Created attachment 244257 [details] A screenshot after reenabling a previously enabled display Some time ago, maybe June or July 2023, after logging in to Plasma (X11): * it was necessary to re-enable an external display (one of two, at work) that was enabled when I last used Plasma at the same location. Since then: every time I wake the computer at work, after bringing it from home, I need to re-enable the same display. Both displays are on DisplayPort. HP EliteBook 8570p, docked; the dock has the two DisplayPort connections. % pkg iinfo drm-510-kmod drm-510-kmod-5.10.163_7 % uname -aKU FreeBSD mowa219-gjp4-8570p-freebsd 14.0-ALPHA2 FreeBSD 14.0-ALPHA2 amd64 1400094 #4 main-n264868-edacf4b4824a-dirty: Fri Aug 18 23:46:09 BST 2023 grahamperrin@mowa219-gjp4-8570p-freebsd:/usr/obj/usr/src/amd64.amd64/sys/GENERIC amd64 1400094 1400094 % From 'About this System — Info Centre': Operating System: FreeBSD 14.0 KDE Plasma Version: 5.27.7 KDE Frameworks Version: 5.108.0 Qt Version: 5.15.8 Kernel Version: 14.0-ALPHA2 (64-bit) Graphics Platform: X11 Memory: 15.9 GiB of RAM Graphics Processor: llvmpipe Manufacturer: Hewlett-Packard Product Name: HP EliteBook 8570p System Version: A1029D1102 I could, or should, have reported sooner, however I chose to wait until other bugs were fixed (or overcome by events); also I wanted to see whether there would appear an upstream bug report. I see nothing related at <https://bugs.kde.org/>, so I assume that this regression is specific to FreeBSD.
I don't get it, what's changed?
> Graphics Processor: llvmpipe Maybe Plasma treats llvmpipe as broken GPU configuration[1] thus disables saving to prevent intermittent issues from overwriting the last good config. plasma5-systemsettings source lacks GPU-related code, so it's hard to confirm/disprove my guess. [1] https://github.com/freebsd/drm-kmod/issues/175
Thank you both, (In reply to Jan Beich from comment #2) > … > > [1] https://github.com/freebsd/drm-kmod/issues/175 Possibly a factor, although I'm fairly certain that I invisibly subscribed to that June 2022 issue long, long before my 31st July 2023 comment there. ---- I wish I had kept a private note of when this bug 273264 became noticeable. It's _possible_ that I noticed it, but ignored it, around the time that Display Configuration gained a new feature: * the 'Keep display configuration' dialogue, which, if ignored, allows settings to revert after a timeout. Whenever that was. (Readers might recall that previous to this new feature, if you virtually rotated the display 90° the wrong way, you'd give yourself an upside-down situation that was amusingly difficult to recover from with a pointing device. That type of thing.) In any case: * when the Keep button is pressed, what file is written? A Plasma-specific file? Or, maybe something X-related that's theoretically usable by some other desktop environment or window manager?
In as few words as possible: * I encountered two other bugs whilst testing whilst typing comment 3 * one of the two involved Dolphin and near-complete exhaustion of both real memory and swap space * the other involved the 'Keep display configuration' dialogue. It's tempting to think of the latter as directly related to this bug 273264, however the latter was _not_ consistently reproducible and AFAIK this bug _is_ consistently reproducible. I might write more about this elsewhere, probably not in Bugzilla. The Dolphin-and-displays-and-memory bug, I took to Matrix: <https://matrix.to/#/!IBdGSejslGivmIcnQs:matrix.org/$169315086532769NNYfR:matrix.org?via=matrix.org&via=mozilla.org&via=tchncs.de>. ---- What's above is a point of reference, nothing more. For my own sanity (on a mental timeline) it belongs here, but I don't want it to pollute/complicate discussion of this bug.
(In reply to Gleb Popov from comment #1) > I don't get it, what's changed? Good question. What has changed, for me, in recent months? Picture me with three HP docks and five identical Philips displays, always using DisplayPort with a dock. Home: one Philips display, primary, to the left of the dock with the notebook display secondary. Work, my usual office: two Philips displays, primary landscape, secondary portrait, dock to the rear with the notebook case not open. Work, an occasional nearby office: two Philips displays, primary landscape, secondary portrait, dock to the left with … logically, I must have either (a) had the notebook case not open, or (b) had the notebook display disabled. This 'occasional' office was effectively disused for a few months, with all IT equipment removed. I'd need to dig through calendars etc. to tell, with certainty, when I resumed occupancy … off the top of my head, it was probably some time after early or mid-May. I wonder whether this bug 273264 became noticeable around the same time.
Graham, you wrote 4 lengthy comments, but I still don't get what the actual problem is. Can you please concisely explain what are the steps to reproduce the issue, what do you expect and what do you actually get?
(In reply to Gleb Popov from comment #6) Sorry, (In reply to Graham Perrin ◐ from comment #0) > … every time I wake the computer at work, after bringing it from home, > I need to re-enable the same display … That's the regression. Previously, to the best of my recollection: I rarely, or never, needed to touch System Settings. I'd use the computer at work and elsewhere, with the portrait display enabled at my main office, never unexpectedly disabled. Now: I find it unexpectedly disabled. HTH