public inbox for netdev@vger.kernel.org 
 help / color / mirror / Atom feed
* ioctl SIOCSHWTSTAMP failed on a vlan interface within non-default network namespace
@ 2023-03-15  5:35 Yan Jiang
  2023-03-15 14:38 ` Vadim Fedorenko
  0 siblings, 1 reply; 3+ messages in thread
From: Yan Jiang @ 2023-03-15  5:35 UTC (permalink / raw)
  To: davem@davemloft•net, edumazet@google•com, kuba@kernel•org,
	pabeni@redhat•com, richardcochran@gmail•com
  Cc: netdev@vger•kernel.org

Hi all,

I'd like to report an issue, which I suspect it's related to kernel network namespace handling. 
It's my first report so please bear with me if I missed something. Feel free to ask for more information please.

What's the issue:
Ptp4l works perfectly on a VLAN interface in default network namespace. But it doesn't work if the VLAN interface is in non-default network namespace: ioctl(fd, SIOCSHWTSTAMP, &ifreq) failed due to error "Operation not supported".
"ethtool -T" shows that the VLAN interface has all required capabilities. And ptp4l works fine on the base interface in that network namespace. 

How to reproduce this issue:
# create a new network namespace for test purpose
root@viavi-PowerEdge-R740:~# ip netns add mytest

# move eno4 into the namespace
root@viavi-PowerEdge-R740:~# ip link set eno4 netns mytest

# add vlan interface and turn interfaces up
root@viavi-PowerEdge-R740:~# ip netns exec mytest ifconfig eno4 up
root@viavi-PowerEdge-R740:~# ip netns exec mytest ip link add link eno4 name eno4.4000 type vlan id 4000
root@viavi-PowerEdge-R740:~# ip netns exec mytest ifconfig eno4.4000 up

# ptp4l runs ok on eno4
root@viavi-PowerEdge-R740:~# ip netns exec mytest ptp4l -2 -m -i eno4
ptp4l[1670814.457]: selected /dev/ptp1 as PTP clock
ptp4l[1670814.491]: port 1: INITIALIZING to LISTENING on INIT_COMPLETE
ptp4l[1670814.492]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE

#ptp4l cannot run on eno4.4000
root@viavi-PowerEdge-R740:~# ip netns exec mytest ptp4l -2 -m -i eno4.4000
ptp4l[1670819.969]: selected /dev/ptp1 as PTP clock
ptp4l[1670820.003]: driver rejected most general HWTSTAMP filter
ptp4l[1670820.003]: ioctl SIOCSHWTSTAMP failed: Operation not supported
ptp4l[1670820.039]: port 1: INITIALIZING to FAULTY on FAULT_DETECTED (FT_UNSPECIFIED)
ptp4l[1670820.039]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE

#ethtool -T shows that eno4.4000 has same capability as eno4:
root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -T eno4.4000
Time stamping parameters for eno4.4000:
Capabilities:
        hardware-transmit     (SOF_TIMESTAMPING_TX_HARDWARE)
        software-transmit     (SOF_TIMESTAMPING_TX_SOFTWARE)
        hardware-receive      (SOF_TIMESTAMPING_RX_HARDWARE)
        software-receive      (SOF_TIMESTAMPING_RX_SOFTWARE)
        software-system-clock (SOF_TIMESTAMPING_SOFTWARE)
        hardware-raw-clock    (SOF_TIMESTAMPING_RAW_HARDWARE)
PTP Hardware Clock: 1
Hardware Transmit Timestamp Modes:
        off                   (HWTSTAMP_TX_OFF)
        on                    (HWTSTAMP_TX_ON)
Hardware Receive Filter Modes:
        none                  (HWTSTAMP_FILTER_NONE)
        ptpv1-l4-event        (HWTSTAMP_FILTER_PTP_V1_L4_EVENT)
        ptpv2-l4-event        (HWTSTAMP_FILTER_PTP_V2_L4_EVENT)
        ptpv2-l2-event        (HWTSTAMP_FILTER_PTP_V2_L2_EVENT)

