On Fri, Oct 11, 2024 at 8:27 AM Jakub Kicinski wrote:
>
> On Thu, 10 Oct 2024 12:05:38 -0700 Mina Almasry wrote:
> > diff --git a/net/core/sock.c b/net/core/sock.c
> > index 083d438d8b6f..cb3d8b19de14 100644
> > --- a/net/core/sock.c
> > +++ b/net/core/s
On Thu, Oct 10, 2024 at 4:17 AM Lai, Yi wrote:
>
> Hi Mina Almasry,
>
> Greetings!
>
> I used Syzkaller and found that there is BUG: soft lockup inqt in linux-next
> tree next-20241008
>
> After bisection and the first bad commit is:
> "
> 678f6e28b5f6 net
Make skb_frag_page() fail in the case where the frag is not backed
by a page, and fix its relevant callers to handle this case.
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
Reviewed-by: Jakub Kicinski
---
v10:
- Fixed newly generated kdoc warnings found by patchwork. While we&#x
Add documentation outlining the usage and details of devmem TCP.
Signed-off-by: Mina Almasry
Reviewed-by: Bagas Sanjaya
Reviewed-by: Donald Hunter
---
v25:
- Doc cleanups (Jakub)
v16:
- Add documentation on unbinding the NIC from dmabuf (Donald).
- Add note that any dmabuf should work
ff-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v23:
- WARN_ON when we don't see the dma_addr in the gen_pool (Jakub)
v20:
- Removed dma_addr field in dmabuf_genpool_chunk_owner not used in th
7;inflight': 1023,
'inflight-mem': 4190208},
And queue stats:
$ ./cli.py --spec ../netlink/specs/netdev.yaml --dump queue-get
...
{'dmabuf': 10, 'id': 8, 'ifindex': 3, 'type': 'rx'},
{'dmabuf': 10, 'id': 9, &
Add an interface for the user to notify the kernel that it is done
reading the devmem dmabuf frags returned as cmsg. The kernel will
drop the reference on the frags to make them available for reuse.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
, ncdevmem has a validation mode,
where it sends a specific pattern and validates this pattern on the
receiver side to ensure data integrity.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
---
v24:
- Add *.d to .gitignore, to stop tracking this file the build generates:
```
git status
ags of
devmem skbs and avoid coalescing devmem skbs with non devmem skbs.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
Reviewed-by: Jakub Kicinski
---
v25:
- Remove readable check in tcp_skb_can_collapse_to().
- Add ch
er_frags, and each page passed to userspace is get_page()'d.
This reference is dropped once the userspace indicates that it is
done reading this page. All pages are released when the socket is
destroyed.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Re
is omitted for simplicity & p.order !=
0.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v25:
- Change page_pool_param netdev_rx_queue struct to queue_idx (Jakub)
- Address nits (Jakub)
hether the underlying type is page or net_iov.
Implement checks for net_iov in netmem helpers which delegate to mm
APIs, to ensure net_iov are never passed to the mm stack.
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Acked-by: Jakub Kicinski
---
v23:
- Fix comment on netmem_is_perf_
The netdev_dmabuf_binding struct is refcounted, and releases its
resources only when all the refs are released.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov # excluding netlink
Acked-by: Daniel Vetter
Reviewed-by: Jakub Kicinski
---
API takes the dma-buf fd as input, and binds it to the netdevice. The
user can specify the rx queues to bind the dma-buf to.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
Reviewed-by: Donald Hunter
Reviewed-by: Jakub Kicinski
---
v16:
- Use subset-of: queue queue-id instead
ff-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v18:
- Add more color to commit message (Xuan Zhuo).
v17:
- Use ASSERT_RTNL() (Jakub).
v13:
- Add reviewed-by from Pavel (thanks!)
