> > > > > > > > > +    description: |
> > > > > > > > > +      DMA configuration for a slave channel. Each channel
> > > > > > > > > + must have an array of
> > > > > > > > > +      3 items as below.
> > > > > > > > > +      first item in the array is MID+RID
> > > > > > > >
> > > > > > > > Already in dmas.
> > > > > > > >
> > > > > > > > > +      second item in the array is slave src or dst
> > > > > > > > > + address
> > > > > > > >
> > > > > > > > As pointed out by Rob, already known by the slave driver.
> > > > > > > >
> > > > > > > > > +      third item in the array is channel configuration value.
> > > > > > > >
> > > > > > > > What exactly is this?
> > > > > >
> > > > > > What would prevent the DMA client from passing the configuration
> > > > > > to the DMA channel through the DMA engine API, just like it
> > > > > > passes the slave source or destination address ?
> > > > >
> > > > > On RZ/G2L, there is 1 case(SSIF ch2) where MID+RID is same for both tx and rx.
> > > > > The only way we can distinguish it is from channel configuration value.
> > > >
> > > > Are those two different hardware DMA channels ? And configuration
> > > > values change between the two ?
> > >
> > > Yes, REQD is different, apart from this Rx have transfer source and Tx
> > > have Transfer destination.
> > > This particular SSIF ch2 is used only for half duplex compared to other
> > > SSIF channels.
> > 
> > Does this mean there's a single DMA channel, used by two clients, but not
> > at the same time as it only supports half-duplex ?
> 
> From hardware perspective, it is 2 channel. For eg:- playback/recording use case.
> You cannot do simultaneous playback, but you can do playback or record separately.

If the two channels have the same MID+RID and only differ by the
direction, I'd add a cell in the dmas property with the direction only.
The source/destination address should be dropped, as it's already known
by the driver.

This being said, in your example below, you have

dmas = <&dmac 0x255 0x10049c18 CH_CFG(0x1,0x0,0x1,0x0,0x2,0x1,0x1,0x0)>,
       <&dmac 0x256 0x10049c1c CH_CFG(0x0,0x0,0x1,0x0,0x2,0x1,0x1,0x0)>;
dma-names = "tx", "rx";

This looks like different MID+RID values for the two channels.

> > > > > > > > Does the R-Car DMAC have this too? If yes, how does its
> > > > > > > > driver handle it?
> > > > > > >
> > > > > > > On R-CAR DMAC, we have only MID + RID values. Where as here we
> > > > > > > have channel configuration value With different set of
> > > > > > > parameter as mentioned in Table 16.4.
> > > > > > >
> > > > > > > Please see Page 569, Table 16.4 On-Chip Module requests section.
> > > > > > >
> > > > > > > For eg:- as per Rob's suggestion, I have modelled the driver
> > > > > > > with the below entries in ALSA driver for playback/record use case.
> > > > > > >
> > > > > > > dmas = <&dmac 0x255 0x10049c18 CH_CFG(0x1,0x0,0x1,0x0,0x2,0x1,0x1,0x0)>,
> > > > > > >        <&dmac 0x256 0x10049c1c CH_CFG(0x0,0x0,0x1,0x0,0x2,0x1,0x1,0x0)>;
> > > > > > > dma-names = "tx", "rx";
> > > > > > >
> > > > > > > Using first parameter, it gets dmac channel. using second and
> > > > > > > third parameter it configures the channel.

-- 
Regards,

Laurent Pinchart



[Index of Archives]     [Device Tree Compilter]     [Device Tree Spec]     [Linux Driver Backports]     [Video for Linux]     [Linux USB Devel]     [Linux PCI Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [XFree86]     [Yosemite Backpacking]


  Powered by Linux