Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1400123

Aw: [PATCH 0/6] hv_netvsc: avoid races on mtu change/set channels

Show all headers | View raw

Hi,

>
> MTU change and set channels operations are implemented as netvsc device
> re-creation destroying internal structures (struct net_device stays). This
> is really unfortunate but there is no support from Hyper-V host to do it
> in a different way. Such re-creation is unsurprisingly racy, Haiyang
> reported a crash when netvsc_change_mtu() is racing with
> netvsc_link_change() but I was able to identify additional races upon
> investigation. Both netvsc_set_channels() and netvsc_change_mtu() race
> against:
> 1) netvsc_link_change()
> 2) netvsc_remove()
> 3) netvsc_send()
> 

after having a look into this driver I got the impression that you are working around an
unfortunate implementation of the shutdown sequence in the remove function:
If you do unregister_netdev() first instead of resource cleanup then neither set_channels()
nor change_mtu() can race with remove(). This is since after unregister_netdev() returns
the netdev is not longer available from userspace and thus neither set_channels nor
change_mtu can be called anymore (note that all of these functions are protected by the 
rtnl_lock).

To avoid the race between netvsc_change_mtu()/netvsc_set_channels() and netvsc_link_change()
you have to stop the concerning worker thread (dwork) before you call netvsc_close() and
restart it once the device is up again.

Regards,
Lino

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH 0/6] hv_netvsc: avoid races on mtu change/set channels Vitaly Kuznetsov <vkuznets@redhat.com> - 2016-05-12 14:50 +0200
  [PATCH 5/6] hv_netvsc: synchronize netvsc_change_mtu()/netvsc_set_channels() with netvsc_remove() Vitaly Kuznetsov <vkuznets@redhat.com> - 2016-05-12 15:00 +0200
  [PATCH 6/6] hv_netvsc: set nvdev link after populating chn_table Vitaly Kuznetsov <vkuznets@redhat.com> - 2016-05-12 15:00 +0200
  [PATCH 4/6] hv_netvsc: get rid of struct net_device pointer in struct netvsc_device Vitaly Kuznetsov <vkuznets@redhat.com> - 2016-05-12 15:00 +0200
  [PATCH 3/6] hv_netvsc: untangle the pointer mess Vitaly Kuznetsov <vkuznets@redhat.com> - 2016-05-12 15:00 +0200
  Aw: [PATCH 0/6] hv_netvsc: avoid races on mtu change/set channels "Lino Sanfilippo" <LinoSanfilippo@gmx.de> - 2016-05-12 17:00 +0200
    Re: Aw: [PATCH 0/6] hv_netvsc: avoid races on mtu change/set channels Vitaly Kuznetsov <vkuznets@redhat.com> - 2016-05-12 17:10 +0200
      Aw: Re:  [PATCH 0/6] hv_netvsc: avoid races on mtu change/set  channels "Lino Sanfilippo" <LinoSanfilippo@gmx.de> - 2016-05-12 18:20 +0200

csiph-web