On 2/12/2018 6:26 PM, Chas Williams wrote: > It's not clear to me that link_properties_valid() is even correct. Nothing > prevents an adapter from later negotiating a lower speed and would fail this > test. If both adapters are set to autoneg, that should be sufficient but > nothing enforces the speed match after the slaves are configured. So what is > the point of this check?
Reminder of this patch. This is waiting in the backlog for a long time. It is not even clear if the patch is still valid or not. Also based on missing response to Chas' clarification request, I am for dropping/rejecting the patch. @Declan, @Radu please chime in if this patch is required/valid. > > On Tue, Feb 6, 2018 at 3:52 PM, Thomas Monjalon <tho...@monjalon.net > <mailto:tho...@monjalon.net>> wrote: > > 17/01/2018 17:02, Ferruh Yigit: > > On 11/29/2017 3:42 PM, Tomasz Kulasek wrote: > > > Some devices needs more time to initialize and bring interface up. > When > > > link is down the link properties are not valid, e.g. link_speed is > > > reported as 0 and this is not a valid speed for slave as well as for > whole > > > bonding. > > > > > > During NIC (and bonding) initialization there's concurrency between > > > updating link status and adding slave to the bonding. > > > > > > This patch: > > > > > > - adds delay before configuring bonding (if link is down) to be sure > that > > > link status of new slave is valid, > > > - propagates information about link status from first slave with > link up > > > instead of first slave at all, to be sure that link speed is valid. > > > > > > Fixes: 6abd94d72ab5 ("net/bonding: fix check slaves link properties") > > > Signed-off-by: Tomasz Kulasek <tomaszx.kula...@intel.com > <mailto:tomaszx.kula...@intel.com>> > > > --- > > > v2 changes: > > > - Checkpatch warnings, > > > - Improved code style > > Hi Declan, > > > > Any comment on this patch? > > Any news? > > >
