On Mon Sep 9, 2024 at 11:13 AM CEST, Sascha Hauer wrote: > On Thu, Sep 05, 2024 at 04:09:58PM +0800, Andy Yan wrote: > > At 2024-09-05 15:10:56, "Sascha Hauer" <s.hauer@xxxxxxxxxxxxxx> wrote: > > >On Wed, Sep 04, 2024 at 08:02:32PM +0800, Andy Yan wrote: > > >> From: Andy Yan <andy.yan@xxxxxxxxxxxxxx> > > >> > > >> There is a version number hardcoded in the VOP VERSION_INFO > > >> register, and the version number increments sequentially based > > >> on the production order of the SOC. > > >> > > >> So using this version number to distinguish different VOP features > > >> will simplify the code. > > >> > > >> Signed-off-by: Andy Yan <andy.yan@xxxxxxxxxxxxxx> > > >> > > >> --- > > >> > > >> diff --git a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h > > >> index 9b269f6e576e..871d9bcd1d80 100644 > > >> --- a/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h > > >> +++ b/drivers/gpu/drm/rockchip/rockchip_drm_vop2.h > > >> @@ -13,6 +13,15 @@ > > >> #include "rockchip_drm_drv.h" > > >> #include "rockchip_drm_vop.h" > > >> > > >> +#define VOP2_VERSION(major, minor, build) ((major) << 24 | (minor) << 16 | (build)) > > >> + > > >> +/* The new SOC VOP version is bigger than the old */ > > >> +#define VOP_VERSION_RK3568 VOP2_VERSION(0x40, 0x15, 0x8023) > > >> +#define VOP_VERSION_RK3588 VOP2_VERSION(0x40, 0x17, 0x6786) > > >> +#define VOP_VERSION_RK3528 VOP2_VERSION(0x50, 0x17, 0x1263) > > >> +#define VOP_VERSION_RK3562 VOP2_VERSION(0x50, 0x17, 0x4350) > > >> +#define VOP_VERSION_RK3576 VOP2_VERSION(0x50, 0x19, 0x9765) > > > > > >What about the RK3566? Does it have the same version code as the RK3568? > > > > > >This new version field replaces the soc_id mechanism we had before to > > >99%. You keep the soc_id around just for distinguishing between RK3566 > > >and RK3568. It would be nice to fully replace it. > > > > > >I see that the VOP_VERSION_RK* numbers are the same as found in the > > >VOP2_SYS_VERSION_INF registers. On the other hand you never read the > > >value from the register which make the VOP_VERSION_RK* just arbitrary > > >numbers. Wouldn't it be possible to make something up for RK3566, like > > >VOP2_VERSION(0x40, 0x15, 0x8022) to get rid of the soc_id thingy? > > Yes，RK3566 and RK3568 share the same VOP IP block， so the version code at VERSION_REGISTER is > > the same, the difference between rk3568 and rk33566 are introduced at soc Integration。 > > So i would still like to keep the soc_id to handle situation like this。As we always have such cause， one > > same IP block， but there are some subtle differences in features across different SOCs. > > Fine with me. You could leave a comment in the code or commit > message that explains why we need both. Also (or especially?) add that to the commit message of patch 6 of this series. Patch 6's commit message talks about RK3576 while it changes code related to RK3566 and I (too?) thought that not using VOP_VERSION was an oversight, while it turns out to be deliberate. Cheers, Diederik
Attachment:
signature.asc
Description: PGP signature
