From: Przemek Kitszel
Date: Tue, 9 Apr 2024 12:58:33 +0200
> On 4/8/24 11:09, Alexander Lobakin wrote:
>> From: Przemek Kitszel
>> Date: Fri, 5 Apr 2024 12:32:55 +0200
>>
>>> On 4/4/24 17:44, Alexander Lobakin wrote:
>>>> Add a couple intuitive helper
From: Przemek Kitszel
Date: Wed, 10 Apr 2024 15:01:32 +0200
> On 4/10/24 13:49, Alexander Lobakin wrote:
>> From: Przemek Kitszel
>> Date: Tue, 9 Apr 2024 12:58:33 +0200
>>
>>> On 4/8/24 11:09, Alexander Lobakin wrote:
>>>> From: Przemek Kitszel
From: Kees Cook
Date: Tue, 9 Apr 2024 09:17:53 -0700
> On Mon, Apr 08, 2024 at 11:45:32AM +0200, Alexander Lobakin wrote:
>> From: Alexander Lobakin
>> Date: Mon, 8 Apr 2024 11:11:12 +0200
>>
>>> From: Jakub Kicinski
>>> Date: Fri, 5 Apr 2024 21:25:13 -
From: Jakub Kicinski
Date: Wed, 10 Apr 2024 17:54:24 -0700
> On Wed, 10 Apr 2024 15:36:13 +0200 Alexander Lobakin wrote:
>> Which tree this should go through? Should I include this patch to this
>> series with libeth or it's better to push this through kees/linux and
>
27;s never gonna happen, right? But you still
can at least try.
PP conversion for iavf lands within the same series as these two are tied
closely. libie will support Page Pool model only, so that a driver can't
use much of the lib until it's converted. iavf is only the example, the
rest will e
ule,
libeth. This is done to easily distinguish pretty generic code ready
for reusing by any other vendor and/or for moving the layer up from
the code useful in Intel's 1-100G drivers only.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/Kconfig| 7 +
drivers/net
exander Duyck
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/iavf/iavf.h| 2 +-
drivers/net/ethernet/intel/iavf/iavf_txrx.h | 27 +---
.../net/ethernet/intel/iavf/iavf_ethtool.c| 140 --
drivers/net/ethernet/intel/iavf/
_pg_*() definitions are left to reduce diffstat.
They will be removed with the conversion to Page Pool.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/iavf/iavf_txrx.h | 65
drivers/net/ethernet/intel/iavf/iavf_type.h | 2 -
drivers/net/ethernet/intel/iavf/iavf
Add NUMA-aware counterparts for kvmalloc_array() and kvcalloc() to be
able to flexibly allocate arrays for a particular node.
Rewrite kvmalloc_array() to kvmalloc_array_node(NUMA_NO_NODE) call.
Acked-by: Vlastimil Babka
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
include
sizes.
Reviewed-by: Ilias Apalodimas
Signed-off-by: Alexander Lobakin
---
include/net/page_pool/types.h | 4 ++--
include/net/page_pool/helpers.h | 10 +-
net/core/page_pool.c| 10 +-
3 files changed, 12 insertions(+), 12 deletions(-)
diff --git a/include/net/page_pool/ty
synchronization for the
size passed from the driver. It can be used even when the pool doesn't
manage DMA-syncs-for-device, just make sure the page has a correct DMA
address set via page_pool_set_dma_addr().
Signed-off-by: Alexander Lobakin
---
include/net/page_pool/helpers.h | 24 ++
ather a shim, but eventually the libeth-enabled drivers will move
to it, with iavf being the first one.
Signed-off-by: Alexander Lobakin
---
kernel-doc for libeth_fq::fp generates an "Excess struct member" warning
currently, here's patch which fixes the script: [0]
[0]
h
out of &iavf_tx_queue_stats -- it doesn't belong there (used for Tx
stall detection). Place it next to the stats on the ring itself to fill
the 4-byte slot.
The result: no holes and all the hot fields fit into the first 64-byte
cacheline.
