diff mbox series

[5.9,580/757] selftests: mptcp: depends on built-in IPv6

Message ID 20201027135517.727716271@linuxfoundation.org
State New
Headers show
Series None | expand

Commit Message

Greg Kroah-Hartman Oct. 27, 2020, 1:53 p.m. UTC
From: Matthieu Baerts <matthieu.baerts@tessares.net>

[ Upstream commit 287d35405989cfe0090e3059f7788dc531879a8d ]

Recently, CONFIG_MPTCP_IPV6 no longer selects CONFIG_IPV6. As a
consequence, if CONFIG_MPTCP_IPV6=y is added to the kconfig, it will no
longer ensure CONFIG_IPV6=y. If it is not enabled, CONFIG_MPTCP_IPV6
will stay disabled and selftests will fail.

We also need CONFIG_IPV6 to be built-in. For more details, please see
commit 0ed37ac586c0 ("mptcp: depends on IPV6 but not as a module").

Note that 'make kselftest-merge' will take all 'config' files found in
'tools/testsing/selftests'. Because some of them already set
CONFIG_IPV6=y, MPTCP selftests were still passing. But they will fail if
MPTCP selftests are launched manually after having executed this command
to prepare the kernel config:

  ./scripts/kconfig/merge_config.sh -m .config \
      ./tools/testing/selftests/net/mptcp/config

Fixes: 010b430d5df5 ("mptcp: MPTCP_IPV6 should depend on IPV6 instead of selecting it")
Signed-off-by: Matthieu Baerts <matthieu.baerts@tessares.net>
Reviewed-by: Mat Martineau <mathew.j.martineau@linux.intel.com>
Link: https://lore.kernel.org/r/20201021155549.933731-1-matthieu.baerts@tessares.net
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 tools/testing/selftests/net/mptcp/config | 1 +
 1 file changed, 1 insertion(+)

Comments

Matthieu Baerts Oct. 27, 2020, 3:59 p.m. UTC | #1
Hi Greg,

On 27/10/2020 14:53, Greg Kroah-Hartman wrote:
> From: Matthieu Baerts <matthieu.baerts@tessares.net>
> 
> [ Upstream commit 287d35405989cfe0090e3059f7788dc531879a8d ]
> 
> Recently, CONFIG_MPTCP_IPV6 no longer selects CONFIG_IPV6. As a
> consequence, if CONFIG_MPTCP_IPV6=y is added to the kconfig, it will no
> longer ensure CONFIG_IPV6=y. If it is not enabled, CONFIG_MPTCP_IPV6
> will stay disabled and selftests will fail.

I think Sasha wanted to drop this patch [1]. But as I said earlier [2], 
this patch is not wrong, it is just not needed in v5.9 because the 
"Fixes" commit 010b430d5df5 ("mptcp: MPTCP_IPV6 should depend on IPV6 
instead of selecting it") is not in v5.9.

I guess it is certainly easier to keep it because it is not wrong and 
doesn't hurt.

[1] https://lore.kernel.org/stable/20201026142329.GJ4060117@sasha-vm/
[2] 
https://lore.kernel.org/stable/1de2bf78-4b47-21b0-9d56-3c8063cdf4bb@tessares.net/

Cheers,
Matt
Sasha Levin Oct. 28, 2020, 10:09 p.m. UTC | #2
On Tue, Oct 27, 2020 at 04:59:45PM +0100, Matthieu Baerts wrote:
>Hi Greg,
>
>On 27/10/2020 14:53, Greg Kroah-Hartman wrote:
>>From: Matthieu Baerts <matthieu.baerts@tessares.net>
>>
>>[ Upstream commit 287d35405989cfe0090e3059f7788dc531879a8d ]
>>
>>Recently, CONFIG_MPTCP_IPV6 no longer selects CONFIG_IPV6. As a
>>consequence, if CONFIG_MPTCP_IPV6=y is added to the kconfig, it will no
>>longer ensure CONFIG_IPV6=y. If it is not enabled, CONFIG_MPTCP_IPV6
>>will stay disabled and selftests will fail.
>
>I think Sasha wanted to drop this patch [1]. But as I said earlier 
>[2], this patch is not wrong, it is just not needed in v5.9 because 
>the "Fixes" commit 010b430d5df5 ("mptcp: MPTCP_IPV6 should depend on 
>IPV6 instead of selecting it") is not in v5.9.
>
>I guess it is certainly easier to keep it because it is not wrong and 
>doesn't hurt.

I dropped a wrong patch by accident :(

Fixing this up... Thanks for pointing it out!
diff mbox series

Patch

diff --git a/tools/testing/selftests/net/mptcp/config b/tools/testing/selftests/net/mptcp/config
index 8df5cb8f71ff9..741a1c4f4ae8f 100644
--- a/tools/testing/selftests/net/mptcp/config
+++ b/tools/testing/selftests/net/mptcp/config
@@ -1,4 +1,5 @@ 
 CONFIG_MPTCP=y
+CONFIG_IPV6=y
 CONFIG_MPTCP_IPV6=y
 CONFIG_INET_DIAG=m
 CONFIG_INET_MPTCP_DIAG=m