On Tue, Dec 29, 2020 at 1:08 AM Laurent Pinchart wrote: > > On Tue, Dec 29, 2020 at 12:55:44AM +0200, Andy Shevchenko wrote: > > On Mon, Dec 28, 2020 at 10:37:38PM +0000, Daniel Scally wrote: > > > On 28/12/2020 17:05, Sakari Ailus wrote: ... > > > Which do you prefer? > > > > Actually ipu3-cio2.h misses a lot of inclusions (like mutex.h which I > > immediately noticed when scrolled over data types). > > Then ipu3-cio2.h should be fixed :-) Below is a draft patch (it is possible mangled, due to Gmail). Can you look at it and tell me what you think? I believe some headers can be removed, but I have no idea about header inclusion guarantees that v4l2 provides. From 10fa6c7ff66ded35a246677ffe20c677e8453f5b3 Mon Sep 17 00:00:00 2001 From: Andy Shevchenko Date: Tue, 29 Dec 2020 01:42:03 +0200 Subject: [PATCH 1/1] media: ipu3-cio2: Add headers that ipu3-cio2.h is direct user of Add headers that ipu3-cio2.h is direct user of. Signed-off-by: Andy Shevchenko --- drivers/media/pci/intel/ipu3/ipu3-cio2.h | 21 +++++++++++++++++++++ 1 file changed, 21 insertions(+) diff --git a/drivers/media/pci/intel/ipu3/ipu3-cio2.h b/drivers/media/pci/intel/ipu3/ipu3-cio2.h index ccf0b85ae36f..9ea154c50ba1 100644 --- a/drivers/media/pci/intel/ipu3/ipu3-cio2.h +++ b/drivers/media/pci/intel/ipu3/ipu3-cio2.h @@ -4,8 +4,25 @@ #ifndef __IPU3_CIO2_H #define __IPU3_CIO2_H +#include +#include +#include +#include #include +#include + +#include + +#include +#include +#include +#include +#include +#include +#include +#include + #define CIO2_NAME "ipu3-cio2" #define CIO2_DEVICE_NAME "Intel IPU3 CIO2" #define CIO2_ENTITY_NAME "ipu3-csi2" @@ -325,6 +342,8 @@ struct csi2_bus_info { u32 lanes; }; +struct cio2_fbpt_entry; + struct cio2_queue { /* mutex to be used by vb2_queue */ struct mutex lock; @@ -355,6 +374,8 @@ struct cio2_queue { atomic_t bufs_queued; }; +struct pci_dev; + struct cio2_device { struct pci_dev *pci_dev; void __iomem *base; -- 2.29.2 -- With Best Regards, Andy Shevchenko