Re: [PATCH 1/6] Rewrite MSI-HOWTO

[Matthew Wilcox <willy@linux.intel.com>]

Thanks for the review!

On Tue, Feb 24, 2009 at 12:00:03PM -0800, Randy Dunlap wrote:
> Matthew Wilcox wrote:
> > +A Message Signalled Interrupt is a write from the device to a special
> 
> Consistent spelling, please.

Ah yes, I thought I'd fixed that.  I'll follow the PCI spec and use one 'l'.

> > +To support this, the kernel must call each interrupt handler associated
> > +with an interrupt which leads to reduced performance for the system as
> 
>    with an interrupt,

Ack.

> > +Device drivers should normally call this function once per device
> > +during the initialization phase.
> 
> Consistent isation/ization (or ised/ized), please.

The Shorter OED is quite amusing on this point.  It lists 'Initialise' as a
variant, but in the main definition under 'Initialize', the example it
gives of the computer usage uses 'initialise'.  I can only conclude that
a foolish consistency is the hobgoblin of little minds ;-)

> > +4.6 How to tell whether MSI/MSI-X is enabled on a device
> > +
> > +Using lspci -v (as root) will show some devices with "Message Signalled
> 
> Oh gosh, does lspci misspell it?

It used to.  The good news is that recent pciutils now just use MSI
instead of spelling it out.  As a bonus, the capability now fits on one
line.

> > +Interrupts" and others with "MSI-X".  Each of these capabilities have an
> 
>                                          Each ..................... has an

You're correct.

> > +'Enable' flag which will be followed with either "+" (enabled) or "-"
> > +(disabled).


> > +including a full lspci -v so we can add the quirks to the kernel.
> 
>    I would say      "lspci -v" ...

Good idea.

> > +Then, lspci -t gives the list of bridges above a device.  Reading
> 
>          "lspci -t"

Likewise.

> > +It is also worth checking whether the device driver supports MSIs.
> > +
> 
> Unneeded ending blank line (nit).

Heh.  I'll take it out.

May I add your reviewed-by tag?