From: Kevin Hilman <khilman@mvista.com>
To: linux-omap-open-source@linux.omap.com
Subject: [PATCH 2/2] ARM: OMAP: MUSB: BUG on potential deadlocks
Date: Thu, 30 Aug 2007 19:54:01 -0700 [thread overview]
Message-ID: <20070831025811.887406517@mvista.com> (raw)
In-Reply-To: 20070831025359.663866890@mvista.com
[-- Attachment #1: musb-lock-debug.patch --]
[-- Type: text/plain, Size: 3458 bytes --]
Help track down locking bugs.
There are several places where locks are assumed to be held and are
unlocked. BUG-out when the lock is not actually held.
Signed-off-by: Kevin Hilman <khilman@mvista.com>
---
drivers/usb/musb/musb_gadget.c | 7 +++++++
drivers/usb/musb/musb_gadget_ep0.c | 4 ++++
drivers/usb/musb/musb_host.c | 1 +
3 files changed, 12 insertions(+)
Index: dev/drivers/usb/musb/musb_gadget.c
===================================================================
--- dev.orig/drivers/usb/musb/musb_gadget.c
+++ dev/drivers/usb/musb/musb_gadget.c
@@ -118,6 +118,7 @@ __acquires(ep->musb->lock)
musb = req->musb;
ep->busy = 1;
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
if (is_dma_capable()) {
if (req->mapped) {
@@ -1812,6 +1813,7 @@ stop_activity(struct musb *musb, struct
}
}
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
driver->disconnect (&musb->g);
spin_lock(&musb->lock);
@@ -1851,6 +1853,8 @@ int usb_gadget_unregister_driver(struct
stop_activity(musb, driver);
DBG(3, "unregistering driver %s\n", driver->function);
+
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock_irqrestore(&musb->lock, flags);
driver->unbind(&musb->g);
spin_lock_irqsave(&musb->lock, flags);
@@ -1891,6 +1895,7 @@ void musb_g_resume(struct musb *musb)
case OTG_STATE_B_PERIPHERAL:
musb->is_active = 1;
if (musb->gadget_driver && musb->gadget_driver->resume) {
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
musb->gadget_driver->resume(&musb->g);
spin_lock(&musb->lock);
@@ -1918,6 +1923,7 @@ void musb_g_suspend(struct musb *musb)
case OTG_STATE_B_PERIPHERAL:
musb->is_suspended = 1;
if (musb->gadget_driver && musb->gadget_driver->suspend) {
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
musb->gadget_driver->suspend(&musb->g);
spin_lock(&musb->lock);
@@ -1954,6 +1960,7 @@ void musb_g_disconnect(struct musb *musb
musb->g.speed = USB_SPEED_UNKNOWN;
if (musb->gadget_driver && musb->gadget_driver->disconnect) {
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
musb->gadget_driver->disconnect(&musb->g);
spin_lock(&musb->lock);
Index: dev/drivers/usb/musb/musb_gadget_ep0.c
===================================================================
--- dev.orig/drivers/usb/musb/musb_gadget_ep0.c
+++ dev/drivers/usb/musb/musb_gadget_ep0.c
@@ -273,7 +273,9 @@ __acquires(musb->lock)
if (!musb_ep->desc)
break;
+
/* REVISIT do it directly, no locking games */
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
musb_gadget_set_halt(&musb_ep->end_point, 0);
spin_lock(&musb->lock);
@@ -586,6 +588,8 @@ __acquires(musb->lock)
int retval;
if (!musb->gadget_driver)
return -EOPNOTSUPP;
+
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
retval = musb->gadget_driver->setup(&musb->g, ctrlrequest);
spin_lock(&musb->lock);
Index: dev/drivers/usb/musb/musb_host.c
===================================================================
--- dev.orig/drivers/usb/musb/musb_host.c
+++ dev/drivers/usb/musb/musb_host.c
@@ -305,6 +305,7 @@ __acquires(musb->lock)
urb->actual_length, urb->transfer_buffer_length
);
+ BUG_ON(!spin_is_locked(&musb->lock));
spin_unlock(&musb->lock);
usb_hcd_giveback_urb(musb_to_hcd(musb), urb);
spin_lock(&musb->lock);
--
next prev parent reply other threads:[~2007-08-31 2:54 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2007-08-31 2:53 [PATCH 0/2] MUSB: tracking down locking bugs Kevin Hilman
2007-08-31 2:54 ` [PATCH 1/2] ARM: OMAP: MUSB: dont exit with spinlock held Kevin Hilman
2007-08-31 5:47 ` David Brownell
2007-08-31 5:51 ` Kevin Hilman
2007-08-31 2:54 ` Kevin Hilman [this message]
2007-08-31 6:05 ` [PATCH 2/2] ARM: OMAP: MUSB: BUG on potential deadlocks David Brownell
2007-08-31 16:18 ` Kevin Hilman
2007-08-31 17:49 ` David Brownell
2007-08-31 18:41 ` Kevin Hilman
2007-08-31 20:10 ` David Brownell
2007-08-31 5:48 ` [PATCH 0/2] MUSB: tracking down locking bugs David Brownell
2007-08-31 6:04 ` Kevin Hilman
2007-08-31 6:38 ` David Brownell
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=20070831025811.887406517@mvista.com \
--to=khilman@mvista.com \
--cc=linux-omap-open-source@linux.omap.com \
/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