From: Zhangjin Wu <falcon@tinylab.org>
To: david.laight@aculab.com
Cc: ammarfaizi2@gnuweeb.org, arnd@arndb.de, falcon@tinylab.org,
linux-kernel@vger.kernel.org, linux-kselftest@vger.kernel.org,
tanyuan@tinylab.org, thomas@t-8ch.de, w@1wt.eu
Subject: RE: [RFC] tools/nolibc: replace duplicated -ENOSYS return with single -ENOSYS return
Date: Thu, 31 Aug 2023 15:41:51 +0800 [thread overview]
Message-ID: <20230831074151.7900-1-falcon@tinylab.org> (raw)
In-Reply-To: <2d52dbd55e6240d5a91ebdce67fe0b7b@AcuMS.aculab.com>
Hi, David, Hi Ammar
> From: Ammar Faizi
> > Sent: 30 August 2023 15:41
> >
> > On 8/28/23 4:51 AM, David Laight wrote:
> > > I just found a(nother) clang bug:
> > > int f(void) { return "a"[2]; }
> > > compiles to just a 'return'.
> >
> > I don't think that's a bug. It's undefined behavior due to an
> > out-of-bound read. What do you expect it to return?
>
> I was actually expecting a warning/error if it didn't just read
> the byte after the end of the string.
>
> Just silently doing nothing didn't seem right for a modern compiler.
>
godbolt.org uses 'orange' color (see its right top) to indicate a
warning, it does report a warning (see Output label in right bottom)
when we access the byte after the end of the string, but gcc doesn't
report a warning ;-)
int test_outofbound(void)
{
return "a"[10];
}
see the last test case in https://godbolt.org/z/9jY4xoWrT
But it is safe if we use the trick like David used for the above
__atoi() macro:
if (str[0]) { /* always make sure the index is safe and stop at the end of string */
if (str[1]) {
if (str[2]) {
....
}
}
}
We also need this style of checking for the delta logic in __atoi_add(). have
randomly tried different clang and gcc versions, seems all of them work
correctly, but the compiling speed is not that good if we want to support the
worst cases like "((0x900000 + 0x0f0000) + 5)", the shorter one
"((0x900000+0x0f0000)+5)" is used by ARM+OABI (not supported by nolibc
currently), therefore, we can strip some tailing branches but it is either not
that fast, of course, the other architectures/variants can use faster
__atoi_add() versions with less branches and without hex detection, comparison
and calculating.
As a short summary, the compling speed should not be a big problem for most of
the architectures but to support the worst case __NR_*, the compiling speed
will be very slow (for these cases, perhaps we can use a C version of
atoi_add() instead or convert them to a more generic style: (6000 + 111), no
hex and no multiple add), and the .i output is a little ugly and the debugging
may be also a problem: for we can not assume the kernel developers always
define a short and a simple style of __NR_* as we expected. So, the __nrtoi()
requires more work, let's delay the whole RFC patchset and work on some more
urgent tasks at first as suggested by Willy, but David's NR_toi() prototype is
really a very valuable base for future work, really appreciate, I will back to
this discussion if have any new progress, thanks!
Thanks very much,
Zhangjin
> David
next prev parent reply other threads:[~2023-08-31 7:42 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2023-08-27 8:32 Zhangjin Wu
2023-08-27 9:17 ` Thomas Weißschuh
2023-08-29 6:29 ` Willy Tarreau
2023-08-30 0:19 ` Zhangjin Wu
2023-08-30 3:25 ` Willy Tarreau
2023-09-01 19:03 ` Zhangjin Wu
2023-09-03 9:31 ` Willy Tarreau
2023-08-27 21:51 ` David Laight
2023-08-28 9:46 ` David Laight
2023-08-29 23:39 ` Zhangjin Wu
2023-08-30 14:40 ` Ammar Faizi
2023-08-30 16:21 ` David Laight
2023-08-31 7:41 ` Zhangjin Wu [this message]
2023-08-31 8:37 ` David Laight
2023-09-01 19:48 ` Zhangjin Wu
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=20230831074151.7900-1-falcon@tinylab.org \
--to=falcon@tinylab.org \
--cc=ammarfaizi2@gnuweeb.org \
--cc=arnd@arndb.de \
--cc=david.laight@aculab.com \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-kselftest@vger.kernel.org \
--cc=tanyuan@tinylab.org \
--cc=thomas@t-8ch.de \
--cc=w@1wt.eu \
/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
all inboxes | Powered by JetHome®