On Wed, Mar 10, 2021 at 09:09:29AM +0100, Björn Töpel wrote:
> From: Björn Töpel
>
> The only user of libbpf_util.h is xsk.h. Move the barriers to xsk.h,
> and remove libbpf_util.h. The barriers are used as an implementation
> detail, and should not be considered part of the stable API.
Does tha
On Thu, Mar 04, 2021 at 08:41:45PM +0100, Eric Dumazet wrote:
> On Thu, Mar 4, 2021 at 8:06 PM Jakub Kicinski wrote:
> >
> > On Thu, 4 Mar 2021 13:51:15 +0100 Eric Dumazet wrote:
> > > I think we are over thinking this really (especially if the fix needs
> > > a change in core networking or driver
On mlx5e fails to compile on the latesst net-next:
CC drivers/net/ethernet/mellanox/mlx5/core/en_main.o
In file included from ../drivers/net/ethernet/mellanox/mlx5/core/en_tc.h:40,
from ../drivers/net/ethernet/mellanox/mlx5/core/en_main.c:45:
../drivers/net/ethernet/mellano
On Mon, Jan 11, 2021 at 06:28:21PM +, Alexander Lobakin wrote:
> skb_cache and skb_count fields are used to store skbuff_heads queued
> for freeing to flush them by bulks, and aren't related to allocation
> path. Give them more obvious names to improve code understanding and
> allow to expand t
.
Patch 12: Replace open-coded assignments with skb_zcopy_init()
Patch 13: Rename skb_zcopy_{get|put} to net_zcopy_{get|put}
Jonathan Lemon (13):
skbuff: remove unused skb_zcopy_abort function
skbuff: simplify sock_zerocopy_put
skbuff: Push status and refcounts into sock_zerocopy_callback
The sock_zerocopy_put_abort function contains logic which is
specific to the current zerocopy implementation. Add a wrapper
which checks the callback and dispatches apppropriately.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 10 ++
net/core/skbuff.c | 12
RX zerocopy fragment pages which are not allocated from the
system page pool require special handling. Give the callback
in skb_zcopy_clear() a chance to process them first.
Signed-off-by: Jonathan Lemon
---
net/core/skbuff.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git
Replace sock_zerocopy_put with the generic skb_zcopy_put()
function. Pass 'true' as the success argument, as this
is identical to no change.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 7 ++-
net/core/skbuff.c | 9 +
net/ipv4/tcp.c | 2 +
Replace direct assignments with skb_zcopy_init() for zerocopy
cases where a new skb is initialized, without changing the
reference counts.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/vhost/net.c| 1 +
include/linux/skbuff.h | 9
All 'struct ubuf_info' users should have a callback defined
as of commit 0a4a060bb204 ("sock: fix zerocopy_success regression
with msg_zerocopy").
Remove the dead code path to consume_skb(), which makes
assumptions about how the structure was allocated.
Signed-off-by: Jona
At Willem's suggestion, rename the sock_zerocopy_* functions
so that they match the MSG_ZEROCOPY flag, which makes it clear
they are specific to this zerocopy implementation.
Signed-off-by: Jonathan Lemon
Acked-by: Willem de Bruijn
---
include/linux/skbuff.h | 16
net
In preparation for expanded zerocopy (TX and RX), move
the zerocopy related bits out of tx_flags into their own
flag word.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/net/xen-netback/interface.c | 4
Add an optional skb parameter to the zerocopy callback parameter,
which is passed down from skb_zcopy_clear(). This gives access
to the original skb, which is needed for upcoming RX zero-copy
error handling.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 2 +-
drivers
Unlike the rest of the skb_zcopy_ functions, these routines
operate on a 'struct ubuf', not a skb. Remove the 'skb_'
prefix from the naming to make things clearer.
Suggested-by: Willem de Bruijn
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 10 +
skb_zcopy_abort() has no in-tree consumers, remove it.
