Currently there are no strict checks while setting SO_TXTIME
from userspace. With the recent development in skb->tstamp_type
clockid with unsupported clocks results in warn_on_once, which causes
unnecessary aborts in some systems which enables panic on warns.
Add validation in setsockopt to support only CLOCK_REALTIME,
CLOCK_MONOTONIC and CLOCK_TAI to be set from userspace.
Link: https://lore.kernel.org/netdev/[email protected]/
Link: https://lore.kernel.org/lkml/[email protected]/
Fixes: 1693c5db6ab8 ("net: Add additional bit to support clockid_t timestamp type")
Reported-by: [email protected]
Closes: https://syzkaller.appspot.com/bug?extid=d7b227731ec589e7f4f0
Reported-by: [email protected]
Closes: https://syzkaller.appspot.com/bug?extid=30a35a2e9c5067cc43fa
Signed-off-by: Abhishek Chauhan <[email protected]>
Acked-by: Martin KaFai Lau <[email protected]>
---
Changes since v1
- Moved from net to net-next since
Fixes tag is available only on net-next
as mentioned by Martin
- Added direct link to design discussion as
mentioned by Willem.
- Parameter in the sockopt_validate_clockid
is of type __kernel_clockid_t so changed it from
int to __kernel_clockid_t as mentioned by
Willem.
- Added Acked-by tag.
net/core/sock.c | 16 ++++++++++++++++
1 file changed, 16 insertions(+)
diff --git a/net/core/sock.c b/net/core/sock.c
index 8629f9aecf91..d497285f283a 100644
--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1083,6 +1083,17 @@ bool sockopt_capable(int cap)
}
EXPORT_SYMBOL(sockopt_capable);
+static int sockopt_validate_clockid(__kernel_clockid_t value)
+{
+ switch (value) {
+ case CLOCK_REALTIME:
+ case CLOCK_MONOTONIC:
+ case CLOCK_TAI:
+ return 0;
+ }
+ return -EINVAL;
+}
+
/*
* This is meant for all protocols to use and covers goings on
* at the socket level. Everything here is generic.
@@ -1497,6 +1508,11 @@ int sk_setsockopt(struct sock *sk, int level, int optname,
ret = -EPERM;
break;
}
+
+ ret = sockopt_validate_clockid(sk_txtime.clockid);
+ if (ret)
+ break;
+
sock_valbool_flag(sk, SOCK_TXTIME, true);
sk->sk_clockid = sk_txtime.clockid;
sk->sk_txtime_deadline_mode =
--
2.25.1
Abhishek Chauhan wrote:
> Currently there are no strict checks while setting SO_TXTIME
> from userspace. With the recent development in skb->tstamp_type
> clockid with unsupported clocks results in warn_on_once, which causes
> unnecessary aborts in some systems which enables panic on warns.
>
> Add validation in setsockopt to support only CLOCK_REALTIME,
> CLOCK_MONOTONIC and CLOCK_TAI to be set from userspace.
>
> Link: https://lore.kernel.org/netdev/[email protected]/
> Link: https://lore.kernel.org/lkml/[email protected]/
> Fixes: 1693c5db6ab8 ("net: Add additional bit to support clockid_t timestamp type")
> Reported-by: [email protected]
> Closes: https://syzkaller.appspot.com/bug?extid=d7b227731ec589e7f4f0
> Reported-by: [email protected]
> Closes: https://syzkaller.appspot.com/bug?extid=30a35a2e9c5067cc43fa
> Signed-off-by: Abhishek Chauhan <[email protected]>
> Acked-by: Martin KaFai Lau <[email protected]>
Reviewed-by: Willem de Bruijn <[email protected]>
> ---
> Changes since v1
> - Moved from net to net-next since
> Fixes tag is available only on net-next
> as mentioned by Martin
> - Added direct link to design discussion as
> mentioned by Willem.
> - Parameter in the sockopt_validate_clockid
> is of type __kernel_clockid_t so changed it from
> int to __kernel_clockid_t as mentioned by
> Willem.
> - Added Acked-by tag.
>
> net/core/sock.c | 16 ++++++++++++++++
> 1 file changed, 16 insertions(+)
>
> diff --git a/net/core/sock.c b/net/core/sock.c
> index 8629f9aecf91..d497285f283a 100644
> --- a/net/core/sock.c
> +++ b/net/core/sock.c
> @@ -1083,6 +1083,17 @@ bool sockopt_capable(int cap)
> }
> EXPORT_SYMBOL(sockopt_capable);
>
> +static int sockopt_validate_clockid(__kernel_clockid_t value)
The __kernel variants are UAPI. It looks odd to use this in kernel
internal code, and I don't see many examples immediately.
But I believe it is correct, as we're passing a sock_txtime.clockid
field, which has this (UAPI) type.
> +{
> + switch (value) {
> + case CLOCK_REALTIME:
> + case CLOCK_MONOTONIC:
> + case CLOCK_TAI:
> + return 0;
> + }
> + return -EINVAL;
> +}
> +
> /*
> * This is meant for all protocols to use and covers goings on
> * at the socket level. Everything here is generic.
> @@ -1497,6 +1508,11 @@ int sk_setsockopt(struct sock *sk, int level, int optname,
> ret = -EPERM;
> break;
> }
> +
> + ret = sockopt_validate_clockid(sk_txtime.clockid);
> + if (ret)
> + break;
> +
> sock_valbool_flag(sk, SOCK_TXTIME, true);
> sk->sk_clockid = sk_txtime.clockid;
> sk->sk_txtime_deadline_mode =
> --
> 2.25.1
>
Hello:
This patch was applied to netdev/net-next.git (main)
by Jakub Kicinski <[email protected]>:
On Wed, 29 May 2024 11:31:30 -0700 you wrote:
> Currently there are no strict checks while setting SO_TXTIME
> from userspace. With the recent development in skb->tstamp_type
> clockid with unsupported clocks results in warn_on_once, which causes
> unnecessary aborts in some systems which enables panic on warns.
>
> Add validation in setsockopt to support only CLOCK_REALTIME,
> CLOCK_MONOTONIC and CLOCK_TAI to be set from userspace.
>
> [...]
Here is the summary with links:
- [net-next,v2] net: validate SO_TXTIME clockid coming from userspace
https://git.kernel.org/netdev/net-next/c/73451e9aaa24
You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html