[PATCH 5.4 22/61] xen: Fix XenStore initialisation for XS_LOCAL

From: Greg Kroah-Hartman
Date: Tue Feb 02 2021 - 13:44:13 EST


From: David Woodhouse <dwmw@xxxxxxxxxxxx>

commit 5f46400f7a6a4fad635d5a79e2aa5a04a30ffea1 upstream.

In commit 3499ba8198ca ("xen: Fix event channel callback via INTX/GSI")
I reworked the triggering of xenbus_probe().

I tried to simplify things by taking out the workqueue based startup
triggered from wake_waiting(); the somewhat poorly named xenbus IRQ
handler.

I missed the fact that in the XS_LOCAL case (Dom0 starting its own
xenstored or xenstore-stubdom, which happens after the kernel is booted
completely), that IRQ-based trigger is still actually needed.

So... put it back, except more cleanly. By just spawning a xenbus_probe
thread which waits on xb_waitq and runs the probe the first time it
gets woken, just as the workqueue-based hack did.

This is actually a nicer approach for *all* the back ends with different
interrupt methods, and we can switch them all over to that without the
complex conditions for when to trigger it. But not in -rc6. This is
the minimal fix for the regression, although it's a step in the right
direction instead of doing a partial revert and actually putting the
workqueue back. It's also simpler than the workqueue.

Fixes: 3499ba8198ca ("xen: Fix event channel callback via INTX/GSI")
Reported-by: Juergen Gross <jgross@xxxxxxxx>
Signed-off-by: David Woodhouse <dwmw@xxxxxxxxxxxx>
Reviewed-by: Juergen Gross <jgross@xxxxxxxx>
Link: https://lore.kernel.org/r/4c9af052a6e0f6485d1de43f2c38b1461996db99.camel@xxxxxxxxxxxxx
Signed-off-by: Juergen Gross <jgross@xxxxxxxx>
Cc: Salvatore Bonaccorso <carnil@xxxxxxxxxx>
Cc: Jason Andryuk <jandryuk@xxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
drivers/xen/xenbus/xenbus_probe.c | 31 +++++++++++++++++++++++++++++++
1 file changed, 31 insertions(+)

--- a/drivers/xen/xenbus/xenbus_probe.c
+++ b/drivers/xen/xenbus/xenbus_probe.c
@@ -714,6 +714,23 @@ static bool xs_hvm_defer_init_for_callba
#endif
}

+static int xenbus_probe_thread(void *unused)
+{
+ DEFINE_WAIT(w);
+
+ /*
+ * We actually just want to wait for *any* trigger of xb_waitq,
+ * and run xenbus_probe() the moment it occurs.
+ */
+ prepare_to_wait(&xb_waitq, &w, TASK_INTERRUPTIBLE);
+ schedule();
+ finish_wait(&xb_waitq, &w);
+
+ DPRINTK("probing");
+ xenbus_probe();
+ return 0;
+}
+
static int __init xenbus_probe_initcall(void)
{
/*
@@ -725,6 +742,20 @@ static int __init xenbus_probe_initcall(
!xs_hvm_defer_init_for_callback()))
xenbus_probe();

+ /*
+ * For XS_LOCAL, spawn a thread which will wait for xenstored
+ * or a xenstore-stubdom to be started, then probe. It will be
+ * triggered when communication starts happening, by waiting
+ * on xb_waitq.
+ */
+ if (xen_store_domain_type == XS_LOCAL) {
+ struct task_struct *probe_task;
+
+ probe_task = kthread_run(xenbus_probe_thread, NULL,
+ "xenbus_probe");
+ if (IS_ERR(probe_task))
+ return PTR_ERR(probe_task);
+ }
return 0;
}
device_initcall(xenbus_probe_initcall);