| Message ID | CADnq5_N6x_vt3TSF8Vkg0j9Gj1UQ5W3V96sX9EESOfTmrY+9mQ@mail.gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
2014-02-21 14:37 keltezéssel, Alex Deucher írta: > > Does the attached patch help? > > Alex Yes, it helped, thank you very much. The compressed Xorg.0.log is attached as proof. To others: the patch is against xf86-video-ati. Two notes, though: 1. Is it normal that "xrandr --listproviders" lists 3 providers? [zozo@localhost ~]$ cat xrandr-providers Providers: number : 3 Provider 0: id: 0x86 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 4 outputs: 3 associated providers: 2 name:radeon Provider 1: id: 0x4f cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 0 outputs: 0 associated providers: 2 name:radeon Provider 2: id: 0x4f cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload crtcs: 0 outputs: 0 associated providers: 2 name:radeon 2. I tried glxgears with and without DRI_PRIME=1 as below: [zozo@localhost ~]$ glxgears Running synchronized to the vertical refresh. The framerate should be approximately the same as the monitor refresh rate. 299 frames in 5.0 seconds = 59.558 FPS [zozo@localhost ~]$ DRI_PRIME=1 glxgears Running synchronized to the vertical refresh. The framerate should be approximately the same as the monitor refresh rate. 5589 frames in 5.0 seconds = 1117.730 FPS 5369 frames in 5.0 seconds = 1073.715 FPS 5699 frames in 5.0 seconds = 1139.670 FPS Obviously, it doesn't sync to the framerate with PRIME. On the other hand, nothing is displayed in the glxgears window. I have llvm 3.4-4 and Mesa 10.1-rc1 from Fedora 21 rawhide. I will go back to plain Fedora 20 (Xorg 1.14, llvm 3.3 and Mesa 9.2.5) and try the same patch to see if there's any difference. Best regards, Zoltán Böszörményi
2014-02-21 15:32 keltezéssel, Boszormenyi Zoltan írta: > 2014-02-21 14:37 keltezéssel, Alex Deucher írta: >> >> Does the attached patch help? >> >> Alex > > Yes, it helped, thank you very much. > The compressed Xorg.0.log is attached as proof. > > To others: the patch is against xf86-video-ati. > > Two notes, though: > > 1. Is it normal that "xrandr --listproviders" lists 3 providers? > > [zozo@localhost ~]$ cat xrandr-providers > Providers: number : 3 > Provider 0: id: 0x86 cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload > crtcs: 4 outputs: 3 associated providers: 2 name:radeon > Provider 1: id: 0x4f cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload > crtcs: 0 outputs: 0 associated providers: 2 name:radeon > Provider 2: id: 0x4f cap: 0xf, Source Output, Sink Output, Source Offload, Sink Offload > crtcs: 0 outputs: 0 associated providers: 2 name:radeon > > 2. I tried glxgears with and without DRI_PRIME=1 as below: > > [zozo@localhost ~]$ glxgears > Running synchronized to the vertical refresh. The framerate should be > approximately the same as the monitor refresh rate. > 299 frames in 5.0 seconds = 59.558 FPS > [zozo@localhost ~]$ DRI_PRIME=1 glxgears > Running synchronized to the vertical refresh. The framerate should be > approximately the same as the monitor refresh rate. > 5589 frames in 5.0 seconds = 1117.730 FPS > 5369 frames in 5.0 seconds = 1073.715 FPS > 5699 frames in 5.0 seconds = 1139.670 FPS > > Obviously, it doesn't sync to the framerate with PRIME. > On the other hand, nothing is displayed in the glxgears window. > I have llvm 3.4-4 and Mesa 10.1-rc1 from Fedora 21 rawhide. > > I will go back to plain Fedora 20 (Xorg 1.14, llvm 3.3 and Mesa 9.2.5) > and try the same patch to see if there's any difference. With Fedora 20 packages X doesn't come up. The last message on the screen is: [ 19.016881] pci_pm_runtime_suspend(): radeon_pmops_runtime_suspend+0x0/0xa0 [radeon] return -22 Xorg crashes and Xorg.0.log now contains the backtrace, attached. Best regards, Zoltán Böszörményi