- Fixed comment (Pavel)
v11:
- Fix not checking dev->queue_mgmt_ops (Pavel
er API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov
Cc: David Wei
Cc: Jason Gunthorpe
Cc: Yunsheng Lin
Cc: Shailend Chand
Cc: Harshitha Ramamurthy
Cc: Shakeel Butt
Cc: Jeroen de Borst
Cc: Pra
On Tue, Sep 10, 2024 at 8:02 AM Jakub Kicinski wrote:
>
> On Mon, 9 Sep 2024 05:43:05 +0000 Mina Almasry wrote:
> > Device memory TCP (devmem TCP) is a proposal for transferring data to and/or
> > from device memory efficiently, without bouncing the data to a host memory
> &
On Mon, Sep 9, 2024 at 4:21 AM Yunsheng Lin wrote:
>
> On 2024/9/9 13:43, Mina Almasry wrote:
>
> >
> > Perf - page-pool benchmark:
> > ---
> >
> > bench_page_pool_simple.ko tests with and without these changes:
> > https:
, ncdevmem has a validation mode,
where it sends a specific pattern and validates this pattern on the
receiver side to ensure data integrity.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
---
v24:
- Add *.d to .gitignore, to stop tracking this file the build generates:
```
git status
7;inflight': 1023,
'inflight-mem': 4190208},
And queue stats:
$ ./cli.py --spec ../netlink/specs/netdev.yaml --dump queue-get
...
{'dmabuf': 10, 'id': 8, 'ifindex': 3, 'type': 'rx'},
{'dmabuf': 10, 'id': 9, &
er_frags, and each page passed to userspace is get_page()'d.
This reference is dropped once the userspace indicates that it is
done reading this page. All pages are released when the socket is
destroyed.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Re
Add documentation outlining the usage and details of devmem TCP.
Signed-off-by: Mina Almasry
Reviewed-by: Bagas Sanjaya
Reviewed-by: Donald Hunter
---
v25:
- Doc cleanups (Jakub)
v16:
- Add documentation on unbinding the NIC from dmabuf (Donald).
- Add note that any dmabuf should work
Add an interface for the user to notify the kernel that it is done
reading the devmem dmabuf frags returned as cmsg. The kernel will
drop the reference on the frags to make them available for reuse.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
ags of
devmem skbs and avoid coalescing devmem skbs with non devmem skbs.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v25:
- Remove readable check in tcp_skb_can_collapse_to().
- Add check in skb_checksum_help (Jakub).
Make skb_frag_page() fail in the case where the frag is not backed
by a page, and fix its relevant callers to handle this case.
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
Reviewed-by: Jakub Kicinski
---
v10:
- Fixed newly generated kdoc warnings found by patchwork. While we&#x
is omitted for simplicity & p.order !=
0.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v25:
- Change page_pool_param netdev_rx_queue struct to queue_idx (Jakub)
- Address nits (Jakub)
- Move mp_dmabuf_devmem.h
hether the underlying type is page or net_iov.
Implement checks for net_iov in netmem helpers which delegate to mm
APIs, to ensure net_iov are never passed to the mm stack.
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Acked-by: Jakub Kicinski
---
v23:
- Fix comment on netmem_is_perf_
ff-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v23:
- WARN_ON when we don't see the dma_addr in the gen_pool (Jakub)
v20:
- Removed dma_addr field in dmabuf_genpool_chunk_owner not used in th
The netdev_dmabuf_binding struct is refcounted, and releases its
resources only when all the refs are released.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov # excluding netlink
Acked-by: Daniel Vetter
Reviewed-by: Jakub Kicinski
---
API takes the dma-buf fd as input, and binds it to the netdevice. The
user can specify the rx queues to bind the dma-buf to.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
Reviewed-by: Donald Hunter
Reviewed-by: Jakub Kicinski
---
v16:
- Use subset-of: queue queue-id instead
ff-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v18:
- Add more color to commit message (Xuan Zhuo).
v17:
- Use ASSERT_RTNL() (Jakub).
v13:
- Add reviewed-by from Pavel (thanks!)
- Fixed comment (Pavel)
v11:
- Fix not checking dev->queue_mgmt_ops (Pavel
end without
a dependency on a specific dmabuf provider.
** Test Setup
Kernel: net-next with this series and memory provider API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov
Cc: David Wei
Cc: Jason Gunth
On Tue, Sep 3, 2024 at 2:19 PM Jakub Kicinski wrote:
>
> On Sat, 31 Aug 2024 00:43:06 +0000 Mina Almasry wrote:
> > diff --git a/include/net/mp_dmabuf_devmem.h b/include/net/mp_dmabuf_devmem.h
> > new file mode 100644
> > index ..6d1cf2a77f6b
> > --- /
On Tue, Sep 3, 2024 at 2:40 PM Jakub Kicinski wrote:
>
> On Sat, 31 Aug 2024 00:43:08 +0000 Mina Almasry wrote:
> > static inline bool tcp_skb_can_collapse_to(const struct sk_buff *skb)
> > {
> > - return likely(!TCP_SKB_CB(skb)->eor);
> > + re
Add an interface for the user to notify the kernel that it is done
reading the devmem dmabuf frags returned as cmsg. The kernel will
drop the reference on the frags to make them available for reuse.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Make skb_frag_page() fail in the case where the frag is not backed
by a page, and fix its relevant callers to handle this case.
