Supermicro Hardware and FreeBSD

[Ted Mittelstaedt]

> -----Original Message-----
> From: owner-freebsd-questions at freebsd.org
> [mailto:owner-freebsd-questions at freebsd.org]On Behalf Of Tm4528 at aol.com
> Sent: Thursday, January 06, 2005 10:54 AM
> To: tedm at toybox.placo.com
> Cc: questions at freebsd.org
> Subject: Re: Supermicro Hardware and FreeBSD
>
>
> In a message dated 1/6/05 2:10:28 AM Eastern Standard Time,
> tedm at toybox.placo.com writes:
> >> Do you really have no contacts at SM or Dell? What kind of a
> development
> >> org has no contacts with major vendors?
> >
> >It's not a question of not having contacts.  It's a question of
> >actually defining the problem in a way that a developer can get
> >a fix on it.
> >
> >Currently, this is done with the PR mechanism on FreeBSD.org.  Doing
> >a search of this shows only PR i386/72579, which claims FreeBSD 4.X
> >doesen't work at all on this chipset, which is contrary to what the
> >OP was saying.
> Thats NOT contrary to what anyone was saying (Im not sure who OP is).

It surprises me that as you imply that you have been around the Internet
for a long time that you don't know that OP stands for Original Poster.

> You're just too busy writing people off as trolls to read whats
> written. He
> said it didnt work at all with the 7520 MB,

No, the PR author said that 5.3 "works OK".  He said that only 4.10 doesen't
work
with the NEW E7520 chipset but that 4.9 works with the OLD E7520 chipset.
The implication is that 5.3 works fine on BOTH the NEW and the OLD
chipset.

Note also that the PR in question was filed OCTOBER 12 which was fully
a MONTH before 5.3 was actually released.  Thus the person filing the
PR was running a BETA version of 5.3.

> and he said his OLD 533Mhz
> MB was faster in 4.9 than the new MB was in 5.3 so it made no sense
> to upgrade.
>

Boris Spirialitious is the poster that said that, and he WASN'T the one
who filed PR i386/72579.  Boris did NOT say WHAT version of 5.3 he was
running - was it RELEASE or an earlier beta?  Nor did he post results
of any testing program that showed a speed difference.

> So, apparently, doing a PR doesnt work, since the PR you cited has
> been largely ignored for 3 months.

PR's that are NOT filed well are going to be ignored.  Please, read the
following:

http://www.freebsd.org/doc/en_US.ISO8859-1/articles/problem-reports/article.
html

Here's a list of things that both Boris and the PR author (and the person
posting a followup to the PR) HAVEN'T done:

1) Boris hasn't posted a followup to the PR explaing what he's seeing.
2) Missing exact versions of FreeBSD in both cases
3) Missing model# of Boris's motherboard.
3a) Poster of followup to PR hasn't included model# of his board.  He's also
using a
lame excuse for not running 5.3 but that's a side issue.
4) dmesg output from the 5.3 system that has the new E7520 chip and can't
run 4.10 is missing
5) dmesg output from 4.9 system that has the old E7520 chip and CAN run 4.X
is missing.
6) Identifying numbers off the E7520 chips or from the two machines BIOS are
missing
7) No posting of whether 4.11 RC runs on the new chipset or not

And overall in summary I have to say this - BOTH Boris and the author of
this PR
state that FreeBSD 5.3 runs on their motherboards, Boris says also that 4.10
runs
on his MB, the PR author says that 4.10 doesen't run on ONE of his
motherboards.
The PR author didn't even say he was having a problem with this, or why he
wanted to
run 4.X instead of 5.3 - although considering the PR was filed about a month
prior
to the release of 5.3 I'll assume the PR author just didn't want to run beta
code.
Only one followup poster to the PR complained about having to run 5.3.  And,
Boris complained.

The one followup post to the PR stated:  "We cannot run 5.x as it is
considered
insecure." which is IMHO a big pile of baloney - who considers it insecure
and why?

And Boris hasn't posted any of his slowness complaint to this PR, so the PR
lacks that.

> So whats else do you recommend, Ted?
>
> PS: the 7520/7530 is required for use of Intel's newest CPUs, so its
> not some random chipset. It should be way higher in the list of priorities
> than the peripheral "fixes" noted for 4.11.

Boris didn't say that 4.11 RC didn't work on his motherboard, he said 4.10
didn't
work.  How do you know 4.11 isn't going to work on this chipset when nobody
has
even tried it?  Why the heck do you think the release team even bothers with
RC releases to begin with?  It's so people like Boris can post their
showstopper
bugs.  Since that hasn't happened here, just a bunch of hand-waving bitching
in -questions, why do you even bother asking why nothing is happening?

This PR is kind of equivalent to someone posting
"My Windows 2000 operating system doesen't work on my new SuperMicro
motherboard
and I want Microsoft to fix it" when the SuperMicro motherboard in question
was designed and manufactured AFTER Windows 2K was released - although to
the
PR authors credit, he wasn't demanding that 4.10 be modified to work on his
hardware.  Perhaps we all should ask why it is OK for SuperMicro to release
a
motherboard that is incompatible with the existing FreeBSD versions?

Ted