On Mon, Dec 11, 2023 at 3:51 AM Yunsheng Lin linyunsheng@huawei.com wrote:
On 2023/12/11 12:04, Mina Almasry wrote:
On Sun, Dec 10, 2023 at 6:26 PM Mina Almasry almasrymina@google.com wrote:
On Sun, Dec 10, 2023 at 6:04 PM Yunsheng Lin linyunsheng@huawei.com wrote:
On 2023/12/9 0:05, Mina Almasry wrote:
On Fri, Dec 8, 2023 at 1:30 AM Yunsheng Lin linyunsheng@huawei.com wrote:
As mentioned before, it seems we need to have the above checking every time we need to do some per-page handling in page_pool core, is there a plan in your mind how to remove those kind of checking in the future?
I see 2 ways to remove the checking, both infeasible:
- Allocate a wrapper struct that pulls out all the fields the page pool needs:
struct netmem { /* common fields */ refcount_t refcount; bool is_pfmemalloc; int nid; ... union { struct dmabuf_genpool_chunk_owner *owner; struct page * page; }; };
The page pool can then not care if the underlying memory is iov or page. However this introduces significant memory bloat as this struct needs to be allocated for each page or ppiov, which I imagine is not acceptable for the upside of removing a few static_branch'd if statements with no performance cost.
- Create a unified struct for page and dmabuf memory, which the mm
folks have repeatedly nacked, and I imagine will repeatedly nack in the future.
So I imagine the special handling of ppiov in some form is critical and the checking may not be removable.
If the above is true, perhaps devmem is not really supposed to be intergated into page_pool.
Adding a checking for every per-page handling in page_pool core is just too hacky to be really considerred a longterm solution.
The only other option is to implement another page_pool for ppiov and have the driver create page_pool or ppiov_pool depending on the state of the netdev_rx_queue (or some helper in the net stack to do that for the driver). This introduces some code duplication. The ppiov_pool & page_pool would look similar in implementation.
I think there is a design pattern already to deal with this kind of problem, refactoring common code used by both page_pool and ppiov into a library to aovid code duplication if most of them have similar implementation.
Code can be refactored if it's identical, not if it is similar. I suspect the page_pools will be only similar, and if you're not willing to take devmem handling into the page pool then refactoring page_pool code into helpers that do devmem handling may also not be an option.
But this was all discussed in detail in RFC v2 and the last response I heard from Jesper was in favor if this approach, if I understand correctly:
https://lore.kernel.org/netdev/7aedc5d5-0daf-63be-21bc-3b724cc1cab9@redhat.c...
Would love to have the maintainer weigh in here.
I should note we may be able to remove some of the checking, but maybe not all.
- Checks that disable page fragging for ppiov can be removed once
ppiov has frag support (in this series or follow up).
- If we use page->pp_frag_count (or page->pp_ref_count) for
refcounting ppiov, we can remove the if checking in the refcounting.
I'm not sure this is actually possible in the short term. The page_pool uses both page->_refcount and page->pp_frag_count for refcounting, and I will not be able to remove the special handling around page->_refcount as i'm not allowed to call page_ref_*() APIs on a non-struct page.
- We may be able to store the dma_addr of the ppiov in page->dma_addr,
but I'm unsure if that actually works, because the dma_buf dmaddr is dma_addr_t (u32 or u64), but page->dma_addr is unsigned long (4 bytes I think). But if it works for pages I may be able to make it work for ppiov as well.
- Checks that obtain the page->pp can work with ppiov if we align the
offset of page->pp and ppiov->pp.
- Checks around page->pp_magic can be removed if we also have offset
aligned ppiov->pp_magic.
Sadly I don't see us removing the checking for these other cases:
- page_is_pfmemalloc(): I'm not allowed to pass a non-struct page into
that helper.
We can do similar trick like above as bit 1 of page->pp_magic is used to indicate that if it is a pfmemalloc page.
Likely yes.
- page_to_nid(): I'm not allowed to pass a non-struct page into that helper.
Yes, this one need special case.
- page_pool_free_va(): ppiov have no va.
Doesn't the skb_frags_readable() checking will protect the page_pool_free_va() from being called on devmem?
This function seems to be only called from veth which doesn't support devmem. I can remove the handling there.
- page_pool_sync_for_dev/page_pool_dma_map: ppiov backed by dma-buf
fundamentally can't get mapped again.
Can we just fail the page_pool creation with PP_FLAG_DMA_MAP and DMA_ATTR_SKIP_CPU_SYNC flags for devmem provider?
Jakub says PP_FLAG_DMA_MAP must be enabled for devmem, such that the page_pool handles the dma mapping of the devmem and the driver doesn't use it on its own.
We may fail creating the page pool on PP_FLAG_DMA_SYNC_DEV maybe, and remove the checking from page_pool_sync_for_dev(), I think.
Are the removal (or future removal) of these checks enough to resolve this?
Yes, that is somewhat similar to my proposal, the biggest objection seems to be that we need to have a safe type checking for it to work correctly.
It is somewhat ironical that devmem is using static_branch to alliviate the performance impact for normal memory at the possible cost of performance degradation for devmem, does it not defeat some purpose of intergating devmem to page_pool?
I don't see the issue. The static branch sets the non-ppiov path as default if no memory providers are in use, and flips it when they are, making the default branch prediction ideal in both cases.
You are assuming the we are not using page pool for both normal memory and devmem at the same. But a generic solution should not have that assumption as my understanding.
Even though a static_branch check is added in page_is_page_pool_iov(), it does not make much sense that a core has tow different 'struct' for its most basic data.
IMHO, the ppiov for dmabuf is forced fitting into page_pool without much design consideration at this point.
...
For now, the above may work for the the rx part as it seems that you are only enabling rx for dmabuf for now.
What is the plan to enable tx for dmabuf? If it is also intergrated into page_pool? There was a attempt to enable page_pool for tx, Eric seemed to have some comment about this: https://lkml.kernel.org/netdev/2cf4b672-d7dc-db3d-ce90-15b4e91c4005@huawei.c...
If tx is not intergrated into page_pool, do we need to create a new layer for the tx dmabuf?
I imagine the TX path will reuse page_pool_iov, page_pool_iov_*() helpers, and page_pool_page_*() helpers, but will not need any core page_pool changes. This is because the TX path will have to piggyback
We may need another bit/flags checking to demux between page_pool owned devmem and non-page_pool owned devmem.
The way I'm imagining the support, I don't see the need for such flags. We'd be re-using generic helpers like page_pool_iov_get_dma_address() and what not that don't need that checking.
Also calling page_pool_*() on non-page_pool owned devmem is confusing enough that we may need a thin layer handling non-page_pool owned devmem in the end.
The page_pool_page* & page_pool_iov* functions can be renamed if confusing. I would think that's no issue (note that the page_pool_*
When you rename those functions, you will have a thin layer automatically.
functions need not be called for TX path).
on MSG_ZEROCOPY (devmem is not copyable), so no memory allocation from the page_pool (or otherwise) is needed or possible. RFCv1 had a TX implementation based on dmabuf pages without page_pool involvement, I imagine I'll do something similar.
It would be good to have a tx implementation for the next version, so that we can have a whole picture of devmem.