On 2/22/2024 05:08, Maxime Ripard wrote:
Hi Daniel,
On Mon, Feb 19, 2024 at 05:02:34PM +0800, Daniel van Vugt wrote:
Until now, deferred console takeover only meant defer until there is
output. But that risks stepping on the toes of userspace splash screens
as console messages may appear before the splash screen.
This becomes more likely the later the splash screen starts, but even
systems whose splash exists in initrd may not be not immune because they
still rely on racing against all possible kernel messages that might
trigger the fbcon takeover. And those kernel messages are hardware
dependent so what boots silently on one machine may not be so quiet on
the next. We also want to shield users from seeing warnings about their
hardware/firmware that they don't always have the power to fix themselves,
and may not be deemed worthy of fixing by the vendor.
So now we check the command line for the expectation of userspace splash
(CONFIG_FRAMEBUFFER_CONSOLE_DEFERRED_TAKEOVER_CONDITION) and if present
then defer fbcon's takeover until the first console switch. In the case
of Plymouth, its value would typically be "splash". This keeps the boot
experience clean and silent so long as the command line requests so.
Closes: https://bugs.launchpad.net/bugs/1970069
Cc: Mario Limonciello <mario.limonciello@xxxxxxx>
Signed-off-by: Daniel van Vugt <daniel.van.vugt@xxxxxxxxxxxxx>
I did test this series on an Ubuntu userspace and it works as you
suggest it should.
Tested-by: Mario Limonciello <mario.limonciello@xxxxxxx>
It's not clear to me why we should want to make it an option? If one
strategy is better than the other, and I guess the new one is if you
consider it fixes a bug and bothered to submit it upstream, why not just
get rid of the old one entirely?
I guess my question is: why do we want the choice, and what are the
tradeoff each strategy brings?
Maxime
The reason for choice is that it keys off a kernel command line
parameter that is inconsistent across distributions.
For example Ubuntu uses "splash", Fedora used "rhgb" etc.
Even the plymouth userspace maintains a list for it's behaviors of what
parameters to look for to start at bootup. So the obvious alternative
is to clone that list in the kernel.
