Please try the attached patch it should fix your problem.
Thanks allot for this bug report, that was just a stupid typo in the
original patch which would probably went unnoticed for years otherwise.
Christian.
Am 12.05.2014 18:32, schrieb Ken Moffat:
On Mon, May 12, 2014 at 09:32:54AM +0200, Christian König wrote:
Hi Ken,
*sigh* did I already mentioned that I hate PLLs? As soon as you fix
something another use case immediately starts to break.
Please provide dmesg output created with drm.debug=0xe with and without the
patch breaking it.
Thanks in advance,
Christian.
The reverted version is from linus's tree after -rc5 with the patch
reverted, I assume the version -00010-gc9a25d0fc393 will NOT match
any public tree because I used git revert in a local branch. That
one works fine.
The bad version is from a random kernel which showed the problem
while I was bisecting, in this case rc2-00086. I first tried
booting vanilla rc5, but for some reason my blind attempt to login
and run 'dmesg >dmesg.bad' failed.
Thanks. Sorry you are having to deal with PLLs.
ĸen
>From 8b5c70b48d73b533f0003639cdb68bcffe7c1293 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Christian=20K=C3=B6nig?= <[email protected]>
Date: Tue, 13 May 2014 12:50:54 +0200
Subject: [PATCH] drm/radeon: fix typo in finding PLL params
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit
Otherwise the limit is raised to high.
Signed-off-by: Christian König <[email protected]>
---
drivers/gpu/drm/radeon/radeon_display.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/radeon/radeon_display.c b/drivers/gpu/drm/radeon/radeon_display.c
index 408b6ac..f00dbbf 100644
--- a/drivers/gpu/drm/radeon/radeon_display.c
+++ b/drivers/gpu/drm/radeon/radeon_display.c
@@ -999,7 +999,7 @@ void radeon_compute_pll_avivo(struct radeon_pll *pll,
/* avoid high jitter with small fractional dividers */
if (pll->flags & RADEON_PLL_USE_FRAC_FB_DIV && (fb_div % 10)) {
- fb_div_min = max(fb_div_min, (9 - (fb_div % 10)) * 20 + 60);
+ fb_div_min = max(fb_div_min, (9 - (fb_div % 10)) * 20 + 50);
if (fb_div < fb_div_min) {
unsigned tmp = DIV_ROUND_UP(fb_div_min, fb_div);
fb_div *= tmp;
--
1.9.1
