On Tue, Sep 19, 2023 at 03:36:30PM +0100, Edward Cree wrote:
On 19/09/2023 12:10, Martin Habets wrote:
quoted
On Tue, Sep 12, 2023 at 03:21:39PM +0100, edward.cree@amd.com wrote:
quoted
+ int (*create_rxfh_context)(struct net_device *,
+ struct ethtool_rxfh_context *ctx,
+ const u32 *indir, const u8 *key,
+ const u8 hfunc, u32 rss_context);
To return the rss_context this creates shouldn't it use a pointer to
rss_context here?
No, the whole point of this new API is that the core, not the
driver, chooses the value of rss_context. Does the commit
message not explain that sufficiently?
Your commit describes it correct, but I had my brain wired the wrong
way around. My mistake.
Martin
(If you look at Patch #7 you'll see that sfc doesn't even use the
value, though other drivers might if their HW has a fixed set of
slots for RSS configs.)
-ed