[ewg] Re: kernel_addons patch for ipath support

[Bryan O'Sullivan]

Michael S. Tsirkin wrote:
> Okay, but where does the following come from, and what does it do?

It's Hoang-Nam's hvcall patch that somehow ended up in the wrong 
directory.  I thought I'd deleted all bad copies of it from my copy of 
Vlad's tree, but apparently not.

> Also: you sure linux/device.h linux/mm.h aren't needed here?
> 
>> diff -urN ofed_1_2/kernel_addons/backport/2.6.5_sles9_sp3/include/linux/dma-mapping.h infinipath-drivers/kernel_addons/backport/2.6.5_sles9_sp3/include/linux/dma-mapping.h

Not on ppc64 or x86_64, at least.

> The following looks like it might break mthca - what will provide
> pci_disable_msi if you remote it from here?
> 
>> diff -urN ofed_1_2/kernel_addons/backport/2.6.5_sles9_sp3/include/linux/pci.h infinipath-drivers/kernel_addons/backport/2.6.5_sles9_sp3/include/linux/pci.h

Those definitions are already present in linux/pci.h on SLES9 SP3, so 
they actually break the build due to repeated definition.  That's why I 
removed them.

	<b