On Fri, Feb 21, 2014 at 9:32 AM, Boszormenyi Zoltan <zboszor@pr.hu> wrote: > 2014-02-21 14:37 keltezéssel, Alex Deucher írta: > >> >> Does the attached patch help? >> >> Alex > > > Yes, it helped, thank you very much. > The compressed Xorg.0.log is attached as proof. > > To others: the patch is against xf86-video-ati. > > Two notes, though: > > 1. Is it normal that "xrandr --listproviders" lists 3 providers? > > [zozo@localhost ~]$ cat xrandr-providers > Providers: number : 3 > Provider 0: id: 0x86 cap: 0xf, Source Output, Sink Output, Source Offload, > Sink Offload crtcs: 4 outputs: 3 associated providers: 2 name:radeon > Provider 1: id: 0x4f cap: 0xf, Source Output, Sink Output, Source Offload, > Sink Offload crtcs: 0 outputs: 0 associated providers: 2 name:radeon > Provider 2: id: 0x4f cap: 0xf, Source Output, Sink Output, Source Offload, > Sink Offload crtcs: 0 outputs: 0 associated providers: 2 name:radeon > > 2. I tried glxgears with and without DRI_PRIME=1 as below: > > [zozo@localhost ~]$ glxgears > Running synchronized to the vertical refresh. The framerate should be > approximately the same as the monitor refresh rate. > 299 frames in 5.0 seconds = 59.558 FPS > [zozo@localhost ~]$ DRI_PRIME=1 glxgears > Running synchronized to the vertical refresh. The framerate should be > approximately the same as the monitor refresh rate. > 5589 frames in 5.0 seconds = 1117.730 FPS > 5369 frames in 5.0 seconds = 1073.715 FPS > 5699 frames in 5.0 seconds = 1139.670 FPS > > Obviously, it doesn't sync to the framerate with PRIME. > On the other hand, nothing is displayed in the glxgears window. > I have llvm 3.4-4 and Mesa 10.1-rc1 from Fedora 21 rawhide. > Make sure you are running a compositor. Alex
2014-02-21 16:12 keltezéssel, Alex Deucher írta: > On Fri, Feb 21, 2014 at 9:32 AM, Boszormenyi Zoltan <zboszor@pr.hu> wrote: >> [zozo@localhost ~]$ DRI_PRIME=1 glxgears >> Running synchronized to the vertical refresh. The framerate should be >> approximately the same as the monitor refresh rate. >> 5589 frames in 5.0 seconds = 1117.730 FPS >> 5369 frames in 5.0 seconds = 1073.715 FPS >> 5699 frames in 5.0 seconds = 1139.670 FPS >> >> Obviously, it doesn't sync to the framerate with PRIME. >> On the other hand, nothing is displayed in the glxgears window. >> I have llvm 3.4-4 and Mesa 10.1-rc1 from Fedora 21 rawhide. >> > Make sure you are running a compositor. Turning on compositing in MATE fixed it. Thank you very much. It seems I will have to keep Rawhide on this notebook. :-) Best regards, Zoltán Böszörményi
From b6aef10258062e2a8ec638b9431a52214b61fdf0 Mon Sep 17 00:00:00 2001 From: Alex Deucher <alexander.deucher@amd.com> Date: Fri, 21 Feb 2014 08:33:21 -0500 Subject: [PATCH] radeon: don't install colormap handling if there are no crtcs Fixes a crash on cards with 0 crtcs. Signed-off-by: Alex Deucher <alexander.deucher@amd.com> --- src/drmmode_display.c | 22 +++++++++++++--------- 1 file changed, 13 insertions(+), 9 deletions(-) diff --git a/src/drmmode_display.c b/src/drmmode_display.c index 76b79d8..641e231 100644 --- a/src/drmmode_display.c +++ b/src/drmmode_display.c @@ -1939,19 +1939,23 @@ static void drmmode_load_palette(ScrnInfoPtr pScrn, int numColors, Bool drmmode_setup_colormap(ScreenPtr pScreen, ScrnInfoPtr pScrn) { + xf86CrtcConfigPtr xf86_config = XF86_CRTC_CONFIG_PTR(pScrn); + xf86DrvMsgVerb(pScrn->scrnIndex, X_INFO, RADEON_LOGLEVEL_DEBUG, "Initializing kms color map\n"); - if (!miCreateDefColormap(pScreen)) - return FALSE; - /* all radeons support 10 bit CLUTs */ - if (!xf86HandleColormaps(pScreen, 256, 10, - drmmode_load_palette, NULL, - CMAP_PALETTED_TRUECOLOR + if (xf86_config->num_crtc) { + if (!miCreateDefColormap(pScreen)) + return FALSE; + /* all radeons support 10 bit CLUTs */ + if (!xf86HandleColormaps(pScreen, 256, 10, + drmmode_load_palette, NULL, + CMAP_PALETTED_TRUECOLOR #if 0 /* This option messes up text mode! (eich@suse.de) */ - | CMAP_LOAD_EVEN_IF_OFFSCREEN + | CMAP_LOAD_EVEN_IF_OFFSCREEN #endif - | CMAP_RELOAD_ON_MODE_SWITCH)) - return FALSE; + | CMAP_RELOAD_ON_MODE_SWITCH)) + return FALSE; + } return TRUE; } -- 1.8.3.1