#OS/kernel version
root@viavi-PowerEdge-R740:~# uname -a
Linux viavi-PowerEdge-R740 5.15.0-60-generic #66~20.04.1-Ubuntu SMP Wed Jan 25 09:41:30 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux

#driver info: 
root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -i eno4
driver: tg3
version: 5.15.0-60-generic
firmware-version: FFV21.40.21 bc 5720-v1.39
expansion-rom-version:
bus-info: 0000:01:00.1
supports-statistics: yes
supports-test: yes
supports-eeprom-access: yes
supports-register-dump: yes
supports-priv-flags: no
root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -i eno4.4000
driver: 802.1Q VLAN Support
version: 1.8
firmware-version: N/A
expansion-rom-version:
bus-info:
supports-statistics: no
supports-test: no
supports-eeprom-access: no
supports-register-dump: no
supports-priv-flags: no

#(PS: can reproduce same issue with a Mellanox NIC, so I guess it's not caused by a specific NIC model)

Could you kindly take a look at this and see if this is a kernel issue? This blocks linuxptp running in cloud environment, if VLAN is required. 
Hope to get feedback from you soon, thanks!

/Yan

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: ioctl SIOCSHWTSTAMP failed on a vlan interface within non-default network namespace
  2023-03-15  5:35 ioctl SIOCSHWTSTAMP failed on a vlan interface within non-default network namespace Yan Jiang
@ 2023-03-15 14:38 ` Vadim Fedorenko
       [not found]   ` <BYAPR18MB2408F4FEDD31152C3C15D5648DBF9@BYAPR18MB2408.namprd18.prod.outlook.com>
  0 siblings, 1 reply; 3+ messages in thread
From: Vadim Fedorenko @ 2023-03-15 14:38 UTC (permalink / raw)
  To: Yan Jiang, davem@davemloft•net, edumazet@google•com,
	kuba@kernel•org, pabeni@redhat•com, richardcochran@gmail•com
  Cc: netdev@vger•kernel.org

On 15/03/2023 05:35, Yan Jiang wrote:
> Hi all,
> 
> I'd like to report an issue, which I suspect it's related to kernel network namespace handling.
> It's my first report so please bear with me if I missed something. Feel free to ask for more information please.
> 
> What's the issue:
> Ptp4l works perfectly on a VLAN interface in default network namespace. But it doesn't work if the VLAN interface is in non-default network namespace: ioctl(fd, SIOCSHWTSTAMP, &ifreq) failed due to error "Operation not supported".
> "ethtool -T" shows that the VLAN interface has all required capabilities. And ptp4l works fine on the base interface in that network namespace.


That was explicitly forbidden in
https://lore.kernel.org/netdev/20190509065507.23991-1-liuhangbin@gmail.com/