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v10:
- Fixed newly generated kdoc warnings found by patchwork. While we're
at it, fix the Return se
is omitted for simplicity & p.order !=
0.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v23:
- Sort includes (Jakub)
- Add missing linux/mm.h include found after sorting.
v21:
- Provide empty definitions of fu
, ncdevmem has a validation mode,
where it sends a specific pattern and validates this pattern on the
receiver side to ensure data integrity.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
---
v24:
- Add *.d to .gitignore, to stop tracking this file the build generates:
```
git status
Add documentation outlining the usage and details of devmem TCP.
Signed-off-by: Mina Almasry
Reviewed-by: Bagas Sanjaya
Reviewed-by: Donald Hunter
---
v16:
- Add documentation on unbinding the NIC from dmabuf (Donald).
- Add note that any dmabuf should work (Donald).
v9:
https
er_frags, and each page passed to userspace is get_page()'d.
This reference is dropped once the userspace indicates that it is
done reading this page. All pages are released when the socket is
destroyed.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Re
7;inflight': 1023,
'inflight-mem': 4190208},
And queue stats:
$ ./cli.py --spec ../netlink/specs/netdev.yaml --dump queue-get
...
{'dmabuf': 10, 'id': 8, 'ifindex': 3, 'type': 'rx'},
{'dmabuf': 10, 'id': 9, &
ags of
devmem skbs and avoid coalescing devmem skbs with non devmem skbs.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v16:
- Fix unreadable handling in skb_split_no_header() (Eric).
v11:
- drop excessive checks for
hether the underlying type is page or net_iov.
Implement checks for net_iov in netmem helpers which delegate to mm
APIs, to ensure net_iov are never passed to the mm stack.
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Acked-by: Jakub Kicinski
---
v23:
- Fix comment on netmem_is_perf_
ff-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v18:
- Add more color to commit message (Xuan Zhuo).
v17:
- Use ASSERT_RTNL() (Jakub).
v13:
- Add reviewed-by from Pavel (thanks!)
- Fixed comment (Pavel)
v11:
- Fix not checking dev->queue_mgmt_ops (Pavel
The netdev_dmabuf_binding struct is refcounted, and releases its
resources only when all the refs are released.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov # excluding netlink
Acked-by: Daniel Vetter
---
v24:
- Fix a coupl
ff-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v23:
- WARN_ON when we don't see the dma_addr in the gen_pool (Jakub)
v20:
- Removed dma_addr field in dmabuf_genpool_chunk_owner not used in th
API takes the dma-buf fd as input, and binds it to the netdevice. The
user can specify the rx queues to bind the dma-buf to.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
Reviewed-by: Donald Hunter
Reviewed-by: Jakub Kicinski
---
v16:
- Use subset-of: queue queue-id instead
dependency on a specific dmabuf provider.
** Test Setup
Kernel: net-next with this series and memory provider API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov
Cc: David Wei
Cc: Jason Gunthorpe
Cc: Y
On Thu, Aug 29, 2024 at 2:24 PM Vadim Fedorenko
wrote:
>
> On 29/08/2024 22:08, Jakub Kicinski wrote:
> > On Thu, 29 Aug 2024 06:01:16 + Mina Almasry wrote:
> >> +err = genlmsg_reply(rsp, info);
> >> +if (err)
> >> +got
> +#include
> #include
>
...
>
> +#include
> #include
>
Darn, I went too overboard with sorting of includes. ynl-regen.sh
wants these in the reverse order, which is unsorted. I'll fix it in
the next iteration, and I added this check as well to my presubmits.
BTW I submitted 2 iterations al
7;inflight': 1023,
'inflight-mem': 4190208},
And queue stats:
$ ./cli.py --spec ../netlink/specs/netdev.yaml --dump queue-get
...
{'dmabuf': 10, 'id': 8, 'ifindex': 3, 'type': 'rx'},
{'dmabuf': 10, 'id': 9,
Add documentation outlining the usage and details of devmem TCP.
