netdev - Re: [net-next PATCH 13/15] eth: fbnic: add basic Rx handling

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: <b725331c-ae88-b9dd-de12-e8e9b9fc020b@huawei.com>
Date: Tue, 16 Apr 2024 21:25:25 +0800
From: Yunsheng Lin <linyunsheng@...wei.com>
To: Jakub Kicinski <kuba@...nel.org>, Alexander Duyck
	<alexander.duyck@...il.com>
CC: <netdev@...r.kernel.org>, Alexander Duyck <alexanderduyck@...com>,
	<davem@...emloft.net>, <pabeni@...hat.com>
Subject: Re: [net-next PATCH 13/15] eth: fbnic: add basic Rx handling

On 2024/4/16 6:01, Jakub Kicinski wrote:
> On Mon, 15 Apr 2024 11:55:37 -0700 Alexander Duyck wrote:
>> It would take a few more changes to make it all work. Basically we
>> would need to map the page into every descriptor entry since the worst
>> case scenario would be that somehow we end up with things getting so
>> tight that the page is only partially mapped and we are working
>> through it as a subset of 4K slices with some at the beginning being
>> unmapped from the descriptor ring while some are still waiting to be
>> assigned to a descriptor and used. What I would probably have to look
>> at doing is adding some sort of cache on the ring to hold onto it
>> while we dole it out 4K at a time to the descriptors. Either that or
>> enforce a hard 16 descriptor limit where we have to assign a full page
>> with every allocation meaning we are at a higher risk for starving the
>> device for memory.
> 
> Hm, that would be more work, indeed, but potentially beneficial. I was
> thinking of separating the page allocation and draining logic a bit
> from the fragment handling logic.
> 
> #define RXPAGE_IDX(idx)		((idx) >> PAGE_SHIFT - 12)
> 
> in fbnic_clean_bdq():
> 
> 	while (RXPAGE_IDX(head) != RXPAGE_IDX(hw_head))
> 
> refer to rx_buf as:
> 
> 	struct fbnic_rx_buf *rx_buf = &ring->rx_buf[idx >> LOSE_BITS];
> 
> Refill always works in batches of multiple of PAGE_SIZE / 4k.

Are we expecting drivers wanting best possible performance doing the
above duplicated trick?

"grep -rn '_reuse_' drivers/net/ethernet/" seems to suggest that we
already have similar trick to do the page spliting in a lot of drivers,
I would rather we do not duplicate the above trick again.