Signed-off-by: Jonathan Lemon
Acked-by: Willem de Bruijn
---
include/linux/skbuff.h | 11 ---
1 file changed, 11 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 333bcdc39635..3ca8d7c7b30c 100644
Currently, when an ubuf is attached to a new skb, the shared
flags word is initialized to a fixed value. Instead of doing
this, set the default flags in the ubuf, and have new skbs
inherit from this default.
This is needed when setting up different zerocopy types.
Signed-off-by: Jonathan Lemon
,
which saves the status and handles its own refcounts.
This makes the behavior of the sock_zerocopy_callback identical
to the tpacket and vhost callbacks.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 3 ---
net/core/skbuff.c | 14 +++---
2 files changed, 11 insertions
Rename the get routines for consistency.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 12 ++--
net/core/skbuff.c | 2 +-
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index a6c86839035b..5b8a53ab51fd
On Tue, Jan 05, 2021 at 03:45:55PM -0800, Florian Fainelli wrote:
> On 1/5/21 3:40 PM, Jonathan Lemon wrote:
> > On Tue, Jan 05, 2021 at 03:11:03PM -0800, Florian Fainelli wrote:
> >> On 1/5/21 2:06 PM, Jonathan Lemon wrote:
> >>> From: Jonathan Lemon
> >>
On Tue, Jan 05, 2021 at 03:11:03PM -0800, Florian Fainelli wrote:
> On 1/5/21 2:06 PM, Jonathan Lemon wrote:
> > From: Jonathan Lemon
> >
> > This is set of cleanup patches for zerocopy which are intended
> > to allow a introduction of a different zerocopy implementat
From: Jonathan Lemon
Rename the get routines for consistency.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 12 ++--
net/core/skbuff.c | 2 +-
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index
From: Jonathan Lemon
Before this change, the caller of sock_zerocopy_callback would
need to save the zerocopy status, decrement and check the refcount,
and then call the callback function - the callback was only invoked
when the refcount reached zero.
Now, the caller just passes the status into
From: Jonathan Lemon
RX zerocopy fragment pages which are not allocated from the
system page pool require special handling. Give the callback
in skb_zcopy_clear() a chance to process them first.
Signed-off-by: Jonathan Lemon
---
net/core/skbuff.c | 3 ++-
1 file changed, 2 insertions(+), 1
From: Jonathan Lemon
Unlike the rest of the skb_zcopy_ functions, these routines
operate on a 'struct ubuf', not a skb. Remove the 'skb_'
prefix from the naming to make things clearer.
Suggested-by: Willem de Bruijn
Signed-off-by: Jonathan Lemon
---
include
From: Jonathan Lemon
This is set of cleanup patches for zerocopy which are intended
to allow a introduction of a different zerocopy implementation.
The top level API will use the skb_zcopy_*() functions, while
the current TCP specific zerocopy ends up using msg_zerocopy_*()
calls.
There should
From: Jonathan Lemon
skb_zcopy_abort() has no in-tree consumers, remove it.
Signed-off-by: Jonathan Lemon
Acked-by: Willem de Bruijn
---
include/linux/skbuff.h | 11 ---
1 file changed, 11 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 333bcdc39635
From: Jonathan Lemon
In preparation for expanded zerocopy (TX and RX), move
the zerocopy related bits out of tx_flags into their own
flag word.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/net/xen-netback
From: Jonathan Lemon
The sock_zerocopy_put_abort function contains logic which is
specific to the current zerocopy implementation. Add a wrapper
which checks the callback and dispatches apppropriately.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 10 ++
net/core
From: Jonathan Lemon
Currently, when an ubuf is attached to a new skb, the shared
flags word is initialized to a fixed value. Instead of doing
this, set the default flags in the ubuf, and have new skbs
inherit from this default.
This is needed when setting up different zerocopy types.
