Hi Seems there is another problem with fbsplash and a framebuffer in directcolor mode.(atleast with radeonfb which im using here). The result is the same color corruption that recently was fixed in fbsplash-0.9.1. So i think it must be somehow a similar problem. It seems as if the corruption occurs when directfb finishes. But i havent looked any further into it so that maybe wrong. I attach a simple directfb app as a testcase. Reproducible: Always Steps to Reproduce: 1. get the testcase, compile and run it on a framebuffer console with enabled fbsplash 2. colors get messed up 3. switch to another console, switch back and everything is ok 4. switch off fbsplash(su -c "splash_util --vc=x -c off) 5. run testapp again 6. no color corruption occurs Actual Results: fbsplash messes up colors. Expected Results: colors should be ok kernel: 2.6.10 + fbsplash-0.9.1-2.6.11-rc1 patch (but problem already exists in previous versions, at least i checked it with 2.6.9) kernel cmdline: root=/dev/evms/root video=radeonfb:1024x768-32@60 splash=silent,theme:dragonmatrix output from fbset -i: mode "1024x768-60" # D: 65.003 MHz, H: 48.365 kHz, V: 60.006 Hz geometry 1024 768 1024 768 32 timings 15384 168 8 29 3 144 6 rgba 8/16,8/8,8/0,0/0 endmode Frame buffer device information: Name : ATI Radeon NP Address : 0xc8000000 Size : 67108864 Type : PACKED PIXELS Visual : DIRECTCOLOR XPanStep : 8 YPanStep : 1 YWrapStep : 0 LineLength : 4096 MMIO Address: 0xd7ef0000 MMIO Size : 16384 Accelerator : No Please say if you need any additional informations.
Created attachment 48647 [details] testcase.c
I think i narrowed down the problem a bit. The problem is that DirectFB tries to backup the color map by the FBIOGETCMAP ioctl and restores it again at the end via FBIOPUTCMAP. See testcase2.c for example. The problem is that the colormap which fbsplash creates in directcolor mode isnt stored anywhere in the memory.(as far as i have seen). Thus the FBIOGETCMAP ioctl only returns the default colormap which then of course leds to some trouble when getting restored. However i didnt managed to solve the problem. Just producing more errors ;)
Created attachment 48867 [details] testcase2.c
I have just released a new version of fbsplash which fixes the problem :) It will be included in next releases of g-d-s.
great! ;) just want to confirm that the problem is gone with your last version. Thx.
Updated version of fbsplash has been included in g-d-s-2.6.10-r7 - closing.
