netdev - Re: [PATCH v2] usbnet: optimize usbnet

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Wed, 21 Dec 2022 16:19:45 +0900
From:   Leesoo Ahn <lsahn@...eel.net>
To:     Greg KH <greg@...ah.com>
Cc:     Oliver Neukum <oneukum@...e.com>,
        "David S. Miller" <davem@...emloft.net>,
        Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>, netdev@...r.kernel.org,
        linux-usb@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2] usbnet: optimize usbnet_bh() to reduce CPU load


On 22. 12. 21. 15:32, Greg KH wrote:
> On Wed, Dec 21, 2022 at 01:42:30PM +0900, Leesoo Ahn wrote:
>> The current source pushes skb into dev->done queue by calling
>> skb_queue_tail() and then pop it by calling skb_dequeue() to branch to
>> rx_cleanup state for freeing urb/skb in usbnet_bh(). It takes extra CPU
>> load, 2.21% (skb_queue_tail) as follows.
>>
>> -   11.58%     0.26%  swapper          [k] usbnet_bh
>>     - 11.32% usbnet_bh
>>        - 6.43% skb_dequeue
>>             6.34% _raw_spin_unlock_irqrestore
>>        - 2.21% skb_queue_tail
>>             2.19% _raw_spin_unlock_irqrestore
>>        - 1.68% consume_skb
>>           - 0.97% kfree_skbmem
>>                0.80% kmem_cache_free
>>             0.53% skb_release_data
>>
>> To reduce the extra CPU load use return values jumping to rx_cleanup
>> state directly to free them instead of calling skb_queue_tail() and
>> skb_dequeue() for push/pop respectively.
>>
>> -    7.87%     0.25%  swapper          [k] usbnet_bh
>>     - 7.62% usbnet_bh
>>        - 4.81% skb_dequeue
>>             4.74% _raw_spin_unlock_irqrestore
>>        - 1.75% consume_skb
>>           - 0.98% kfree_skbmem
>>                0.78% kmem_cache_free
>>             0.58% skb_release_data
>>          0.53% smsc95xx_rx_fixup
>>
>> Signed-off-by: Leesoo Ahn <lsahn@...eel.net>
>> ---
>> v2:
>>    - Replace goto label with return statement to reduce goto entropy
>>    - Add CPU load information by perf in commit message
>>
>> v1 at:
>>    https://patchwork.kernel.org/project/netdevbpf/patch/20221217161851.829497-1-lsahn@ooseel.net/
>> ---
>>   drivers/net/usb/usbnet.c | 19 +++++++++----------
>>   1 file changed, 9 insertions(+), 10 deletions(-)
>>
>> diff --git a/drivers/net/usb/usbnet.c b/drivers/net/usb/usbnet.c
>> index 64a9a80b2309..6e82fef90dd9 100644
>> --- a/drivers/net/usb/usbnet.c
>> +++ b/drivers/net/usb/usbnet.c
>> @@ -555,32 +555,30 @@ static int rx_submit (struct usbnet *dev, struct urb *urb, gfp_t flags)
>>   
>>   /*-------------------------------------------------------------------------*/
>>   
>> -static inline void rx_process (struct usbnet *dev, struct sk_buff *skb)
>> +static inline int rx_process(struct usbnet *dev, struct sk_buff *skb)
>>   {
>>   	if (dev->driver_info->rx_fixup &&
>>   	    !dev->driver_info->rx_fixup (dev, skb)) {
>>   		/* With RX_ASSEMBLE, rx_fixup() must update counters */
>>   		if (!(dev->driver_info->flags & FLAG_RX_ASSEMBLE))
>>   			dev->net->stats.rx_errors++;
>> -		goto done;
>> +		return 1;
> "1" means that you processed 1 byte, not that this is an error, which is
> what you want to say here, right?
No not at all..
> Please return a negative error value
> like I asked this to be changed to last time :(
Could you help me to decide the message type at this point please? I am 
confused.

The return value totally depends on how rx_fixup() is. For instance, in 
smsc95xx.c, smsc95xx_rx_fixup() function returns 0 in two cases that

1) frame size is greater than ETH_FRAME_LEN(1526 bytes) as follows

  1853             /* ETH_FRAME_LEN + 4(CRC) + 2(COE) + 4(Vlan) */
  1854             if (unlikely(size > (ETH_FRAME_LEN + 12))) {
  1855                 netif_dbg(dev, rx_err, dev->net,
  1856                       "size err header=0x%08x\n", header);
  1857                 return 0;
  1858             }

2) it is failed for skb allocation, but memory?

  1870             ax_skb = skb_clone(skb, GFP_ATOMIC);
  1871             if (unlikely(!ax_skb)) {
  1872                 netdev_warn(dev->net, "Error allocating skb\n");
  1873                 return 0;
  1874             }

I guess EPROTO or ENOMEM, one of them could be the value at the point 
but I have no ideas..

Best regards,
Leesoo