Skip to contentSkip to conversation

InfiniBand/RDMA merge plans for 2.6.29

DISCUSSION Roland Dreier 4 participants 7 messages Dec 22, 2008 – Dec 31, 2008

lore
Share & tools
lore
#0 lore raw

I guess now is probably a good time to send out my plans for the 2.6.29 cycle. I'm not operating at full speed (since I'm kind of on paternity leave and kind of on holiday break), but I'll try to get through as much of my backlog as possible.

I should mention that I've changed the way I manage my git tree slightly. I've started using "topic" branches to organize the patches I have queued; so for example I'll put IPoIB changes in my "ipoib" branch and ehca changes in my "ehca" branch. Then I merge everything that I consider ready for the next merge window into my "for-next" branch (which is included as part of Stephen Rothwell's linux-next tree). Any topics that I'm not sure are ready get to stay on their own branch until they are ready (or get dropped); for example, I've had an "xrc" branch for XRC work for a while now.

Anyway, here are all the pending things that I'm aware of. As usual, if something isn't already in my tree and isn't listed below, I probably missed it or dropped it by mistake. Please remind me again in that case.

As usual, when submitting a patch:

- Give a good changelog that explains what issue your patch addresses, how you address the issue, how serious the issue is, and any other information that would be useful to someone evaluating your patch or reading it years from now.

- Please make sure that you include a "Signed-off-by:" line, and put any extra junk that should not go into the final kernel log *after* the "---" line so that git tools strip it off automatically. Make the subject line be appropriate for inclusion in the kernel log as well once the leading "[PATCH ...]" stuff is stripped off. I waste a lot of time fixing patches by hand that could otherwise be spent doing something productive like watching youtube.

- Run your patch through checkpatch.pl so I don't have to nag you to fix trivial issues (or spend time fixing them myself).

- Read your patch over so I don't see a memory leak or deadlock as soon as I look at it.

- Build your patch with sparse checking ("C=2 CF=-D__CHECK_ENDIAN__") and make sure it doesn't introduce new warnings. (A big bonus in goodwill for sending patches that fix old warnings)

- Test your patch on a kernel with things like slab debugging and lockdep turned on.

And while you're waiting for me to get to your patch, I sure wouldn't mind if you read and commented on someone else's patch. None of this means you shouldn't remind me about pending patches, since I often lose track of things and drop them accidentally.

Core:

- Multiple CQ event vector support. I don't see much progress on making this actually usable or useful, but maybe the best way to make progress on getting the correct interface for applications to actually figure out which vector to use is to merge the multiple vector support for some low-level drivers. I have the mlx4 changes merged, and I plan to try and do some analogous changes for mthca. A resurrection of the ehca multiple vector support would be welcome as well.

- I plan to merge Aleksey's patches for RDMA CM IPv6 support when I have a version that applies to the current kernel. Might be a good idea for cxgb3 and nes guys to look at this and make sure that we at least don't have any kernel crashes caused if someone tries to make an IPv6 connection.

ULPs:

- I sincerely appreciate everyone who resent IPoIB patches recently. I will review them and apply this week most likely.

HW specific:

- A bunch of ipath fixes.

- A bunch of nes cleanups and fixes.

- A few ehca fixes and cleanups.

Here are a few topics that I believe will not be ready in time for the 2.6.29 window and will need to wait for 2.6.30 at least:

- Jack's XRC patch set. I think we're getting closer to converging here, but I still want to get a clean solution for the "free XRC domain with other objects still associated because multiple contexts share XRCDs" problem.

Here all the patches I already have in my for-next branch:

Chien Tung (2):
RDMA/nes: Add loopback check to make_cm_node() RDMA/nes: Cleanup warnings

Dave Olson (4):
IB/ipath: Don't count IB symbol and link errors unless link is UP IB/ipath: Only do 1X workaround on rev1 chips
IB/ipath: Fix spi_pioindex value
IB/ipath: Add locking for interrupt use of ipath_pd contexts vs free

David Disseldorp (1):
IB/iser: Avoid recv buffer exhaustion caused by unexpected PDUs

