qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Hans de Goede <hdegoede@redhat.com>
To: qemu-devel@nongnu.org
Cc: Hans de Goede <hdegoede@redhat.com>, Gerd Hoffmann <kraxel@redhat.com>
Subject: [Qemu-devel] [PATCH 15/22] combined-packet: Add a workaround for Linux usbfs + live migration
Date: Mon, 15 Oct 2012 12:38:24 +0200	[thread overview]
Message-ID: <1350297511-25437-16-git-send-email-hdegoede@redhat.com> (raw)
In-Reply-To: <1350297511-25437-1-git-send-email-hdegoede@redhat.com>

Older versions (anything but the latest) of Linux usbfs + libusb(x),
will submit larger (bulk) transfers split into multiple 16k submissions,
which means that rather then all tds getting linked into the queue in
one atomic operarion they get linked in a bunch at a time, which could
cause problems if:
1) We scan the queue while libusb is in the middle of submitting a split
   bulk transfer
2) While this bulk transfer is pending we migrate to another host.

The problem is that after 2, the new host will rescan the queue and
combine the packets in one large transfer, where as 1) has caused the
original host to see them as 2 transfers. This patch fixes this by stopping
combinging if we detect a 16k transfer with its int_req flag set.

This should not adversely effect performance for other cases as:
1) Linux never sets the interrupt flag on packets other then the last
2) Windows does set the in_req flag on each td, but will submit large
transfers in 20k tds thus never triggering the check

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
---
 hw/usb/combined-packet.c | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/hw/usb/combined-packet.c b/hw/usb/combined-packet.c
index b6fcb2c..d347324 100644
--- a/hw/usb/combined-packet.c
+++ b/hw/usb/combined-packet.c
@@ -120,7 +120,7 @@ void usb_ep_combine_input_packets(USBEndpoint *ep)
     USBPacket *p, *u, *next, *prev = NULL, *first = NULL;
     USBPort *port = ep->dev->port;
     USBDeviceClass *klass = USB_DEVICE_GET_CLASS(ep->dev);
-    int ret;
+    int ret, totalsize;
 
     assert(ep->pipeline);
     assert(ep->pid == USB_TOKEN_IN);
@@ -165,8 +165,11 @@ void usb_ep_combine_input_packets(USBEndpoint *ep)
         }
 
         /* Is this packet the last one of a (combined) transfer? */
+        totalsize = (p->combined) ? p->combined->iov.size : p->iov.size;
         if ((p->iov.size % ep->max_packet_size) != 0 || !p->short_not_ok ||
-                next == NULL) {
+                next == NULL ||
+                /* Work around for Linux usbfs bulk splitting + migration */
+                (totalsize == 16348 && p->int_req)) {
             ret = klass->handle_combined_data(ep->dev, first);
             assert(ret == USB_RET_ASYNC);
             if (first->combined) {
-- 
1.7.12.1

  parent reply	other threads:[~2012-10-15 10:38 UTC|newest]

Thread overview: 33+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2012-10-15 10:38 [Qemu-devel] RFC: usb input pipelining / speedup patchset v2 Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 01/22] uhci: Properly unmap packets on cancel / invalid pid Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 02/22] uhci: Move checks to continue queuing to uhci_fill_queue() Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 03/22] ehci: Get rid of packet tbytes field Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 04/22] ehci: Set int flag on a short input packet Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 05/22] ehci: Improve latency of interrupt delivery and async schedule scanning Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 06/22] ehci: Speed up the timer of raising int from the async schedule Hans de Goede
2012-10-15 11:17   ` Gerd Hoffmann
2012-10-15 13:00     ` Hans de Goede
2012-10-17 11:01       ` Gerd Hoffmann
2012-10-17 11:11         ` Hans de Goede
2012-10-17 11:37           ` Gerd Hoffmann
2012-10-15 10:38 ` [Qemu-devel] [PATCH 07/22] ehci: Detect going in circles when filling the queue Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 08/22] xhci: Add a xhci_ep_nuke_one_xfer helper function Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 09/22] usb: Rename __usb_packet_complete to usb_packet_complete_one Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 10/22] usb: Add USB_RET_ADD_TO_QUEUE packet result code Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 11/22] usb: Move clearing of queue on halt to the core Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 12/22] usb: Move short-not-ok handling " Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 13/22] usb: Add an int_req flag to USBPacket Hans de Goede
2012-10-17 11:04   ` Gerd Hoffmann
2012-10-17 11:11     ` Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 14/22] usb: Add packet combining functions Hans de Goede
2012-10-17 11:29   ` Gerd Hoffmann
2012-10-17 14:41     ` Hans de Goede
2012-10-18  6:00       ` Gerd Hoffmann
2012-10-15 10:38 ` Hans de Goede [this message]
2012-10-15 10:38 ` [Qemu-devel] [PATCH 16/22] usb-redir: When a packet contains data on a stall, ignore the stall Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 17/22] usb-redir: Add support for 32 bits bulk packet length Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 18/22] usb-redir: Add support for input pipelining Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 19/22] usb-redir: Add an usbredir_setup_usb_eps() helper function Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 20/22] usb-redir: Use reject rather the disconnect on bad ep info Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 21/22] usb-redir: Allow to attach USB 2.0 devices to 1.1 host controller Hans de Goede
2012-10-15 10:38 ` [Qemu-devel] [PATCH 22/22] usb-redir: Allow redirecting super speed devices to high speed controllers Hans de Goede

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=1350297511-25437-16-git-send-email-hdegoede@redhat.com \
    --to=hdegoede@redhat.com \
    --cc=kraxel@redhat.com \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).