Signed-off-by: Alexander Lobakin
---
drivers
to allocating a page
every second descriptor.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/iavf/iavf_txrx.h | 34 +--
include/linux/net/intel/libie/rx.h| 17 ++
drivers/net/ethernet/intel/iavf/iavf_main.c | 7 +-
drivers/net/ethernet/intel/iavf/iavf_txrx.c
Add myself as a maintainer/supporter for libeth and libie. Let they have
separate entries from the Intel ethernet code as it's a bit different
case and all patches will go through me rather than Tony.
Signed-off-by: Alexander Lobakin
---
MAINTAINERS | 20
1 file change
: Michal Kubiak
Reviewed-by: Wojciech Drewek
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 1 +
drivers/net/ethernet/intel/idpf/idpf_lib.c | 1 +
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 12 +++-
3 files changed, 9 insertions(+), 5 deletions(-
From: Andy Shevchenko
Date: Wed, 8 May 2024 16:23:15 +0300
> *-objs suffix is reserved rather for (user-space) host programs while
> usually *-y suffix is used for kernel drivers (although *-objs works
> for that purpose for now).
>
> Let's correct the old usages of *-objs in Makefiles.
Wait, I
From: Andy Shevchenko
Date: Wed, 8 May 2024 17:25:31 +0300
> On Wed, May 08, 2024 at 03:35:26PM +0200, Alexander Lobakin wrote:
>>> *-objs suffix is reserved rather for (user-space) host programs while
>>> usually *-y suffix is used for kernel drivers (although *-objs works
From: Alexander Lobakin
Date: Wed, 8 May 2024 16:39:21 +0200
> From: Andy Shevchenko
> Date: Wed, 8 May 2024 17:25:31 +0300
>
>> On Wed, May 08, 2024 at 03:35:26PM +0200, Alexander Lobakin wrote:
>>>> *-objs suffix is reserved rather for (user-space) host programs wh
uses
napi_build_skb() instead of in-place short skb allocation.
Most idpf patches, except for the queue split, removes more lines
than adds.
Expect far better memory utilization and +5-8% on Rx depending on
the case (+17% on skb XDP_DROP :>).
Alexander Lobakin (12):
libeth: add cacheline / str
they don't cross the specified bound.
Signed-off-by: Alexander Lobakin
---
include/net/libeth/cache.h | 64 ++
1 file changed, 64 insertions(+)
create mode 100644 include/net/libeth/cache.h
diff --git a/include/net/libeth/cache.h b/include/net/libeth/ca
In C, we have structures and unions.
Casting `void *` via macros is not only error-prone, but also looks
confusing and awful in general.
In preparation for splitting the queue structs, replace it with a
union and direct array dereferences.
Signed-off-by: Alexander Lobakin
---
drivers/net
masks and put &idpf_q_vector on a good diet.
Also remove redundant pointer to the interrupt name from the structure.
request_irq() saves it and free_irq() returns it on deinit, so that you
can free the memory.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_tx
this gives up to 2% of perf bump on Rx.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 370 +++-
1 file changed, 205 insertions(+), 165 deletions(-)
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.h
b/drivers/net/ethernet/intel
ng is the
same, just call either idpf_tx_splitq_frame() or idpf_tx_singleq_frame()
depending on the active model to actually map and send the skb.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 7 ++---
drivers/net/ethernet/intel/idpf/idpf_lib.
's enabled manually. This corresponds to
-10 Kb of object code size and a good bunch of hotpath checks.
idpf_is_queue_model_split() works as a gate and compiles out to `true`
when the config option is disabled.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/Kconfig
TATS unconditionally,
while the latter is often turned off for better performance.
Remove all the references to PP stats from the Ethtool code. The stats
are still available in their full via the generic Netlink interface.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/Kconfig
ed in
&idpf_rx_queue.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/Kconfig | 1 +
drivers/net/ethernet/intel/idpf/idpf.h| 2 +-
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 88 +---
drivers/net/ethernet/intel/idpf/idpf_lib.c| 3 +
accordingly. All the HW-related details like buffer alignment are still
accounted. For the header buffers, pick 256 bytes as in most places in
the kernel (have you ever seen frames with bigger headers?).
