On 10/01, Joe Damato wrote: > Greetings: > > Welcome to RFC v4. > > Very important and significant changes have been made since RFC v3 [1], > please see the changelog below for details. > > A couple important call outs for this revision for reviewers: > > 1. idpf embeds a napi_struct in an internal data structure and > includes an assertion on the size of napi_struct. The maintainers > have stated that they think anyone touching napi_struct should update > the assertion [2], so I've done this in patch 3. > > Even though the assertion has been updated, I've given the > cacheline placement of napi_struct within idpf's internals no > thought or consideration. > > Would appreciate other opinions on this; I think idpf should be > fixed. It seems unreasonable to me that anyone changing the size of > a struct in the core should need to think about cachelines in idpf.
[..] > 2. This revision seems to work (see below for a full walk through). Is > this the behavior we want? Am I missing some use case or some > behavioral thing other folks need? The walk through looks good! > 3. Re a previous point made by Stanislav regarding "taking over a NAPI > ID" when the channel count changes: mlx5 seems to call napi_disable > followed by netif_napi_del for the old queues and then calls > napi_enable for the new ones. In this RFC, the NAPI ID generation > is deferred to napi_enable. This means we won't end up with two of > the same NAPI IDs added to the hash at the same time (I am pretty > sure). [..] > Can we assume all drivers will napi_disable the old queues before > napi_enable the new ones? If yes, we might not need to worry about > a NAPI ID takeover function. With the explicit driver opt-in via netif_napi_add_config, this shouldn't matter? When somebody gets to converting the drivers that don't follow this common pattern they'll have to solve the takeover part :-)
