linux-kernel - Re: [PATCH] misc: kgdbts: silence array underflow warning

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]

Date:   Thu, 17 May 2018 14:31:51 +0100
From:   Daniel Thompson <daniel.thompson@...aro.org>
To:     Dan Carpenter <dan.carpenter@...cle.com>
Cc:     Jason Wessel <jason.wessel@...driver.com>,
        Arnd Bergmann <arnd@...db.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        kgdb-bugreport@...ts.sourceforge.net, linux-kernel@...r.kernel.org,
        kernel-janitors@...r.kernel.org
Subject: Re: [PATCH] misc: kgdbts: silence array underflow warning

On Thu, May 17, 2018 at 03:22:29PM +0300, Dan Carpenter wrote:
> Smatch distrusts simple_strtol().  I don't know the code well enough
> to say if the distrust is justified here, but it seems harmless to
> silence the warning.

What warning does this fix? I'd prefer to have it in the description.


> Signed-off-by: Dan Carpenter <dan.carpenter@...cle.com>
> 
> diff --git a/drivers/misc/kgdbts.c b/drivers/misc/kgdbts.c
> index 6193270e7b3d..e0508acaedaa 100644
> --- a/drivers/misc/kgdbts.c
> +++ b/drivers/misc/kgdbts.c
> @@ -400,13 +400,15 @@ static void skip_back_repeat_test(char *arg)
>  	int go_back = simple_strtol(arg, NULL, 10);

If go_back is out of range then this is a serious error in the test
plans found in the module. Something simple and clear such as
BUG_ON(go_back <= 0 || go_back > ts.idx) is probably sufficient.


>  	repeat_test--;
> -	if (repeat_test <= 0) {
> +	if (repeat_test <= 0 || go_back < 0) {

The BUG_ON() will mess things up because whatever breakpoints
the test is using are still enabled.

If you really want to recover cleanly then perhaps:

+	if (repeat_test <= 0 || WARN_ON(go_back < 0 || go_back > ts.idx)) {


>  		ts.idx++;
>  	} else {
>  		if (repeat_test % 100 == 0)
>  			v1printk("kgdbts:RUN ... %d remaining\n", repeat_test);
>  
>  		ts.idx -= go_back;
> +		if (ts.idx < 0)
> +			ts.idx = 0;

Not sure about this. If we know the opcodes are bad then re-executing
them doesn't seem like a good idea (hence covering it in the WARN_ON
branch above).


Daniel.