On Tue, Jul 21, 2020 at 12:05:18AM +0300, Vladimir Oltean wrote: > > Yes, the problematic cases have to do with stacked PHCs (DSA, PHY). The > pattern is that: > - DSA sets SKBTX_IN_PROGRESS Nit: DSA should _not_ set this bit (but a PHY/MII device should). > - calls dev_queue_xmit towards the MAC driver > - MAC driver sees SKBTX_IN_PROGRESS, thinks it's the one who set it > - MAC driver delivers TX timestamp > - DSA ends poll or receives TX interrupt, collects its timestamp, and > delivers a second TX timestamp > So it's very far from obvious that setting this flag is 'the prevailing > convention'. For a MAC driver, that might well be, but for DSA/PHY, > there seem to be risks associated with doing that, and driver writers > should know what they're signing up for. The flag only exists to prevent the stack from delivering SW time stamps when HW time stamps are active. If an interface doesn't provide SW time stamps (like DSA interfaces), then there is no need to set the flag. For MAC and PHY/MII time stamping, they must co-operate, meaning that the MAC driver must be prepared to deal with the fact that the PHY/MII might set this flag. Many MAC drivers don't do this correctly, but there are very few PHY/MII actually in use, and so very few authors of MAC drivers have paid attention to this. Thanks, Richard