Signed
From: Jonathan Lemon
Replace direct assignments with skb_zcopy_init() for zerocopy
cases where a new skb is initialized, without changing the
reference counts.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/vhost/net.c| 1
From: Jonathan Lemon
Replace sock_zerocopy_put with the generic skb_zcopy_put()
function. Pass 'true' as the success argument, as this
is identical to no change.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 7 ++-
net/core/skbuff.c | 9 +
net/
From: Jonathan Lemon
At Willem's suggestion, rename the sock_zerocopy_* functions
so that they match the MSG_ZEROCOPY flag, which makes it clear
they are specific to this zerocopy implementation.
Signed-off-by: Jonathan Lemon
Acked-by: Willem de Bruijn
---
include/linux/skbuff.h
From: Jonathan Lemon
All 'struct ubuf_info' users should have a callback defined
as of commit 0a4a060bb204 ("sock: fix zerocopy_success regression
with msg_zerocopy").
Remove the dead code path to consume_skb(), which makes
assumptions about how the structure was alloc
From: Jonathan Lemon
Add an optional skb parameter to the zerocopy callback parameter,
which is passed down from skb_zcopy_clear(). This gives access
to the original skb, which is needed for upcoming RX zero-copy
error handling.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c
On Mon, Jan 04, 2021 at 11:22:35PM -0500, Willem de Bruijn wrote:
> On Mon, Jan 4, 2021 at 11:17 PM Jonathan Lemon
> wrote:
> >
> > On Mon, Jan 04, 2021 at 12:39:35PM -0500, Willem de Bruijn wrote:
> > > On Wed, Dec 30, 2020 at 2:12 PM Jonathan Lemon
> > > w
On Mon, Jan 04, 2021 at 12:39:35PM -0500, Willem de Bruijn wrote:
> On Wed, Dec 30, 2020 at 2:12 PM Jonathan Lemon
> wrote:
> >
> > From: Jonathan Lemon
> >
> > This is set of cleanup patches for zerocopy which are intended
> > to allow a introduction of
From: Jonathan Lemon
This is set of cleanup patches for zerocopy which are intended
to allow a introduction of a different zerocopy implementation.
The top level API will use the skb_zcopy_*() functions, while
the current TCP specific zerocopy ends up using msg_zerocopy_*()
calls.
There should
From: Jonathan Lemon
Replace direct assignments with skb_zcopy_init() for zerocopy
cases where a new skb is initialized, without changing the
reference counts.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/vhost/net.c| 1
From: Jonathan Lemon
Currently, when an ubuf is attached to a new skb, the shared
flags word is initialized to a fixed value. Instead of doing
this, set the default flags in the ubuf, and have new skbs
inherit from this default.
This is needed when setting up different zerocopy types.
Signed
From: Jonathan Lemon
RX zerocopy fragment pages which are not allocated from the
system page pool require special handling. Give the callback
in skb_zcopy_clear() a chance to process them first.
Signed-off-by: Jonathan Lemon
---
net/core/skbuff.c | 3 ++-
1 file changed, 2 insertions(+), 1
From: Jonathan Lemon
All 'struct ubuf_info' users should have a callback defined
as of commit 0a4a060bb204 ("sock: fix zerocopy_success regression
with msg_zerocopy").
Remove the dead code path to consume_skb(), which makes
assumptions about how the structure was alloc
From: Jonathan Lemon
Replace sock_zerocopy_put with the generic skb_zcopy_put()
function. Pass 'true' as the success argument, as this
is identical to no change.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 7 ++-
net/core/skbuff.c | 9 +
net/
From: Jonathan Lemon
At Willem's suggestion, rename the sock_zerocopy_* functions
so that they match the MSG_ZEROCOPY flag, which makes it clear
they are specific to this zerocopy implementation.
Signed-off-by: Jonathan Lemon
Acked-by: Willem de Bruijn
---
include/linux/skbuff.h
From: Jonathan Lemon
The sock_zerocopy_put_abort function contains logic which is
specific to the current zerocopy implementation. Add a wrapper
which checks the callback and dispatches apppropriately.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 10 ++
net/core
From: Jonathan Lemon
Before this change, the caller of sock_zerocopy_callback would
need to save the zerocopy status, decrement and check the refcount,
and then call the callback function - the callback was only invoked
when the refcount reached zero.