Signed-off-by: Alexander Lobakin
---
include/net/libeth/rx.h| 19
more from the data buffer later. This W/A will hopefully
be removed one day.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 54 ++--
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 1 +
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 237 +++--
nly after 4 HW writes rather than two.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/Kconfig | 1 -
drivers/net/ethernet/intel/idpf/idpf.h| 1 -
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 88 ++-
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 25 +-
ame as Rx). This HT
is used only for OOO completions, which aren't really hotpath anyway.
Note that this change must be done atomically, otherwise it's really
easy to get lost and miss something.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf.h| 3 +-
ported-by: Xu Du
> Closes: https://issues.redhat.com/browse/RHEL-36182
> Signed-off-by: Michal Schmidt
Reviewed-by: Alexander Lobakin
> ---
> drivers/net/ethernet/intel/idpf/idpf_ethtool.c | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/dr
From: Tantilov, Emil S
Date: Mon, 13 May 2024 19:01:30 -0700
> On 5/10/2024 8:26 AM, Alexander Lobakin wrote:
>> Currently, there's no HW supporting idpf in the singleq model. Still,
>> this dead code is supported by the driver and often times add hotpath
> The driver su
From: Mina Almasry
Date: Fri, 10 May 2024 09:22:05 -0700
> On Fri, May 10, 2024 at 8:30 AM Alexander Lobakin
> wrote:
>>
>> idpf's in-kernel parsed ptype structure is almost identical to the one
>> used in the previous Intel drivers, which means it can be
nt DMA uses on hotpath, uses
napi_build_skb() instead of in-place short skb allocation.
Most idpf patches, except for the queue split, removes more lines
than adds.
Expect far better memory utilization and +5-8% on Rx depending on
the case (+17% on skb XDP_DROP :>).
Alexander Lobakin (12):
libe
they don't cross the specified bound.
Signed-off-by: Alexander Lobakin
---
scripts/kernel-doc | 1 +
include/net/libeth/cache.h | 100 +
2 files changed, 101 insertions(+)
create mode 100644 include/net/libeth/cache.h
diff --git a/scripts/kerne
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf.h| 1 -
.../net/ethernet/intel/idpf/idpf_lan_txrx.h | 2 +
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 47 ++-
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 20
drivers/net/ethernet
ng is the
same, just call either idpf_tx_splitq_frame() or idpf_tx_singleq_frame()
depending on the active model to actually map and send the skb.
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 9 ++
drivers/net/ethernet/i
masks and put &idpf_q_vector on a good diet.
Also remove redundant pointer to the interrupt name from the structure.
request_irq() saves it and free_irq() returns it on deinit, so that you
can free the memory.
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net
this gives up to 2% of perf bump on Rx.
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 443 +++-
1 file changed, 250 insertions(+), 193 deletions(-)
diff --git a/drivers/net/ethernet/intel/idpf/idpf_txrx.h
b/dr
in
&idpf_rx_queue.
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/Kconfig | 1 +
drivers/net/ethernet/intel/idpf/idpf.h| 2 +-
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 88 +---
drivers/net/ethernet/intel/id
y: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/Kconfig| 13 +
drivers/net/ethernet/intel/idpf/Kconfig | 27 +++
drivers/net/ethernet/intel/idpf/Makefile | 3 ++-
drivers/net/ethernet/intel/idpf/idpf.h| 3 ++-
d
accordingly. All the HW-related details like buffer alignment are still
accounted. For the header buffers, pick 256 bytes as in most places in
the kernel (have you ever seen frames with bigger headers?).
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
include/net/libeth/rx.h
TATS unconditionally,
while the latter is often turned off for better performance.
Remove all the references to PP stats from the Ethtool code. The stats
are still available in their full via the generic Netlink interface.
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net/ethe
more from the data buffer later. This W/A will hopefully
be removed one day.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 52 ++--
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 1 +
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 253 -
nly after 4 HW writes rather than two.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/Kconfig | 1 -
drivers/net/ethernet/intel/idpf/idpf.h| 2 -
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 86 +--
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 25 +-
ame as Rx). This HT
is used only for OOO completions, which aren't really hotpath anyway.
Note that this change must be done atomically, otherwise it's really
easy to get lost and miss something.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf.h| 3 +-
to an OOO or due to that
something went bad with the HW?
Note that there are several patches fixing Tx (incl. timeouts) in my
tree, including yours (Joshua's) which you somehow didn't send yet ._.
