netdev - Re: [net-next 11/15] net/tls: Multi-threaded calls to TX tls_dev

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]

Date:   Wed, 6 Jul 2022 19:37:35 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     Saeed Mahameed <saeed@...nel.org>
Cc:     "David S. Miller" <davem@...emloft.net>,
        Paolo Abeni <pabeni@...hat.com>,
        Eric Dumazet <edumazet@...gle.com>,
        Saeed Mahameed <saeedm@...dia.com>, netdev@...r.kernel.org,
        Tariq Toukan <tariqt@...dia.com>,
        Maxim Mikityanskiy <maximmi@...dia.com>
Subject: Re: [net-next 11/15] net/tls: Multi-threaded calls to TX
 tls_dev_del

On Wed,  6 Jul 2022 16:24:17 -0700 Saeed Mahameed wrote:
> diff --git a/include/net/tls.h b/include/net/tls.h
> index 4fc16ca5f469..c4be74635502 100644
> --- a/include/net/tls.h
> +++ b/include/net/tls.h
> @@ -163,6 +163,11 @@ struct tls_record_info {
>  	skb_frag_t frags[MAX_SKB_FRAGS];
>  };
>  
> +struct destruct_work {
> +	struct work_struct work;
> +	struct tls_context *ctx;

Pretty strange to bundle the back-pointer with the work.
Why not put it directly in struct tls_offload_context_tx?

Also now that we have the backpointer, can we move the list member of
struct tls_context to the offload context? (I haven't checked if its
used in other places)

>  
>  	up_write(&device_offload_lock);
>  
> -	flush_work(&tls_device_gc_work);
> -
>  	return NOTIFY_DONE;
>  }
>  
> @@ -1435,6 +1416,5 @@ void __init tls_device_init(void)
>  void __exit tls_device_cleanup(void)
>  {
>  	unregister_netdevice_notifier(&tls_dev_notifier);
> -	flush_work(&tls_device_gc_work);
>  	clean_acked_data_flush();
>  }

Why don't we need the flush any more? The module reference is gone as
soon as destructor runs (i.e. on ULP cleanup), the work can still be
pending, no?