Signed-off-by: Mina Almasry
Reviewed-by: Bagas Sanjaya
Reviewed-by: Donald Hunter
---
v16:
- Add documentation on unbinding the NIC from dmabuf (Donald).
- Add note that any dmabuf should work (Donald).
v9:
https
, ncdevmem has a validation mode,
where it sends a specific pattern and validates this pattern on the
receiver side to ensure data integrity.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
---
v22:
- Add run_command helper. It reduces boiler plate and prints the
commands it is running
Add an interface for the user to notify the kernel that it is done
reading the devmem dmabuf frags returned as cmsg. The kernel will
drop the reference on the frags to make them available for reuse.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
ags of
devmem skbs and avoid coalescing devmem skbs with non devmem skbs.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v16:
- Fix unreadable handling in skb_split_no_header() (Eric).
v11:
- drop excessive checks for
is omitted for simplicity & p.order !=
0.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v23:
- Sort includes (Jakub)
- Add missing linux/mm.h include found after sorting.
v21:
- Provide empty definitions of fu
er_frags, and each page passed to userspace is get_page()'d.
This reference is dropped once the userspace indicates that it is
done reading this page. All pages are released when the socket is
destroyed.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Re
Make skb_frag_page() fail in the case where the frag is not backed
by a page, and fix its relevant callers to handle this case.
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v10:
- Fixed newly generated kdoc warnings found by patchwork. While we're
at it, fix the Return se
hether the underlying type is page or net_iov.
Implement checks for net_iov in netmem helpers which delegate to mm
APIs, to ensure net_iov are never passed to the mm stack.
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Acked-by: Jakub Kicinski
---
v23:
- Fix comment on netmem_is_perf_
ff-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v23:
- WARN_ON when we don't see the dma_addr in the gen_pool (Jakub)
v20:
- Removed dma_addr field in dmabuf_genpool_chunk_owner not used in th
API takes the dma-buf fd as input, and binds it to the netdevice. The
user can specify the rx queues to bind the dma-buf to.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
Reviewed-by: Donald Hunter
Reviewed-by: Jakub Kicinski
---
v16:
- Use subset-of: queue queue-id instead
The netdev_dmabuf_binding struct is refcounted, and releases its
resources only when all the refs are released.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov # excluding netlink
Acked-by: Daniel Vetter
---
v23:
- Create CONFIG_NET_DE
ff-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v18:
- Add more color to commit message (Xuan Zhuo).
v17:
- Use ASSERT_RTNL() (Jakub).
v13:
- Add reviewed-by from Pavel (thanks!)
- Fixed comment (Pavel)
v11:
- Fix not checking dev->queue_mgmt_ops (Pavel
f provider.
** Test Setup
Kernel: net-next with this series and memory provider API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov
Cc: David Wei
Cc: Jason Gunthorpe
Cc: Yunsheng Lin
Cc: Shailend Chan
On Tue, Aug 27, 2024 at 7:15 PM Jakub Kicinski wrote:
>
> On Sun, 25 Aug 2024 04:15:02 +0000 Mina Almasry wrote:
> > +void net_devmem_free_dmabuf(struct net_iov *niov)
> > +{
> > + struct net_devmem_dmabuf_binding *binding = net_iov_binding(niov);
> >
Add an interface for the user to notify the kernel that it is done
reading the devmem dmabuf frags returned as cmsg. The kernel will
drop the reference on the frags to make them available for reuse.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
7;inflight': 1023,
'inflight-mem': 4190208},
And queue stats:
$ ./cli.py --spec ../netlink/specs/netdev.yaml --dump queue-get
...
{'dmabuf': 10, 'id': 8, 'ifindex': 3, 'type': 'rx'},
{'dmabuf': 10, 'id': 9,
, ncdevmem has a validation mode,
where it sends a specific pattern and validates this pattern on the
receiver side to ensure data integrity.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
---
v22:
- Add run_command helper. It reduces boiler plate and prints the
commands it is running
ags of
devmem skbs and avoid coalescing devmem skbs with non devmem skbs.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v16:
- Fix unreadable handling in skb_split_no_header() (Eric).
v11:
- drop excessive checks for
er_frags, and each page passed to userspace is get_page()'d.