Now, the caller just passes the status into
From: Jonathan Lemon
skb_zcopy_abort() has no in-tree consumers, remove it.
Signed-off-by: Jonathan Lemon
Acked-by: Willem de Bruijn
---
include/linux/skbuff.h | 11 ---
1 file changed, 11 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 333bcdc39635
From: Jonathan Lemon
Rename the get routines for consistency.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 12 ++--
net/core/skbuff.c | 2 +-
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index
From: Jonathan Lemon
Add an optional skb parameter to the zerocopy callback parameter,
which is passed down from skb_zcopy_clear(). This gives access
to the original skb, which is needed for upcoming RX zero-copy
error handling.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c
From: Jonathan Lemon
In preparation for expanded zerocopy (TX and RX), move
the zerocopy related bits out of tx_flags into their own
flag word.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/net/xen-netback
On Tue, Dec 22, 2020 at 05:26:15PM -0500, Willem de Bruijn wrote:
> On Tue, Dec 22, 2020 at 12:22 PM Jonathan Lemon
> wrote:
> >
> > On Tue, Dec 22, 2020 at 09:43:15AM -0500, Willem de Bruijn wrote:
> > > On Mon, Dec 21, 2020 at 7:09 PM Jonathan Lemon
> > > w
On Tue, Dec 22, 2020 at 10:00:37AM -0500, Willem de Bruijn wrote:
> On Mon, Dec 21, 2020 at 7:09 PM Jonathan Lemon
> wrote:
> >
> > From: Jonathan Lemon
> >
> > Currently, an ubuf is attached to a new skb, the skb zc_flags
> > is initialized to a fixed value.
On Tue, Dec 22, 2020 at 09:43:39AM -0500, Willem de Bruijn wrote:
> On Mon, Dec 21, 2020 at 7:09 PM Jonathan Lemon
> wrote:
> >
> > From: Jonathan Lemon
> >
> > Before this change, the caller of sock_zerocopy_callback would
> > need to save the zerocopy stat
On Tue, Dec 22, 2020 at 09:43:15AM -0500, Willem de Bruijn wrote:
> On Mon, Dec 21, 2020 at 7:09 PM Jonathan Lemon
> wrote:
> >
> > From: Jonathan Lemon
> >
> > In preparation for expanded zerocopy (TX and RX), move
> > the ZC related bits out of t
On Tue, Dec 22, 2020 at 09:42:40AM -0500, Willem de Bruijn wrote:
> On Mon, Dec 21, 2020 at 7:09 PM Jonathan Lemon
> wrote:
> >
> > From: Jonathan Lemon
> >
> > Replace sock_zerocopy_put with the generic skb_zcopy_put()
> > function. Pass 'true
From: Jonathan Lemon
Before this change, the caller of sock_zerocopy_callback would
need to save the zerocopy status, decrement and check the refcount,
and then call the callback function - the callback was only invoked
when the refcount reached zero.
Now, the caller just passes the status into
From: Jonathan Lemon
At Willem's suggestion, rename the sock_zerocopy_* functions
so that they match the MSG_ZEROCOPY flag, which makes it clear
they are specific to this zerocopy implementation.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 16
net/core/skb
From: Jonathan Lemon
Replace direct assignments with skb_zcopy_set() for clarity.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/vhost/net.c | 1 +
3 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/net/tap.c b/drivers
From: Jonathan Lemon
This is set of cleanup patches for zerocopy which are intended
to allow a introduction of a different zerocopy implementation.
The top level API will use the skb_zcopy_*() functions, while
the current TCP specific zerocopy ends up using msg_zerocopy_*()
calls.
There should
From: Jonathan Lemon
Add an optional skb parameter to the zerocopy callback parameter,
which is passed down from skb_zcopy_clear(). This gives access
to the original skb, which is needed for upcoming RX zero-copy
error handling.
