Hi Niklas, CC'ing Nelson Costa from Synopsys. On Thu, Sep 21, 2023 at 08:45:37PM +0200, Niklas Söderlund wrote: > Hello, > > On 2023-09-18 00:25:09 +0300, Laurent Pinchart wrote: > > [snip] > > > > > diff --git a/drivers/media/platform/nxp/dwc-mipi-csi2.h > > > > b/drivers/media/platform/nxp/dwc-mipi-csi2.h > > > > new file mode 100644 > > > > index 000000000000..470023787c25 > > > > --- /dev/null > > > > +++ b/drivers/media/platform/nxp/dwc-mipi-csi2.h > > > > @@ -0,0 +1,299 @@ > > > > +/* SPDX-License-Identifier: GPL-2.0 */ > > > > +/* > > > > + * Copyright 2023 NXP > > > > + */ > > > > + > > > > +#ifndef __DWC_MIPI_CSI2_H__ > > > > +#define __DWC_MIPI_CSI2_H__ > > > > + > > > > +#include <linux/device.h> > > > > +#include <linux/delay.h> > > > > +#include <linux/io.h> > > > > + > > > > +#include <media/v4l2-ctrls.h> > > > > +#include <media/v4l2-device.h> > > > > +#include <media/v4l2-fwnode.h> > > > > +#include <media/v4l2-mc.h> > > > > +#include <media/v4l2-subdev.h> > > > > > > I don't think you need many of these headers here. Please move to the .c > > > file those you don't need here. > > > > > > > + > > > > +/* MIPI CSI-2 Host Controller Registers Define */ > > > > + > > > > Niklas, is it me, or do the registers here look very familiar ? The > > R-Car V4H may have the same CSI-2 RX. Should we have a single driver ? > > I agree it looks very familiar. The register space to indeed start out > aligned between this driver and R-Car V4H (not checked every bit tho). > But once we get to later registers it seems the two diverge a bit. If > this is due to usage or something else I don't know. I have very sparse > documentation for the V4H. > > It becomes even more interesting. Have a look at this drivers > hsfreqrange lookup table, dwc_csi_mbps_table[]. That overlaps with R-Car > V3U (rcar-csi2.c hsfreqrange_v3u[]) instead of V4H, but V3U appears to > support faster bus. I expect different SoCs to have different frequency range tables, depending on how the IP core is synthesized (the manufacturing process may play a role too). > I think we can agree these two drivers are very similar, but not > identical. The most interesting register here is the first one > CSI2RX_VERSION. Maybe if we can figure that out we could start to think > about a generic driver? Unfortunately I have no documentation for the > content of this register. And the V4H driver do not use it, so I don't > even know if it contains anything. > > Is there any publicly documentation about this device? I have very > sparse documentation about V4H, and nothing I can share. As to not block > anyone maybe two drivers are OK and if we can figure out what parts, if > any can be shared? I don't think there's any public documentation available from Synopsys. Various SoC manuals, public or private, include different level of details. Nelson, do you have any insight you could share, or maybe point us to someone familiar with this IP core ? -- Regards, Laurent Pinchart
