linux-kernel - Re: [PATCH] driver: of: Properly truncate command line if too long

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] [day] [month] [year] [list]

Message-ID: <f587170c-ab83-8b38-d59e-c0f51daa5f3c@ghiti.fr>
Date:   Tue, 6 Apr 2021 10:53:17 -0400
From:   Alex Ghiti <alex@...ti.fr>
To:     Rob Herring <robh+dt@...nel.org>
Cc:     Frank Rowand <frowand.list@...il.com>,
        Dmitry Vyukov <dvyukov@...gle.com>, devicetree@...r.kernel.org,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH] driver: of: Properly truncate command line if too long

Le 4/6/21 à 9:40 AM, Rob Herring a écrit :
> On Sat, Apr 3, 2021 at 7:09 AM Alex Ghiti <alex@...ti.fr> wrote:
>>
>> Hi,
>>
>> Le 3/16/21 à 3:38 PM, Alexandre Ghiti a écrit :
>>> In case the command line given by the user is too long, warn about it
>>> and truncate it to the last full argument.
>>>
>>> This is what efi already does in commit 80b1bfe1cb2f ("efi/libstub:
>>> Don't parse overlong command lines").
>>>
>>> Reported-by: Dmitry Vyukov <dvyukov@...gle.com>
>>> Signed-off-by: Alexandre Ghiti <alex@...ti.fr>
>>> ---
>>>    drivers/of/fdt.c | 21 ++++++++++++++++++++-
>>>    1 file changed, 20 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/drivers/of/fdt.c b/drivers/of/fdt.c
>>> index dcc1dd96911a..de4c6f9bac39 100644
>>> --- a/drivers/of/fdt.c
>>> +++ b/drivers/of/fdt.c
>>> @@ -25,6 +25,7 @@
>>>    #include <linux/serial_core.h>
>>>    #include <linux/sysfs.h>
>>>    #include <linux/random.h>
>>> +#include <linux/ctype.h>
>>>
>>>    #include <asm/setup.h>  /* for COMMAND_LINE_SIZE */
>>>    #include <asm/page.h>
>>> @@ -1050,9 +1051,27 @@ int __init early_init_dt_scan_chosen(unsigned long node, const char *uname,
>>>
>>>        /* Retrieve command line */
>>>        p = of_get_flat_dt_prop(node, "bootargs", &l);
>>> -     if (p != NULL && l > 0)
>>> +     if (p != NULL && l > 0) {
>>>                strlcpy(data, p, min(l, COMMAND_LINE_SIZE));
>>>
>>> +             /*
>>> +              * If the given command line size is larger than
>>> +              * COMMAND_LINE_SIZE, truncate it to the last complete
>>> +              * parameter.
>>> +              */
>>> +             if (l > COMMAND_LINE_SIZE) {
>>> +                     char *cmd_p = (char *)data + COMMAND_LINE_SIZE - 1;
>>> +
>>> +                     while (!isspace(*cmd_p))
>>> +                             cmd_p--;
>>> +
>>> +                     *cmd_p = '\0';
>>> +
>>> +                     pr_err("Command line is too long: truncated to %d bytes\n",
>>> +                            (int)(cmd_p - (char *)data + 1));
>>> +             }
>>> +     }
>>> +
>>>        /*
>>>         * CONFIG_CMDLINE is meant to be a default in case nothing else
>>>         * managed to set the command line, unless CONFIG_CMDLINE_FORCE
>>>
>>
>> Any thought about that ?
> 
> It looks fine to me, but this will need to be adapted to the generic
> command line support[1][2] when that is merged. So I've been waiting
> to see if that's going to happen this cycle.

Ok I'll take a look then, thanks.

Alex

> 
> Rob
> 
> [1] https://lore.kernel.org/lkml/cover.1616765869.git.christophe.leroy@csgroup.eu/
> [2] https://lore.kernel.org/lkml/41021d66db2ab427c14255d2a24bb4517c8b58fd.1617126961.git.danielwa@cisco.com/
>