Maybe start from them first?
I don't buy 30 seconds, at least for now. Maybe I'm missing something.
Nacked-by: Alexander Lobakin
Thanks,
Olek
From: Mateusz Polchlopek
Date: Tue, 4 Jun 2024 09:13:57 -0400
> From: Jacob Keller
>
> Using VIRTCHNL_VF_OFFLOAD_FLEX_DESC, the iAVF driver is capable of
> negotiating to enable the advanced flexible descriptor layout. Add the
> flexible NIC layout (RXDID=2) as a member of the Rx descriptor un
From: Maciej Fijalkowski
Date: Tue, 4 Jun 2024 15:21:48 +0200
> From: Michal Kubiak
>
> Address a scenario in which XSK ZC Tx produces descriptors to XDP Tx
> ring when link is either not yet fully initialized or process of
> stopping the netdev has already started. To avoid this, add checks
>
From: Maciej Fijalkowski
Date: Tue, 11 Jun 2024 16:21:27 +0200
> On Tue, Jun 11, 2024 at 01:59:37PM +0200, Alexander Lobakin wrote:
>> From: Maciej Fijalkowski
>> Date: Tue, 4 Jun 2024 15:21:48 +0200
>>
>>> From: Michal Kubiak
>>>
>>> Address a
From: Alexander Lobakin
Date: Wed, 12 Jun 2024 11:09:10 +0200
> From: Maciej Fijalkowski
> Date: Tue, 11 Jun 2024 16:21:27 +0200
[...]
>>>> diff --git a/drivers/net/ethernet/intel/ice/ice_xsk.c
>>>> b/drivers/net/ethernet/intel/ice/ice_xsk.c
>>>&g
From: Josh Hay
Date: Tue, 11 Jun 2024 11:13:53 -0700
>
>
> On 6/11/2024 3:44 AM, Alexander Lobakin wrote:
>> From: David Decotigny
>> Date: Tue, 4 Jun 2024 16:34:48 -0700
[...]
>> Note that there are several patches fixing Tx (incl. timeouts) in my
>> tree,
From: Jacob Keller
Date: Tue, 11 Jun 2024 13:52:57 -0700
>
>
> On 6/11/2024 4:47 AM, Alexander Lobakin wrote:
>> From: Mateusz Polchlopek
>> Date: Tue, 4 Jun 2024 09:13:57 -0400
>>
>>> From: Jacob Keller
>>>
>>> Using VIRTCH
From: Jacob Keller
Date: Tue, 28 May 2024 17:43:34 -0700
>
>
> On 5/28/2024 6:48 AM, Alexander Lobakin wrote:
>> Now that the queue and queue vector structures are separated and laid
>> out optimally, group the fields as read-mostly, read-write, and cold
>> cacheli
From: Alexander Lobakin
Date: Wed, 12 Jun 2024 15:03:07 +0200
> From: Jacob Keller
> Date: Tue, 28 May 2024 17:43:34 -0700
>
>>
>>
>> On 5/28/2024 6:48 AM, Alexander Lobakin wrote:
>>> Now that the queue and queue vector structures are separated and laid
&g
From: Jacob Keller
Date: Tue, 28 May 2024 17:47:32 -0700
>
>
> On 5/28/2024 6:48 AM, Alexander Lobakin wrote:
>> Currently, all HW supporting idpf supports the singleq model, but none
>> of it advertises it by default, as splitq is supported and preferred
>> for mu
From: David Decotigny
Date: Wed, 12 Jun 2024 11:01:46 -0700
> On Tue, Jun 11, 2024 at 11:13 AM Josh Hay wrote:
>>
>>
>>
>> On 6/11/2024 3:44 AM, Alexander Lobakin wrote:
>>> From: David Decotigny
>>> Date: Tue, 4 Jun 2024 16:34:48 -0700
>>>
From: Jakub Kicinski
Date: Wed, 29 May 2024 18:34:09 -0700
> On Tue, 28 May 2024 15:48:35 +0200 Alexander Lobakin wrote:
>> diff --git a/scripts/kernel-doc b/scripts/kernel-doc
>> index 95a59ac78f82..d0cf9a2d82de 100755
>> --- a/scripts/kernel-doc
>> +++ b/scrip
From: Jakub Kicinski
Date: Wed, 29 May 2024 18:40:12 -0700
> On Tue, 28 May 2024 15:48:45 +0200 Alexander Lobakin wrote:
>> Currently, idpf uses the following model for the header buffers:
>>
>> * buffers are allocated via dma_alloc_coherent();
>> * when receiving, nap
From: Simon Horman
Date: Sat, 1 Jun 2024 09:53:08 +0100
> On Tue, May 28, 2024 at 03:48:37PM +0200, Alexander Lobakin wrote:
>> Currently, sizeof(struct idpf_queue) is 32 Kb.
