On 7/8/25 07:21, Johannes Berg wrote:
On Tue, 2025-07-08 at 06:58 -0700, Ben Greear wrote:
On 7/8/25 05:31, Johannes Berg wrote:
On Wed, 2024-12-18 at 15:25 -0800, greearb@xxxxxxxxxxxxxxx wrote:
From: Ben Greear <greearb@xxxxxxxxxxxxxxx>
For drivers that can report the tx link-id, account tx
stats against that link. If we cannot determine tx link,
then use deflink.
According to all the RX stats discussion [1] you need some changes here,
so I'd appreciate if you could take a look and rebase/resend.
[1] https://lore.kernel.org/linux-wireless/c22a9e7e-d0f7-477b-b732-c2454a0ac904@xxxxxxxxxxx/
And please, as I very frequently keep asking you, don't mix different
things into a single patch (such as 'rep_packets'/'rep_bytes' in this
patch). By insisting on not splitting your patches properly before
submitting them you're effectively saying your time is more important
than mine, and I don't appreciate that. All that achieves is that I
don't even want to look at your patches.
As far as I can tell, I split them properly. I added new counters in one
patch you reference above, and I named them as I think they should be named.
In another patch, I renamed existing variables with a commit message as to why.
This patch said:
For drivers that can report the tx link-id, account tx
stats against that link. If we cannot determine tx link,
then use deflink.
which I think we can agree should be only a refactoring/fix of the link
that the counters go on?
But you also have
+ /* Packets and bytes reported transmitted (per link) */
+ u64 rep_packets;
+ u64 rep_bytes;
which was probably meant to be in the third patch?
In today's linus tree, the code looks like this:
/* Updated from TX path only, no locking requirements */
struct {
u64 packets[IEEE80211_NUM_ACS];
u64 bytes[IEEE80211_NUM_ACS];
struct ieee80211_tx_rate last_rate;
struct rate_info last_rate_info;
u64 msdu[IEEE80211_NUM_TIDS + 1];
} tx_stats;
There are no counters for tx bytes or tx packets (I think the ACS classification
counters are already used for other purposes).
So I added new counters, and I wanted them named to indicate it is what is reported by
the driver, not something specified by the mac80211 stack.
Thanks,
Ben
johannes
--
Ben Greear <greearb@xxxxxxxxxxxxxxx>
Candela Technologies Inc http://www.candelatech.com
