Re: kern/54976: radeon framebuffer fails initialization in NetBSD-9

[Roy Bixler <rcbixler%nyx.net@localhost>]

On Sat, Feb 22, 2020 at 01:10:33PM +1100, matthew green wrote:
> Roy Bixler writes:
> > I tried that one, but I still got the "mixing old and new" message.  I
> > tried also downloading the firmware from kernel.org and I replaced the
> > entire "radeon" directory.  This boots with the original kernel.
> 
> OK, this is some sort of success.  :)
> 
> can you tell me the sizes/sha256 of the bad/good kaveri_mec2.bin?
> just so i can be sure i figure out a way to make this right with
> the right files.

They match.  The one I downloaded from you:

17024 / 0be34c61ce8d5c077d607fff205e6a13

The one from kernel.org:

17024 / 0be34c61ce8d5c077d607fff205e6a13

The difference is that the one from kernel.org contains a bunch of
lower-case files such as:

-rw-r--r--  1 rcb  users    8832 Jan 22 11:26 kaveri_ce.bin
-rw-r--r--  1 rcb  users    8832 Jan 22 11:26 kaveri_me.bin
-rw-r--r--  1 rcb  users   17024 Jan 22 11:26 kaveri_mec.bin
-rw-r--r--  1 rcb  users   17024 Jan 22 11:26 kaveri_mec2.bin
-rw-r--r--  1 rcb  users    8832 Jan 22 11:26 kaveri_pfp.bin
-rw-r--r--  1 rcb  users   10496 Jan 22 11:26 kaveri_rlc.bin
-rw-r--r--  1 rcb  users    4456 Jan 22 11:26 kaveri_sdma.bin
-rw-r--r--  1 rcb  users    4456 Jan 22 11:26 kaveri_sdma1.bin
-rw-r--r--  1 rcb  users  232752 Jan 22 11:26 kaveri_uvd.bin
-rw-r--r--  1 rcb  users  101072 Jan 22 11:26 kaveri_vce.bin

and upper case files:

-rw-r--r--  1 rcb  users    8576 Jan 22 11:26 KAVERI_ce.bin
-rw-r--r--  1 rcb  users    8576 Jan 22 11:26 KAVERI_me.bin
-rw-r--r--  1 rcb  users   16768 Jan 22 11:26 KAVERI_mec.bin
-rw-r--r--  1 rcb  users    8576 Jan 22 11:26 KAVERI_pfp.bin
-rw-r--r--  1 rcb  users   10240 Jan 22 11:26 KAVERI_rlc.bin
-rw-r--r--  1 rcb  users    4200 Jan 22 11:26 KAVERI_sdma.bin

whereas the original only has the upper-case files (with the exception
of the one you had me try.)

-- 
Roy Bixler <rcbixler%nyx.net@localhost>
"The fundamental principle of science, the definition almost, is this: the
sole test of the validity of any idea is experiment."
-- Richard P. Feynman