public inbox for netdev@vger.kernel.org 
 help / color / mirror / Atom feed
From: Timur Tabi <timur@freescale•com>
To: Andy Fleming <afleming@gmail•com>
Cc: David Daney <david.daney@cavium•com>,
	"netdev@vger•kernel.org" <netdev@vger•kernel.org>,
	"davem@davemloft•net" <davem@davemloft•net>
Subject: Re: [PATCH] netdev/phy: Use mdiobus_read() so that proper locks are taken.
Date: Mon, 14 Nov 2011 11:10:42 -0600	[thread overview]
Message-ID: <4EC14B92.70607@freescale.com> (raw)
In-Reply-To: <CAKWjMd4WjDNHH8dOr9GPxQTC69SGgjh+euWh_DSjhtRxquB8qA@mail.gmail.com>

Andy Fleming wrote:

> Yes, that is correct. I have a patch, I just have to resend it. Will do
> next time I come to a break in what I'm working on.

Can you point me to that patch, so that I can try it out now?

I fixed some lockdep-related code in my audio driver.  The driver was not initializing a sysfs attr structure, so I added a call to sysfs_attr_init().  But when I do that, I get the following bug report.  I don't understand the connection.

=============================================                                   
[ INFO: possible recursive locking detected ]                                   
3.2.0-10b-00093-gebea711-dirty #10                                              
---------------------------------------------                                   
kworker/1:1/271 is trying to acquire lock:                                      
 (&(&(priv->tx_queue[i]->txlock))->rlock){......}, at: [<c02b2e28>] lock_tx_qs+0
x34/0x54                                                                        
                                                                                
but task is already holding lock:                                               
 (&(&(priv->tx_queue[i]->txlock))->rlock){......}, at: [<c02b2e28>] lock_tx_qs+0
x34/0x54                                                                        
                                                                                
other info that might help us debug this:                                       
 Possible unsafe locking scenario:                                              
                                                                                
       CPU0                                                                     
       ----                                                                     
  lock(&(&(priv->tx_queue[i]->txlock))->rlock);                                 
  lock(&(&(priv->tx_queue[i]->txlock))->rlock);                                 
                                                                                
 *** DEADLOCK ***                                                               
                                                                                
 May be due to missing lock nesting notation                                    
                                                                                
4 locks held by kworker/1:1/271:                                                
 #0:  (events){.+.+.+}, at: [<c005b7c8>] process_one_work+0x138/0x490           
 #1:  ((&(&dev->state_queue)->work)){+.+...}, at: [<c005b7c8>] process_one_work+
0x138/0x490                                                                     
 #2:  (&dev->lock){+.+...}, at: [<c02ab3cc>] phy_state_machine+0x30/0x580       
 #3:  (&(&(priv->tx_queue[i]->txlock))->rlock){......}, at: [<c02b2e28>] lock_tx
_qs+0x34/0x54                                                                   
                                                                                
stack backtrace:                                                                
Call Trace:                                                                     
[e69d5d80] [c0008c7c] show_stack+0x44/0x154 (unreliable)                        
[e69d5dc0] [c007bafc] __lock_acquire+0xefc/0x1824                               
[e69d5e70] [c007c870] lock_acquire+0x4c/0x68                                    
[e69d5e90] [c04575d8] _raw_spin_lock+0x44/0x60                                  
[e69d5ea0] [c02b2e28] lock_tx_qs+0x34/0x54                                      
[e69d5ec0] [c02b2f24] adjust_link+0x34/0x1d8                                    
[e69d5ef0] [c02ab73c] phy_state_machine+0x3a0/0x580                             
[e69d5f10] [c005b83c] process_one_work+0x1ac/0x490                              
[e69d5f50] [c005e4c0] worker_thread+0x18c/0x35c                                 
[e69d5f90] [c00631d4] kthread+0x7c/0x80                                         
[e69d5ff0] [c000e588] kernel_thread+0x4c/0x68                                   

-- 
Timur Tabi
Linux kernel developer at Freescale

      parent reply	other threads:[~2011-11-14 17:10 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-09-30 21:51 [PATCH] netdev/phy: Use mdiobus_read() so that proper locks are taken David Daney
2011-09-30 22:54 ` David Miller
2011-11-11  0:29 ` Tabi Timur-B04825
2011-11-11  0:48   ` David Daney
     [not found]     ` <CAKWjMd4WjDNHH8dOr9GPxQTC69SGgjh+euWh_DSjhtRxquB8qA@mail.gmail.com>
2011-11-14 17:10       ` Timur Tabi [this message]

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=4EC14B92.70607@freescale.com \
    --to=timur@freescale$(echo .)com \
    --cc=afleming@gmail$(echo .)com \
    --cc=davem@davemloft$(echo .)net \
    --cc=david.daney@cavium$(echo .)com \
    --cc=netdev@vger$(echo .)kernel.org \
    /path/to/YOUR_REPLY

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