tcp: fix mid stream window clamp.
authorPaolo Abeni <pabeni@redhat.com>
Mon, 4 Dec 2023 16:08:05 +0000 (17:08 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Wed, 13 Dec 2023 17:45:07 +0000 (18:45 +0100)
commite4a2a4328c85f1830c8caf5bc8fa566057404fe6
tree605897ff85e0e2a574a3fe987ab65cd10bcdf51e
parentdc903ddc826f020baf193d8befb6c822baa530e3
tcp: fix mid stream window clamp.

[ Upstream commit 58d3aade20cdddbac6c9707ac0f3f5f8c1278b74 ]

After the blamed commit below, if the user-space application performs
window clamping when tp->rcv_wnd is 0, the TCP socket will never be
able to announce a non 0 receive window, even after completely emptying
the receive buffer and re-setting the window clamp to higher values.

Refactor tcp_set_window_clamp() to address the issue: when the user
decreases the current clamp value, set rcv_ssthresh according to the
same logic used at buffer initialization, but ensuring reserved mem
provisioning.

To avoid code duplication factor-out the relevant bits from
tcp_adjust_rcv_ssthresh() in a new helper and reuse it in the above
scenario.

When increasing the clamp value, give the rcv_ssthresh a chance to grow
according to previously implemented heuristic.

Fixes: 3aa7857fe1d7 ("tcp: enable mid stream window clamp")
Reported-by: David Gibson <david@gibson.dropbear.id.au>
Reported-by: Stefano Brivio <sbrivio@redhat.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Reviewed-by: Eric Dumazet <edumazet@google.com>
Link: https://lore.kernel.org/r/705dad54e6e6e9a010e571bf58e0b35a8ae70503.1701706073.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
include/net/tcp.h
net/ipv4/tcp.c