Re: [PATCH net-next v10 08/16] tls: Inline do_tcp_sendpages()

[Tariq Toukan <ttoukan.linux@gmail.com>]

On 07/06/2023 18:03, David Howells wrote:
> Tariq Toukan <ttoukan.linux@gmail.com> wrote:
> 
>> My team spotted a new degradation in TLS TX device offload, bisected to this
>> patch.
> 
> I presume you're using some hardware (I'm guessing Mellanox?) that can
> actually do TLS offload?  Unfortunately, I don't have any hardware that can do
> this, so I can't test the tls_device stuff.
> 
>>  From a quick look at the patch, it's not clear to me what's going wrong.
>> Please let us know of any helpful information that we can provide to help in
>> the debug.
> 
> Can you find out what source line this corresponds to?
> 
> 	RIP: 0010:skb_splice_from_iter+0x102/0x300
> 
> Assuming you're building your own kernel, something like the following might
> do the trick:
> 
> 	echo "RIP: 0010:skb_splice_from_iter+0x102/0x300" |
> 	./scripts/decode_stacktrace.sh /my/built/vmlinux /my/build/tree
> 

Hi,

It's:
RIP: 0010:skb_splice_from_iter (/usr/linux/net/core/skbuff.c:6957)

which coresponds to this line:
                         if (WARN_ON_ONCE(!sendpage_ok(page)))

> if you run it in the kernel source tree you're using and substitute the
> paths to vmlinux and the build tree for modules.
> 
> David
>