First, the environment. It's a build-your-own system, with an ASUS
H81M-C motherboard, and using on-board (in-CPU?) graphics. The CPU is
identified by cpuctl(8) as
cpu0: "Intel(R) Core(TM) i5-4460 CPU @ 3.20GHz"
cpu0: Intel 4th gen Core, Xeon E3-12xx v3 (Haswell) (686-class), 3192.75
MHz
cpu0: family 0x6 model 0x3c stepping 0x3 (id 0x306c3)
This is using i915drmkms - here's the relevant dmesg lines:
i915drmkms0 at pci0 dev 2 function 0: vendor 8086 product 0412 (rev. 0x06)
drm: Memory usable by graphics device = 2048M
drm: Supports vblank timestamp caching Rev 2 (21.10.2013).
drm: Driver supports precise vblank timestamp query.
i915drmkms0: interrupting at ioapic0 pin 16 (i915)
intelfb0 at i915drmkms0
i915drmkms0: info: registered panic notifier
i915drmkms0: More than 8 outputs detected via ACPI
drm: Enabling RC6 states: RC6 on, RC6p off, RC6pp off
My display is a Samsung 20-inch device, with a default/native resolution
of 1600x900
The box is running NetBSD amd64/7.99.8 (installed from the sets that
were at one time available on nyftp.n.o from 201503300400Z).
Numerous self-compiled packages are installed, including shells/fvwm2
and www/firefow, along with all dependencies.
The problem appears to be mis-aligned data being written to the graphics
device. By this I mean, it looks like rows are offset, with row 2 being
some number of pixels to the left or right of row 1. The problem occurs
mostly in the window title bars displayed by fvwm2, but also in various
images displayed in firefox. Sometimes I can force a successful redraw
of the corrupted region (for example, shifting window focus causes the
title bar's highlighting to be changed).
If there's more info needed, please let me know!
As I said, it's not critical, just annoying. But it would be nice to
fix. I'm not at all sure if this issue also occurs on 7.0 BETA, but if
so, IMHO we really should fix it before release.