Re: [libvirt] PATCH: 9/12: HAL/DevitKit node device impl

2008-11-20 Thread Daniel P. Berrange
On Fri, Nov 14, 2008 at 02:57:09PM -0500, David Lively wrote: +(void)get_int_prop(ctx, udi, pci.vendor_id, (int *)d-pci_dev.vendor); +if (get_str_prop(ctx, udi, pci.vendor, d-pci_dev.vendor_name) != 0) +(void)get_str_prop(ctx, udi, info.vendor,

Re: [libvirt] PATCH: 9/12: HAL/DevitKit node device impl

2008-11-14 Thread Daniel Veillard
On Thu, Nov 13, 2008 at 05:33:20PM +, Daniel P. Berrange wrote: This is the main implementation of the local device enumation driver. The main change since David's patch is that we hav a single nodedevRegister() API call, instead of initializing HAL/DevKit separtely. This was needed to

Re: [libvirt] PATCH: 9/12: HAL/DevitKit node device impl

2008-11-14 Thread Mark McLoughlin
On Thu, 2008-11-13 at 17:33 +, Daniel P. Berrange wrote: This is the main implementation of the local device enumation driver. The main change since David's patch is that we hav a single nodedevRegister() API call, instead of initializing HAL/DevKit separtely. This was needed to make the

Re: [libvirt] PATCH: 9/12: HAL/DevitKit node device impl

2008-11-14 Thread David Lively
On Fri, 2008-11-14 at 13:41 +, Mark McLoughlin wrote: The DeviceKit implementation seems to be much more of a work-in-progress than the HAL implementation - maybe disable it by default in configure? (i.e. with_devkit=no instead of with_devkit=check) That's a good idea. I don't think the

Re: [libvirt] PATCH: 9/12: HAL/DevitKit node device impl

2008-11-13 Thread Daniel P. Berrange
This is the main implementation of the local device enumation driver. The main change since David's patch is that we hav a single nodedevRegister() API call, instead of initializing HAL/DevKit separtely. This was needed to make the dlopen() support easier to implement. Functionally the end result