RE: [PATCH net v6 1/4] net: ethernet: oa_tc6: Protect skb pointer used by two different kernel instances

[Selvamani Rajagopal <Selvamani.Rajagopal@onsemi.com>]

Andrew, Paolo,

> 
> > I think this is giving a warning that when net is merged to net-next,
> > which happens most Thursdays, there is going to be a conflict. How bad
> > is that conflict? If it is trivial, don't worry, it will get
> > handled. If the conflict resolution is not obvious, maybe you can look


After little investigation, looks like it is trivial issue and mainly it can't be handled by 
rearranging/changing the patches. Issue is the context differs between "net" and "net-next" branch. 
I am hoping that this can be resolved while merging.

In the following example, "net" branch has INT_MASK0_ALL_INTERRUPTS. But it is changed to
OA_TC6_INT_MASK0_ALL_INTERRUPTS in net-next branch. So, no matter how I move the code around,
this will not apply cleanly.

static void oa_tc6_disable_traffic(struct oa_tc6 *tc6)
 {
        u32 regval = INT_MASK0_ALL_INTERRUPTS;
+       struct sk_buff *skb;


Second place, STATUS0_RX_BUFFER_OVERFLOW_ERROR in net differs from 
OA_TC6_ STATUS0_RX_BUFFER_OVERFLOW_ERROR that is in net-next branch. 

        if (FIELD_GET(STATUS0_RX_BUFFER_OVERFLOW_ERROR, value)) {
-               tc6->rx_buf_overflow = true;
-               oa_tc6_cleanup_ongoing_rx_skb(tc6);
+               oa_tc6_look_for_new_frame(tc6);


> > at it now, and provide the answer?
> 
> Out of 24 hunks, 2 fails. Let me check if there is a way to resolve those
> by splitting the changes in to smaller patches.
> 
> >
> > Andrew