Comment # 28 on bug 835975 from
(In reply to Gary Greene from comment #26)
> I'm thinking that the issue is in the lead up to the following block of code
> in xf86platformBus.c, starting at line 382:
> 
> static Bool
> probeSingleDevice(struct xf86_platform_device *dev, DriverPtr drvp, GDevPtr
> gdev, int flags)
> 
> When I loaded this in kdbg over a remote X connection on SSH and stepped
> through the code, the comparison fell apart with the following values:
> 
> drvp->supported_devices.device_id: int 12624 hex value: 0x3150
> dev->pdev->device_id: int 38275 hex value: 0x9583
> 
> Obviously these will fail. 

This is a whole long list of supported devices. Your's definitely in the list.
The code will iterate over the list so it should find this one matching entry.
What you can do once you are there is to add another breakpoint to
doPlatformProbe:
(gdb) break doPlatformProbe
and let the program run until the end of the present function (ie
probeSingleDevice)
(gdb) finish
This way, you don't have to step thru the long list, will end up in
doPlatformProbe() (this is called one a hit is found) but don't run off if none
is found.

> It looks like the supported_devices entry is being populated with the wrong
> value somewhere along the way.

I guess we can do this as well.


You are receiving this mail because: