Re: [PATCH v5] i40e: Look up MAC address in Open Firmware or IDPROM

[Jeff Kirsher <jeffrey.t.kirsher@intel.com>]

[-- Attachment #1: Type: text/plain, Size: 1711 bytes --]

On Thu, 2015-12-03 at 10:48 -0500, Sowmini Varadhan wrote:
> The patch here: http://patchwork.ozlabs.org/patch/540218/
> is marked "Awaiting Upstream", I think that means it has to first
> show up in some other repo first (which one?)?
> 
> On addiitonal testing, we found a bug in the patch: if
> we did not find the macaddr from Open Firmwre or IDPROM (i.e.,
> defaults
> were ok) then you dont want to be doing i40e_macaddr_init again, else
> you will get a failure like this (truncated dump_stack shown
> below)
> 
> [ 8127.050926] WARNING: CPU: 18 PID: 878 at kernel/irq/manage.c:1346
> __free_irq+0x9f/0x230()
> [ 8127.050927] Trying to free already-free IRQ 177
>   :
> [ 8127.051013]  [<ffffffffa043ddd0>]
> i40e_clear_interrupt_scheme+0xb0/0xc0 [i40e]
> [ 8127.051018]  [<ffffffffa044b538>] i40e_probe.part.64+0x1018/0x1320
> [i40e]
>   :
> [ 8127.051057]  [<ffffffffa044b862>] i40e_probe+0x22/0x30 [i40e]
> 
> I can think of a couple of ways to fix this- one (the ugly
> way) is to ifdef the i40e_macaddr_init invocation for CONFIG_OF or
> CONFIG_SPARC
> only. Another way to solve this is to track some bit in struct
> i40e_hw that 
> indicates that the macaddr is not the default, thus
> i40e_macaddr_init() should
> be called before register_netdev only if that bit is set. (Dont know
> if there are cache-line considerations that exist in i40e_hw that
> need to be taken into account for the second version)
> 
> In order to send a fix out for review, what should I clone? 
> should I just apply the patch/540218 to net-next and send the update?

I will drop your current patch in my next-queue tree (dev-queue branch)
and will await an updated patch.  

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 819 bytes --]