if idle is used as a boolean, why not be consequent about it and declare it as such also? To prevent future misuses with numeric calculations...

In D44092#1013944, @dim wrote:

KMSAN is @markj I believe? Though it looks like you're getting a clang assertion there. So what is the clang version being used?

• some compiler trickery to the correct struct member size

In D44362#1011765, @kib wrote:

I do not understand the change. You re-wrote the code as if the TCP_RFC7413 option is always set. This is not true, at least from the fast grepping of the sources.

thx

only minor whitespace nits. i'm ok with removing this macro mechanistically.

i've read through the TCP related changes, which look good to me. But that's only a minute part of this change. Others should comment on the remaining aspects of this major diff.

addressed all comments.

• sort prototype definitions alphabetically

• clean key material in the error path too.

Discussed transport

Can I ping @kib like this?

As at least two, maybe three different NIC vendors have similar capabilities, wouldn't it make sense to have global, driver independent sysctls?

• clean before free all crypto material

whitespace fix 2

whitespace fix

In D44092#1005867, @emaste wrote:

Makes sense to me. Have you tested make tinderbox?

MAIN e0dfecadf5946b6a8ad423be4eb1c14017b3a8a7 errors on make tinderbox with the same error:

In D44092#1005867, @emaste wrote:

Makes sense to me. Have you tested make tinderbox?

In D44092#1005876, @karels wrote:

Does -Wswitch require a default if all values of an enum are explicitly included? I don't know if that is common though.

Unfortunately, my box here is too small to build serveral universes.

Apparently I've only checked clang or gccc12 but ont gcc13 - where there are subtle differences in what these compilers consider acceptable;

I suspect this is not the final patch, but suffices for now to stop excessive inflation of cwnd on the first ACK when transistioning from slow-start to congestion avoidance.
The guidance given was that t_epoch should be reset at the time cc transitions from slow-start to cong-avoidance.
However, I would like to revisit this anew, once the major overhaul in D41715 is in...

• do max(1, div) instead max(maxseg, cwnd)/div

max(div, 1) prior of outer division.

• rebase main

• rebase main

• rebase main

• rebase main

• delta to D43355_sack_after_rto

• rebase to main

• rebase main

Better revert to the prior state (with tcp_timer_stop() in tcp_discardcb()) - D43855