>> This is due to the 12-bit hashtable declaration at the end of the queue.
>> This HT is needed on
From: Simon Horman
Date: Sat, 1 Jun 2024 10:08:42 +0100
> + Dan Carpenter
>
> On Tue, May 28, 2024 at 03:48:46PM +0200, Alexander Lobakin wrote:
>> idpf uses Page Pool for data buffers with hardcoded buffer lengths of
>> 4k for "classic" buffers and 2k for &qu
From: Josh Hay
Date: Wed, 12 Jun 2024 23:36:29 -0700
>
>
> On 6/12/2024 2:34 AM, Alexander Lobakin wrote:
>> From: Josh Hay
>> Date: Tue, 11 Jun 2024 11:13:53 -0700
>>
>>>
>>>
>>> On 6/11/2024 3:44 AM, Alexander Lobakin wrote:
>>&
From: Maciej Fijalkowski
Date: Thu, 13 Jun 2024 18:07:53 +0200
> On Thu, Jun 13, 2024 at 05:51:25PM +0200, Maciej Fijalkowski wrote:
>> On Wed, Jun 12, 2024 at 11:15:31AM +0200, Alexander Lobakin wrote:
>>> From: Alexander Lobakin
>>> Date: Wed, 12 Jun 2024 11:09:10
From: Willem De Bruijn
Date: Thu, 30 May 2024 09:46:46 -0400
> Alexander Lobakin wrote:
>> Currently, idpf uses the following model for the header buffers:
>>
>> * buffers are allocated via dma_alloc_coherent();
>> * when receiving, napi_alloc_skb() is called and the
d internal temperature sensor")
>> Signed-off-by: Petr Oros
>> Acked-by: Ivan Vecera
To be added to v2:
Reviewed-by: Alexander Lobakin
Thanks,
Olek
From: Marcin Szycik
Date: Tue, 18 Jun 2024 16:11:54 +0200
> From: Michal Swiatkowski
>
> Remove unnecessary size checks when copying bitmaps in ice_add_sw_recipe()
> and replace them with compile time assert. Check if the bitmaps are equal
> size, as they are copied both ways.
>
> Signed-off-b
From: Willem De Bruijn
Date: Mon, 17 Jun 2024 14:13:07 -0400
> Alexander Lobakin wrote:
>> From: Willem De Bruijn
>> Date: Thu, 30 May 2024 09:46:46 -0400
>>
>>> Alexander Lobakin wrote:
>>>> Currently, idpf uses the following model for the header buff
nt DMA uses on hotpath, uses
napi_build_skb() instead of in-place short skb allocation.
Most idpf patches, except for the queue split, removes more lines
than adds.
Expect far better memory utilization and +5-8% on Rx depending on
the case (+17% on skb XDP_DROP :>).
Alexander Lobakin (
E_BYTES`` instead of ``L1_CACHE_BYTES`` as the former
also accounts L2, needed in some cases.
Signed-off-by: Alexander Lobakin
---
include/linux/cache.h | 59 +++
1 file changed, 59 insertions(+)
diff --git a/include/linux/cache.h b/include/linux/cache.
Instead of doing __cacheline_group_begin() __aligned(), use the new
__cacheline_group_{begin,end}_aligned(), so that it will take care
of the group alignment itself.
Also replace open-coded `4 * sizeof(long)` in two places with
a definition.
Signed-off-by: Alexander Lobakin
---
include/net
possible, as the size can then be easily predicted.
For the rest, make sure they don't cross the specified bound.
