rndis gadget: Inconsistent locking

On 2011-01-10 13:14, David Brownell wrote:
> 
> 
>> I have just retested Michals patch but I have found another
>> lockdep failure.
>>
>> It looks like rndis_msg_parser() can call dev_get_stats> too:
> 
> Can someone provide a fully working patch then?
> Or maybe just revert the one that caused all the
> breakage??
> 
> Rememeber that this driver was working fine for
> years before netdev changes added multiple bugs
> because (at least) they didn't update all callers.

Maybe I miss something, but if it's like Michal described something
like this should be enough (not tested nor compiled).

Jarek P.
---

 drivers/usb/gadget/f_phonet.c |    6 ++++++
 drivers/usb/gadget/u_ether.c  |    6 ++++++
 2 files changed, 12 insertions(+), 0 deletions(-)

--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Le mercredi 12 janvier 2011 à 12:28 +0000, Jarek Poplawski a écrit :
> On 2011-01-10 13:14, David Brownell wrote:
> > 
> > 
> >> I have just retested Michals patch but I have found another
> >> lockdep failure.
> >>
> >> It looks like rndis_msg_parser() can call dev_get_stats> too:
> > 
> > Can someone provide a fully working patch then?
> > Or maybe just revert the one that caused all the
> > breakage??
> > 
> > Rememeber that this driver was working fine for
> > years before netdev changes added multiple bugs
> > because (at least) they didn't update all callers.
> 
> Maybe I miss something, but if it's like Michal described something
> like this should be enough (not tested nor compiled).
> 
> Jarek P.
> ---
> 
>  drivers/usb/gadget/f_phonet.c |    6 ++++++
>  drivers/usb/gadget/u_ether.c  |    6 ++++++
>  2 files changed, 12 insertions(+), 0 deletions(-)
> 
> diff --git a/drivers/usb/gadget/f_phonet.c b/drivers/usb/gadget/f_phonet.c
> index 3c6e1a0..bef4622 100644
> --- a/drivers/usb/gadget/f_phonet.c
> +++ b/drivers/usb/gadget/f_phonet.c
> @@ -207,6 +207,11 @@ static int pn_net_close(struct net_device *dev)
>  	return 0;
>  }
>  
> +static struct net_device_stats *pn_net_stats(struct net_device *dev)
> +{
> +	return &dev->stats;
> +}
> +
>  static void pn_tx_complete(struct usb_ep *ep, struct usb_request *req)
>  {
>  	struct f_phonet *fp = ep->driver_data;
> @@ -279,6 +284,7 @@ static int pn_net_mtu(struct net_device *dev, int new_mtu)
>  static const struct net_device_ops pn_netdev_ops = {
>  	.ndo_open	= pn_net_open,
>  	.ndo_stop	= pn_net_close,
> +	.ndo_get_stats	= pn_net_stats,
>  	.ndo_start_xmit	= pn_net_xmit,
>  	.ndo_change_mtu	= pn_net_mtu,
>  };
> diff --git a/drivers/usb/gadget/u_ether.c b/drivers/usb/gadget/u_ether.c
> index 1eda968..f21520f 100644
> --- a/drivers/usb/gadget/u_ether.c
> +++ b/drivers/usb/gadget/u_ether.c
> @@ -702,6 +702,11 @@ static int eth_stop(struct net_device *net)
>  	return 0;
>  }
>  
> +static struct net_device_stats *eth_get_stats(struct net_device *net)
> +{
> +	return &net->stats;
> +}
> +
>  /*-------------------------------------------------------------------------*/
>  
>  /* initial value, changed by "ifconfig usb0 hw ether xx:xx:xx:xx:xx:xx" */
> @@ -740,6 +745,7 @@ static struct eth_dev *the_dev;
>  static const struct net_device_ops eth_netdev_ops = {
>  	.ndo_open		= eth_open,
>  	.ndo_stop		= eth_stop,
> +	.ndo_get_stats		= eth_get_stats,
>  	.ndo_start_xmit		= eth_start_xmit,
>  	.ndo_change_mtu		= ueth_change_mtu,
>  	.ndo_set_mac_address 	= eth_mac_addr,
> --

Hmm... 

So all net devices in gen_ndis_query_resp() should have a
ndo_get_stats() or ndo_get_stats64() method, not allowed to use
spin_lock_bh() / spin_unlock_bh()

If yes, we should add big fat comments to pn_net_stats()/eth_get_stats() so that nobody
tries to revert your patch ;)



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Wed, Jan 12, 2011 at 02:07:12PM +0100, Eric Dumazet wrote:
...
> 
> Hmm... 
> 
> So all net devices in gen_ndis_query_resp() should have a
> ndo_get_stats() or ndo_get_stats64() method, not allowed to use
> spin_lock_bh() / spin_unlock_bh()
> 
> If yes, we should add big fat comments to pn_net_stats()/eth_get_stats() so that nobody
> tries to revert your patch ;)

I'm not sure I got your point: my patch could be replaced with
ndo_get_stats64() implementing irq safe locking or by changing
gen_ndis_query_resp() calling context. It's intended as a fast
(compatible) fix.

