RE: [RFC][PATCH 04/12] digest_lists: Objects

From: Roberto Sassu
Date: Mon Jun 28 2021 - 04:14:46 EST


> From: Greg KH [mailto:gregkh@xxxxxxxxxxxxxxxxxxx]
> Sent: Sunday, June 27, 2021 12:56 PM
> On Fri, Jun 25, 2021 at 06:56:06PM +0200, Roberto Sassu wrote:
> > +++ b/security/integrity/digest_lists/digest_lists.h
> > @@ -0,0 +1,117 @@
> > +/* SPDX-License-Identifier: GPL-2.0 */
> > +/*
> > + * Copyright (C) 2005,2006,2007,2008 IBM Corporation
> > + * Copyright (C) 2017-2021 Huawei Technologies Duesseldorf GmbH
> > + *
> > + * Author: Roberto Sassu <roberto.sassu@xxxxxxxxxx>
> > + *
> > + * This program is free software; you can redistribute it and/or
> > + * modify it under the terms of the GNU General Public License as
> > + * published by the Free Software Foundation, version 2 of the
> > + * License.
> > + *
> > + * File: digest_lists.h
> > + * Unexported definitions for digest lists.
>
> Unexported to whom?

Hi Greg

I meant not placed in include/linux.

> > + */
> > +
> > +#ifndef __DIGEST_LISTS_INTERNAL_H
> > +#define __DIGEST_LISTS_INTERNAL_H
> > +
> > +#include <linux/types.h>
> > +#include <linux/crypto.h>
> > +#include <linux/fs.h>
> > +#include <linux/security.h>
> > +#include <linux/hash.h>
> > +#include <linux/tpm.h>
> > +#include <linux/audit.h>
> > +#include <crypto/hash_info.h>
> > +#include <linux/hash_info.h>
> > +#include <uapi/linux/digest_lists.h>
> > +
> > +#define MAX_DIGEST_SIZE 64
> > +#define HASH_BITS 10
> > +#define MEASURE_HTABLE_SIZE (1 << HASH_BITS)
> > +
> > +struct digest_list_item {
> > + loff_t size;
> > + u8 *buf;
> > + u8 actions;
> > + u8 digest[64];
> > + enum hash_algo algo;
> > + const char *label;
> > +};
> > +
> > +struct digest_list_item_ref {
> > + struct digest_list_item *digest_list;
> > + loff_t digest_offset;
> > + loff_t hdr_offset;
> > +};
> > +
> > +struct digest_item {
> > + /* hash table pointers */
> > + struct hlist_node hnext;
> > + /* digest list references (protected by RCU) */
> > + struct digest_list_item_ref *refs;
> > +};
> > +
> > +struct h_table {
> > + atomic_long_t len;
>
> Why is this atomic? Why would that matter?

Yes, it shouldn't be. There are not concurrent updates.

> > + struct hlist_head queue[MEASURE_HTABLE_SIZE];
> > +};
> > +
> > +static inline unsigned int hash_key(u8 *digest)
> > +{
> > + return (digest[0] | digest[1] << 8) % MEASURE_HTABLE_SIZE;
> > +}
>
> Don't we have hashing functions in the kernel already?

We had a discussion before:

https://lore.kernel.org/linux-integrity/1587739544.5190.14.camel@xxxxxxxxxxxxx/

It seems there is no real advantage in hashing a digest.

> > +
> > +static inline struct compact_list_hdr *get_hdr(
> > + struct digest_list_item *digest_list,
> > + loff_t hdr_offset)
> > +{
> > + return (struct compact_list_hdr *)(digest_list->buf + hdr_offset);
> > +}
>
> pointer math feels rough, are you shure you want to do this this way?

Maybe, I could change digest_list_item_ref to:

struct digest_list_item_ref {
struct digest_list_item *digest_list;
u8 *digest;
struct compact_list_hdr *hdr;
};

where digest and hdr are calculated in the same way.

Or you have a different suggestion?

Thanks

Roberto

HUAWEI TECHNOLOGIES Duesseldorf GmbH, HRB 56063
Managing Director: Li Peng, Li Jian, Shi Yanli

> thanks,
>
> greg k-h