Signed-off-by: Alexander Lobakin
---
include/net/libeth/cache.h | 66 ++
1 file changed, 66 insertions(+)
create mode 100644 include/net/libeth/ca
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf.h| 1 -
.../net/ethernet/intel/idpf/idpf_lan_txrx.h | 2 +
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 47 ++-
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 20
drivers/net/ethernet
masks and put &idpf_q_vector on a good diet.
Also remove redundant pointer to the interrupt name from the structure.
request_irq() saves it and free_irq() returns it on deinit, so that you
can free the memory.
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net
this gives up to 2% of perf bump on Rx.
Reviewed-by: Przemek Kitszel
Reviewed-by: Jacob Keller
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 185 +---
1 file changed, 118 insertions(+), 67 deletions(-)
diff --git a/drivers/net/ethernet/intel
ng is the
same, just call either idpf_tx_splitq_frame() or idpf_tx_singleq_frame()
depending on the active model to actually map and send the skb.
Reviewed-by: Przemek Kitszel
Reviewed-by: Jacob Keller
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 9 ++
y: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/Kconfig| 13 +
drivers/net/ethernet/intel/idpf/Kconfig | 27 +++
drivers/net/ethernet/intel/idpf/Makefile | 3 ++-
drivers/net/ethernet/intel/idpf/idpf.h| 3 ++-
d
in
&idpf_rx_queue.
Reviewed-by: Przemek Kitszel
Reviewed-by: Jacob Keller
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/Kconfig | 1 +
drivers/net/ethernet/intel/idpf/idpf.h| 2 +-
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 88 +---
TATS unconditionally,
while the latter is often turned off for better performance.
Remove all the references to PP stats from the Ethtool code. The stats
are still available in their full via the generic Netlink interface.
Reviewed-by: Przemek Kitszel
Reviewed-by: Jacob Keller
Signed-off-by: Alexander Lob
accordingly. All the HW-related details like buffer alignment are still
accounted. For the header buffers, pick 256 bytes as in most places in
the kernel (have you ever seen frames with bigger headers?).
Reviewed-by: Przemek Kitszel
Signed-off-by: Alexander Lobakin
---
include/net/libeth/rx.h
more from the data buffer later. This W/A will hopefully
be removed one day.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 52 ++--
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 1 +
drivers/net/ethernet/intel/idpf/idpf_txrx.c | 253 -
nly after 4 HW writes rather than two.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/Kconfig | 1 -
drivers/net/ethernet/intel/idpf/idpf.h| 2 -
drivers/net/ethernet/intel/idpf/idpf_txrx.h | 86 +--
.../ethernet/intel/idpf/idpf_singleq_txrx.c | 27 +-
ame as Rx). This HT
is used only for OOO completions, which aren't really hotpath anyway.
Note that this change must be done atomically, otherwise it's really
easy to get lost and miss something.
Signed-off-by: Alexander Lobakin
---
drivers/net/ethernet/intel/idpf/idpf.h| 3 +-
From: Yan Zhai
Date: Thu, 20 Jun 2024 15:19:22 -0700
> Add a common point to transfer offloading info from XDP context to skb.
>
> Signed-off-by: Yan Zhai
> Signed-off-by: Jesper Dangaard Brouer
> ---
> drivers/net/ethernet/intel/ice/ice_txrx.c | 2 ++
> drivers/net/ethernet/intel/ice/ice_xsk
From: Alexander Lobakin
Date: Wed, 12 Jun 2024 14:33:17 +0200
> From: Jacob Keller
> Date: Tue, 11 Jun 2024 13:52:57 -0700
>
>>
>>
>> On 6/11/2024 4:47 AM, Alexander Lobakin wrote:
>>> From: Mateusz Polchlopek
>>> Date: Tue, 4 Jun 20
From: Przemek Kitszel
Date: Thu, 20 Jun 2024 17:15:53 +0200
> On 6/20/24 15:53, Alexander Lobakin wrote:
>> __cacheline_group_begin(), unfortunately, doesn't align the group
>> anyhow. If it is wanted, then you need to do something like
>>
>> __cacheline_
after the conversion is applied
anyway.
