lkml.org 
[lkml]   [2008]   [Sep]   [27]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: [PATCH 5/6] file capabilities: remove needless inline functions
Quoting Andrew G. Morgan (morgan@kernel.org):
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Serge,
>
> I'd much rather simply remove the target argument from the
> security_capset_check() call. Relying on the caller to not do something

Yes, I did do that in a patch on top of all of these. It increased the
kernel size by 8 (or was it 16) bytes :) I assume that was because
now there were more references to 'current' which is inlined? So I
should be able to fix that by saving current() away at the top of
the fn.

Will try that.

thanks,
-serge

> bad seems fragile... If the code internally operates on current only,
> then it doesn't need a target argument... No? (Evidently, such a change
> is also needed to selinux_capset_check() too, but this doesn't look like
> it will pose a problem for the selinux code.)
>
> Cheers
>
> Andrew
>
> Serge E. Hallyn wrote:
> > cap_limit_ptraced_target always returns 1, so nix it.
> >
> > cap_block_setpcap can't return 1 any more, because
> > kernel/capabilities.c:sys_capset() will return -EPERM
> > if it is called on a task other than current, and will
> > never get to cap_capset_check.
> >
> > This brings the vmlinux size with my config down another
> > 16 bytes (making up for the 8 byte increase from the
> > last patch).
> >
> > Signed-off-by: Serge E. Hallyn <serue@us.ibm.com>
> > ---
> > security/commoncap.c | 22 +++-------------------
> > 1 files changed, 3 insertions(+), 19 deletions(-)
> >
> > diff --git a/security/commoncap.c b/security/commoncap.c
> > index d48fdd8..e5afb7c 100644
> > --- a/security/commoncap.c
> > +++ b/security/commoncap.c
> > @@ -93,15 +93,6 @@ int cap_capget (struct task_struct *target, kernel_cap_t *effective,
> > return 0;
> > }
> >
> > -static inline int cap_block_setpcap(struct task_struct *target)
> > -{
> > - /*
> > - * No support for remote process capability manipulation with
> > - * filesystem capability support.
> > - */
> > - return (target != current);
> > -}
> > -
> > static inline int cap_inh_is_capped(void)
> > {
> > /*
> > @@ -112,14 +103,9 @@ static inline int cap_inh_is_capped(void)
> > return (cap_capable(current, CAP_SETPCAP) != 0);
> > }
> >
> > -static inline int cap_limit_ptraced_target(void) { return 1; }
> > -
> > int cap_capset_check (struct task_struct *target, kernel_cap_t *effective,
> > kernel_cap_t *inheritable, kernel_cap_t *permitted)
> > {
> > - if (cap_block_setpcap(target)) {
> > - return -EPERM;
> > - }
> > if (cap_inh_is_capped()
> > && !cap_issubset(*inheritable,
> > cap_combine(target->cap_inheritable,
> > @@ -343,11 +329,9 @@ void cap_bprm_apply_creds (struct linux_binprm *bprm, int unsafe)
> > bprm->e_uid = current->uid;
> > bprm->e_gid = current->gid;
> > }
> > - if (cap_limit_ptraced_target()) {
> > - bprm->cap_post_exec_permitted = cap_intersect(
> > - bprm->cap_post_exec_permitted,
> > - current->cap_permitted);
> > - }
> > + bprm->cap_post_exec_permitted = cap_intersect(
> > + bprm->cap_post_exec_permitted,
> > + current->cap_permitted);
> > }
> > }
> >
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.7 (Darwin)
> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>
> iD8DBQFI3bjz+bHCR3gb8jsRAqJpAJ9Ca1pADkG5BnGoOVZA+EmZbuRPfgCgoQ95
> ljvsvj7Ssp+0mXDuCy0/TnU=
> =79ni
> -----END PGP SIGNATURE-----


\
 
 \ /
  Last update: 2008-09-27 15:43    [W:0.073 / U:0.856 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site