Re: Regression in -current

[Chavdar Ivanov <ci4ic4%gmail.com@localhost>]

On Wed, 19 Jul 2023 at 13:23, Chavdar Ivanov <ci4ic4%gmail.com@localhost> wrote:
>
> Hi,
>
> Among others, I have an HP Envy 17 laptop with Intel 530 graphics, as
> well as NVidia GeForce 980M. I have been running NetBSD-current on it
> since I bought it seven years ago, without much trouble. nouveau0 was
> recognized, but never worked, Intel graphics used to work just fine
> with accelerated OpenGL etc., dri2 was fine, mode-switching and font
> loading on bootstrap OK.  It boots in EFI mode, I have to manually
> select the requisite .efi file to do it.
>
> The last kernel I am able to boot on it is from the 12th of July;
> subsequent ones hard-reset immediately after the boot menu. I was able
> to boot the old kernel initially, and, thinking that it is a problem
> with the latest updates to the nouveau driver and I am perhaps missing
> the right firmware files, I added 'userconf=disable nouveau*' to
> /boot.cfg. This didn't make any difference - the system resets
> immediately - but if now I select the kernel which used to boot OK,
> after the initial kernel messages, upon the mode switch, the screen
> gets corrupted - as if it is zoomed out four times and then repeated
> close to the top of the screen. Since then, I am unable to boot NetBSD
> on it (I will repair it using a usb stick later). On top of that, the
> system has a second monitor conected via HDMI, which worked fine
> before, still left as a mirror to the original screen. After the above
> attempt the HDMI port appeared completely out of order and stopped
> working under Windows 11 and Linux... I had to reset the laptop to the
> point of disconnecting the battery, after which the HDMI port started
> to work again.

I was able to boot 10.99.6 from today by using the live system - which
is, for some reason, only BIOS based. It wouldn't boot from the USB
stick, as it does not see USB3 device past the kernel boot, but it
showed me the original disk slice, which booted further without a
problem. So it is some issue related to the EFI boot on this machine.
I will try tomorrow with amd64-install.img, which is EFI, to see if it
will fail the same way.

>
>
> Any ideas as to what may have caused this?
>
>
> Chavdar
>
>
> --
> ----



-- 
----