unsupported prctl option 1398164801

[PMc]

kernel: [4025] linux: jid 5 pid 88403 (dart:io EventHa): unsupported prctl option 1398164801edge kernel: [4025] linux: jid 5 pid 88403 (dart:io EventHa): unsupported prctl option 1398164801

Anybody have seen this before? I've not (yet) noticed anything harmful, but it is flooding my system logs.

 

[PMc]

I found out a bit more. That 1398164801 seems to be PR_SET_VMA.
(That's exactly why I went away from linux in 1995: in order to map that number to a symbol name that can be found, one would need the headerfile/sourcecode, and with linux one just does not find the proper source, it's all dispersed any- and everywhere.)

PR_SET_VMA(2const) - Linux manual page

I don't get an idea what that might be good for.

⚙ D127326 [Sanitizers] prctl interception update for the PR_SET_VMA option case

reviews.llvm.org

creepy... apparently a googlism.

 

[ZioMario]

That's exactly why I went away from linux in 1995...

excusme,but why are you blaming Linux ? I'm pretty sure that's a bug in the Linuxulator. I've found the same error while I was using Ubuntu emulated with the Linuxulator. Not sure if using a jail or not :

Ubuntu Pastebin

pastebin.ubuntu.com

I'm not sure if that bug should be reported.

 

[PMc]

excusme,but why are you blaming Linux ?

Read above:

with linux one just does not find the proper source, it's all dispersed any- and everywhere.)

In order to figure out what this is about, one needs to read the headerfile in /usr/include, which would contain a list of defined symbols and their numbers. On Berkeley one can easily find the authoritative source, in /usr/include (at https://cgit.freebsd.org/ or such). For Linux, one can search a day and not find it - because there is no authoritative source anywhere.
This does not concern this specific issue, it concerns all issues.

I'm not sure if that bug should be reported.

I'm currently not up to date with this issue. Probably I just patched the source to suppress the message.

 

[PMc]

Yepp, I patched it to report only once and then let it be.

As it is now "concerns some people", it would indeed warrant a bug report/change request.

diff --git a/sys/compat/linux/linux_misc.c b/sys/compat/linux/linux_misc.c
index fa6b7eff601a..8573375b022e 100644
--- a/sys/compat/linux/linux_misc.c
+++ b/sys/compat/linux/linux_misc.c
@@ -1853,6 +1853,10 @@ linux_prctl(struct thread *td, struct linux_prctl_args *a
rgs)
                linux_msg(td, "unsupported prctl PR_SET_PTRACER");
                error = EINVAL;
                break;
+       case LINUX_PR_SET_VMA:
+               LINUX_RATELIMIT_MSG("unsupported prctl PR_SET_VMA");
+               error = EINVAL;
+               break;
        default:
                linux_msg(td, "unsupported prctl option %d", args->option);
                error = EINVAL;
diff --git a/sys/compat/linux/linux_misc.h b/sys/compat/linux/linux_misc.h
index 51c0ee40f5f8..8174a8a8a657 100644
--- a/sys/compat/linux/linux_misc.h
+++ b/sys/compat/linux/linux_misc.h
@@ -59,6 +59,7 @@
 #define        LINUX_PR_CAPBSET_READ   23
 #define        LINUX_PR_SET_NO_NEW_PRIVS       38
 #define        LINUX_PR_SET_PTRACER    1499557217
+#define        LINUX_PR_SET_VMA        1398164801

 #define        LINUX_MAX_COMM_LEN      16      /* Maximum length of the process name. */