Make the name of the the zerocopy success pararmeter consistent
From: Jonathan Lemon
The sock_zerocopy_put_abort function contains logic which is
specific to the current zerocopy implementation. Add a wrapper
which checks the callback and dispatches apppropriately.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 10 ++
net/core
From: Jonathan Lemon
RX zerocopy fragment pages which are not allocated from the
system page pool require special handling. Give the callback
in skb_zcopy_clear() a chance to process them first.
Signed-off-by: Jonathan Lemon
---
net/core/skbuff.c | 3 ++-
1 file changed, 2 insertions(+), 1
From: Jonathan Lemon
Currently, an ubuf is attached to a new skb, the skb zc_flags
is initialized to a fixed value. Instead of doing this, set
the default zc_flags in the ubuf, and have new skb's inherit
from this default.
This is needed when setting up different zerocopy types.
Signed-o
From: Jonathan Lemon
Rename the get routines for consistency.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 12 ++--
net/core/skbuff.c | 2 +-
2 files changed, 7 insertions(+), 7 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index
From: Jonathan Lemon
Replace sock_zerocopy_put with the generic skb_zcopy_put()
function. Pass 'true' as the success argument, as this
is identical to no change.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 7 ++-
net/core/skbuff.c | 9 +
net/
From: Jonathan Lemon
In preparation for expanded zerocopy (TX and RX), move
the ZC related bits out of tx_flags into their own flag
word.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/net/xen-netback
From: Jonathan Lemon
All 'struct ubuf_info' users should have a callback defined.
Remove the dead code path to consume_skb(), which makes
unwarranted assumptions about how the structure was allocated.
Signed-off-by: Jonathan Lemon
---
net/core/skbuff.c | 8 ++--
1 file
From: Jonathan Lemon
skb_zcopy_abort() has no in-tree consumers, remove it.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 11 ---
1 file changed, 11 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 69588b304f83..fb6dd6af0f82 100644
--- a
On Mon, Dec 21, 2020 at 05:52:08PM -0500, Willem de Bruijn wrote:
> > > > - marking the skb data as inaccessible so skb_condense()
> > > > and skb_zeroocopy_clone() leave it alone.
> > >
> > > Yep. Skipping content access on the Rx path will be interesting. I
> > > wonder if that should be a
On Sat, Dec 19, 2020 at 02:00:55PM -0500, Willem de Bruijn wrote:
> On Fri, Dec 18, 2020 at 4:27 PM Jonathan Lemon
> wrote:
> >
> > On Fri, Dec 18, 2020 at 03:49:44PM -0500, Willem de Bruijn wrote:
> > > On Fri, Dec 18, 2020 at 3:23 PM Jonathan Lemon
> > > w
On Sat, Dec 19, 2020 at 01:46:13PM -0500, Willem de Bruijn wrote:
> On Fri, Dec 18, 2020 at 3:20 PM Jonathan Lemon
> wrote:
> >
> > From: Jonathan Lemon
> >
> > In preparation for further work, the zcopy* routines will
> > become basic building blocks,
On Fri, Dec 18, 2020 at 03:49:44PM -0500, Willem de Bruijn wrote:
> On Fri, Dec 18, 2020 at 3:23 PM Jonathan Lemon
> wrote:
> >
> > From: Jonathan Lemon
> >
> > This is set of cleanup patches for zerocopy which are intended
> > to allow a introduction of
From: Jonathan Lemon
Replace direct assignments with skb_zcopy_set() for clarity.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/vhost/net.c | 1 +
3 files changed, 3 insertions(+), 4 deletions(-)
diff --git a/drivers/net/tap.c b/drivers
From: Jonathan Lemon
The sock_zerocopy_put_abort function contains logic which is
specific to the current zerocopy implementation. Add a wrapper
which checks the callback and dispatches apppropriately.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 10 ++
net/core
From: Jonathan Lemon
In preparation for expanded zerocopy (TX and RX), move
the ZC related bits out of tx_flags into their own flag
word.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c | 3 +--
drivers/net/tun.c | 3 +--
drivers/net/xen-netback
From: Jonathan Lemon
Add an optional skb parameter to the zerocopy callback parameter,
which is passed down from skb_zcopy_clear(). This gives access
to the original skb, which is needed for upcoming RX zero-copy
error handling.
