Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]
Groups > linux.kernel > #1422288 > unrolled thread
| Started by | Andrew Morton <akpm@linux-foundation.org> |
|---|---|
| First post | 2016-06-14 22:20 +0200 |
| Last post | 2016-06-14 23:10 +0200 |
| Articles | 6 — 4 participants |
Back to article view | Back to linux.kernel
This discussion starts older than the indexed window; earlier articles aren't shown. The article labeled Started by
below is the oldest one visible, not the original post.
Re: [PATCH 1/1] kernel/sysctl.c: avoid overflow Andrew Morton <akpm@linux-foundation.org> - 2016-06-14 22:20 +0200
Re: [PATCH 1/1] kernel/sysctl.c: avoid overflow Willy Tarreau <w@1wt.eu> - 2016-06-14 22:50 +0200
Re: [PATCH 1/1] kernel/sysctl.c: avoid overflow Dave Young <dyoung@redhat.com> - 2016-06-15 10:40 +0200
Re: [PATCH 1/1] kernel/sysctl.c: avoid overflow Willy Tarreau <w@1wt.eu> - 2016-06-15 10:50 +0200
Re: [PATCH 1/1] kernel/sysctl.c: avoid overflow Dave Young <dyoung@redhat.com> - 2016-06-15 11:00 +0200
Re: [PATCH 1/1] kernel/sysctl.c: avoid overflow Kees Cook <keescook@chromium.org> - 2016-06-14 23:10 +0200
| From | Andrew Morton <akpm@linux-foundation.org> |
|---|---|
| Date | 2016-06-14 22:20 +0200 |
| Subject | Re: [PATCH 1/1] kernel/sysctl.c: avoid overflow |
| Message-ID | <rK2Yy-1dW-3@gated-at.bofh.it> |
On Sat, 11 Jun 2016 03:33:08 +0200 Heinrich Schuchardt <xypron.glpk@gmx.de> wrote:
> An undetected overflow may occur in do_proc_dointvec_minmax_conv_param.
>
> ...
>
> --- a/kernel/sysctl.c
> +++ b/kernel/sysctl.c
> @@ -2313,7 +2313,17 @@ static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
> {
> struct do_proc_dointvec_minmax_conv_param *param = data;
> if (write) {
> - int val = *negp ? -*lvalp : *lvalp;
> + int val;
> +
> + if (*negp) {
> + if (*lvalp > (unsigned long) INT_MAX + 1)
> + return -EINVAL;
> + val = -*lvalp;
> + } else {
> + if (*lvalp > (unsigned long) INT_MAX)
> + return -EINVAL;
> + val = *lvalp;
> + }
> if ((param->min && *param->min > val) ||
> (param->max && *param->max < val))
> return -EINVAL;
hm.
What happens if someone does
echo -1 > /proc/foo
expecting to get 0xffffffff? That's a reasonable shorthand, and if we
change that to spit out EINVAL then people's stuff may break.
[toc] | [next] | [standalone]
| From | Willy Tarreau <w@1wt.eu> |
|---|---|
| Date | 2016-06-14 22:50 +0200 |
| Message-ID | <rK3rz-1nX-7@gated-at.bofh.it> |
| In reply to | #1422288 |
On Tue, Jun 14, 2016 at 01:19:06PM -0700, Andrew Morton wrote:
> On Sat, 11 Jun 2016 03:33:08 +0200 Heinrich Schuchardt <xypron.glpk@gmx.de> wrote:
>
> > An undetected overflow may occur in do_proc_dointvec_minmax_conv_param.
> >
> > ...
> >
> > --- a/kernel/sysctl.c
> > +++ b/kernel/sysctl.c
> > @@ -2313,7 +2313,17 @@ static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
> > {
> > struct do_proc_dointvec_minmax_conv_param *param = data;
> > if (write) {
> > - int val = *negp ? -*lvalp : *lvalp;
> > + int val;
> > +
> > + if (*negp) {
> > + if (*lvalp > (unsigned long) INT_MAX + 1)
> > + return -EINVAL;
> > + val = -*lvalp;
> > + } else {
> > + if (*lvalp > (unsigned long) INT_MAX)
> > + return -EINVAL;
> > + val = *lvalp;
> > + }
> > if ((param->min && *param->min > val) ||
> > (param->max && *param->max < val))
> > return -EINVAL;
>
> hm.
>
> What happens if someone does
>
> echo -1 > /proc/foo
>
> expecting to get 0xffffffff? That's a reasonable shorthand, and if we
> change that to spit out EINVAL then people's stuff may break.
I'd go even further, I don't see anymore how it becomes possible
to actually *write* 0xffffffff at all! This function is used by
proc_dointvec_minmax() which is used with extra1=&zero and extra2
not set with some unsigned ints to allow the full range to be
configured (eg: dirty_expire_interval is the first I found by a
quick random look).
So for me this change is bogus.
