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>] [day] [month] [year] [list]
Date:   Fri, 11 Sep 2020 08:57:50 -0300
From:   Jason Gunthorpe <jgg@...pe.ca>
To:     Hillf Danton <hdanton@...a.com>
Cc:     syzbot <syzbot+cc6fc752b3819e082d0c@...kaller.appspotmail.com>,
        dledford@...hat.com, leon@...nel.org, linux-kernel@...r.kernel.org,
        linux-rdma@...r.kernel.org, syzkaller-bugs@...glegroups.com
Subject: Re: KASAN: use-after-free Read in ucma_close (2)

On Fri, Sep 11, 2020 at 12:16:40PM +0800, Hillf Danton wrote:
> Detect race destroying ctx in order to avoid UAF.
> 
> +++ b/drivers/infiniband/core/ucma.c
> @@ -625,6 +625,10 @@ static ssize_t ucma_destroy_id(struct uc
>  		return PTR_ERR(ctx);
>  
>  	mutex_lock(&ctx->file->mut);
> +	if (ctx->destroying == 1) {
> +		mutex_unlock(&ctx->file->mut);
> +		return -ENXIO;
> +	}
>  	ctx->destroying = 1;
>  	mutex_unlock(&ctx->file->mut);
>  
> @@ -1826,6 +1830,8 @@ static int ucma_close(struct inode *inod
>  
>  	mutex_lock(&file->mut);
>  	list_for_each_entry_safe(ctx, tmp, &file->ctx_list, list) {
> +		if (ctx->destroying == 1)
> +			continue;
>  		ctx->destroying = 1;
>  		mutex_unlock(&file->mut);
>  

ucma_destroy_id() is called from write() and ucma_close is release(),
so there is no way these can race?

Jason

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