linux-kernel - Re: [RESEND][PATCH] ima: Set and clear FMODE_CAN_READ in ima_calc_file

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: <20201114111057.GA16415@infradead.org>
Date:   Sat, 14 Nov 2020 11:10:57 +0000
From:   Christoph Hellwig <hch@...radead.org>
To:     Roberto Sassu <roberto.sassu@...wei.com>
Cc:     zohar@...ux.ibm.com, linux-integrity@...r.kernel.org,
        linux-security-module@...r.kernel.org,
        linux-kernel@...r.kernel.org, silviu.vlasceanu@...wei.com,
        stable@...r.kernel.org, torvalds@...ux-foundation.org,
        viro@...iv.linux.org.uk, linux-fsdevel@...r.kernel.org
Subject: Re: [RESEND][PATCH] ima: Set and clear FMODE_CAN_READ in
 ima_calc_file_hash()

On Fri, Nov 13, 2020 at 09:01:32AM +0100, Roberto Sassu wrote:
> Commit a1f9b1c0439db ("integrity/ima: switch to using __kernel_read")
> replaced the __vfs_read() call in integrity_kernel_read() with
> __kernel_read(), a new helper introduced by commit 61a707c543e2a ("fs: add
> a __kernel_read helper").
> 
> Since the new helper requires that also the FMODE_CAN_READ flag is set in
> file->f_mode, this patch saves the original f_mode and sets the flag if the
> the file descriptor has the necessary file operation. Lastly, it restores
> the original f_mode at the end of ima_calc_file_hash().

This looks bogus.  FMODE_CAN_READ has a pretty clear definition and
you can't just go and read things if it is not set.  Also f_mode
manipulations on a life file are racy.

> 
> Cc: stable@...r.kernel.org # 5.8.x
> Fixes: a1f9b1c0439db ("integrity/ima: switch to using __kernel_read")
> Signed-off-by: Roberto Sassu <roberto.sassu@...wei.com>
> ---
>  security/integrity/ima/ima_crypto.c | 8 ++++++--
>  1 file changed, 6 insertions(+), 2 deletions(-)
> 
> diff --git a/security/integrity/ima/ima_crypto.c b/security/integrity/ima/ima_crypto.c
> index 21989fa0c107..22ed86a0c964 100644
> --- a/security/integrity/ima/ima_crypto.c
> +++ b/security/integrity/ima/ima_crypto.c
> @@ -537,6 +537,7 @@ int ima_calc_file_hash(struct file *file, struct ima_digest_data *hash)
>  	loff_t i_size;
>  	int rc;
>  	struct file *f = file;
> +	fmode_t saved_mode;
>  	bool new_file_instance = false, modified_mode = false;
>  
>  	/*
> @@ -550,7 +551,7 @@ int ima_calc_file_hash(struct file *file, struct ima_digest_data *hash)
>  	}
>  
>  	/* Open a new file instance in O_RDONLY if we cannot read */
> -	if (!(file->f_mode & FMODE_READ)) {
> +	if (!(file->f_mode & FMODE_READ) || !(file->f_mode & FMODE_CAN_READ)) {
>  		int flags = file->f_flags & ~(O_WRONLY | O_APPEND |
>  				O_TRUNC | O_CREAT | O_NOCTTY | O_EXCL);
>  		flags |= O_RDONLY;
> @@ -562,7 +563,10 @@ int ima_calc_file_hash(struct file *file, struct ima_digest_data *hash)
>  			 */
>  			pr_info_ratelimited("Unable to reopen file for reading.\n");
>  			f = file;
> +			saved_mode = f->f_mode;
>  			f->f_mode |= FMODE_READ;
> +			if (likely(file->f_op->read || file->f_op->read_iter))
> +				f->f_mode |= FMODE_CAN_READ;
>  			modified_mode = true;
>  		} else {
>  			new_file_instance = true;
> @@ -582,7 +586,7 @@ int ima_calc_file_hash(struct file *file, struct ima_digest_data *hash)
>  	if (new_file_instance)
>  		fput(f);
>  	else if (modified_mode)
> -		f->f_mode &= ~FMODE_READ;
> +		f->f_mode = saved_mode;
>  	return rc;
>  }
>  
> -- 
> 2.27.GIT
> 
---end quoted text---