<div dir="ltr">Reviewed-by: Marek Olšák <<a href="mailto:marek.olsak@amd.com">marek.olsak@amd.com</a>></div> <div class="gmail_quote gmail_quote_container"><div dir="ltr" class="gmail_attr">On Tue, Jun 18, 2019 at 3:19 AM Richard Thier <<a href="mailto:u9vata@gmail.com">u9vata@gmail.com</a>> wrote: </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">num_gb_pipes was set to a wrong value using r420_pipe_config This have lead to HyperZ glitches on fast Z clearing. See: <a href="https://bugs.freedesktop.org/show_bug.cgi?id=110897" rel="noreferrer" target="_blank">https://bugs.freedesktop.org/show_bug.cgi?id=110897</a> Signed-off-by: Richard Thier <<a href="mailto:u9vata@gmail.com" target="_blank">u9vata@gmail.com</a>> --- drivers/gpu/drm/radeon/r300.c | 3 ++- drivers/gpu/drm/radeon/rs400.c | 21 +++++++++++++++++++-- 2 files changed, 21 insertions(+), 3 deletions(-) diff --git a/drivers/gpu/drm/radeon/r300.c b/drivers/gpu/drm/radeon/r300.c index 652126f..6724c15 100644 --- a/drivers/gpu/drm/radeon/r300.c +++ b/drivers/gpu/drm/radeon/r300.c @@ -355,7 +355,8 @@ int r300_mc_wait_for_idle(struct radeon_device *rdev) return -1; } -static void r300_gpu_init(struct radeon_device *rdev) +/* rs400_gpu_init also calls this! */ +void r300_gpu_init(struct radeon_device *rdev) { uint32_t gb_tile_config, tmp; diff --git a/drivers/gpu/drm/radeon/rs400.c b/drivers/gpu/drm/radeon/rs400.c index 4121209..4117572 100644 --- a/drivers/gpu/drm/radeon/rs400.c +++ b/drivers/gpu/drm/radeon/rs400.c @@ -32,6 +32,9 @@ #include "radeon_asic.h" #include "rs400d.h" +/* Needed for rs400_gpu_init that forwards to the r300 one */ +void r300_gpu_init(struct radeon_device *rdev); + /* This files gather functions specifics to : rs400,rs480 */ static int rs400_debugfs_pcie_gart_info_init(struct radeon_device *rdev); @@ -252,8 +255,22 @@ int rs400_mc_wait_for_idle(struct radeon_device *rdev) static void rs400_gpu_init(struct radeon_device *rdev) { - /* FIXME: is this correct ? */ - r420_pipes_init(rdev); + /* Earlier code was calling r420_pipes_init and then + * rs400_mc_wait_for_idle(rdev). The problem is that + * at least on my Mobility Radeon Xpress 200M RC410 card + * that ends up in this code path ends up num_gb_pipes == 3 + * while the card seems to have only one pipe. With the + * r420 pipe initialization method. + * + * Problems shown up as HyperZ glitches, see: + * <a href="https://bugs.freedesktop.org/show_bug.cgi?id=110897" rel="noreferrer" target="_blank">https://bugs.freedesktop.org/show_bug.cgi?id=110897</a> + * + * Delegating initialization to r300 code seems to work + * and results in proper pipe numbers. The rs400 cards + * are said to be not r400, but r300 kind of cards. + */ + r300_gpu_init(rdev); + if (rs400_mc_wait_for_idle(rdev)) { pr_warn("rs400: Failed to wait MC idle while programming pipes. Bad things might happen. %08x\n", RREG32(RADEON_MC_STATUS)); -- 2.21.0 _______________________________________________ amd-gfx mailing list <a href="mailto:amd-gfx@lists.freedesktop.org" target="_blank">amd-gfx@lists.freedesktop.org</a> <a href="https://lists.freedesktop.org/mailman/listinfo/amd-gfx" rel="noreferrer" target="_blank">https://lists.freedesktop.org/mailman/listinfo/amd-gfx</a></blockquote></div>