Differential D6688

net: Use M_HASHTYPE_OPAQUE_HASH if the mbuf flowid has hash properties
ClosedPublic
Actions

Authored by sepherosa_gmail.com on Jun 2 2016, 3:02 AM.

Tags

None

Referenced Files

	Unknown Object (File)
	Wed, Oct 30, 4:22 AM

	Unknown Object (File)
	Oct 22 2024, 8:30 AM

	Unknown Object (File)
	Oct 22 2024, 8:29 AM

	Unknown Object (File)
	Oct 22 2024, 8:29 AM

	Unknown Object (File)
	Oct 20 2024, 3:22 PM

	Unknown Object (File)
	Oct 2 2024, 8:20 AM

	Unknown Object (File)
	Oct 1 2024, 11:37 AM

	Unknown Object (File)
	Sep 29 2024, 11:24 PM

View All 62 Files

Subscribers

freebsd-net-list

Details

Reviewers

kmacy
gallatin
bryanv
erj
• hselasky
np
delphij
rrs
glebius
tuexen
gnn
davidcs
sbruno
bz
adrian
rwatson

Group Reviewers

Intel Networking

Commits

rS301538: net: Use M_HASHTYPE_OPAQUE_HASH if the mbuf flowid has hash properties

Diff Detail

Repository

rS FreeBSD src repository - subversion

Lint

Lint Not Applicable

Unit

Tests Not Applicable

Event Timeline

sepherosa_gmail.com updated this revision to Diff 17213.Jun 2 2016, 3:02 AM

sepherosa_gmail.com retitled this revision from to net: Use M_HASHTYPE_OPAQUE_HASH if the mbuf flowid has hash properties.

sepherosa_gmail.com updated this object.

sepherosa_gmail.com edited the test plan for this revision. (Show Details)

sepherosa_gmail.com added reviewers: rrs, gallatin, glebius, rwatson, gnn, adrian, delphij, • hselasky, tuexen, bryanv, kmacy, davidcs, bz, sbruno, erj, np.

Herald added a reviewer: Intel Networking. · View Herald TranscriptJun 2 2016, 3:02 AM

sepherosa_gmail.com edited edge metadata.Jun 2 2016, 3:03 AM

sepherosa_gmail.com added a subscriber: freebsd-net-list.

OK with regard to mlx5

This revision is now accepted and ready to land.Jun 2 2016, 7:35 AM

OK for the SCTP changes

The intel driver changes look ok to me, but what's the reason behind indicating that the hash that comes out of the card has hash properties? (I guess this might be covered in previous reviews or commit messages)

In D6688#141142, @erj wrote:

The intel driver changes look ok to me, but what's the reason behind indicating that the hash that comes out of the card has hash properties? (I guess this might be covered in previous reviews or commit messages)

At first, it's added mainly for the patch in this review:
https://reviews.freebsd.org/D6689

But it turns out that it also corrects certain types of mbuf hash type checks or makes the checks resilient to future hash type changes, e.g. the if_vxlan change in this patch.

And I think it could also be used in other places in the future :)

Closed by commit rS301538: net: Use M_HASHTYPE_OPAQUE_HASH if the mbuf flowid has hash properties (authored by sephe). · Explain WhyJun 7 2016, 4:52 AM

This revision was automatically updated to reflect the committed changes.

Revision Contents
Changeset List

Path

Size

head/

sys/

dev/

cxgb/

3 lines

e1000/

2 lines

hyperv/

netvsc/

hv_netvsc_drv_freebsd.c

12 lines

ixgbe/

2 lines

ixl/

6 lines

mlx5/

mlx5_en/

4 lines

qlxgbe/

4 lines

qlxge/

2 lines

net/

2 lines

3 lines

netinet/

2 lines

2 lines

ofed/

drivers/

net/

mlx4/

2 lines

Diff 17373

head/sys/dev/cxgb/cxgb_sge.c

Loading...

head/sys/dev/e1000/if_igb.c

Loading...

head/sys/dev/hyperv/netvsc/hv_netvsc_drv_freebsd.c

Loading...

head/sys/dev/ixgbe/ix_txrx.c

Loading...

head/sys/dev/ixl/ixl_txrx.c

Loading...

head/sys/dev/mlx5/mlx5_en/mlx5_en_rx.c

Loading...

head/sys/dev/qlxgbe/ql_isr.c

Loading...

head/sys/dev/qlxge/qls_isr.c

Loading...

head/sys/net/flowtable.c

Loading...

head/sys/net/if_vxlan.c

Loading...

head/sys/netinet/sctp_input.c

Loading...

head/sys/netinet/sctp_pcb.c

Loading...

head/sys/ofed/drivers/net/mlx4/en_rx.c

Loading...