| Message ID | 20220222001817.2264967-2-marmarek@invisiblethingslab.com (mailing list archive) |
|---|---|
| State | Accepted |
| Commit | e8240addd0a3919e0fd7436416afe9aa6429c484 |
| Delegated to: | Netdev Maintainers |
| Headers | show |
| Series | [v2,1/2] Revert "xen-netback: remove 'hotplug-status' once it has served its purpose" | expand |
On Tue, Feb 22, 2022 at 01:18:17AM +0100, Marek Marczykowski-Górecki wrote: > This reverts commit 2afeec08ab5c86ae21952151f726bfe184f6b23d. > > The reasoning in the commit was wrong - the code expected to setup the > watch even if 'hotplug-status' didn't exist. In fact, it relied on the > watch being fired the first time - to check if maybe 'hotplug-status' is > already set to 'connected'. Not registering a watch for non-existing > path (which is the case if hotplug script hasn't been executed yet), > made the backend not waiting for the hotplug script to execute. This in > turns, made the netfront think the interface is fully operational, while > in fact it was not (the vif interface on xen-netback side might not be > configured yet). > > This was a workaround for 'hotplug-status' erroneously being removed. > But since that is reverted now, the workaround is not necessary either. > > More discussion at > https://lore.kernel.org/xen-devel/afedd7cb-a291-e773-8b0d-4db9b291fa98@ipxe.org/T/#u > > Cc: stable@vger.kernel.org > Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> > --- > I believe this is the same issue as discussed at > https://lore.kernel.org/xen-devel/20220113111946.GA4133739@dingwall.me.uk/ Right - I believe we need to leave that workaround in place in libxl in order to deal with bogus Linux netbacks? Thanks, Roger.
On Tue, Feb 22, 2022 at 09:31:31AM +0100, Roger Pau Monné wrote: > On Tue, Feb 22, 2022 at 01:18:17AM +0100, Marek Marczykowski-Górecki wrote: > > This reverts commit 2afeec08ab5c86ae21952151f726bfe184f6b23d. > > > > The reasoning in the commit was wrong - the code expected to setup the > > watch even if 'hotplug-status' didn't exist. In fact, it relied on the > > watch being fired the first time - to check if maybe 'hotplug-status' is > > already set to 'connected'. Not registering a watch for non-existing > > path (which is the case if hotplug script hasn't been executed yet), > > made the backend not waiting for the hotplug script to execute. This in > > turns, made the netfront think the interface is fully operational, while > > in fact it was not (the vif interface on xen-netback side might not be > > configured yet). > > > > This was a workaround for 'hotplug-status' erroneously being removed. > > But since that is reverted now, the workaround is not necessary either. > > > > More discussion at > > https://lore.kernel.org/xen-devel/afedd7cb-a291-e773-8b0d-4db9b291fa98@ipxe.org/T/#u > > > > Cc: stable@vger.kernel.org > > Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> > > --- > > I believe this is the same issue as discussed at > > https://lore.kernel.org/xen-devel/20220113111946.GA4133739@dingwall.me.uk/ > > Right - I believe we need to leave that workaround in place in libxl > in order to deal with bogus Linux netbacks? I'm afraid so, yes.
