lkml.org 
[lkml]   [2023]   [Jul]   [5]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
SubjectRe: [PATCH RFC net-next 2/4] net: page_pool: avoid calling no-op externals when possible
From
From: Jakub Kicinski <kuba@kernel.org>
Date: Mon, 3 Jul 2023 13:32:07 -0700

> On Fri, 30 Jun 2023 17:34:02 +0200 Alexander Lobakin wrote:
>>> I am not a fan of having the page pool force the syncing either. Last
>>> I knew I thought the PP_FLAG_DMA_SYNC_DEV was meant to be set by the
>>
>> Please follow the logics of the patch.
>>
>> 1. The driver sets DMA_SYNC_DEV.
>> 2. PP tries to shortcut and replaces it with MAYBE_SYNC.
>> 3. If dma_need_sync() returns true for some page, it gets replaced back
>> to DMA_SYNC_DEV, no further dma_need_sync() calls for that pool.
>>
>> OR
>>
>> 1. The driver doesn't set DMA_SYNC_DEV.
>> 2. PP doesn't turn on MAYBE_SYNC.
>> 3. No dma_need_sync() tests.
>>
>> Where does PP force syncs for drivers which don't need them?
>
> I think both Alex and I got confused about what's going on here.
>
> Could you reshuffle the code somehow to make it more obvious?
> Rename the flag, perhaps put it in a different field than
> the driver-set PP flags?

PP currently doesn't have a field for internal flags or so, so I reused
the existing one :s But you're probably right, that would make it more
obvious.

1. Driver sets PP_SYNC_DEV.
2. PP doesn't set its internal one until dma_need_sync() returns false.
3. PP-sync-for-dev checks for the internal flag.

Although needs more lines to be changed :D

Thanks,
Olek

\
 
 \ /
  Last update: 2023-07-05 16:43    [W:0.771 / U:0.028 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site