[PATCH v2] selinux: use GFP_ATOMIC in convert_context()

[Paul Moore]

On Tue, Oct 18, 2022 at 10:56 PM GONG, Ruiqi <gongruiqi1 at huawei.com> wrote:
>
> The following warning was triggered on a hardware environment:
>
>   SELinux: Converting 162 SID table entries...
>   BUG: sleeping function called from invalid context at __might_sleep+0x60/0x74 0x0
>   in_atomic(): 1, irqs_disabled(): 128, non_block: 0, pid: 5943, name: tar
>   CPU: 7 PID: 5943 Comm: tar Tainted: P O 5.10.0 #1
>   Call trace:
>    dump_backtrace+0x0/0x1c8
>    show_stack+0x18/0x28
>    dump_stack+0xe8/0x15c
>    ___might_sleep+0x168/0x17c
>    __might_sleep+0x60/0x74
>    __kmalloc_track_caller+0xa0/0x7dc
>    kstrdup+0x54/0xac
>    convert_context+0x48/0x2e4
>    sidtab_context_to_sid+0x1c4/0x36c
>    security_context_to_sid_core+0x168/0x238
>    security_context_to_sid_default+0x14/0x24
>    inode_doinit_use_xattr+0x164/0x1e4
>    inode_doinit_with_dentry+0x1c0/0x488
>    selinux_d_instantiate+0x20/0x34
>    security_d_instantiate+0x70/0xbc
>    d_splice_alias+0x4c/0x3c0
>    ext4_lookup+0x1d8/0x200 [ext4]
>    __lookup_slow+0x12c/0x1e4
>    walk_component+0x100/0x200
>    path_lookupat+0x88/0x118
>    filename_lookup+0x98/0x130
>    user_path_at_empty+0x48/0x60
>    vfs_statx+0x84/0x140
>    vfs_fstatat+0x20/0x30
>    __se_sys_newfstatat+0x30/0x74
>    __arm64_sys_newfstatat+0x1c/0x2c
>    el0_svc_common.constprop.0+0x100/0x184
>    do_el0_svc+0x1c/0x2c
>    el0_svc+0x20/0x34
>    el0_sync_handler+0x80/0x17c
>    el0_sync+0x13c/0x140
>   SELinux: Context system_u:object_r:pssp_rsyslog_log_t:s0:c0 is not valid (left unmapped).
>
> It was found that within a critical section of spin_lock_irqsave in
> sidtab_context_to_sid(), convert_context() (hooked by
> sidtab_convert_params.func) might cause the process to sleep via
> allocating memory with GFP_KERNEL, which is problematic.
>
> As Ondrej pointed out [1], convert_context()/sidtab_convert_params.func
> has another caller sidtab_convert_tree(), which is okay with GFP_KERNEL.
> Therefore, fix this problem by adding a gfp_t argument for
> convert_context()/sidtab_convert_params.func and pass GFP_KERNEL/_ATOMIC
> properly in individual callers.
>
> Link: https://lore.kernel.org/all/20221018120111.1474581-1-gongruiqi1@huawei.com/ [1]
> Reported-by: Tan Ninghao <tanninghao1 at huawei.com>
> Fixes: ee1a84fdfeed ("selinux: overhaul sidtab to fix bug and improve performance")
> Signed-off-by: GONG, Ruiqi <gongruiqi1 at huawei.com>
> ---
>
> v2: change as Ondrej suggests & redraft commit message
>
>  security/selinux/ss/services.c | 5 +++--
>  security/selinux/ss/sidtab.c   | 4 ++--
>  security/selinux/ss/sidtab.h   | 2 +-
>  3 files changed, 6 insertions(+), 5 deletions(-)

Merged into selinux/stable-6.1, thank you.  Normally I would send this
to Linus in a day or two, but due to some personal logistical
challenges I may be a bit delayed in sending this up.  I would ask for
your patience and that everyone take this opportunity to do some
additional testing :)

Thanks everyone.

-- 
paul-moore.com