Bad way then [LWN.net]

Monta Vista does hold your hand

Posted May 7, 2008 7:51 UTC (Wed) by rvfh (subscriber, #31018) [Link]

I have had to work with Montavista in a previous job because they were writing drivers for us,
and they do provide tools and doc. They also found bugs in our hardware and came up with
solutions.

I can see that in two sentences I have sold their stuff better than with that FUD stuff from
Ready, and that's real technical stuff, stuff the customer is asking for!

All in all, we did not really need it for the distrib and tools, and went on using the one I
was maintaining on a previous version (not a product so no real security issue at stake), but
for a company with no Linux expert around or just no time for that, I can see clear benefits
in going with a MV or now WindRiver.

Thanks Jon for pointing this BTW, it's very instructive on how FUD is always bad, and might
serve interests opposite to yours ITE.

Monta Vista does hold your hand

Posted May 7, 2008 9:20 UTC (Wed) by man_ls (subscriber, #15091) [Link]

Maybe all this FUD is directed to pointy-haired bosses, and maybe it is effective with them. But any possible positive effect is offset by the negative effect on technical people. Which side has more decision power (management or techies) depends on how technical a company is; in the end, if only Dilbert-esque companies use your product then your reputation will probably suffer.

Monta Vista does hold your hand

Posted May 7, 2008 14:01 UTC (Wed) by sepreece (subscriber, #19270) [Link]

Ready's comments are largely aimed at companies that build embedded products and have been
using specialized real-time OSs, but are now looking at Linux and thinking "Hmm - get rid of
those licensing costs and I could make more money per device."

Those folks are used to buying an OS from a vendor that provides support and it's reasonably
important to make sure they understand that while you don't need to license Linux, you can't
expect to use it in products without some kind of support costs - either an in-house
development OS-development team or a specialist embedded-Linux distributor.

At a previous employer we determined that using an embedded distributor (MontaVista, in fact)
cost less than it would cost to staff the work we relied on them to do.

[Aside to Jon - you make fun of "the assertion that MontaVista and Wind River are "Linux's
strongest proponents." However, the actual assertion was that they were "embedded Linux's
strongest proponents", which is a claim MV could probably make a reasonable argument for.]

Data

Posted May 7, 2008 14:08 UTC (Wed) by alex (subscriber, #1355) [Link]

I would like to see how much of mvista's code is in the mainline kernel and how much is in
their own distro kernels. Granted they have a better record than WR but it still doesn't look
like much in the grans scheme of things:


[linux-2.6.git] >git-log --pretty=format:"%ae" | wc -l
87767
[linux-2.6.git] >git-log --pretty=format:"%ae" | grep "mvista" | wc -l
569

Data

Posted May 7, 2008 14:21 UTC (Wed) by rvfh (subscriber, #31018) [Link]

The O(1) scheduler was developed by Robert Love while at Montavista IIRC.

Data

Posted May 8, 2008 21:43 UTC (Thu) by jzbiciak (✭ supporter ✭, #5246) [Link]

Huh? The O(1) scheduler was developed by Ingo Molnar while he was at RedHat.

Data

Posted May 9, 2008 7:08 UTC (Fri) by rvfh (subscriber, #31018) [Link]

http://www.linuxdevices.com/news/NS8778096476.html

O(1) != CONFIG_PREEMPT

Posted May 9, 2008 7:21 UTC (Fri) by jzbiciak (✭ supporter ✭, #5246) [Link]

I guess I missed where this is the O(1) scheduler? In fact, that press release is largely content free. It just says "It's Not RTLinux."

I think you're thinking of CONFIG_PREEMPT, not the O(1) scheduler.

Perhaps this patch (ingo-O1-sched/preempt-kernel-rml-2.4.18-rc1-ingo-K3-1.patch) is what confused you? It's Robert Love's preempt patch against Ingo Molnar's O(1) scheduler.... :-)