Tetsuo Handa <[EMAIL PROTECTED]> writes:
> Hello.
>
> Andrew Morton wrote:
>> I believe (args->nlen > CTL_MAXNAME) was correct.
> I'll leave it to you.
> But if you want to allow args->nlen == CTL_MAXNAME,
> you also need to update do_sysctl().
Which has been that way since before I decided to
Tetsuo Handa [EMAIL PROTECTED] writes:
Hello.
Andrew Morton wrote:
I believe (args-nlen CTL_MAXNAME) was correct.
I'll leave it to you.
But if you want to allow args-nlen == CTL_MAXNAME,
you also need to update do_sysctl().
Which has been that way since before I decided to touch it.
Hello.
Andrew Morton wrote:
> I believe (args->nlen > CTL_MAXNAME) was correct.
I'll leave it to you.
But if you want to allow args->nlen == CTL_MAXNAME,
you also need to update do_sysctl().
int do_sysctl(int __user *name, int nlen, void __user *oldval, size_t __user
*oldlenp,
On Tue, 13 Nov 2007 12:07:23 +0900 Tetsuo Handa <[EMAIL PROTECTED]> wrote:
> Andrew, please replace previous patch with this one.
> This one returns -ENOTDIR.
> --
>
> Original patch forgot to check args->nlen.
> I don't know why args->nlen == CTL_MAXNAME is rejected,
> but it has been
Hello.
Eric W. Biederman wrote:
> name[CTL_MAXNAME} is not valid.
> name[0...CTL_MAXNAME-1] is valid.
Yes.
> The check that got lost in the refactoring was specfically:
>
> - if (tmp.nlen <= 0 || tmp.nlen >= CTL_MAXNAME)
> - return -ENOTDIR;
Thus I think tmp.nlen ==
Tetsuo Handa <[EMAIL PROTECTED]> writes:
> Hello.
>
> Thanks for reformatting my patch
> and sorry for surprising you with directory name
> (I meant to type linux-2.6.24-rc2, not linux-2.6.22-rc2).
>
> According to linux-2.6.23,
> it seems that I should return -ENOTDIR
> for invalid args->nlen
Tetsuo Handa [EMAIL PROTECTED] writes:
Hello.
Thanks for reformatting my patch
and sorry for surprising you with directory name
(I meant to type linux-2.6.24-rc2, not linux-2.6.22-rc2).
According to linux-2.6.23,
it seems that I should return -ENOTDIR
for invalid args-nlen value.
I got
Hello.
Eric W. Biederman wrote:
name[CTL_MAXNAME} is not valid.
name[0...CTL_MAXNAME-1] is valid.
Yes.
The check that got lost in the refactoring was specfically:
- if (tmp.nlen = 0 || tmp.nlen = CTL_MAXNAME)
- return -ENOTDIR;
Thus I think tmp.nlen == CTL_MAXNAME
On Tue, 13 Nov 2007 12:07:23 +0900 Tetsuo Handa [EMAIL PROTECTED] wrote:
Andrew, please replace previous patch with this one.
This one returns -ENOTDIR.
--
Original patch forgot to check args-nlen.
I don't know why args-nlen == CTL_MAXNAME is rejected,
but it has been rejected
Hello.
Andrew Morton wrote:
I believe (args-nlen CTL_MAXNAME) was correct.
I'll leave it to you.
But if you want to allow args-nlen == CTL_MAXNAME,
you also need to update do_sysctl().
int do_sysctl(int __user *name, int nlen, void __user *oldval, size_t __user
*oldlenp,
void
Hello.
Thanks for reformatting my patch
and sorry for surprising you with directory name
(I meant to type linux-2.6.24-rc2, not linux-2.6.22-rc2).
According to linux-2.6.23,
it seems that I should return -ENOTDIR
for invalid args->nlen value.
I got a question here regarding interpretation of
Hello.
Thanks for reformatting my patch
and sorry for surprising you with directory name
(I meant to type linux-2.6.24-rc2, not linux-2.6.22-rc2).
According to linux-2.6.23,
it seems that I should return -ENOTDIR
for invalid args-nlen value.
I got a question here regarding interpretation of
> On Thu, 08 Nov 2007 11:57:26 +0900 Tetsuo Handa <[EMAIL PROTECTED]> wrote:
> Original patch assumed args->nlen < CTL_MAXNAME, but it can be false.
>
> Signed-off-by: Tetsuo Handa <[EMAIL PROTECTED]>
>
>
> --- linux-2.6.22-rc2.orig/kernel/sysctl.c 2007-11-08 10:38:17.0
> +0900
>
On Thu, 08 Nov 2007 11:57:26 +0900 Tetsuo Handa [EMAIL PROTECTED] wrote:
Original patch assumed args-nlen CTL_MAXNAME, but it can be false.
Signed-off-by: Tetsuo Handa [EMAIL PROTECTED]
--- linux-2.6.22-rc2.orig/kernel/sysctl.c 2007-11-08 10:38:17.0
+0900
+++
14 matches
Mail list logo