LKML Archive on
help / color / mirror / Atom feed
From: Vaidyanathan Srinivasan <>
To: Soeren Sonnenburg <>
Cc: Linux Kernel <>
Subject: Re: prioritize PCI traffic ?
Date: Mon, 15 Jan 2007 20:23:30 +0530	[thread overview]
Message-ID: <> (raw)
In-Reply-To: <1168869630.15294.44.camel@localhost>

Soeren Sonnenburg wrote:
> On Mon, 2007-01-15 at 19:23 +0530, Vaidyanathan Srinivasan wrote:
>> Soeren Sonnenburg wrote:
>>> Dear all,
>>> is it possible to explicitly tell the kernel to prioritize PCI traffic
>>> for a number of cards in pci slots x,y,z ?
>>> I am asking as severe ide traffic causes lost frames when watching TV
>>> using 2 DVB cards + vdr... This is simply due to the fact that the PCI
>>> bus is saturated...
>> How do you know that the bus is saturated?
> I simply dd if=/dev/sd? of=/dev/null from four brand new sata-harddisks.
>> Are you streaming data to/from the ide hard disks/CDROM?
> yes.
>> Do you have DMAs 'ON' for the hard disks?
> yes.


>> Is everything just fine if there are no IDE traffic?
> yes.


>> Are you running 2.6 kernel with preempt 'ON'?

Try CONFIG_PREEMPT=y it may help. It made the system more responsive
for me.

>> Are all hardware on the same IRQ line? (shared interrupts)
> no: libata devices are on IRQ 16 and DVB devices on IRQ 20

Watch the interrupt count (/proc/interrupts) and interrupt rate,
that may give some clue on what is happening on the PCI bus.

>>> So, is any prioritizing of the PCI bus possible ?
>> The drivers + application indirectly can control priority on the
>> bus.  Just reduce the priority of the application that uses IDE and
>> see if adjusting nice values of applications can change the scenario.
> That unfortunately did not help... no change...

33Mhz 32-bit PCI bus on typical PC can do around 100MB/sec... watch
your individual disk transfer rates and data rate to each DVB.
dd from 4 new SATA drive will probably fillup the bus.  However,
depending on your motherboard, the integrated SATA controllers may
be on a different PCI bus than the one for the slots.

You can watch the lspci output to determine which bus the devices
are located.


> Soeren
> --
> Sometimes, there's a moment as you're waking, when you become aware of
> the real world around you, but you're still dreaming.

  reply	other threads:[~2007-01-15 14:54 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2007-01-15 11:07 Soeren Sonnenburg
2007-01-15 12:01 ` Andreas Mohr
2007-01-15 12:07   ` Soeren Sonnenburg
2007-01-15 13:53 ` Vaidyanathan Srinivasan
2007-01-15 14:00   ` Soeren Sonnenburg
2007-01-15 14:53     ` Vaidyanathan Srinivasan [this message]
2007-01-17 20:22       ` Hans-Peter Jansen
2007-01-27  0:32         ` Lee Revell
     [not found] <>
2007-01-15 14:24 ` Robert Hancock

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \ \ \ \ \
    --subject='Re: prioritize PCI traffic ?' \

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).