lkml.org 
[lkml]   [2018]   [Feb]   [15]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
From
Date
SubjectRe: [RFC PATCH 1/3] audit: remove arch_f pointer from struct audit_krule
On Mon, Feb 12, 2018 at 7:29 AM, Richard Guy Briggs <rgb@redhat.com> wrote:
> The arch_f pointer was added to the struct audit_krule in commit:
> e54dc2431d740a79a6bd013babade99d71b1714f ("audit signal recipients")
>
> This is only used on addition and deletion of rules which isn't time
> critical and the arch field is likely to be one of the first fields,
> easily found iterating over the field type. This isn't worth the
> additional complexity and storage. Delete the field.
>
> Signed-off-by: Richard Guy Briggs <rgb@redhat.com>
> ---
> include/linux/audit.h | 1 -
> kernel/auditfilter.c | 12 ++++++++----
> 2 files changed, 8 insertions(+), 5 deletions(-)

I haven't decided if I like the removal of arch_f or not, but I think
I might know where your oops/panic is coming from, thoughts below ...

> diff --git a/include/linux/audit.h b/include/linux/audit.h
> index af410d9..64a3b0e 100644
> --- a/include/linux/audit.h
> +++ b/include/linux/audit.h
> @@ -58,7 +58,6 @@ struct audit_krule {
> u32 field_count;
> char *filterkey; /* ties events to rules */
> struct audit_field *fields;
> - struct audit_field *arch_f; /* quick access to arch field */
> struct audit_field *inode_f; /* quick access to an inode field */
> struct audit_watch *watch; /* associated watch */
> struct audit_tree *tree; /* associated watched tree */
> diff --git a/kernel/auditfilter.c b/kernel/auditfilter.c
> index 739a6d2..3343d1c 100644
> --- a/kernel/auditfilter.c
> +++ b/kernel/auditfilter.c
> @@ -220,7 +220,14 @@ static inline int audit_match_class_bits(int class, u32 *mask)
>
> static int audit_match_signal(struct audit_entry *entry)
> {
> - struct audit_field *arch = entry->rule.arch_f;
> + int i;
> + struct audit_field *arch;
> +
> + for (i = 0; i < entry->rule.field_count; i++)
> + if (entry->rule.fields[i].type == AUDIT_ARCH) {
> + arch = &entry->rule.fields[i];
> + break;
> + }

In the original code arch_f was initialized to NULL via the allocator
so the arch local variable was guaranteed to have a valid value or
NULL. Unfortunately, in your code if there is no AUDIT_ARCH field
arch could remain uninitialized which I believe could lead to the
oops/panic you are seeing.

> if (!arch) {
> /* When arch is unspecified, we must check both masks on biarch
> @@ -496,9 +503,6 @@ static struct audit_entry *audit_data_to_entry(struct audit_rule_data *data,
> if (!gid_valid(f->gid))
> goto exit_free;
> break;
> - case AUDIT_ARCH:
> - entry->rule.arch_f = f;
> - break;
> case AUDIT_SUBJ_USER:
> case AUDIT_SUBJ_ROLE:
> case AUDIT_SUBJ_TYPE:
> --
> 1.8.3.1

--
paul moore
www.paul-moore.com

\
 
 \ /
  Last update: 2018-02-15 21:43    [W:0.076 / U:0.280 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site