Thread (31 messages) 31 messages, 3 authors, 2025-02-14

Re: [PATCH 1/6] fbcon: Make cursor_blink=0 work when configured before fb devices appear

flat view

From: Helge Deller <deller@gmx.de>
Date: 2025-02-14 18:18:02
Also in: dri-devel

On 2/14/25 18:41, Ville Syrjälä wrote:
On Thu, Feb 13, 2025 at 11:47:42PM +0100, Helge Deller wrote:
quoted
On 2/13/25 15:42, Ville Syrjälä wrote:
quoted
On Thu, Sep 26, 2024 at 12:13:04PM +0200, Helge Deller wrote:
quoted
On 9/26/24 11:57, Ville Syrjälä wrote:
quoted
On Thu, Sep 26, 2024 at 08:08:07AM +0200, Helge Deller wrote:
quoted
Hi Ville,

On 9/23/24 17:57, Ville Syrjala wrote:
quoted
Currently setting cursor_blink attribute to 0 before any fb
devices are around does absolutely nothing. When fb devices appear
and fbcon becomes active the cursor starts blinking. Fix the problem
by recoding the desired state of the attribute even if no fb devices
are present at the time.

Also adjust the show() method such that it shows the current
state of the attribute even when no fb devices are in use.

Note that store_cursor_blink() is still a bit dodgy:
- seems to be missing some of the other checks that we do
      elsewhere when deciding whether the cursor should be
      blinking or not
- when set to 0 when the cursor is currently invisible due
      to blinking, the cursor will remains invisible. We should
      either explicitly make it visible here, or wait until the
      full blink cycle has finished.
are you planning to send follow-up patches to address those shortcomings?
Nope. I don't really care about those as I never plan to
turn the cursor blinking back on after turning it off via
udev.
Sad, but OK. I will look into this when I find time.
I'd hoped to push those patches upstream during this merge window,
but then I think I have to delay them at least until kernel 6.13.
What happened to these? Not seeing them anywhere...
The issues above are not fixed yet, so they are still parked in my for-next-next tree:
https://git.kernel.org/pub/scm/linux/kernel/git/deller/linux-fbdev.git/log/?h=for-next-next
Those issues are already present in the current code, so
I'm sure what's the point of holding this up due to them.
Let's try to fix it while we touch it.

Helge
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help