diff options
| author |   Davide Caratti <dcaratti@redhat.com> | 2021-05-22 15:14:45 +0200 |
|---|---|---|
| committer |   David S. Miller <davem@davemloft.net> | 2021-05-23 17:16:09 -0700 |
| commit | 3a62fed2fd7b6fea96d720e779cafc30dfb3a22e (patch) | |
| tree | 3e6a1db4a98b22e8bc060ddf280b895a2ed10a8e /tools/testing/radix-tree | |
| parent | net: macb: ensure the device is available before accessing GEMGXL control registers (diff) | |
| download | wireguard-linux-3a62fed2fd7b6fea96d720e779cafc30dfb3a22e.tar.xz wireguard-linux-3a62fed2fd7b6fea96d720e779cafc30dfb3a22e.zip | |
net/sched: fq_pie: re-factor fix for fq_pie endless loop
the patch that fixed an endless loop in_fq_pie_init() was not considering
that 65535 is a valid class id. The correct bugfix for this infinite loop
is to change 'idx' to become an u32, like Colin proposed in the past [1].
Fix this as follows:
- restore 65536 as maximum possible values of 'flows_cnt'
- use u32 'idx' when iterating on 'q->flows'
- fix the TDC selftest
This reverts commit bb2f930d6dd708469a587dc9ed1efe1ef969c0bf.
[1] https://lore.kernel.org/netdev/20210407163808.499027-1-colin.king@canonical.com/
CC: Colin Ian King <colin.king@canonical.com>
CC: stable@vger.kernel.org
Fixes: bb2f930d6dd7 ("net/sched: fix infinite loop in sch_fq_pie")
Fixes: ec97ecf1ebe4 ("net: sched: add Flow Queue PIE packet scheduler")
Signed-off-by: Davide Caratti <dcaratti@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'tools/testing/radix-tree')
0 files changed, 0 insertions, 0 deletions