From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([208.118.235.92]:50335) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TZgq1-00065X-C9 for qemu-devel@nongnu.org; Sat, 17 Nov 2012 06:45:32 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1TZgpy-000652-AA for qemu-devel@nongnu.org; Sat, 17 Nov 2012 06:45:29 -0500 Received: from mx1.redhat.com ([209.132.183.28]:7590) by eggs.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1TZgpy-00064w-13 for qemu-devel@nongnu.org; Sat, 17 Nov 2012 06:45:26 -0500 Received: from int-mx10.intmail.prod.int.phx2.redhat.com (int-mx10.intmail.prod.int.phx2.redhat.com [10.5.11.23]) by mx1.redhat.com (8.14.4/8.14.4) with ESMTP id qAHBjNCN005758 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Sat, 17 Nov 2012 06:45:25 -0500 From: Hans de Goede Date: Sat, 17 Nov 2012 12:47:15 +0100 Message-Id: <1353152838-26886-3-git-send-email-hdegoede@redhat.com> In-Reply-To: <1353152838-26886-1-git-send-email-hdegoede@redhat.com> References: <1353152838-26886-1-git-send-email-hdegoede@redhat.com> Subject: [Qemu-devel] [PATCH 2/5] usb: Don't allow USB_RET_ASYNC for interrupt packets List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Gerd Hoffmann Cc: Hans de Goede , qemu-devel@nongnu.org It is tempting to use USB_RET_ASYNC for interrupt packets, rather then the current NAK + polling approach, but this causes issues for migration, as an async completed packet will not getting written back to guest memory until the next poll time, and if a migration happens in between it will get lost! Make an exception for host devices, because: 1) host-linux actually uses async completion for interrupt endpoints 2) host devices don't migrate anyways Ideally we would convert host-linux.c to handle (input) interrupt endpoints in a buffered manner like it does for isoc endpoints, keeping multiple urbs submitted to ensure the devices timing requirements are met, as well as making its interrupt ep handling the same as other usb-devices. Signed-off-by: Hans de Goede --- hw/usb.h | 1 + hw/usb/core.c | 4 ++++ hw/usb/host-bsd.c | 1 + hw/usb/host-linux.c | 1 + 4 files changed, 7 insertions(+) diff --git a/hw/usb.h b/hw/usb.h index 7d6de69..58f812f 100644 --- a/hw/usb.h +++ b/hw/usb.h @@ -197,6 +197,7 @@ struct USBEndpoint { enum USBDeviceFlags { USB_DEV_FLAG_FULL_PATH, + USB_DEV_FLAG_IS_HOST, }; /* definition of a USB device */ diff --git a/hw/usb/core.c b/hw/usb/core.c index 52b5310..8e360d3 100644 --- a/hw/usb/core.c +++ b/hw/usb/core.c @@ -406,7 +406,11 @@ void usb_handle_packet(USBDevice *dev, USBPacket *p) if (QTAILQ_EMPTY(&p->ep->queue) || p->ep->pipeline) { usb_process_one(p); if (p->status == USB_RET_ASYNC) { + /* hcd drivers cannot handle async for isoc */ assert(p->ep->type != USB_ENDPOINT_XFER_ISOC); + /* using async for interrupt packets breaks migration */ + assert(p->ep->type != USB_ENDPOINT_XFER_INT || + (dev->flags & USB_DEV_FLAG_IS_HOST)); usb_packet_set_state(p, USB_PACKET_ASYNC); QTAILQ_INSERT_TAIL(&p->ep->queue, p, queue); } else if (p->status == USB_RET_ADD_TO_QUEUE) { diff --git a/hw/usb/host-bsd.c b/hw/usb/host-bsd.c index 6473e8b..dae0009 100644 --- a/hw/usb/host-bsd.c +++ b/hw/usb/host-bsd.c @@ -292,6 +292,7 @@ static void usb_host_handle_destroy(USBDevice *opaque) static int usb_host_initfn(USBDevice *dev) { + dev->flags |= (1 << USB_DEV_FLAG_IS_HOST); return 0; } diff --git a/hw/usb/host-linux.c b/hw/usb/host-linux.c index aa77b77..bdafb6b 100644 --- a/hw/usb/host-linux.c +++ b/hw/usb/host-linux.c @@ -1476,6 +1476,7 @@ static int usb_host_initfn(USBDevice *dev) { USBHostDevice *s = DO_UPCAST(USBHostDevice, dev, dev); + dev->flags |= (1 << USB_DEV_FLAG_IS_HOST); dev->auto_attach = 0; s->fd = -1; s->hub_fd = -1; -- 1.8.0