[PATCH] yenta: irq-routing for TI bridges...again

[Pavel Roskin]

On Tue, 9 Mar 2004, Daniel Ritz wrote:

> On Tuesday 09 March 2004 23:01, David Hinds wrote:
> > The pcmcia-cs drivers have always tested the PCI interrupt as well as
> > the ISA ones; this caught a not-insignificant number of BIOS problems
> > related to incorrect interrupt routing tables, in addition to bridge
> > configuration issues.  The testing itself never caused any problems,
> > as far as I know.
>
> thanks for acknowledge. btw guess where the idea (and code parts)
> comes from :)

OK, good to know that probing is not an issue.  Sorry for making you so
angry, but some questions just had to be asked.

I have tested the patch and it's working for me (TI PCI1410).

It would be nice to clean it up and integrate it better into the
surrounding code.  I believe patches to the PCI ID database should be
submitted separately to different people.  Be prepared that the PCI ID
database may be patches after the PCMCIA part.  Many drivers provide
replacement definitions, e.g.

#ifndef PCI_VENDOR_ID_APPLICOM
#define PCI_VENDOR_ID_APPLICOM                0x1389
#endif

Kernel addresses should not be printed unless you expect the
subsequent stack dump and panic.  There are better ideas for the prefix,
such as the source file name, function name and pci_name().

All hex numbers should be preceded by "0x" unless it's impractical.  It
may be better to use words rather than numbers - not everyone knows
meaning of bits in mfunc.

Please make a single patch (except the PCI ID changes), look at it, think
what can be done better, then post it.  I'll be glad to test it on more
systems.

-- 
Regards,
Pavel Roskin