[LAU] HW upgrade - new computer

Len Ovens len at ovenwerks.net
Wed Apr 30 23:41:23 UTC 2014


On Wed, 30 Apr 2014, Clemens Ladisch wrote:

> Len Ovens wrote:
>> PCI bridge: Intel Corporation 8 Series/C220 Series Chipset Family PCI Express Root Port #3 (rev d5)
>> 00:1c.3 PCI bridge: Intel Corporation 82801 PCI Bridge (rev d5)
>>
>> I am not sure which is which. The first may be the express port.
>
> The ASM1083 (which the manual says you have) should be listed as "ASM1083".
> What is the output of "lspci -t"?

len at music:~$ lspci -t
-[0000:00]-+-00.0
            +-01.0-[01]--
            +-02.0
            +-14.0
            +-16.0
            +-1a.0
            +-1c.0-[02]--
            +-1c.2-[03]----00.0
            +-1c.3-[04-05]----00.0-[05]--+-00.0
            |                            \-02.0
            +-1d.0
            +-1f.0
            +-1f.2
            \-1f.3

00:1c.3 would be the PCI bridge then as my two audio cards are listed as 
05:00.0 and 05:02.0. I would guess that the third PCI slot would be 01. I 
see that the bridge is listed as [04-05] I am guessing that means 04 is 
the PCIe side and 05 is the PCI side? 04 is:
04:00.0 PCI bridge: ASMedia Technology Inc. ASM1083/1085 PCIe to PCI 
Bridge (rev 03)

>> On Mon, 28 Apr 2014, Clemens Ladisch wrote:
>>> Anyway, PCIe devices that support MSI use a completely separate
>>> interrupt line.  (USB3 ports, graphics, and built-in audio should be
>>> harmless then.)
>>
>> I have no PCIe devices at this time.
>
> All your motherboard devices are likely to be PCIe.

Not what I meant  :) (I was meaning only that I had nothing in the slots) 
So it sounds like you are saying that the PCIe devices are assigned an irq 
to be compatible with the PCI standard, but most PCIe devices don't use 
the irq having another way to get the OS attention. Looking at 
/proc/interrupts, I see that ahci (which from lspci seems to be the SATA 
controller) shows irq 42 even though the book says it is sharing an irq 
with PCI slot 1 which has an ens1370 at irq 19. So I would guess that this 
irq 42 is a "soft" irq? (and why do I have an ISA bridge?) So that would 
follow that irq 41 to 45 would be onboard soft irqs... They are called 
PCI-MSI-edge.

I would guess that if I installed a PCIe card that was PCI HW with a 
PCIe->PCI bridge on the card to update it, that it would still use the old 
irq and could clash with one of my other PCI audio cards. I can't see me 
getting anything other than an ide, audio or firewire card. (I have a 
number of ide drives around, but may instead put them in the server and 
mount over network)

This is starting to make sense. I have seen a number of people have xrun 
trouble as soon as they got new a MB that were faster than trouble free 
slower/older HW, but my experience has been totally the opposite. The only 
thing that has given any xruns so far as I can see is cpu speed changes 
when using ondemand (I already don't have hyperthreading to worry about). 
I did choose my MB/CPU based on what I felt would be good for audio, not 
for a fast desktop (though I have gained there too). For example, I chose 
the i5 cpu over the i7 because I didn't want hyprthreading, not because it 
was cheaper.

--
Len Ovens
www.ovenwerks.net



More information about the Linux-audio-user mailing list