linux-kernel - Re: [PATCH 2/2] powerpc/xmon: use KSYM_NAME

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] [thread-next>] [day] [month] [year] [list]

Message-ID: <87ilc8ym6v.fsf@mail.lhotse>
Date:   Thu, 01 Jun 2023 12:02:16 +1000
From:   Michael Ellerman <mpe@...erman.id.au>
To:     Miguel Ojeda <miguel.ojeda.sandonis@...il.com>,
        Maninder Singh <maninder1.s@...sung.com>
Cc:     bcain@...cinc.com, npiggin@...il.com, christophe.leroy@...roup.eu,
        keescook@...omium.org, nathanl@...ux.ibm.com, ustavoars@...nel.org,
        alex.gaynor@...il.com, gary@...yguo.net, ojeda@...nel.org,
        pmladek@...e.com, wedsonaf@...gle.com,
        linux-hexagon@...r.kernel.org, linux-kernel@...r.kernel.org,
        linuxppc-dev@...ts.ozlabs.org, Onkarnath <onkarnath.1@...sung.com>
Subject: Re: [PATCH 2/2] powerpc/xmon: use KSYM_NAME_LEN in array size

Miguel Ojeda <miguel.ojeda.sandonis@...il.com> writes:
> On Mon, May 29, 2023 at 1:14 PM Maninder Singh <maninder1.s@...sung.com> wrote:
>>
>> +static char tmpstr[KSYM_NAME_LEN];
>
> Reviewed-by: Miguel Ojeda <ojeda@...nel.org>
>
> Side-note: in `get_function_bounds()`, I see `kallsyms_lookup()` being
> used, but the name seems discarded? Can
> `kallsyms_lookup_size_offset()` be used instead, thus avoiding the
> usage of the buffer there to begin with?

A few lines below it uses the modname, and AFAICS there's no (easy) way
to lookup the modname without also looking up the name.

> Side-note 2: in `scanhex()`, I see a loop `i<63` using `tmpstr` which
> then is used to do a `kallsyms_lookup_name()`, so I guess symbols
> larger than 64 couldn't be found. I have no idea about what are the
> external constraints here, but perhaps it is possible to increase the
> `line` buffer etc. to then allow for bigger symbols to be found.

Yeah that looks wrong. I don't see any symbols that long in current
kernels, but we should fix it.

Thanks for looking.

cheers