Patchwork usb: selective endpoint initialization

login
register
mail settings
Submitter Gerd Hoffmann
Date Aug. 9, 2012, 8:59 a.m.
Message ID <50237BDA.8030906@redhat.com>
Download mbox | patch
Permalink /patch/176030/
State New
Headers show

Comments

Gerd Hoffmann - Aug. 9, 2012, 8:59 a.m.
> Hi Gerd,
> 
> sorry for the delays, I tested the latest pulled patch queue and it's
> now fine on my Intel board, too. The dongle gets detected again without
> assertions. Thanks for your work.
> 
> Still remaining are the multiple usb resets on host side before the
> dongle gets finally detected / usable on the guest.

Can you try if the attached patch makes a difference?

thanks,
  Gerd
From d51ea0d80009eb58a09e8e571a7eca5da562becc Mon Sep 17 00:00:00 2001
From: Gerd Hoffmann <kraxel@redhat.com>
Date: Thu, 9 Aug 2012 10:57:32 +0200
Subject: [PATCH] async control xfer fixup

Signed-off-by: Gerd Hoffmann <kraxel@redhat.com>
---
 hw/usb/core.c       |    1 +
 hw/usb/host-linux.c |    1 +
 2 files changed, 2 insertions(+), 0 deletions(-)
Erik Rull - Aug. 13, 2012, 1:36 p.m.
On August 9, 2012 at 10:59 AM Gerd Hoffmann <kraxel@redhat.com> wrote:
> > Hi Gerd,
> >
> > sorry for the delays, I tested the latest pulled patch queue and it's
> > now fine on my Intel board, too. The dongle gets detected again without
> > assertions. Thanks for your work.
> >
> > Still remaining are the multiple usb resets on host side before the
> > dongle gets finally detected / usable on the guest.
>
> Can you try if the attached patch makes a difference?
>
> thanks,
>   Gerd

Hi Gerd,

sorry, no difference. Still the same behavior.

Best regards,

Erik
Erik Rull - Aug. 24, 2012, 10:11 a.m.
On August 9, 2012 at 10:59 AM Gerd Hoffmann <kraxel@redhat.com> wrote:

> > Hi Gerd,
> >
> > sorry for the delays, I tested the latest pulled patch queue and it's
> > now fine on my Intel board, too. The dongle gets detected again without
> > assertions. Thanks for your work.
> >
> > Still remaining are the multiple usb resets on host side before the
> > dongle gets finally detected / usable on the guest.
>
> Can you try if the attached patch makes a difference?
>
> thanks,
>   Gerd

Hi Gerd,

it still has a lot of USB host resets with the latest QEMU master.

Any ideas how to proceed?

Best regards,

Erik

Patch

diff --git a/hw/usb/core.c b/hw/usb/core.c
index 01a7622..c7e5bc0 100644
--- a/hw/usb/core.c
+++ b/hw/usb/core.c
@@ -107,6 +107,7 @@  static int do_token_setup(USBDevice *s, USBPacket *p)
     }
 
     usb_packet_copy(p, s->setup_buf, p->iov.size);
+    p->result = 0;
     s->setup_len   = (s->setup_buf[7] << 8) | s->setup_buf[6];
     s->setup_index = 0;
 
diff --git a/hw/usb/host-linux.c b/hw/usb/host-linux.c
index d55be87..8df9207 100644
--- a/hw/usb/host-linux.c
+++ b/hw/usb/host-linux.c
@@ -1045,6 +1045,7 @@  static int usb_host_handle_control(USBDevice *dev, USBPacket *p,
 
     /* Note request is (bRequestType << 8) | bRequest */
     trace_usb_host_req_control(s->bus_num, s->addr, p, request, value, index);
+    assert(p->result == 0);
 
     switch (request) {
     case DeviceOutRequest | USB_REQ_SET_ADDRESS: