10/20/2023 12:49 PM, Mickaël Salaün пишет:
> On Fri, Oct 20, 2023 at 07:08:33AM +0300, Konstantin Meskhidze (A) wrote:
> >
> >
> > 10/18/2023 3:29 PM, Mickaël Salaün пишет:
> > > On Mon, Oct 16, 2023 at 09:50:26AM +0800, Konstantin Meskhidze wrote:
> > > > diff --git a/security/landlock/net.h b/security/landlock/net.h
> > > > new file mode 100644
> > > > index 000000000000..588a49fd6907
> > > > --- /dev/null
> > > > +++ b/security/landlock/net.h
> > > > @@ -0,0 +1,33 @@
> > > > +/* SPDX-License-Identifier: GPL-2.0-only */
> > > > +/*
> > > > + * Landlock LSM - Network management and hooks
> > > > + *
> > > > + * Copyright © 2022-2023 Huawei Tech. Co., Ltd.
> > > > + */
> > > > +
> > > > +#ifndef _SECURITY_LANDLOCK_NET_H
> > > > +#define _SECURITY_LANDLOCK_NET_H
> > > > +
> > > > +#include "common.h"
> > > > +#include "ruleset.h"
> > > > +#include "setup.h"
> > > > +
> > > > +#if IS_ENABLED(CONFIG_INET)
> > > > +__init void landlock_add_net_hooks(void);
> > > > +
> > > > +int landlock_append_net_rule(struct landlock_ruleset *const ruleset,
> > > > + const u16 port, access_mask_t access_rights);
> > > > +#else /* IS_ENABLED(CONFIG_INET) */
> > > > +static inline void landlock_add_net_hooks(void)
> > > > +{
> > > > +}
> > > > +
> > > > +static inline int
> > > > +landlock_append_net_rule(struct landlock_ruleset *const ruleset, const u16 port,
> > > > + access_mask_t access_rights);
> > > > +{
> > > > + return -EAFNOSUPPORT;
> > > > +}
> > > > +#endif /* IS_ENABLED(CONFIG_INET) */
> > > > +
> > > > +#endif /* _SECURITY_LANDLOCK_NET_H */
> > > > diff --git a/security/landlock/ruleset.c b/security/landlock/ruleset.c
> > > > index 4c209acee01e..1fe4298ff4a7 100644
> > > > --- a/security/landlock/ruleset.c
> > > > +++ b/security/landlock/ruleset.c
> > > > @@ -36,6 +36,11 @@ static struct landlock_ruleset *create_ruleset(const u32 num_layers)
> > > > refcount_set(&new_ruleset->usage, 1);
> > > > mutex_init(&new_ruleset->lock);
> > > > new_ruleset->root_inode = RB_ROOT;
> > > > +
> > > > +#if IS_ENABLED(CONFIG_INET)
> > > > + new_ruleset->root_net_port = RB_ROOT;
> > > > +#endif /* IS_ENABLED(CONFIG_INET) */
> > > > +
> > > > new_ruleset->num_layers = num_layers;
> > > > /*
> > > > * hierarchy = NULL
> > > > @@ -46,16 +51,21 @@ static struct landlock_ruleset *create_ruleset(const u32 num_layers)
> > > > }
> > > > > > struct landlock_ruleset *
> > > > -landlock_create_ruleset(const access_mask_t fs_access_mask)
> > > > +landlock_create_ruleset(const access_mask_t fs_access_mask,
> > > > + const access_mask_t net_access_mask)
> > > > {
> > > > struct landlock_ruleset *new_ruleset;
> > > > > > /* Informs about useless ruleset. */
> > > > - if (!fs_access_mask)
> > > > + if (!fs_access_mask && !net_access_mask)
> > > > return ERR_PTR(-ENOMSG);
> > > > new_ruleset = create_ruleset(1);
> > > > - if (!IS_ERR(new_ruleset))
> > > > + if (IS_ERR(new_ruleset))
> > > > + return new_ruleset;
> > > > + if (fs_access_mask)
> > > > landlock_add_fs_access_mask(new_ruleset, fs_access_mask, 0);
> > > > + if (net_access_mask)
> > > > + landlock_add_net_access_mask(new_ruleset, net_access_mask, 0);
> > > > This is good, but it is not tested: we need to add a test that
> > both
> > > handle FS and net restrictions. You can add one in net.c, just handling
> > > LANDLOCK_ACCESS_FS_READ_DIR and LANDLOCK_ACCESS_NET_BIND_TCP, add one
> > > rule with path_beneath (e.g. /dev) and another with net_port, and check
> > > that open("/") is denied, open("/dev") is allowed, and and only the
> > > allowed port is allowed with bind(). This test should be simple and can
> > > only check against an IPv4 socket, i.e. using ipv4_tcp fixture, just
> > > after port_endianness. fcntl.h should then be included by net.c
> >
> > Ok.
> > > > I guess that was the purpose of layout1.with_net (in fs_test.c)
> > but it
> >
> > Yep. I added this kind of nest in fs_test.c to test both fs and network
> > rules together.
> > > is not complete. You can revamp this test and move it to net.c
> > > following the above suggestions, keeping it consistent with other tests
> > > in net.c . You don't need the test_open() nor create_ruleset() helpers.
> > > > This test must failed if we change
> > "ruleset->access_masks[layer_level] |="
> > > to "ruleset->access_masks[layer_level] =" in
> > > landlock_add_fs_access_mask() or landlock_add_net_access_mask().
> >
> > Do you want to change it? Why?
>
> The kernel code is correct and must not be changed. However, if by
> mistake we change it and remove the OR, a test should catch that. We
> need a test to assert this assumption.
>
> > Fs and network masks are ORed to not intersect with each other.
>
> Yes, they are ORed, and we need a test to check that. Noting is
> currently testing this OR (and the different rule type consistency).
> I'm suggesting to revamp the layout1.with_net test into
> ipv4_tcp.with_fs and make it check ruleset->access_masks[] and rule
> addition of different types.
About my previous comment.
Checking the code we can notice that adding fs mask goes first:
...
if (fs_access_mask)
landlock_add_fs_access_mask(new_ruleset, fs_access_mask, 0);
if (net_access_mask)
landlock_add_net_access_mask(new_ruleset, net_access_mask, 0);
....
So with we change "ruleset->access_masks[layer_level] |="
>> > to "ruleset->access_masks[layer_level] =" in
landlock_add_fs_access_mask() nothing bad will happen.
But if we do that in landlock_add_net_access_mask()
fs mask will be overwritten and adding fs rule will fail
(as unhandled allowed_accesss).
