Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 8 Jul 2011 01:22:08 +0300
From:      George Liaskos <geo.liaskos@gmail.com>
To:        Kostik Belousov <kostikbel@gmail.com>
Cc:        x11@freebsd.org
Subject:   Re: Intel GPU kernel driver
Message-ID:  <CANcjpOD_LM=uT%2BrPW=9-WTrRVjJOh7cGuenXi0-tTsKdyCbC=Q@mail.gmail.com>
In-Reply-To: <20110630162315.GV48734@deviant.kiev.zoral.com.ua>
References:  <20110630162315.GV48734@deviant.kiev.zoral.com.ua>

next in thread | previous in thread | raw e-mail | index | archive | help
On Thu, Jun 30, 2011 at 7:23 PM, Kostik Belousov <kostikbel@gmail.com> wrote:
> [Please remove current@ when replying.]
>
> I created the first code drop for the ongoing GEM/KMS project. Please
> note that this is not an end-user release, and even _not_ a call for
> testing. The project is not finished yet, and I expect quite more
> efforts from me even after the scheduled project end, and from ports/x11
> people, before the driver and usermode infrastructure will be ready for
> the general public consumption.
>
> That said, the patch is only of use for you now if you want to review,
> debug or otherwise help the project. The driver is known to be unstable,
> some parts are missing, some (esp. VM changes) are under the discussion
> and propably will be changed.
>
> If you have fix or useful bug analisys or suggestions for improvements,
> you are welcome. I will not answer to the support requests for this
> code now, please do not waste your time asking for it.
>
> The pointers to the patches, useful hints for debugging and bug
> reporting, and some notes are at the http://wiki.freebsd.org/Intel_GPU.
> I will maintain this page further.
>
> Current patch is ~50KLOC, it took quite an efforts to bring the code to
> the state where there is something to debug. Thanks for everybody who
> waited for it, and please be patient while the further work is done.
>

Hello Kostik,

I am trying to get my gm45 [0x2a428086] working with your
patches and i found a bug.

A call to agp_gen5_get_gtt_total_entries returns ENXIO, what i
found is that in my case (gcc1 & AGP_G4x_GCC1_SIZE_MASK) == 0xb00.

I did a cross check with Linux and there isn't such a valid value defined
but the Linux code does not error on default case, it continues with our
agp_i965_get_gtt_total_entries() equivalent.

After this change, X starts but the image has artifacts. It's progress :)
Thank you for your work!

xorg log: http://pastebin.com/a0zak5LY



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?CANcjpOD_LM=uT%2BrPW=9-WTrRVjJOh7cGuenXi0-tTsKdyCbC=Q>