[ofa-general] Folks is this a known problem / already fixed ?

[Eli Cohen]

On Fri, 2008-05-16 at 14:28 -0400, Richard Frank wrote:
> We see the following failure for our ConnetX HCAs.. with 1.3.1 Daily 
> 20080512 done on vanilla OEL5U1.
> 
> They are failing to initialize with the following:
> 
> mlx4_core: Mellanox ConnectX core driver v1.0 (February 28, 2008)
> mlx4_core: Initializing 0000:05:00.0
> mlx4_core 0000:05:00.0: Failed to initialize queue pair table, aborting.
> mlx4_core 0000:05:00.0: Failed to initialize queue pair table, aborting.
> mlx4_core: probe of 0000:05:00.0 failed with error -16
> 
> And lspci shows:
> 
> 05:00.0 InfiniBand: Mellanox Technologies MT25418 [ConnectX IB DDR] (rev 
> a0)
>        Subsystem: Mellanox Technologies MT25418 [ConnectX IB DDR]
>        Flags: fast devsel, IRQ 169
>        Memory at fcc00000 (64-bit, non-prefetchable) [disabled] [size=1M]
>        Memory at fff000000 (64-bit, prefetchable) [disabled] [size=8M]
>        Memory at fcbfe000 (64-bit, non-prefetchable) [disabled] [size=8K]
>        Capabilities: [40] Power Management version 3
>        Capabilities: [48] Vital Product Data
>        Capabilities: [9c] MSI-X: Enable- Mask- TabSize=256
>        Capabilities: [60] Express Endpoint IRQ 0
> 
Can you send the output of lspci for the bridge connecting the ConnectX
with the upstream PCI bus? I guess the problem would be that the bridge
blocks memory writes to ConnectX's UAR area thus causing a failure to
arm the EQ and eventually resulting in failure to load the driver. Now
it could be a failure of the kernel to configure the bridge properly.
Could you try with the latest kernel?