Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 22 Aug 1995 12:13:56 -0700 (PDT)
From:      "Rodney W. Grimes" <rgrimes@gndrsh.aac.dev.com>
To:        se@ZPR.Uni-Koeln.DE (Stefan Esser)
Cc:        tinguely@plains.nodak.edu, current@freebsd.org
Subject:   Re: Matrox Meteor Video Capture Card Driver
Message-ID:  <199508221913.MAA07292@gndrsh.aac.dev.com>
In-Reply-To: <199508221312.AA01823@Sysiphos> from "Stefan Esser" at Aug 22, 95 03:12:58 pm

next in thread | previous in thread | raw e-mail | index | archive | help
> 
> In article <DDo1EM.oqq@ns1.nodak.edu>, Mark Tinguely writes:
> * Warning: the Meteor does not work with all PCI chipsets. A MERCURY chipset
>   motherboard is almost certain to NOT work with the Meteor. A NEPTUNE chipset
>   motherboard with a PCI video card may hang with heavy Meteor use. So it is
>   wise to use a "SATURN" or "TRITON" chipset motherboard and also limit the
>   number of PCI bus-mastering devices on that bus. We had difficulties getting
>   the Meteor capture card to work with systems that also use the NCR PCI SCSI
>   boards.  It is also wise to have 16 or more Megabytes of RAM in the computer.
> 
> Since I'm currently maintaining the PCI and NCR code,
> we might try to get the Matrox Meteor supported on 
> more motherboards ...

I doubt you are going to fix the ``hardware'' bugs in the Mercury
chipset that causes the problems with boards like the Meteor.
 
> Do you have a complete set of technical documents for
> the Meteor, especially regarding the PCI bus interface ?
> 
> I currently do not have much time to spend on this, but 
> if you have some idea what prevents the Meteor from 
> working, I might be able to change the PCI or NCR code 
> accordingly ...
> 
> There have been problems with PCI bursts of more than one 
> cache line of data with the NCR, but I had thought the 
> Saturn was one of the chip sets that need more attention
> than other Intel chip sets.

Saturn-I may have some problems here, don't know, don't have access to
a Saturn-I based machine.  Saturn-II seems to handle 4 bus masters
without any problem, and I suspect a Saturn-II can run the Meteor just
fine.  But then, doing video capture on a 486 class machine does not
make a whole lot of since given the CPU demand needed to do mpeg.

Of the Intel PCI chip sets the following is a list of brokenness from
worst to best and a short description of brokenness.

a)  Mercury - Cache coherency problems, especially if there are ISA
    bus masters behind the ISA to PCI bridge chip.  Hardware flaw, only
    known work around is to turn the cache off.

b)  Saturn-I (ie, 82424ZX at rev 0, 1 or 2) - write back cache coherency
    problems.  Hardware flaw, only known work around is to set the external
    cache to write-through mode.  Upgrade to Saturn-II.

c)  Saturn-II (ie 82424ZX at rev 3 or 4) - works fine, but many MB manufactures
    leave out the external dirty bit SRAM needed for write back operation.
    Work arounds are either run it in write through mode, or get the dirty
    bit SRAM installed.  (I have these for the ASUS PCI/I-486SP3G rev 1.6
    and later boards).

d)  Neptune - Can not run more than 2 bus master devices.  Admitted Intel
    design flaw.  Workarounds include don't run more than 2 bus masters,
    special hardware design to replace the PCI bus arbiter (appears on
    Intel Altair board and several other Intel server group MB's).  And
    of course Intel's official answer, move to the Triton chip set, we
    ``fixed it there''.

e)  Triton - No known cache coherency or bus master problems, chip set does
    not implement parity checking.  Workaround for parity issue - wait for
    Triton-II :-(.

f)  Triton-II - Unknown, not yet shipping :-(.



-- 
Rod Grimes                                      rgrimes@gndrsh.aac.dev.com
Accurate Automation Company                 Reliable computers for FreeBSD



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199508221913.MAA07292>