Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting

From	Johannes Berg <johannes@sipsolutions.net>
Newsgroups	linux.kernel
Subject	Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting
Date	2015-08-25 11:40 +0200
Message-ID	<q1iRZ-2x2-17@gated-at.bofh.it> (permalink)
References	(1 earlier) <q1ifh-1xW-1@gated-at.bofh.it> <q1ioW-1YL-15@gated-at.bofh.it> <q1iyE-2ao-69@gated-at.bofh.it> <q1iRZ-2x2-19@gated-at.bofh.it> <q1iyE-2ao-69@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Show all headers | View raw

On Tue, 2015-08-25 at 11:17 +0200, Ingo Molnar wrote:
> 
> If we do that then we don't even have to introduce per system call error code 
> conversion, but could unconditionally save the last extended error info in the 
> task struct and continue - this could be done very cheaply with the linker trick 
> driven integer ID.
> 
> I.e. system calls could opt in to do:
> 
> 	> return err_str(-EBUSY, "perf/x86: BTS conflicts with active events");
> 
> and the overhead of this would be minimal, we'd essentially do something like this 
> to save the error:
> 
> 	> current->err_code = code;
> 
> where 'code' is a build time constant in essence.
> 
> We could use this even in system calls where the error path is performance 
> critical, as all the string recovery and copying overhead would be triggered by 
> applications that opt in via the new system call:
> 
> 	> struct err_desc {
> 	>        const char              *message;
> 	>        const char              *owner;
> 	>        const int               code;
> 	> };
> 
> 	> SyS_err_get_desc(struct err_desc *err_desc __user);
> 
> [ Which could perhaps be a prctl() extension as well (PR_GET_ERR_DESC): finally 
>   some truly matching functionality for prctl(). ]
> 
> Hm?

That's neat in a way, but doesn't work in general I think.

Considering the wifi case, or more generally any netlink based
protocol, the syscall (sendmsg) won't return an error, but a subsequent
recvmsg() (which also won't return an error) returns an error message
[in the sense of a protocol message, not a human readable message] to a
buffer provided by the application.
However, this message can be extended relatively easily to include the
string information, but the syscall/prctl wouldn't work since the
syscalls didn't actually fail.

However, it could possibly help with the namespace/module issue if you
also store THIS_MODULE (or perhaps instead a pointer to the module's
error table) in the task. Again not in the netlink case though, I
think, that will always require special handling [although there it
could be stored away in the socket or so, similar to the task]

johannes
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/

Thread

[PATCH v2 0/6] perf: Introduce extended syscall error reporting Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-24 16:40 +0200
  [PATCH v2 5/6] perf/x86/intel/pt: Use extended error reporting in event initialization Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-24 16:40 +0200
  [PATCH v2 6/6] perf/x86/intel/bts: Use extended error reporting in event initialization Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-24 16:40 +0200
  [PATCH v2 1/6] perf: Introduce extended syscall error reporting Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-24 16:40 +0200
    Re: [PATCH v2 1/6] perf: Introduce extended syscall error reporting Andy Shevchenko <andy.shevchenko@gmail.com> - 2015-08-31 20:50 +0200
      Re: [PATCH v2 1/6] perf: Introduce extended syscall error reporting Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-09-01 08:40 +0200
  [PATCH v2 3/6] perf: Annotate some of the error codes with perf_err() Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-24 16:40 +0200
  [PATCH v2 2/6] perf: Add file name and line number to perf extended error reports Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-24 16:40 +0200
  [PATCH v2 4/6] perf/x86: Annotate some of the error codes with perf_err() Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-24 16:40 +0200
  Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-25 10:30 +0200
  Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Johannes Berg <johannes@sipsolutions.net> - 2015-08-25 11:00 +0200
    Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-25 11:10 +0200
      Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-25 11:20 +0200
        Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Johannes Berg <johannes@sipsolutions.net> - 2015-08-25 11:40 +0200
          Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-25 12:10 +0200
            Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Johannes Berg <johannes@sipsolutions.net> - 2015-08-25 12:30 +0200
              Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-26 07:00 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Johannes Berg <johannes@sipsolutions.net> - 2015-08-26 09:10 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-26 09:30 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-26 19:00 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Arnaldo Carvalho de Melo <acme@kernel.org> - 2015-08-26 23:00 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Andrew Morton <akpm@linux-foundation.org> - 2015-08-26 20:50 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Peter Zijlstra <peterz@infradead.org> - 2015-08-26 22:10 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Andrew Morton <akpm@linux-foundation.org> - 2015-08-26 22:30 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error  reporting Vince Weaver <vince@deater.net> - 2015-08-26 23:00 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Andrew Morton <akpm@linux-foundation.org> - 2015-08-26 23:00 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error  reporting Vince Weaver <vince@deater.net> - 2015-08-26 23:20 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-28 12:10 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Arnaldo Carvalho de Melo <acme@infradead.org> - 2015-08-26 23:10 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-26 09:30 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Johannes Berg <johannes@sipsolutions.net> - 2015-08-26 09:40 +0200
                Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Ingo Molnar <mingo@kernel.org> - 2015-08-26 09:10 +0200
        Re: [PATCH v2 0/6] perf: Introduce extended syscall error reporting Alexander Shishkin <alexander.shishkin@linux.intel.com> - 2015-08-26 13:40 +0200

csiph-web