Willy
[toc] | [prev] | [next] | [standalone]
| From | Dave Young <dyoung@redhat.com> |
|---|---|
| Date | 2016-06-15 10:40 +0200 |
| Message-ID | <rKewG-a9-31@gated-at.bofh.it> |
| In reply to | #1422316 |
On 06/14/16 at 10:41pm, Willy Tarreau wrote:
> On Tue, Jun 14, 2016 at 01:19:06PM -0700, Andrew Morton wrote:
> > On Sat, 11 Jun 2016 03:33:08 +0200 Heinrich Schuchardt <xypron.glpk@gmx.de> wrote:
> >
> > > An undetected overflow may occur in do_proc_dointvec_minmax_conv_param.
> > >
> > > ...
> > >
> > > --- a/kernel/sysctl.c
> > > +++ b/kernel/sysctl.c
> > > @@ -2313,7 +2313,17 @@ static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
> > > {
> > > struct do_proc_dointvec_minmax_conv_param *param = data;
> > > if (write) {
> > > - int val = *negp ? -*lvalp : *lvalp;
> > > + int val;
> > > +
> > > + if (*negp) {
> > > + if (*lvalp > (unsigned long) INT_MAX + 1)
> > > + return -EINVAL;
> > > + val = -*lvalp;
> > > + } else {
> > > + if (*lvalp > (unsigned long) INT_MAX)
> > > + return -EINVAL;
> > > + val = *lvalp;
> > > + }
> > > if ((param->min && *param->min > val) ||
> > > (param->max && *param->max < val))
> > > return -EINVAL;
> >
> > hm.
> >
> > What happens if someone does
> >
> > echo -1 > /proc/foo
> >
> > expecting to get 0xffffffff? That's a reasonable shorthand, and if we
> > change that to spit out EINVAL then people's stuff may break.
>
> I'd go even further, I don't see anymore how it becomes possible
> to actually *write* 0xffffffff at all! This function is used by
> proc_dointvec_minmax() which is used with extra1=&zero and extra2
> not set with some unsigned ints to allow the full range to be
> configured (eg: dirty_expire_interval is the first I found by a
> quick random look).
sysctl_writes_strict use extra1 = -1 and extra2 = 1
But I do not get why -1 does not work, 1 < (unsigned long) INT_MAX + 1
so val = -1, it is still right?
>
> So for me this change is bogus.
>
> Willy
>
Thanks
Dave
[toc] | [prev] | [next] | [standalone]
| From | Willy Tarreau <w@1wt.eu> |
|---|---|
| Date | 2016-06-15 10:50 +0200 |
| Message-ID | <rKeGm-ez-3@gated-at.bofh.it> |
| In reply to | #1422781 |
On Wed, Jun 15, 2016 at 04:33:32PM +0800, Dave Young wrote:
> On 06/14/16 at 10:41pm, Willy Tarreau wrote:
> > On Tue, Jun 14, 2016 at 01:19:06PM -0700, Andrew Morton wrote:
> > > On Sat, 11 Jun 2016 03:33:08 +0200 Heinrich Schuchardt <xypron.glpk@gmx.de> wrote:
> > >
> > > > An undetected overflow may occur in do_proc_dointvec_minmax_conv_param.
> > > >
> > > > ...
> > > >
> > > > --- a/kernel/sysctl.c
> > > > +++ b/kernel/sysctl.c
> > > > @@ -2313,7 +2313,17 @@ static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
> > > > {
> > > > struct do_proc_dointvec_minmax_conv_param *param = data;
> > > > if (write) {
> > > > - int val = *negp ? -*lvalp : *lvalp;
> > > > + int val;
> > > > +
> > > > + if (*negp) {
> > > > + if (*lvalp > (unsigned long) INT_MAX + 1)
> > > > + return -EINVAL;
> > > > + val = -*lvalp;
> > > > + } else {
> > > > + if (*lvalp > (unsigned long) INT_MAX)
> > > > + return -EINVAL;
> > > > + val = *lvalp;
> > > > + }
> > > > if ((param->min && *param->min > val) ||
> > > > (param->max && *param->max < val))
> > > > return -EINVAL;
> > >
> > > hm.
> > >
> > > What happens if someone does
> > >
> > > echo -1 > /proc/foo
> > >
> > > expecting to get 0xffffffff? That's a reasonable shorthand, and if we
> > > change that to spit out EINVAL then people's stuff may break.
> >
> > I'd go even further, I don't see anymore how it becomes possible
> > to actually *write* 0xffffffff at all! This function is used by
> > proc_dointvec_minmax() which is used with extra1=&zero and extra2
> > not set with some unsigned ints to allow the full range to be
> > configured (eg: dirty_expire_interval is the first I found by a
> > quick random look).
>
> sysctl_writes_strict use extra1 = -1 and extra2 = 1
>
> But I do not get why -1 does not work, 1 < (unsigned long) INT_MAX + 1
> so val = -1, it is still right?
-1 should indeed work but 0xffffffff will definitely not.