> How to reproduce this issue:
> # create a new network namespace for test purpose
> root@viavi-PowerEdge-R740:~# ip netns add mytest
> 
> # move eno4 into the namespace
> root@viavi-PowerEdge-R740:~# ip link set eno4 netns mytest
> 
> # add vlan interface and turn interfaces up
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ifconfig eno4 up
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ip link add link eno4 name eno4.4000 type vlan id 4000
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ifconfig eno4.4000 up
> 
> # ptp4l runs ok on eno4
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ptp4l -2 -m -i eno4
> ptp4l[1670814.457]: selected /dev/ptp1 as PTP clock
> ptp4l[1670814.491]: port 1: INITIALIZING to LISTENING on INIT_COMPLETE
> ptp4l[1670814.492]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE
> 
> #ptp4l cannot run on eno4.4000
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ptp4l -2 -m -i eno4.4000
> ptp4l[1670819.969]: selected /dev/ptp1 as PTP clock
> ptp4l[1670820.003]: driver rejected most general HWTSTAMP filter
> ptp4l[1670820.003]: ioctl SIOCSHWTSTAMP failed: Operation not supported
> ptp4l[1670820.039]: port 1: INITIALIZING to FAULTY on FAULT_DETECTED (FT_UNSPECIFIED)
> ptp4l[1670820.039]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE
> 
> #ethtool -T shows that eno4.4000 has same capability as eno4:
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -T eno4.4000
> Time stamping parameters for eno4.4000:
> Capabilities:
>          hardware-transmit     (SOF_TIMESTAMPING_TX_HARDWARE)
>          software-transmit     (SOF_TIMESTAMPING_TX_SOFTWARE)
>          hardware-receive      (SOF_TIMESTAMPING_RX_HARDWARE)
>          software-receive      (SOF_TIMESTAMPING_RX_SOFTWARE)
>          software-system-clock (SOF_TIMESTAMPING_SOFTWARE)
>          hardware-raw-clock    (SOF_TIMESTAMPING_RAW_HARDWARE)
> PTP Hardware Clock: 1
> Hardware Transmit Timestamp Modes:
>          off                   (HWTSTAMP_TX_OFF)
>          on                    (HWTSTAMP_TX_ON)
> Hardware Receive Filter Modes:
>          none                  (HWTSTAMP_FILTER_NONE)
>          ptpv1-l4-event        (HWTSTAMP_FILTER_PTP_V1_L4_EVENT)
>          ptpv2-l4-event        (HWTSTAMP_FILTER_PTP_V2_L4_EVENT)
>          ptpv2-l2-event        (HWTSTAMP_FILTER_PTP_V2_L2_EVENT)
> 
> #OS/kernel version
> root@viavi-PowerEdge-R740:~# uname -a
> Linux viavi-PowerEdge-R740 5.15.0-60-generic #66~20.04.1-Ubuntu SMP Wed Jan 25 09:41:30 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
> 
> #driver info:
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -i eno4
> driver: tg3
> version: 5.15.0-60-generic
> firmware-version: FFV21.40.21 bc 5720-v1.39
> expansion-rom-version:
> bus-info: 0000:01:00.1
> supports-statistics: yes
> supports-test: yes
> supports-eeprom-access: yes
> supports-register-dump: yes
> supports-priv-flags: no
> root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -i eno4.4000
> driver: 802.1Q VLAN Support
> version: 1.8
> firmware-version: N/A
> expansion-rom-version:
> bus-info:
> supports-statistics: no
> supports-test: no
> supports-eeprom-access: no
> supports-register-dump: no
> supports-priv-flags: no
> 
> #(PS: can reproduce same issue with a Mellanox NIC, so I guess it's not caused by a specific NIC model)
> 
> Could you kindly take a look at this and see if this is a kernel issue? This blocks linuxptp running in cloud environment, if VLAN is required.
> Hope to get feedback from you soon, thanks!
> 
> /Yan


^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: ioctl SIOCSHWTSTAMP failed on a vlan interface within non-default network namespace
       [not found]   ` <BYAPR18MB2408F4FEDD31152C3C15D5648DBF9@BYAPR18MB2408.namprd18.prod.outlook.com>
@ 2023-03-15 15:56     ` Vadim Fedorenko
  0 siblings, 0 replies; 3+ messages in thread
From: Vadim Fedorenko @ 2023-03-15 15:56 UTC (permalink / raw)
  To: Yan Jiang, davem@davemloft•net, edumazet@google•com,
	kuba@kernel•org, pabeni@redhat•com, richardcochran@gmail•com
  Cc: netdev@vger•kernel.org

On 15/03/2023 15:17, Yan Jiang wrote:
> HI Vadim,
> 
> Thank you for your reply. This explains why ioctl failed on vlan interface.
> 
> I am ignorant here, but in my example below both base interface and vlan 
> interface are in same namespace,  and I feel that the defense you 
> pointed to me is not protecting anything.
> We can even take the default namespace as example: why we allow ioctl 
> SIOCSHWTSTAMP to work on a VLAN interface in default namespace?
> 
> I think it is common to assign a sriov VF to a container. In this case, 
> both base interface and VLAN interface created on top of the VF, are in 
> same network namespace.

