From: Rasmus Villemoes <linux@rasmusvillemoes.dk>
To: "Uwe Kleine-König" <uwe@kleine-koenig.org>,
"Rasmus Villemoes" <linux@rasmusvillemoes.dk>,
"Andrew Morton" <akpm@linux-foundation.org>,
"Jonathan Corbet" <corbet@lwn.net>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>,
Joe Perches <joe@perches.com>, Petr Mladek <pmladek@suse.com>,
Sergey Senozhatsky <sergey.senozhatsky.work@gmail.com>,
Jani Nikula <jani.nikula@linux.intel.com>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
linux-doc@vger.kernel.org
Subject: Re: [PATCH v2] printf: add support for printing symbolic error codes
Date: Mon, 16 Sep 2019 15:23:40 +0200 [thread overview]
Message-ID: <53331f01-dd77-b448-33ab-cb26dd4b9380@rasmusvillemoes.dk> (raw)
In-Reply-To: <33592303-091c-c480-460f-e2acd89e99be@kleine-koenig.org>
On 16/09/2019 14.23, Uwe Kleine-König wrote:
> Hello Rasmus,
>
> On 9/9/19 10:38 PM, Rasmus Villemoes wrote:
>> It has been suggested several times to extend vsnprintf() to be able
>> to convert the numeric value of ENOSPC to print "ENOSPC". This is yet
>> another attempt. Rather than adding another %p extension, simply teach
>> plain %p to convert ERR_PTRs. While the primary use case is
>>
>> if (IS_ERR(foo)) {
>> pr_err("Sorry, can't do that: %p\n", foo);
>> return PTR_ERR(foo);
>> }
>>
>> it is also more helpful to get a symbolic error code (or, worst case,
>> a decimal number) in case an ERR_PTR is accidentally passed to some
>> %p<something>, rather than the (efault) that check_pointer() would
>> result in.
>>
>> With my embedded hat on, I've made it possible to remove this.
>>
>> I've tested that the #ifdeffery in errcode.c is sufficient to make
>> this compile on arm, arm64, mips, powerpc, s390, x86 - I'm sure the
>> 0day bot will tell me which ones I've missed.
>>
>> The symbols to include have been found by massaging the output of
>>
>> find arch include -iname 'errno*.h' | xargs grep -E 'define\s*E'
>>
>> In the cases where some common aliasing exists
>> (e.g. EAGAIN=EWOULDBLOCK on all platforms, EDEADLOCK=EDEADLK on most),
>> I've moved the more popular one (in terms of 'git grep -w Efoo | wc)
>> to the bottom so that one takes precedence.
>>
>> Acked-by: Uwe Kleine-König <uwe@kleine-koenig.org>
>> Signed-off-by: Rasmus Villemoes <linux@rasmusvillemoes.dk>
>
> Even with my ack already given I still think having %pE (or %pe) for
> ints holding an error code is sensible.
I don't understand why you'd want an explicit %p<something> to do what
%p does by itself - in fact, with the current vsnprintf implementation,
"%pe", ERR_PTR(-EFOO) would already do what you want (since after %p is
processed, all alphanumeric are skipped whether they were interpreted or
not). So we could "reserve" %pe perhaps in order to make the call sites
a little more readable, but no code change in vsnprintf.c would be
necessary.
Or did you mean %pe with the argument being an (int*), so one would do
if (err < 0)
pr_err("bad: %pe\n", &err);
Maybe I'd buy that one, though I don't think it's much worse to do
if (err < 0)
pr_err("bad: %p\n", ERR_PTR(err));
Also, the former has less type safety/type genericity than the latter;
if err happens to be a long (or s8 or s16) the former won't work while
the latter will.
Or perhaps you meant introduce a %d<something> extension? I still think
that's a bad idea, and I've in the meantime found another reason
(covering %d in particular): Netdevices can be given a name containing
exactly one occurrence of %d (or no % at all), and then the actual name
will be determined based on that pattern. These patterns are settable
from userspace. And everything of course breaks horribly if somebody set
a name to "bla%deth" and that got turned into "blaEPERMth".