This reference is dropped once the userspace indicates that it is
done reading this page. All pages are released when the socket is
destroyed.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Re
Make skb_frag_page() fail in the case where the frag is not backed
by a page, and fix its relevant callers to handle this case.
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v10:
- Fixed newly generated kdoc warnings found by patchwork. While we're
at it, fix the Return se
Add documentation outlining the usage and details of devmem TCP.
Signed-off-by: Mina Almasry
Reviewed-by: Bagas Sanjaya
Reviewed-by: Donald Hunter
---
v16:
- Add documentation on unbinding the NIC from dmabuf (Donald).
- Add note that any dmabuf should work (Donald).
v9:
https
ff-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v20:
- Removed dma_addr field in dmabuf_genpool_chunk_owner not used in this
patch (moved to later patch where it's used).
v19:
- Don't reset dma_addr on allocation/fr
is omitted for simplicity & p.order !=
0.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v21:
- Provide empty definitions of functions moved to page_pool_priv.h, so
that the build still succeeds when CONFIG_PAGE_
hether the underlying type is page or net_iov.
Implement checks for net_iov in netmem helpers which delegate to mm
APIs, to ensure net_iov are never passed to the mm stack.
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v19:
- Move page_pool_set_dma_addr(_netmem) to page_pool_priv.h
- D
API takes the dma-buf fd as input, and binds it to the netdevice. The
user can specify the rx queues to bind the dma-buf to.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
Reviewed-by: Donald Hunter
Reviewed-by: Jakub Kicinski
---
v16:
- Use subset-of: queue queue-id instead
ff-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v18:
- Add more color to commit message (Xuan Zhuo).
v17:
- Use ASSERT_RTNL() (Jakub).
v13:
- Add reviewed-by from Pavel (thanks!)
- Fixed comment (Pavel)
v11:
- Fix not checking dev->queue_mgmt_ops (Pavel
The netdev_dmabuf_binding struct is refcounted, and releases its
resources only when all the refs are released.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov # excluding netlink
Acked-by: Daniel Vetter
---
v22:
- Disable binding xdp t
IC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov
Cc: David Wei
Cc: Jason Gunthorpe
Cc: Yunsheng Lin
Cc: Shailend Chand
Cc: Harshitha Ramamurthy
Cc: Shakeel Butt
Cc: Jeroen de Borst
Cc: Praveen Kaligineedi
Cc: Bagas Sanjaya
Cc: Steven Rostedt
Cc: Christoph Hel
On Wed, Aug 21, 2024 at 5:15 AM Taehee Yoo wrote:
>
> On Tue, Aug 20, 2024 at 1:01 PM Mina Almasry wrote:
> >
> > On Mon, Aug 19, 2024 at 6:53 PM Jakub Kicinski wrote:
> > >
> > > On Mon, 19 Aug 2024 00:44:27 +0900 Taehee Yoo wrote:
> > > > &g
On Tue, Aug 20, 2024 at 11:19 AM Jakub Kicinski wrote:
>
> On Tue, 20 Aug 2024 00:01:02 -0400 Mina Almasry wrote:
> > Took a bit of a look here. Forgive me, I'm not that familiar with XDP
> > and virtual interfaces, so I'm a bit unsure what to do here.
> >
Add an interface for the user to notify the kernel that it is done
reading the devmem dmabuf frags returned as cmsg. The kernel will
drop the reference on the frags to make them available for reuse.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
7;inflight': 1023,
'inflight-mem': 4190208},
And queue stats:
$ ./cli.py --spec ../netlink/specs/netdev.yaml --dump queue-get
...
{'dmabuf': 10, 'id': 8, 'ifindex': 3, 'type': 'rx'},
{'dmabuf': 10, 'id': 9,
, ncdevmem has a validation mode,
where it sends a specific pattern and validates this pattern on the
receiver side to ensure data integrity.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
---
v20:
- Remove unnecessary sleep(1)
- Add test to ensure dmabuf binding fails if header split is
ff-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v20:
- Removed dma_addr field in dmabuf_genpool_chunk_owner not used in this
patch (moved to later patch where it's used).
v19:
- Don't reset dma_addr on allocation/fr
er_frags, and each page passed to userspace is get_page()'d.
This reference is dropped once the userspace indicates that it is
done reading this page. All pages are released when the socket is
destroyed.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Re
Add documentation outlining the usage and details of devmem TCP.