I've just proposed the patch to address the situation, hope it will find 
some support :)

> 
> Possibly there are some situations that I am not aware. Hope to have 
> your thoughts. Thank you!
> 
> 
> ------------------------------------------------------------------------
> *From:* Vadim Fedorenko <vadim.fedorenko@linux•dev>
> *Sent:* Wednesday, March 15, 2023 10:38 PM
> *To:* Yan Jiang <Yan.Jiang@viavisolutions•com>; davem@davemloft•net 
> <davem@davemloft•net>; edumazet@google•com <edumazet@google•com>; 
> kuba@kernel•org <kuba@kernel•org>; pabeni@redhat•com 
> <pabeni@redhat•com>; richardcochran@gmail•com <richardcochran@gmail•com>
> *Cc:* netdev@vger•kernel.org <netdev@vger•kernel.org>
> *Subject:* Re: ioctl SIOCSHWTSTAMP failed on a vlan interface within 
> non-default network namespace
> EXTERNAL EMAIL: Do not click links or open attachments unless you know 
> and trust the sender.
> 
> 
> On 15/03/2023 05:35, Yan Jiang wrote:
>> Hi all,
>>
>> I'd like to report an issue, which I suspect it's related to kernel network namespace handling.
>> It's my first report so please bear with me if I missed something. Feel free to ask for more information please.
>>
>> What's the issue:
>> Ptp4l works perfectly on a VLAN interface in default network namespace. But it doesn't work if the VLAN interface is in non-default network namespace: ioctl(fd, SIOCSHWTSTAMP, &ifreq) failed due to error "Operation not supported".
>> "ethtool -T" shows that the VLAN interface has all required capabilities. And ptp4l works fine on the base interface in that network namespace.
> 
> 
> That was explicitly forbidden in
> https://urldefense.com/v3/__https://lore.kernel.org/netdev/20190509065507.23991-1-liuhangbin@gmail.com/__;!!Niha4SQ!_NTn9qN2nCq-QapNeIln9DvJwmpQWK7Y4EJOdXjbOq6w7Bhh-ZQp-vJu8TPyI9QMigOIDPeN2IuZSJamNFZI4DAfQvMIWyjqEMj9$ <https://urldefense.com/v3/__https://lore.kernel.org/netdev/20190509065507.23991-1-liuhangbin@gmail.com/__;!!Niha4SQ!_NTn9qN2nCq-QapNeIln9DvJwmpQWK7Y4EJOdXjbOq6w7Bhh-ZQp-vJu8TPyI9QMigOIDPeN2IuZSJamNFZI4DAfQvMIWyjqEMj9$>
> 
> 
>> How to reproduce this issue:
>> # create a new network namespace for test purpose
>> root@viavi-PowerEdge-R740:~# ip netns add mytest
>>
>> # move eno4 into the namespace
>> root@viavi-PowerEdge-R740:~# ip link set eno4 netns mytest
>>
>> # add vlan interface and turn interfaces up
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ifconfig eno4 up
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ip link add link eno4 name eno4.4000 type vlan id 4000
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ifconfig eno4.4000 up
>>
>> # ptp4l runs ok on eno4
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ptp4l -2 -m -i eno4
>> ptp4l[1670814.457]: selected /dev/ptp1 as PTP clock
>> ptp4l[1670814.491]: port 1: INITIALIZING to LISTENING on INIT_COMPLETE
>> ptp4l[1670814.492]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE
>>
>> #ptp4l cannot run on eno4.4000
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ptp4l -2 -m -i eno4.4000