Alexander Lobakin (2):
idpf: fix memory leaks and crashes while performing a soft reset
idpf: fix UAFs when destroying the queues
Michal Kubiak (1):
idpf: fix memleak in vport interrupt configuration
drivers/net/ethernet/intel/idpf/idpf_lib.c | 48
e number and reopen the
interface on rollback - that solves limbo states when the device is left
disabled and/or without HW queues enabled.
Fixes: 02cbfba1add5 ("idpf: add ethtool callbacks")
Fixes: e4891e4687c8 ("idpf: split &idpf_queue into 4 strictly-typed queue
str
base name separately, considering that the
function 2) is never called out of idpf_vport_intr_init() context.
Fixes: d4d558718266 ("idpf: initialize interrupts and enable vport")
Cc: sta...@vger.kernel.org # 6.7
Signed-off-by: Michal Kubiak
Reviewed-by: Pavan Kumar Linga
Signed-off-b
the queues (+ both are freed within one function,
so it's not clear why nullify the pointers at all).
Fixes: 1c325aac10a8 ("idpf: configure resources for TX queues")
Fixes: 90912f9f4f2d ("idpf: convert header split mode to libeth +
napi_build_skb()")
Reported-by: Mich
From: Karol Kolacinski
Date: Thu, 25 Jul 2024 11:34:48 +0200
> From: Michal Michalik
>
> Add specific functions and definitions for E830 devices to enable
> PTP support.
> Introduce new PHY model ICE_PHY_E830.
> E830 devices support direct write to GLTSYN_ registers without shadow
> registers a
From: Karol Kolacinski
Date: Thu, 25 Jul 2024 11:34:49 +0200
> Simplify TSYN IRQ processing by moving it to a separate function and
> having appropriate behavior per PHY model, instead of multiple
> conditions not related to HW, but to specific timestamping modes.
>
> Reviewed-by: Przemek Kitsze
From: Karol Kolacinski
Date: Thu, 25 Jul 2024 11:34:50 +0200
> E830 PHY supports timestamp ready bitmap.
> Enable the bitmap by refactoring tx init function.
>
> Reviewed-by: Przemek Kitszel
> Signed-off-by: Karol Kolacinski
[...]
> @@ -987,24 +965,25 @@ ice_ptp_init_tx_e82x(struct ice_pf *p
From: Karol Kolacinski
Date: Thu, 25 Jul 2024 11:34:51 +0200
> From: Jacob Keller
>
> The E830 and E82x devices use essentially the same logic for performing
> a crosstimestamp. The only difference is that E830 hardware has
> different offsets. Instead of having two implementations, combine the
From: Simon Horman
Date: Fri, 26 Jul 2024 17:09:54 +0100
> On Wed, Jul 24, 2024 at 03:40:22PM +0200, Alexander Lobakin wrote:
>> The second tagged commit introduced a UAF, as it removed restoring
>> q_vector->vport pointers after reinitializating the structures.
>> This
From: Mateusz Polchlopek
Date: Tue, 30 Jul 2024 05:14:55 -0400
> [PATCH iwl-next v8 00/14] Add support for Rx timestamping for both ice
> and iavf drivers.
Oops, please remove that period at the end of the sentence when sending
next version.
> Initially, during VF creation it registers the PTP
From: Mateusz Polchlopek
Date: Tue, 30 Jul 2024 05:14:57 -0400
> From: Simei Su
>
> To support Rx timestamp offload, VIRTCHNL_OP_1588_PTP_CAPS is sent by
> the VF to request PTP capability and responded by the PF what capability
> is enabled for that VF.
[...]
> diff --git a/drivers/net/ether
From: Mateusz Polchlopek
Date: Tue, 30 Jul 2024 05:14:58 -0400
> From: Jacob Keller
>
> Support for allowing VF to negotiate the descriptor format requires that
> the VF specify which descriptor format to use when requesting Rx queues.
> The VF is supposed to request the set of supported format
From: Mateusz Polchlopek
Date: Tue, 30 Jul 2024 05:14:59 -0400
> From: Jacob Keller
>
> Enable support for VIRTCHNL_VF_OFFLOAD_RX_FLEX_DESC, to enable the VF
> driver the ability to determine what Rx descriptor formats are
> available. This requires sending an additional message during
> initia
1 - 100 of 443 matches
Mail list logo
