[dpdk-stable] [dpdk-dev] [PATCH v2] net/ixgbe: fix busy polling while fiber link update

[Burakov, Anatoly]

On 01-Nov-18 4:04 PM, Ilya Maximets wrote:
> If the multispeed fiber link is in DOWN state, ixgbe_setup_link
> could take around a second of busy polling. This is highly
> inconvenient for the case where single thread periodically
> checks the link statuses. For example, OVS main thread
> periodically updates the link statuses and hangs for a really
> long time busy waiting on ixgbe_setup_link() for a DOWN fiber
> ports. For case with 3 down ports it hangs for a 3 seconds and
> unable to do anything including packet processing.
> Fix that by shifting that workaround to a separate thread by
> alarm handler that will try to set up link if it is DOWN.
> 
> Fixes: c12d22f65b13 ("net/ixgbe: ensure link status is updated")
> CC: stable at dpdk.org
> 
> Signed-off-by: Ilya Maximets <i.maximets at samsung.com>
> ---

On my setup, this commit breaks ixgbe init in pktgen 3.5.7:

ixgbe_dev_start(): failure in ixgbe_dev_start(): -15
!PANIC!: rte_eth_dev_start: port=0, Input/output error
PANIC in pktgen_config_ports():
rte_eth_dev_start: port=0, Input/output error6: 
[build/DPDK/pktgen(_start+0x2a) [0x560880ec838a]]
5: [/lib/x86_64-linux-gnu/libc.so.6(__libc_start_main+0xe7) 
[0x7fabc654cb97]]
4: [build/DPDK/pktgen(main+0xe77) [0x560880ec0357]]
3: [build/DPDK/pktgen(pktgen_config_ports+0x1cf0) [0x560880ef53e0]]
2: [build/DPDK/pktgen(__rte_panic+0xc5) [0x560880eb11b4]]
1: [build/DPDK/pktgen(rte_dump_stack+0x2e) [0x560880fde69e]]
Aborted

-- 
Thanks,
Anatoly