> So I wonder if it would be a
> good idea to split this patch into one that introduces errcode() and
> then the patch that teaches vsprintf about emitting its return value for
> error valued pointers. Then I could rebase my initial patch for %pe on
> top of your first one.
Well, I think my patch as-is is simple enough, there's not much point
separating the few lines in vsnprintf() from the introduction of
errcode() (which, realistically, will never have other callers).
> Other than that I wonder how we can go forward from here. So I think it
> is time for v3 which picks up the few suggestions.
Yes, I have actually prepared a v3, was just waiting for additional
comments on my responses to the v2 review comments.
Rasmus
next prev parent reply other threads:[~2019-09-16 13:23 UTC|newest]
Thread overview: 42+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-08-30 21:46 [PATCH] printf: add support for printing symbolic error codes Rasmus Villemoes
2019-08-30 21:53 ` Joe Perches
2019-08-30 22:03 ` Rasmus Villemoes
2019-08-30 22:21 ` Joe Perches
2019-08-30 22:50 ` Rasmus Villemoes
2019-09-02 9:07 ` David Laight
2019-08-31 9:38 ` kbuild test robot
2019-09-02 15:29 ` Uwe Kleine-König
2019-09-04 9:13 ` Rasmus Villemoes
2019-09-04 9:21 ` Uwe Kleine-König
2019-09-04 16:19 ` Andy Shevchenko
2019-09-04 16:28 ` Uwe Kleine-König
2019-09-05 11:40 ` Rasmus Villemoes
2019-09-09 20:38 ` [PATCH v2] " Rasmus Villemoes
2019-09-10 15:26 ` Andy Shevchenko
2019-09-11 0:15 ` Joe Perches
2019-09-11 6:43 ` Rasmus Villemoes
2019-09-11 9:37 ` Uwe Kleine-König
2019-09-11 10:14 ` Rasmus Villemoes
2019-09-15 9:43 ` Pavel Machek
2019-09-16 12:23 ` Uwe Kleine-König
2019-09-16 13:23 ` Rasmus Villemoes [this message]
2019-09-16 13:36 ` Uwe Kleine-König
2019-09-17 6:59 ` [PATCH v3] " Rasmus Villemoes
2019-09-25 14:36 ` Petr Mladek
2019-09-29 20:09 ` Rasmus Villemoes
2019-10-02 8:34 ` Petr Mladek
2019-10-05 21:48 ` Andrew Morton
2019-10-11 13:36 ` [PATCH v4 0/1] printf: add support for printing symbolic error names Rasmus Villemoes
2019-10-11 13:36 ` [PATCH v4 1/1] " Rasmus Villemoes
2019-10-14 5:51 ` Uwe Kleine-König
2019-10-14 13:02 ` Petr Mladek
2019-10-14 13:10 ` Andy Shevchenko
2019-10-15 12:17 ` Rasmus Villemoes
2019-10-15 13:44 ` David Laight
2019-10-15 19:07 ` [PATCH v5] " Rasmus Villemoes
2019-10-16 13:49 ` Andy Shevchenko
2019-10-16 14:52 ` Petr Mladek
2019-10-16 16:31 ` Andy Shevchenko
2019-10-17 15:02 ` Petr Mladek
2019-11-26 14:04 ` [PATCH v4 0/1] " Geert Uytterhoeven
2019-11-27 8:54 ` Petr Mladek
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=53331f01-dd77-b448-33ab-cb26dd4b9380@rasmusvillemoes.dk \
--to=linux@rasmusvillemoes.dk \
--cc=akpm@linux-foundation.org \
--cc=andy.shevchenko@gmail.com \
--cc=corbet@lwn.net \
--cc=jani.nikula@linux.intel.com \
--cc=joe@perches.com \
--cc=linux-doc@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=pmladek@suse.com \
--cc=sergey.senozhatsky.work@gmail.com \
--cc=uwe@kleine-koenig.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).