Signed-off-by: Jonathan Lemon
---
drivers/net/tap.c
From: Jonathan Lemon
Rename the get routines for consistency.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 4 ++--
net/core/skbuff.c | 2 +-
2 files changed, 3 insertions(+), 3 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index df98d61e8c51
From: Jonathan Lemon
RX zerocopy fragment pages which are not allocated from the
system page pool require special handling. Give the callback
in skb_zcopy_clear() a chance to process them first.
Signed-off-by: Jonathan Lemon
---
net/core/skbuff.c | 3 ++-
1 file changed, 2 insertions(+), 1
From: Jonathan Lemon
This is set of cleanup patches for zerocopy which are intended
to allow a introduction of a different zerocopy implementation.
The top level api will use the skb_zcopy_*() functions, while
the current TCP specific zerocopy would use the sock_zerocopy_*()
calls.
There
From: Jonathan Lemon
skb_zcopy_abort() has no in-tree consumers, remove it.
Signed-off-by: Jonathan Lemon
---
include/linux/skbuff.h | 11 ---
1 file changed, 11 deletions(-)
diff --git a/include/linux/skbuff.h b/include/linux/skbuff.h
index 69588b304f83..fb6dd6af0f82 100644
--- a
From: Jonathan Lemon
In preparation for further work, the zcopy* routines will
become basic building blocks, while the zerocopy* ones will
be specific for the existing zerocopy implementation.
All uargs should have a callback function, (unless nouarg
is set), so push all special case logic
From: Jonathan Lemon
Currently, an ubuf is attached to a new skb, the skb zc_flags
is initialized to a fixed value. Instead of doing this, set
the default zc_flags in the ubuf, and have new skb's inherit
from this default.
This is needed when setting up different zerocopy types.
Signed-o
From: Jonathan Lemon
Instead of directly comparing task->tgid and task->pid, use the
thread_group_leader() helper. This helps with readability, and
there should be no functional change.
Signed-off-by: Jonathan Lemon
---
kernel/bpf/task_iter.c | 2 +-
1 file changed, 1 insertion
From: Jonathan Lemon
Only obtain the task reference count at the end of the RCU section
instead of repeatedly obtaining/releasing it when iterating though
a thread group.
Jump to the correct branch when it is known that the task is NULL.
Signed-off-by: Jonathan Lemon
---
kernel/bpf
From: Jonathan Lemon
On some systems, some variant of the following splat is
repeatedly seen. The common factor in all traces seems
to be the entry point to task_file_seq_next(). With the
patch, all warnings go away.
rcu: INFO: rcu_sched self-detected stall on CPU
rcu: \x0926
From: Jonathan Lemon
v3->v4:
Split commit into separate patches.
v2->v3:
Add splat to commitlog descriptions
v1->v2
Use Fixes: shas from correct tree
Jonathan Lemon (3):
bpf: save correct stopping point in file seq iteration.
bpf: Use thread_group_leader()
bpf: opti
On Fri, Dec 18, 2020 at 08:53:22AM -0800, Yonghong Song wrote:
>
>
> On 12/11/20 9:11 AM, Jonathan Lemon wrote:
> > From: Jonathan Lemon
> >
> > On some systems, some variant of the following splat is
> > repeatedly seen. The common factor in all traces s
On Fri, Dec 11, 2020 at 12:59:52PM -0700, David Ahern wrote:
> On 12/11/20 11:45 AM, Jakub Kicinski wrote:
> > Ack, these patches are not exciting (to me), so I'm wondering if there
> > is a better way. The only reason NIC would have to understand a ULP for
> > ZC is to parse out header/message len
On Fri, Dec 11, 2020 at 12:23:34PM -0800, Andrii Nakryiko wrote:
> > @@ -164,7 +164,7 @@ task_file_seq_get_next(struct
> > bpf_iter_seq_task_file_info *info)
> > curr_files = get_files_struct(curr_task);
> > if (!curr_files) {
> > put_task_st
From: Jonathan Lemon
Reposting with one of the many splats seen.
