Re: [PATCH v4] kni: fix device address set

["Min Hu (Connor)" <humin29@huawei.com>]

Hi, Andrew,

在 2022/6/1 17:02, Andrew Rybchenko 写道:
> On 6/1/22 04:59, Min Hu (Connor) wrote:
>> Currently, run KNI APP When Kernel version is 5.17. When quit the APP,
>> dmesg buffer get calltrace, info like:
>> [ 5965.847401] rte_kni: Creating kni...
>> [ 6225.627205] vEth0 (unregistered): Current addr:  70 fd 45 d0 72 a7 
>> 00..
>> [ 6225.640113] vEth0 (unregistered): Expected addr: 00 00 00 00 00 00 
>> 00..
>> [ 6225.653010] ------------[ cut here ]------------
>> [ 6225.657609] netdevice: vEth0 (unregistered): Incorrect 
>> netdev->dev_addr
>> [ 6225.832647] Call trace:
>> [ 6225.835083]  dev_addr_check+0xa0/0x144
>> [ 6225.838816]  dev_addr_flush+0x30/0x9c
>> [ 6225.842462]  free_netdev+0x8c/0x1e0
>> [ 6225.845939]  kni_release+0xc0/0x1d0 [rte_kni]
>> [ 6225.850281]  __fput+0x78/0x220
>> [ 6225.853327]  ____fput+0x1c/0x30
>> [ 6225.856455]  task_work_run+0x88/0xc0
>> [ 6225.860017]  do_exit+0x2fc/0x940
>> [ 6225.863232]  do_group_exit+0x40/0xac
>> [ 6225.866791]  get_signal+0x190/0x960
>> [ 6225.870265]  do_notify_resume+0x26c/0x1360
>> [ 6225.874346]  el0_interrupt+0x60/0xe0
>> [ 6225.877910]  __el0_irq_handler_common+0x18/0x24
>> [ 6225.882420]  el0t_64_irq_handler+0x14/0x20
>> [ 6225.886499]  el0t_64_irq+0x1a0/0x1a4
>> [ 6225.890059] ---[ end trace 0000000000000000 ]---
>> [ 6245.598157] rte_kni: Creating kni...
>>
>> The reason is that 5.17 kernel introduce 'dev_addr_shadow' in function
>> 'dev_addr_set' to ensure that netdev->dev_addr should only be modified
>> via helpers('dev_addr_set'). 'dev_addr_check' will check if
>> netdev->dev_addr is modified by other ways, like 'memcpy'.
>>
>> More info could get by referring to kernel patch:
>> https://patchwork.kernel.org/project/netdevbpf/patch/
>> 20211118041501.3102861-8-kuba@kernel.org/
>> https://www.spinics.net/lists/netdev/msg764992.html
>>
>> Fixes: ea6b39b5b847 ("kni: remove ethtool support")
>>
>> Signed-off-by: Min Hu (Connor) <humin29@huawei.com>
>> ---
>> v3,v4:
>> * fix compiling errors.
>>
>> v2:
>> * change 'memcpy' to 'ether_addr_copy' to copy device addr.
>> ---
>>   kernel/linux/kni/kni_misc.c | 10 ++++++++--
>>   1 file changed, 8 insertions(+), 2 deletions(-)
>>
>> diff --git a/kernel/linux/kni/kni_misc.c b/kernel/linux/kni/kni_misc.c
>> index 780187d8bf..9c5d8c5576 100644
>> --- a/kernel/linux/kni/kni_misc.c
>> +++ b/kernel/linux/kni/kni_misc.c
>> @@ -299,6 +299,7 @@ kni_ioctl_create(struct net *net, uint32_t ioctl_num,
>>       struct kni_net *knet = net_generic(net, kni_net_id);
>>       int ret;
>>       struct rte_kni_device_info dev_info;
>> +    unsigned char mac_addr[ETH_ALEN];
>>       struct net_device *net_dev = NULL;
>>       struct kni_dev *kni, *dev, *n;
>> @@ -403,10 +404,15 @@ kni_ioctl_create(struct net *net, uint32_t 
>> ioctl_num,
>>       /* if user has provided a valid mac address */
>>       if (is_valid_ether_addr(dev_info.mac_addr))
>> -        memcpy(net_dev->dev_addr, dev_info.mac_addr, ETH_ALEN);
>> +        ether_addr_copy(mac_addr, dev_info.mac_addr);
>>       else
>>           /* Generate random MAC address. */
>> -        eth_random_addr(net_dev->dev_addr);
>> +        eth_random_addr(mac_addr);
> 
> Isn't it better to use function provided by the kernel, if available?
> I mean eth_hw_addr_random() used in [1]
Well, here 'mac_addr' is my object, this object will be used in
'dev_addr_set' in the next code.
using eth_hw_addr_random will directly operate net-dev->dev_addr.

> 
> [1] 
> https://patches.dpdk.org/project/dpdk/patch/20220601054525.7573-1-ke1x.zhang@intel.com/ 
> 
> 
>> +#if KERNEL_VERSION(5, 17, 0) > LINUX_VERSION_CODE
> 
> I think usage of HAVE_* defines like in [1] look much better.
> It is easier to read and will limit scope if condition is finally
> more complicated.
Yes, fixed in v5.
> 
> [1] 
> https://patches.dpdk.org/project/dpdk/patch/20220525102641.20982-1-jslaby@suse.cz/ 
> 
> 
>> +    ether_addr_copy(net_dev->dev_addr, mac_addr);
>> +#else
>> +    dev_addr_set(net_dev, mac_addr);
>> +#endif
>>       if (dev_info.mtu)
>>           net_dev->mtu = dev_info.mtu;
> 
> .