Jarek P.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Le mercredi 12 janvier 2011 à 13:23 +0000, Jarek Poplawski a écrit :
> On Wed, Jan 12, 2011 at 02:07:12PM +0100, Eric Dumazet wrote:
> ...
> > 
> > Hmm... 
> > 
> > So all net devices in gen_ndis_query_resp() should have a
> > ndo_get_stats() or ndo_get_stats64() method, not allowed to use
> > spin_lock_bh() / spin_unlock_bh()
> > 
> > If yes, we should add big fat comments to pn_net_stats()/eth_get_stats() so that nobody
> > tries to revert your patch ;)
> 
> I'm not sure I got your point: my patch could be replaced with
> ndo_get_stats64() implementing irq safe locking or by changing
> gen_ndis_query_resp() calling context. It's intended as a fast
> (compatible) fix.

I was mentioning that we tried in past months to remove useless
ndo_get_stats() methods that were only doing :

return &net->stats;

random commit : b27d50a9ff5cf2775b7a4daf5

Another possibility would be to use u64_stats_sync.h for these txq
counters.

(no locking needed to read counters, only a seqcount fetch/retry)

As a bonus, no overhead on 64bit arches.


--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

On Wed, Jan 12, 2011 at 02:32:20PM +0100, Eric Dumazet wrote:
> Le mercredi 12 janvier 2011 ?? 13:23 +0000, Jarek Poplawski a écrit :
> > On Wed, Jan 12, 2011 at 02:07:12PM +0100, Eric Dumazet wrote:
> > ...
> > > 
> > > Hmm... 
> > > 
> > > So all net devices in gen_ndis_query_resp() should have a
> > > ndo_get_stats() or ndo_get_stats64() method, not allowed to use
> > > spin_lock_bh() / spin_unlock_bh()
> > > 
> > > If yes, we should add big fat comments to pn_net_stats()/eth_get_stats() so that nobody
> > > tries to revert your patch ;)
> > 
> > I'm not sure I got your point: my patch could be replaced with
> > ndo_get_stats64() implementing irq safe locking or by changing
> > gen_ndis_query_resp() calling context. It's intended as a fast
> > (compatible) fix.
> 
> I was mentioning that we tried in past months to remove useless
> ndo_get_stats() methods that were only doing :
> 
> return &net->stats;
> 
> random commit : b27d50a9ff5cf2775b7a4daf5
> 
> Another possibility would be to use u64_stats_sync.h for these txq
> counters.
> 
> (no locking needed to read counters, only a seqcount fetch/retry)
> 
> As a bonus, no overhead on 64bit arches.

Well, I'm OK with anything, but this thread gets older and older ;-)

Jarek P.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Le mercredi 12 janvier 2011 à 14:32 +0100, Eric Dumazet a écrit :
> Le mercredi 12 janvier 2011 à 13:23 +0000, Jarek Poplawski a écrit :
> > On Wed, Jan 12, 2011 at 02:07:12PM +0100, Eric Dumazet wrote:
> > ...
> > > 
> > > Hmm... 
> > > 
> > > So all net devices in gen_ndis_query_resp() should have a
> > > ndo_get_stats() or ndo_get_stats64() method, not allowed to use
> > > spin_lock_bh() / spin_unlock_bh()
> > > 
> > > If yes, we should add big fat comments to pn_net_stats()/eth_get_stats() so that nobody
> > > tries to revert your patch ;)
> > 
> > I'm not sure I got your point: my patch could be replaced with
> > ndo_get_stats64() implementing irq safe locking or by changing
> > gen_ndis_query_resp() calling context. It's intended as a fast
> > (compatible) fix.
> 
> I was mentioning that we tried in past months to remove useless
> ndo_get_stats() methods that were only doing :
> 
> return &net->stats;
> 
> random commit : b27d50a9ff5cf2775b7a4daf5
> 
> Another possibility would be to use u64_stats_sync.h for these txq
> counters.
> 
> (no locking needed to read counters, only a seqcount fetch/retry)
> 
> As a bonus, no overhead on 64bit arches.
> 

Or even better, remove these counters since there is no users left but
ixgbe.

(vlans, tunnels, ... now use percpu stats)



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html