On 22/02/2022 00:18, Marek Marczykowski-Górecki wrote: > This reverts commit 2afeec08ab5c86ae21952151f726bfe184f6b23d. > > The reasoning in the commit was wrong - the code expected to setup the > watch even if 'hotplug-status' didn't exist. In fact, it relied on the > watch being fired the first time - to check if maybe 'hotplug-status' is > already set to 'connected'. Not registering a watch for non-existing > path (which is the case if hotplug script hasn't been executed yet), > made the backend not waiting for the hotplug script to execute. This in > turns, made the netfront think the interface is fully operational, while > in fact it was not (the vif interface on xen-netback side might not be > configured yet). > > This was a workaround for 'hotplug-status' erroneously being removed. > But since that is reverted now, the workaround is not necessary either. > > More discussion at > https://lore.kernel.org/xen-devel/afedd7cb-a291-e773-8b0d-4db9b291fa98@ipxe.org/T/#u > > Cc: stable@vger.kernel.org > Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> Reviewed-by: Paul Durrant <paul@xen.org>
On 24/02/2022 07:56, Durrant, Paul wrote: > On 22/02/2022 00:18, Marek Marczykowski-Górecki wrote: >> This reverts commit 2afeec08ab5c86ae21952151f726bfe184f6b23d. >> >> The reasoning in the commit was wrong - the code expected to setup the >> watch even if 'hotplug-status' didn't exist. In fact, it relied on the >> watch being fired the first time - to check if maybe 'hotplug-status' is >> already set to 'connected'. Not registering a watch for non-existing >> path (which is the case if hotplug script hasn't been executed yet), >> made the backend not waiting for the hotplug script to execute. This in >> turns, made the netfront think the interface is fully operational, while >> in fact it was not (the vif interface on xen-netback side might not be >> configured yet). >> >> This was a workaround for 'hotplug-status' erroneously being removed. >> But since that is reverted now, the workaround is not necessary either. >> >> More discussion at >> https://lore.kernel.org/xen-devel/afedd7cb-a291-e773-8b0d-4db9b291fa98@ipxe.org/T/#u >> >> >> Cc: stable@vger.kernel.org >> Signed-off-by: Marek Marczykowski-Górecki >> <marmarek@invisiblethingslab.com> > > Reviewed-by: Paul Durrant <paul@xen.org> In conjunction with patch 1/2 (which reverts the patch that caused the original problem): Reviewed-by: Michael Brown <mbrown@fensystems.co.uk> Thanks, Michael
diff --git a/drivers/net/xen-netback/xenbus.c b/drivers/net/xen-netback/xenbus.c index 3fad58d22155..990360d75cb6 100644 --- a/drivers/net/xen-netback/xenbus.c +++ b/drivers/net/xen-netback/xenbus.c @@ -824,15 +824,11 @@ static void connect(struct backend_info *be) xenvif_carrier_on(be->vif); unregister_hotplug_status_watch(be); - if (xenbus_exists(XBT_NIL, dev->nodename, "hotplug-status")) { - err = xenbus_watch_pathfmt(dev, &be->hotplug_status_watch, - NULL, hotplug_status_changed, - "%s/%s", dev->nodename, - "hotplug-status"); - if (err) - goto err; + err = xenbus_watch_pathfmt(dev, &be->hotplug_status_watch, NULL, + hotplug_status_changed, + "%s/%s", dev->nodename, "hotplug-status"); + if (!err) be->have_hotplug_status_watch = 1; - } netif_tx_wake_all_queues(be->vif->dev);
This reverts commit 2afeec08ab5c86ae21952151f726bfe184f6b23d. The reasoning in the commit was wrong - the code expected to setup the watch even if 'hotplug-status' didn't exist. In fact, it relied on the watch being fired the first time - to check if maybe 'hotplug-status' is already set to 'connected'. Not registering a watch for non-existing path (which is the case if hotplug script hasn't been executed yet), made the backend not waiting for the hotplug script to execute. This in turns, made the netfront think the interface is fully operational, while in fact it was not (the vif interface on xen-netback side might not be configured yet). This was a workaround for 'hotplug-status' erroneously being removed. But since that is reverted now, the workaround is not necessary either. More discussion at https://lore.kernel.org/xen-devel/afedd7cb-a291-e773-8b0d-4db9b291fa98@ipxe.org/T/#u Cc: stable@vger.kernel.org Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com> --- I believe this is the same issue as discussed at https://lore.kernel.org/xen-devel/20220113111946.GA4133739@dingwall.me.uk/ Cc: James Dingwall <james-xen@dingwall.me.uk Cc: Michael Brown <mcb30@ipxe.org> --- drivers/net/xen-netback/xenbus.c | 12 ++++-------- 1 file changed, 4 insertions(+), 8 deletions(-)