Signed-off-by: Mina Almasry
Reviewed-by: Bagas Sanjaya
Reviewed-by: Donald Hunter
---
v16:
- Add documentation on unbinding the NIC from dmabuf (Donald).
- Add note that any dmabuf should work (Donald).
v9:
https
Make skb_frag_page() fail in the case where the frag is not backed
by a page, and fix its relevant callers to handle this case.
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v10:
- Fixed newly generated kdoc warnings found by patchwork. While we're
at it, fix the Return se
ags of
devmem skbs and avoid coalescing devmem skbs with non devmem skbs.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Eric Dumazet
---
v16:
- Fix unreadable handling in skb_split_no_header() (Eric).
v11:
- drop excessive checks for
is omitted for simplicity & p.order !=
0.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v21:
- Provide empty definitions of functions moved to page_pool_priv.h, so
that the build still succeeds when CONFIG_PAGE_
hether the underlying type is page or net_iov.
Implement checks for net_iov in netmem helpers which delegate to mm
APIs, to ensure net_iov are never passed to the mm stack.
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov
---
v19:
- Move page_pool_set_dma_addr(_netmem) to page_pool_priv.h
- D
The netdev_dmabuf_binding struct is refcounted, and releases its
resources only when all the refs are released.
Signed-off-by: Willem de Bruijn
Signed-off-by: Kaiyuan Zhang
Signed-off-by: Mina Almasry
Reviewed-by: Pavel Begunkov # excluding netlink
Acked-by: Daniel Vetter
---
v21:
- Move definitio
API takes the dma-buf fd as input, and binds it to the netdevice. The
user can specify the rx queues to bind the dma-buf to.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
Reviewed-by: Donald Hunter
Reviewed-by: Jakub Kicinski
---
v16:
- Use subset-of: queue queue-id instead
ff-by: Mina Almasry
Reviewed-by: Pavel Begunkov
Reviewed-by: Jakub Kicinski
---
v18:
- Add more color to commit message (Xuan Zhuo).
v17:
- Use ASSERT_RTNL() (Jakub).
v13:
- Add reviewed-by from Pavel (thanks!)
- Fixed comment (Pavel)
v11:
- Fix not checking dev->queue_mgmt_ops (Pavel
d memory provider API cherry-picked
locally.
Hardware: Google Cloud A3 VMs.
NIC: GVE with header split & RSS & flow steering support.
Cc: Pavel Begunkov
Cc: David Wei
Cc: Jason Gunthorpe
Cc: Yunsheng Lin
Cc: Shailend Chand
Cc: Harshitha Ramamurthy
Cc: Shakeel Butt
Cc: Jeroen de Borst
Cc: P
On Mon, Aug 19, 2024 at 6:53 PM Jakub Kicinski wrote:
>
> On Mon, 19 Aug 2024 00:44:27 +0900 Taehee Yoo wrote:
> > > @@ -9537,6 +9540,10 @@ static int dev_xdp_attach(struct net_device *dev,
> > > struct netlink_ext_ack *extack
> > > NL_SET_ERR_MSG(extack, "Native and gener
On Sun, Aug 18, 2024 at 11:54 PM Mina Almasry wrote:
>
> v20:
> https://patchwork.kernel.org/project/netdevbpf/list/?series=879373&state=*
>
>
> v20 aims to resolve a couple of bug reports against v19, and addresses
> some review comments around the page_p
7;inflight': 1023,
'inflight-mem': 4190208},
And queue stats:
$ ./cli.py --spec ../netlink/specs/netdev.yaml --dump queue-get
...
{'dmabuf': 10, 'id': 8, 'ifindex': 3, 'type': 'rx'},
{'dmabuf': 10, 'id': 9,
, ncdevmem has a validation mode,
where it sends a specific pattern and validates this pattern on the
receiver side to ensure data integrity.
Suggested-by: Stanislav Fomichev
Signed-off-by: Mina Almasry
---
v20:
- Remove unnecessary sleep(1)
- Add test to ensure dmabuf binding fails if header split is
Add documentation outlining the usage and details of devmem TCP.
Signed-off-by: Mina Almasry
Reviewed-by: Bagas Sanjaya
Reviewed-by: Donald Hunter
---
v16:
- Add documentation on unbinding the NIC from dmabuf (Donald).
- Add note that any dmabuf should work (Donald).
v9:
https
1 - 100 of 528 matches
Mail list logo
