summary refs log tree commit diff stats
path: root/results/classifier/deepseek-2-tmp/output/graphic/1752646
diff options
context:
space:
mode:
Diffstat (limited to 'results/classifier/deepseek-2-tmp/output/graphic/1752646')
-rw-r--r--results/classifier/deepseek-2-tmp/output/graphic/175264623
1 files changed, 0 insertions, 23 deletions
diff --git a/results/classifier/deepseek-2-tmp/output/graphic/1752646 b/results/classifier/deepseek-2-tmp/output/graphic/1752646
deleted file mode 100644
index 1f6e6a608..000000000
--- a/results/classifier/deepseek-2-tmp/output/graphic/1752646
+++ /dev/null
@@ -1,23 +0,0 @@
-
-Freezing VNC screen on some the UEFI framebuffer applications
-
-Hi folks!
-
-I use TianCore (UEFI) formware on the qemu (master version last commit a6e0344).
-When kernel/linux is start, it using UEFI Framebuffer. Then I run UEFI application (which writes directly to the framebuffer) my VNS screen is freezing. Then I restart vnclient I see only one frame.
-
-When I run application, I getting in the file hw/display/vga.c on function 'vga_ioport_write' some commands, it change "s->ar_index" from 0x20 -> 0x10 
-
-In the function vga_update_display:
-1751         if (!(s->ar_index & 0x20)) {
-1752             graphic_mode = GMODE_BLANK;
-1753         } else {
-
-And I got GMODE_BLANK mode. If I patch it:
-1751         if (0) {
-
-my VNC not freezing.
-
-From "Hardware Level VGA and SVGA Video Programming Information Page" I saw, what ar_index is 0x3C0 (Attribute Controller Data Write Register), 0x20(5-bit) is PAS -- Palette Address Source
-
-If there is a output via the UEFI framebuffer, does the difference have a PAS or not? Why do we need to pause the output if the PAS is exposed? Especially when the application outputs via framebuffer.
\ No newline at end of file