linux-kernel - Re: [PATCH v2 bpf-next 2/4] bpf: Make bpf inode storage available to tracing program

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <2cb0cd4f-5d78-4b7f-b280-2a3377ffbc21@linux.dev>
Date: Tue, 12 Nov 2024 15:00:48 -0800
From: Martin KaFai Lau <martin.lau@...ux.dev>
To: Song Liu <song@...nel.org>
Cc: bpf@...r.kernel.org, linux-fsdevel@...r.kernel.org,
 linux-kernel@...r.kernel.org, linux-security-module@...r.kernel.org,
 kernel-team@...a.com, andrii@...nel.org, eddyz87@...il.com, ast@...nel.org,
 daniel@...earbox.net, viro@...iv.linux.org.uk, brauner@...nel.org,
 jack@...e.cz, kpsingh@...nel.org, mattbobrowski@...gle.com,
 amir73il@...il.com, repnop@...gle.com, jlayton@...nel.org,
 josef@...icpanda.com, mic@...ikod.net, gnoack@...gle.com
Subject: Re: [PATCH v2 bpf-next 2/4] bpf: Make bpf inode storage available to
 tracing program

On 11/12/24 12:36 AM, Song Liu wrote:
>   void __destroy_inode(struct inode *inode)
>   {
>   	BUG_ON(inode_has_buffers(inode));
> +	bpf_inode_storage_free(inode);

Not sure if this is done in the rcu callback (i.e. after the rcu gp). Please check.

>   	inode_detach_wb(inode);
>   	security_inode_free(inode);
>   	fsnotify_inode_delete(inode);

[ ... ]

> @@ -136,12 +119,7 @@ BPF_CALL_5(bpf_inode_storage_get, struct bpf_map *, map, struct inode *, inode,
>   	if (flags & ~(BPF_LOCAL_STORAGE_GET_F_CREATE))
>   		return (unsigned long)NULL;
>   
> -	/* explicitly check that the inode_storage_ptr is not
> -	 * NULL as inode_storage_lookup returns NULL in this case and
> -	 * bpf_local_storage_update expects the owner to have a
> -	 * valid storage pointer.
> -	 */
> -	if (!inode || !inode_storage_ptr(inode))
> +	if (!inode)
>   		return (unsigned long)NULL;

There is an atomic_read in this function:

	/* only allocate new storage, when the inode is refcounted */
	if (atomic_read(&inode->i_count) &&
	    flags & BPF_LOCAL_STORAGE_GET_F_CREATE) {

If the bpf_inode_storage_free is not done after rcu gp, this will need a 
inc_not_zero like how the sk storage does. I think moving the storage_free to 
the inode rcu call back may be easier if it is not the case now.