v2->v3:
Add splat to commitlog descriptions
v1->v2
Use Fixes: shas from correct tree
Jonathan Lemon (1):
bpf: increment and use correct thread iterator
kernel/bpf/task_iter.c | 8
1 file changed, 4 inse
From: Jonathan Lemon
On some systems, some variant of the following splat is
repeatedly seen. The common factor in all traces seems
to be the entry point to task_file_seq_next(). With the
patch, all warnings go away.
rcu: INFO: rcu_sched self-detected stall on CPU
rcu: \x0926
On Wed, Dec 09, 2020 at 11:02:54AM -0800, Yonghong Song wrote:
>
>
> Maybe you can post v3 of the patch with the above information in the
> commit description so people can better understand what the problem
> you are trying to solve here?
>
> Also, could you also send to b...@vger.kernel.org?
On Fri, Dec 04, 2020 at 12:01:53AM -0800, Yonghong Song wrote:
>
>
> On 12/3/20 7:43 PM, Jonathan Lemon wrote:
> > From: Jonathan Lemon
>
> Could you explain in the commit log what problem this patch
> tries to solve? What bad things could happen without this patch?
The OpenCompute time card is an atomic clock along with
a GPS receiver that provides a Grandmaster clock source
for a PTP enabled network.
More information is available at http://www.timingcard.com/
Signed-off-by: Jonathan Lemon
---
drivers/ptp/Kconfig | 14 ++
drivers/ptp/Makefile | 1
From: Jonathan Lemon
v2->v3:
remove dev_info() informational lines
v1->v2:
make the driver dependent on CONFIG_PCI, for the test robot.
move the config option under PTP_1588_CLOCK hierarcy
v1:
initial submission
Jonathan Lemon (1):
ptp: Add clock driver for the OpenCompute Ti
From: Jonathan Lemon
If unable to obtain the file structure for the current task,
proceed to the next task number after the one returned from
task_seq_get_next(), instead of the next task number from the
original iterator.
Use thread_group_leader() instead of comparing tgid vs pid, which
might
On Thu, Dec 03, 2020 at 04:56:24PM -0800, Richard Cochran wrote:
> On Thu, Dec 03, 2020 at 10:29:25AM -0800, Jonathan Lemon wrote:
> > The OpenCompute time card is an atomic clock along with
> > a GPS receiver that provides a Grandmaster clock source
> > for a PTP enabled n
From: Jonathan Lemon
If unable to obtain the file structure for the current task,
proceed to the next task number after the one returned from
task_seq_get_next(), instead of the next task number from the
original iterator.
Use thread_group_leader() instead of comparing tgid vs pid, which
might
The OpenCompute time card is an atomic clock along with
a GPS receiver that provides a Grandmaster clock source
for a PTP enabled network.
More information is available at http://www.timingcard.com/
Signed-off-by: Jonathan Lemon
---
drivers/ptp/Kconfig | 14 ++
drivers/ptp/Makefile | 1
The OpenCompute time card is an atomic clock along with
a GPS receiver that provides a Grandmaster clock source
for a PTP enabled network.
More information is available at http://www.timingcard.com/
Signed-off-by: Jonathan Lemon
---
drivers/ptp/Kconfig | 13 ++
drivers/ptp/Makefile | 1
On Wed, Nov 11, 2020 at 12:20:22AM +, Victor Stewart wrote:
> On Wed, Nov 11, 2020 at 12:09 AM Jonathan Lemon
> wrote:
> >
> > On Sun, Nov 08, 2020 at 05:04:41PM +, Victor Stewart wrote:
> > > hi all,
> > >
> > > i'm seeking input
1 - 100 of 387 matches
Mail list logo
