summary refs log tree commit diff stats
path: root/gitlab/issues_text/target_sparc/host_missing/accel_missing/2518
diff options
context:
space:
mode:
authorChristian Krinitsin <mail@krinitsin.com>2025-06-01 21:35:14 +0200
committerChristian Krinitsin <mail@krinitsin.com>2025-06-01 21:35:14 +0200
commit3e4c5a6261770bced301b5e74233e7866166ea5b (patch)
tree9379fddaba693ef8a045da06efee8529baa5f6f4 /gitlab/issues_text/target_sparc/host_missing/accel_missing/2518
parente5634e2806195bee44407853c4bf8776f7abfa4f (diff)
downloadqemu-analysis-3e4c5a6261770bced301b5e74233e7866166ea5b.tar.gz
qemu-analysis-3e4c5a6261770bced301b5e74233e7866166ea5b.zip
clean up repository
Diffstat (limited to 'gitlab/issues_text/target_sparc/host_missing/accel_missing/2518')
-rw-r--r--gitlab/issues_text/target_sparc/host_missing/accel_missing/251814
1 files changed, 0 insertions, 14 deletions
diff --git a/gitlab/issues_text/target_sparc/host_missing/accel_missing/2518 b/gitlab/issues_text/target_sparc/host_missing/accel_missing/2518
deleted file mode 100644
index ace94a5d1..000000000
--- a/gitlab/issues_text/target_sparc/host_missing/accel_missing/2518
+++ /dev/null
@@ -1,14 +0,0 @@
-Incorrect vertical mouse leaps on qemu-system-sparc
-Description of problem:
-In openwin (i.e. X) if you turn the scrollwheel on the mouse 1 click the cursor jumps almost all of the way down the screen. Similarly, just pressing the scroll wheel (middle click) multiple times will sometimes produce similar behavior but the cursor doesn't jump as far.
-Steps to reproduce:
-- Boot Solaris and log in
-- capture the mouse by clicking on the screen
-- position the mouse cursor near the top of the screen (just so you can see how far it jumps)
-- click the scroll wheel up or down one click and observe the cursor jump downward
-Additional information:
-The issue is independent of which graphic display is used -- sdl, gtk and vnc all do the same thing. Debugging so far suggests that the problem is related to the fact that `sunmouse_event` in `escc.c` is sending a flood of duplicate events in response to the mouse scroll action. My surmise is that this is causing one byte to be dropped from the 5 byte mouse protocol expected by the Solaris kernel and that it is interpreting a sync byte as a vertical motion byte.
-
-`sunmouse_event` should not send events with only `dz` non-zero and no button changes. The Mouse Systems Corp spec for the the Sun mouse says that it never sends duplicate events (i.e. an event is produced only if there is non-zero `dx` or `dy` or there has been a button state change), and the mouse protocol has no support for `dz` events.
-
-I will propose a patch shortly to enforce this (and which has seemingly fixed the problem).