>> ptp4l[1670819.969]: selected /dev/ptp1 as PTP clock
>> ptp4l[1670820.003]: driver rejected most general HWTSTAMP filter
>> ptp4l[1670820.003]: ioctl SIOCSHWTSTAMP failed: Operation not supported
>> ptp4l[1670820.039]: port 1: INITIALIZING to FAULTY on FAULT_DETECTED (FT_UNSPECIFIED)
>> ptp4l[1670820.039]: port 0: INITIALIZING to LISTENING on INIT_COMPLETE
>>
>> #ethtool -T shows that eno4.4000 has same capability as eno4:
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -T eno4.4000
>> Time stamping parameters for eno4.4000:
>> Capabilities:
>>          hardware-transmit     (SOF_TIMESTAMPING_TX_HARDWARE)
>>          software-transmit     (SOF_TIMESTAMPING_TX_SOFTWARE)
>>          hardware-receive      (SOF_TIMESTAMPING_RX_HARDWARE)
>>          software-receive      (SOF_TIMESTAMPING_RX_SOFTWARE)
>>          software-system-clock (SOF_TIMESTAMPING_SOFTWARE)
>>          hardware-raw-clock    (SOF_TIMESTAMPING_RAW_HARDWARE)
>> PTP Hardware Clock: 1
>> Hardware Transmit Timestamp Modes:
>>          off                   (HWTSTAMP_TX_OFF)
>>          on                    (HWTSTAMP_TX_ON)
>> Hardware Receive Filter Modes:
>>          none                  (HWTSTAMP_FILTER_NONE)
>>          ptpv1-l4-event        (HWTSTAMP_FILTER_PTP_V1_L4_EVENT)
>>          ptpv2-l4-event        (HWTSTAMP_FILTER_PTP_V2_L4_EVENT)
>>          ptpv2-l2-event        (HWTSTAMP_FILTER_PTP_V2_L2_EVENT)
>>
>> #OS/kernel version
>> root@viavi-PowerEdge-R740:~# uname -a
>> Linux viavi-PowerEdge-R740 5.15.0-60-generic #66~20.04.1-Ubuntu SMP Wed Jan 25 09:41:30 UTC 2023 x86_64 x86_64 x86_64 GNU/Linux
>>
>> #driver info:
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -i eno4
>> driver: tg3
>> version: 5.15.0-60-generic
>> firmware-version: FFV21.40.21 bc 5720-v1.39
>> expansion-rom-version:
>> bus-info: 0000:01:00.1
>> supports-statistics: yes
>> supports-test: yes
>> supports-eeprom-access: yes
>> supports-register-dump: yes
>> supports-priv-flags: no
>> root@viavi-PowerEdge-R740:~# ip netns exec mytest ethtool -i eno4.4000
>> driver: 802.1Q VLAN Support
>> version: 1.8
>> firmware-version: N/A
>> expansion-rom-version:
>> bus-info:
>> supports-statistics: no
>> supports-test: no
>> supports-eeprom-access: no
>> supports-register-dump: no
>> supports-priv-flags: no
>>
>> #(PS: can reproduce same issue with a Mellanox NIC, so I guess it's not caused by a specific NIC model)
>>
>> Could you kindly take a look at this and see if this is a kernel issue? This blocks linuxptp running in cloud environment, if VLAN is required.
>> Hope to get feedback from you soon, thanks!
>>
>> /Yan
> 


^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2023-03-15 15:57 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2023-03-15  5:35 ioctl SIOCSHWTSTAMP failed on a vlan interface within non-default network namespace Yan Jiang
2023-03-15 14:38 ` Vadim Fedorenko
     [not found]   ` <BYAPR18MB2408F4FEDD31152C3C15D5648DBF9@BYAPR18MB2408.namprd18.prod.outlook.com>
2023-03-15 15:56     ` Vadim Fedorenko

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox