[Toybox] ASan freaks out when using tsort in multicall binaries

[Oliver Webb]

------- Original Message -------
On Monday, October 9th, 2023 at 17:29, Rob Landley <rob at landley.net> wrote:


> I installed a mint vm (xfce flavor) and reproduced your issues. Sent in fixes
> for some of them, although your gcc still produces warnings about "unused return
> result" even when I typecast the return to (void). (Still, the commit comment
> says why each of those return results are not useful to try to do anything
> with, and LLVM isn't broken that way. If this regression recurs every few
> upgrades, "does not produce spurious warnings on gcc" may not be a defensible
> target, especially with all the "is not actually used uninitialized" warnings
> still cropping up in 2023.)

Yeah, I tried making those warnings go away too, I couldn't get anything that looked clean to work

> 
> > > # 255 bytes, longest VFS name
> > > LONG=0123456789abcdef0123456789abcdef
> > > LONG=$LONG$LONG$LONG$LONG$LONG$LONG$LONG$LONG
> > > LONG=${LONG:1:255}
> > > 
> > > # 4+96=100 (biggest short name), 4+97=101 (shortest long name)
> > > touch dir/${LONG:1:96} dir/${LONG:1:97}
> > > testing "create long fname" "$TAR dir/${LONG:1:97} dir/${LONG:1:96} | SUM 3" \
> > > "d70018505fa5df19ae73498cfc74d0281601e42e\n" "" ""
> > > 
> > > And what I was trying to test was the border condition of the tar internals
> > > where it switches over to an adjunct record to record an overlength field that
> > > won't fit in the structure, and it sounds like what's failing is the
> > > filesystem's ability to have two adjacent directories of length 96 and 97 that
> > > differ only by that final character. Except I didn't add a check for failure to
> > > the "touch" because it wasn't supposed to be part of the test, I just naively
> > > assumed that would portably work...
> 
> 
> Commit 71197a67521d

Test suite now passes on tar, thanks

> 
> > > > Oh, another one I forgot to mention is "truncate sparse" fails on ecryptfs as well, but works on ext4
> > > 
> > > Do you mean the tests/truncate.test entry:
> > > 
> > > testing "is sparse" "truncate -s 1g freep && [ $(stat -c %b freep) -le 8 ] &&
> > > echo okay" "okay\n" "" ""
> > > 
> > > Which is doing a "truncate -s 1g file" and then asking state if the file with
> > > literally no contents used less than 8 512-byte blocks of storage?
> > 
> > That's the one
> 
> 
> Commit 19a1a11f305c

Test suite passes on truncate, thanks

> 
> > > It's not generating a megabyte of random data, it's generating a megabyte of the
> > > same character, and then asking sed search-and-replace replace one byte at a
> > > time in that string a million times. The search-and-replace is s/x/y/ meaning
> > > each x gets replaced with y. The output of "seq" does not contain any "x"
> > > characters, so the search and replace will trigger zero times instead of
> > > triggering a million times.
> 
> 
> I tested this. If I build toybox sed without ASAN the 1 megabyte search and
> replace completes in 0.15 seconds. If I build it with ASAN, it takes 29.35
> seconds. The gcc address sanitizer is slowing sed down 200 times.

Huh, now sed passes it's tests by being faster.
And _also_ mkpasswd passes it's tests too. You'd think having a program counter of 0 would
cause a segfault, or at least the program wouldn't pass it's tests. But apparently not
 
> Elliott and I went over this a year or so back when LLVM introduced (then fixed)
> the same regression.
> 
> > > Is it triggering ASAN in mkpasswd or in a different toybox command out of the
> > > $PATH? (Yay reproduction sequence, but WHAT did it reproduce?)
> > 
> > Putting toybox commands into my PATH runs test_mkpasswd fine, So I don't think it's some other toybox command.
> > And since the lib/password.c plumbing got rewritten and needs to be audited that naturally seemed
> > like the issue.
> 
> 
> I'm seeing the ASAN dumps,

Huh, pulling the latest commit and running "VERBOSE=allnopass make tests" doesn't give the ASAN dumps anymore for me.
Thanks for fixing the errors

> haven't dug into them yet. I chmod -x the mkpasswd
> test ("make tests" runs all the tests with the executable bit, because there's
> no tests/pending directory) until I've circled back around to this.
> 
> Rob

The only command that fails for me now is du,
The 3 tests that fail are "du counts symlinks without following", 
"du -LH does not follow unspecified symlinks", and 
"du -H does not follow unspecified symlinks"

All 3 give the same diff:

@@ -1 +1 @@
-8      du_test
+12     du_test


> _______________________________________________
> Toybox mailing list
> Toybox at lists.landley.net
> http://lists.landley.net/listinfo.cgi/toybox-landley.net

- Oliver Webb <aquahobbyist at proton.me>