Current-Users archive

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Old Index]

Re: debugging memory issues with asan?

On 23.09.2019 00:21, Brett Lymn wrote:
> Folks,
> I have been sloppy with my memory allocations and use in hte libcurses
> testframe which, though the code seems to work, some tests fail when
> jemalloc is enabled.  I am trying to work out where to look but it is
> difficult becuse there are lots of little memory allocations all over
> the place.  It was suggested to me to set MKSANITIZER=yes to try and
> catch the errors.  Sure enough, if I build the curses test frame with
> MKSANITIZER=yes then the test process gets terminated and I get a memory
> map dumped like this:
> [blymn@rover] ./debug_test -v mvscanw
> ==15885==Shadow memory range interleaves with an existing memory
> mapping. ASan cannot proceed correctly. ABORTING.
> ==15885==ASan shadow was supposed to be located in the
> [0x00007fff7000-0x10007fff7fff] range.
> ==15885==Process memory map follows:
>         0x0001f0e00000-0x0001f0e25000
> /usr/tests/lib/libcurses/director
>         0x0001f1025000-0x0001f1026000
> /usr/tests/lib/libcurses/director
>         0x0001f1026000-0x0001f102c000
> /usr/tests/lib/libcurses/director
>         0x0001f102c000-0x0001f102e000
>         0x7c6c0aaae000-0x7c6c0ae00000
>         0x7c6c0ae00000-0x7c6c0ae13000   /lib/
>         0x7c6c0ae13000-0x7c6c0b012000   /lib/
>         0x7c6c0b012000-0x7c6c0b013000   /lib/
>         0x7c6c0b013000-0x7c6c0b014000   /lib/
>         0x7c6c0b200000-0x7c6c0b202000   /usr/lib/
>         0x7c6c0b202000-0x7c6c0b401000   /usr/lib/
>         0x7c6c0b401000-0x7c6c0b402000   /usr/lib/
>         0x7c6c0b402000-0x7c6c0b403000   /usr/lib/
>         0x7c6c0b600000-0x7c6c0b610000   /usr/lib/
>         0x7c6c0b610000-0x7c6c0b80f000   /usr/lib/
>         0x7c6c0b80f000-0x7c6c0b810000   /usr/lib/
>         0x7c6c0b810000-0x7c6c0b811000   /usr/lib/
>         0x7c6c0b811000-0x7c6c0b812000
> .... and so on...
> So, yes, I have a problem but I don't know where.  How can I get more
> information to help me find out where the issue is?
> I am running a -current from around the end of April.  I can update if I
> really need to (I lag behind because this laptop is the machine I use
> most so I like it stable...)

This issue is caused by ASLR. Please disable it globally
(security.pax.aslr.enabled) or per-application.

Newer (including the MKLLVM ones) sanitizers will bail out with a
verbose message here.

Ideally please upgrade to -current and use LLVM, but maybe it is not
needed in this case.

Attachment: signature.asc
Description: OpenPGP digital signature

Home | Main Index | Thread Index | Old Index