eth: fbnic: remove the debugging trick of super high page bias
authorJakub Kicinski <kuba@kernel.org>
Fri, 1 Aug 2025 17:07:54 +0000 (10:07 -0700)
committerJakub Kicinski <kuba@kernel.org>
Tue, 5 Aug 2025 23:01:26 +0000 (16:01 -0700)
commite407fceeaf1b2959892b4fc9b584843d3f2bfc05
treee54d4f743eca0c07ee53691dae72dfca79dd2c05
parente88fbc30dda1cb7438515303704ceddb3ade4ecd
eth: fbnic: remove the debugging trick of super high page bias

Alex added page bias of LONG_MAX, which is admittedly quite
a clever way of catching overflows of the pp ref count.
The page pool code was "optimized" to leave the ref at 1
for freed pages so it can't catch basic bugs by itself any more.
(Something we should probably address under DEBUG_NET...)

Unfortunately for fbnic since commit f7dc3248dcfb ("skbuff: Optimization
of SKB coalescing for page pool") core _may_ actually take two extra
pp refcounts, if one of them is returned before driver gives up the bias
the ret < 0 check in page_pool_unref_netmem() will trigger.

While at it add a FBNIC_ to the name of the driver constant.

Fixes: 0cb4c0a13723 ("eth: fbnic: Implement Rx queue alloc/start/stop/free")
Link: https://patch.msgid.link/20250801170754.2439577-1-kuba@kernel.org
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
drivers/net/ethernet/meta/fbnic/fbnic_txrx.c
drivers/net/ethernet/meta/fbnic/fbnic_txrx.h