Willy
[toc] | [prev] | [next] | [standalone]
| From | Dave Young <dyoung@redhat.com> |
|---|---|
| Date | 2016-06-15 11:00 +0200 |
| Message-ID | <rKeQ2-iF-15@gated-at.bofh.it> |
| In reply to | #1422790 |
On 06/15/16 at 10:40am, Willy Tarreau wrote:
> On Wed, Jun 15, 2016 at 04:33:32PM +0800, Dave Young wrote:
> > On 06/14/16 at 10:41pm, Willy Tarreau wrote:
> > > On Tue, Jun 14, 2016 at 01:19:06PM -0700, Andrew Morton wrote:
> > > > On Sat, 11 Jun 2016 03:33:08 +0200 Heinrich Schuchardt <xypron.glpk@gmx.de> wrote:
> > > >
> > > > > An undetected overflow may occur in do_proc_dointvec_minmax_conv_param.
> > > > >
> > > > > ...
> > > > >
> > > > > --- a/kernel/sysctl.c
> > > > > +++ b/kernel/sysctl.c
> > > > > @@ -2313,7 +2313,17 @@ static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
> > > > > {
> > > > > struct do_proc_dointvec_minmax_conv_param *param = data;
> > > > > if (write) {
> > > > > - int val = *negp ? -*lvalp : *lvalp;
> > > > > + int val;
> > > > > +
> > > > > + if (*negp) {
> > > > > + if (*lvalp > (unsigned long) INT_MAX + 1)
> > > > > + return -EINVAL;
> > > > > + val = -*lvalp;
> > > > > + } else {
> > > > > + if (*lvalp > (unsigned long) INT_MAX)
> > > > > + return -EINVAL;
> > > > > + val = *lvalp;
> > > > > + }
> > > > > if ((param->min && *param->min > val) ||
> > > > > (param->max && *param->max < val))
> > > > > return -EINVAL;
> > > >
> > > > hm.
> > > >
> > > > What happens if someone does
> > > >
> > > > echo -1 > /proc/foo
> > > >
> > > > expecting to get 0xffffffff? That's a reasonable shorthand, and if we
> > > > change that to spit out EINVAL then people's stuff may break.
> > >
> > > I'd go even further, I don't see anymore how it becomes possible
> > > to actually *write* 0xffffffff at all! This function is used by
> > > proc_dointvec_minmax() which is used with extra1=&zero and extra2
> > > not set with some unsigned ints to allow the full range to be
> > > configured (eg: dirty_expire_interval is the first I found by a
> > > quick random look).
> >
> > sysctl_writes_strict use extra1 = -1 and extra2 = 1
> >
> > But I do not get why -1 does not work, 1 < (unsigned long) INT_MAX + 1
> > so val = -1, it is still right?
>
> -1 should indeed work but 0xffffffff will definitely not.
Hmm, right, indeed.
Below echo works originally:
echo 0xffffffff > /proc/sys/kernel/sysctl_writes_strict
>
> Willy
Thanks
Dave
[toc] | [prev] | [next] | [standalone]
| From | Kees Cook <keescook@chromium.org> |
|---|---|
| Date | 2016-06-14 23:10 +0200 |
| Message-ID | <rK3KW-1Jw-21@gated-at.bofh.it> |
| In reply to | #1422288 |
On Tue, Jun 14, 2016 at 1:19 PM, Andrew Morton
<akpm@linux-foundation.org> wrote:
> On Sat, 11 Jun 2016 03:33:08 +0200 Heinrich Schuchardt <xypron.glpk@gmx.de> wrote:
>
>> An undetected overflow may occur in do_proc_dointvec_minmax_conv_param.
>>
>> ...
>>
>> --- a/kernel/sysctl.c
>> +++ b/kernel/sysctl.c
>> @@ -2313,7 +2313,17 @@ static int do_proc_dointvec_minmax_conv(bool *negp, unsigned long *lvalp,
>> {
>> struct do_proc_dointvec_minmax_conv_param *param = data;
>> if (write) {
>> - int val = *negp ? -*lvalp : *lvalp;
>> + int val;
>> +
>> + if (*negp) {
>> + if (*lvalp > (unsigned long) INT_MAX + 1)
>> + return -EINVAL;
>> + val = -*lvalp;
>> + } else {
>> + if (*lvalp > (unsigned long) INT_MAX)
>> + return -EINVAL;
>> + val = *lvalp;
>> + }
>> if ((param->min && *param->min > val) ||
>> (param->max && *param->max < val))
>> return -EINVAL;
>
> hm.
>
> What happens if someone does
>
> echo -1 > /proc/foo
>
> expecting to get 0xffffffff? That's a reasonable shorthand, and if we
> change that to spit out EINVAL then people's stuff may break.
If we expect the interface to allow overflows, we should at least add
comments to do_proc_dointvec_minmax_conv_param()...
-Kees
--
Kees Cook
Chrome OS & Brillo Security
[toc] | [prev] | [standalone]
Back to top | Article view | linux.kernel
csiph-web