hv: vmbus_post_msg: retry the hypercall on some transient errors
authorDexuan Cui <decui@microsoft.com>
Fri, 27 Feb 2015 19:25:59 +0000 (11:25 -0800)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Mon, 2 Mar 2015 03:30:07 +0000 (19:30 -0800)
commit89f9f6796d41e10e224b0cb0027ddd78cb881f65
treedb64755f71d4c714a1ca83d5bba30bdcd1c97279
parent1896566372579939eb86414ea3d664f18d5e909c
hv: vmbus_post_msg: retry the hypercall on some transient errors

I got HV_STATUS_INVALID_CONNECTION_ID on Hyper-V 2008 R2 when keeping running
"rmmod hv_netvsc; modprobe hv_netvsc; rmmod hv_utils; modprobe hv_utils"
in a Linux guest. Looks the host has some kind of throttling mechanism if
some kinds of hypercalls are sent too frequently.
Without the patch, the driver can occasionally fail to load.

Also let's retry HV_STATUS_INSUFFICIENT_MEMORY, though we didn't get it
before.

Removed 'case -ENOMEM', since the hypervisor doesn't return this.

CC: "K. Y. Srinivasan" <kys@microsoft.com>
Reviewed-by: Jason Wang <jasowang@redhat.com>
Signed-off-by: Dexuan Cui <decui@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arch/x86/include/uapi/asm/hyperv.h
drivers/hv/connection.c