linux-kernel - Re: [PATCH 3/3] refscale: always log the error message

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives
Hash Suite: Windows password security audit tool. GUI, reports in PDF.
[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Date:   Thu, 28 Oct 2021 06:30:49 -0700
From:   "Paul E. McKenney" <paulmck@...nel.org>
To:     "lizhijian@...itsu.com" <lizhijian@...itsu.com>
Cc:     "dave@...olabs.net" <dave@...olabs.net>,
        "josh@...htriplett.org" <josh@...htriplett.org>,
        "rostedt@...dmis.org" <rostedt@...dmis.org>,
        "mathieu.desnoyers@...icios.com" <mathieu.desnoyers@...icios.com>,
        "jiangshanlai@...il.com" <jiangshanlai@...il.com>,
        "joel@...lfernandes.org" <joel@...lfernandes.org>,
        "rcu@...r.kernel.org" <rcu@...r.kernel.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH 3/3] refscale: always log the error message

On Wed, Oct 27, 2021 at 07:21:34AM +0000, lizhijian@...itsu.com wrote:
> Hi Paul
> 
> During adding the missing '\n', i also found there are some verbose control error message
> Should we convert it like [PATCH 3/3] refscale: always log the error message

These do indeed look like they should always be logged.  Good catches!

							Thanx, Paul

> lizj@...TPC: linux$ git grep VERBOSE.*_ERRSTRING
> include/linux/torture.h:#define VERBOSE_TOROUT_ERRSTRING(s) \
> kernel/locking/locktorture.c: VERBOSE_TOROUT_ERRSTRING("writer_tasks: Out of memory");
> kernel/locking/locktorture.c: VERBOSE_TOROUT_ERRSTRING("reader_tasks: Out of memory");
> kernel/rcu/rcuscale.c:#define VERBOSE_SCALEOUT_ERRSTRING(s) \
> kernel/rcu/rcuscale.c:          VERBOSE_SCALEOUT_ERRSTRING("All grace periods expedited, no normal ones to measure!");
> kernel/rcu/rcuscale.c:          VERBOSE_SCALEOUT_ERRSTRING("All grace periods normal, no expedited ones to measure!");
> kernel/rcu/rcuscale.c:          VERBOSE_SCALEOUT_ERRSTRING("No expedited async GPs, so went with async!");
> kernel/rcu/rcuscale.c:          VERBOSE_SCALEOUT_ERRSTRING("out of memory");
> kernel/rcu/rcuscale.c:          VERBOSE_SCALEOUT_ERRSTRING("out of memory");
> kernel/rcu/rcutorture.c: VERBOSE_TOROUT_ERRSTRING("out of memory");
> kernel/rcu/rcutorture.c: VERBOSE_TOROUT_ERRSTRING("out of memory");
> kernel/rcu/rcutorture.c: VERBOSE_TOROUT_ERRSTRING("out of memory");
> kernel/rcu/rcutorture.c: VERBOSE_TOROUT_ERRSTRING("out of memory");
> kernel/scftorture.c:#define VERBOSE_SCFTORTOUT_ERRSTRING(s, x...) \
> kernel/scftorture.c:            VERBOSE_SCFTORTOUT_ERRSTRING("all zero weights makes no sense");
> kernel/scftorture.c:            VERBOSE_SCFTORTOUT_ERRSTRING("built as module, weight_resched ignored");
> kernel/scftorture.c:            VERBOSE_SCFTORTOUT_ERRSTRING("out of memory");
> kernel/torture.c:               VERBOSE_TOROUT_ERRSTRING("Failed to alloc mask");
> kernel/torture.c:               VERBOSE_TOROUT_ERRSTRING(f);
> 
> 
> 
> Thanks
> 
> 
> 
> On 27/10/2021 13:17, Li Zhijian wrote:
> >
> >
> > On 26/10/2021 22:06, Paul E. McKenney wrote:
> >> On Mon, Oct 25, 2021 at 06:12:40AM +0000, lizhijian@...itsu.com wrote:
> >>>
> >>> On 25/10/2021 11:26, Li Zhijian wrote:
> >>>> Generally, error message should be logged anyhow.
> >>>>
> >>>> Signed-off-by: Li Zhijian <lizhijian@...fujitsu.com>
> >>>> ---
> >>>>    kernel/rcu/refscale.c | 8 ++++----
> >>>>    1 file changed, 4 insertions(+), 4 deletions(-)
> >>>>
> >>>> diff --git a/kernel/rcu/refscale.c b/kernel/rcu/refscale.c
> >>>> index a4479f00dcdc..f055d168365a 100644
> >>>> --- a/kernel/rcu/refscale.c
> >>>> +++ b/kernel/rcu/refscale.c
> >>>> @@ -58,8 +58,8 @@ do {                                            \
> >>>>        }                                        \
> >>>>    } while (0)
> >>>>    -#define VERBOSE_SCALEOUT_ERRSTRING(s, x...) \
> >>>> -    do { if (verbose) pr_alert("%s" SCALE_FLAG "!!! " s, scale_type, ## x); } while (0)
> >>>> +#define SCALEOUT_ERRSTRING(s, x...) \
> >>>> +    do { pr_alert("%s" SCALE_FLAG "!!! " s, scale_type, ## x); } while (0)
> >>>>       MODULE_LICENSE("GPL");
> >>>>    MODULE_AUTHOR("Joel Fernandes (Google) <joel@...lfernandes.org>");
> >>>> @@ -651,7 +651,7 @@ static int main_func(void *arg)
> >>>>        result_avg = kzalloc(nruns * sizeof(*result_avg), GFP_KERNEL);
> >>>>        buf = kzalloc(800 + 64, GFP_KERNEL);
> >>>>        if (!result_avg || !buf) {
> >>>> -        VERBOSE_SCALEOUT_ERRSTRING("out of memory");
> >>>> +        SCALEOUT_ERRSTRING("out of memory");
> >>> '\n' should be added to the last to flush it.
> >> And there might well be other missing "\n" instances in similar messages
> >> in rcuscale.c, rcutorture.c, scftorture.c, locktorture.c, and torture.c.
> >> Please feel free to send a patch for each file needing this help.
> >
> > Sure, i will do that.
> >
> > Thanks
> > Zhijian
> >
> >> I queued your other three patches for v5.17 (not this coming merge window,
> >> but the one after that), thank you!  I did wordsmith the commit logs,
> >> so please check to see if I messed anything up.
> >>
> >>                             Thanx, Paul
> >>
> >>>>            goto oom_exit;
> >>>>        }
> >>>>        if (holdoff)
> >>>> @@ -837,7 +837,7 @@ ref_scale_init(void)
> >>>>        reader_tasks = kcalloc(nreaders, sizeof(reader_tasks[0]),
> >>>>                       GFP_KERNEL);
> >>>>        if (!reader_tasks) {
> >>>> -        VERBOSE_SCALEOUT_ERRSTRING("out of memory");
> >>>> +        SCALEOUT_ERRSTRING("out of memory");
> >>> ditto
> >>>
> >>> Thanks
> >>> Zhijian
> >>>>            firsterr = -ENOMEM;
> >>>>            goto unwind;
> >>>>        }
> >