Faisal Latif (6):
RDMA/nes: Cleanup cqp_request list usage
RDMA/nes: Lock down connected_nodes list while processing it RDMA/nes: Avoid race between MPA request and reset event to rdma_cm RDMA/nes: Forward packets for a new connection with stale APBVT entry RDMA/nes: Fix TCP compliance test failures
RDMA/nes: Check cqp_avail_reqs is empty after locking the list

Joachim Fenkes (1):
IB/ehca: Fix locking for shca_list_lock

Julia Lawall (1):
IB/ehca: Remove redundant test of vpage

Michael Ellerman (1):
IB/ipath: Fix pointer-to-pointer thinko in ipath_fs.c

Ralph Campbell (3):
IB/ipath: Improve UD loopback performance by allocating temp array only once IB/ipath: Fix PSN of send WQEs after an RDMA read resend IB/ipath: Check return value of dma_map_single()

Roland Dreier (1):
mlx4_core: Delete incorrect comment

Stefan Roscher (1):
IB/ehca: Replace modulus operations in flush error completion path

Yevgeny Petrilin (1):
mlx4_core: Add support for multiple completion event vectors

#1 lore raw

Roland,

Any chance create_qp_flags will get in?

Thanks,
Ron

Quoted 144 lines · Roland Dreier

I guess now is probably a good time to send out my plans for the 2.6.29 cycle. I'm not operating at full speed (since I'm kind of on paternity leave and kind of on holiday break), but I'll try to get through as much of my backlog as possible.

I should mention that I've changed the way I manage my git tree slightly. I've started using "topic" branches to organize the patches I have queued; so for example I'll put IPoIB changes in my "ipoib" branch and ehca changes in my "ehca" branch. Then I merge everything that I consider ready for the next merge window into my "for-next" branch (which is included as part of Stephen Rothwell's linux-next tree). Any topics that I'm not sure are ready get to stay on their own branch until they are ready (or get dropped); for example, I've had an "xrc" branch for XRC work for a while now.

Anyway, here are all the pending things that I'm aware of. As usual, if something isn't already in my tree and isn't listed below, I probably missed it or dropped it by mistake. Please remind me again in that case.

As usual, when submitting a patch:

- Give a good changelog that explains what issue your patch addresses, how you address the issue, how serious the issue is, and any other information that would be useful to someone evaluating your patch or reading it years from now.

- Please make sure that you include a "Signed-off-by:" line, and put any extra junk that should not go into the final kernel log *after* the "---" line so that git tools strip it off automatically. Make the subject line be appropriate for inclusion in the kernel log as well once the leading "[PATCH ...]" stuff is stripped off. I waste a lot of time fixing patches by hand that could otherwise be spent doing something productive like watching youtube.

- Run your patch through checkpatch.pl so I don't have to nag you to fix trivial issues (or spend time fixing them myself).

- Read your patch over so I don't see a memory leak or deadlock as soon as I look at it.

- Build your patch with sparse checking ("C=2 CF=-D__CHECK_ENDIAN__") and make sure it doesn't introduce new warnings. (A big bonus in goodwill for sending patches that fix old warnings)

- Test your patch on a kernel with things like slab debugging and lockdep turned on.

And while you're waiting for me to get to your patch, I sure wouldn't mind if you read and commented on someone else's patch. None of this means you shouldn't remind me about pending patches, since I often lose track of things and drop them accidentally.

Core:

- Multiple CQ event vector support. I don't see much progress on making this actually usable or useful, but maybe the best way to make progress on getting the correct interface for applications to actually figure out which vector to use is to merge the multiple vector support for some low-level drivers. I have the mlx4 changes merged, and I plan to try and do some analogous changes for mthca. A resurrection of the ehca multiple vector support would be welcome as well.

- I plan to merge Aleksey's patches for RDMA CM IPv6 support when I have a version that applies to the current kernel. Might be a good idea for cxgb3 and nes guys to look at this and make sure that we at least don't have any kernel crashes caused if someone tries to make an IPv6 connection.

ULPs:

- I sincerely appreciate everyone who resent IPoIB patches recently. I will review them and apply this week most likely.

HW specific:

- A bunch of ipath fixes.

- A bunch of nes cleanups and fixes.

- A few ehca fixes and cleanups.

Here are a few topics that I believe will not be ready in time for the 2.6.29 window and will need to wait for 2.6.30 at least:

- Jack's XRC patch set. I think we're getting closer to converging here, but I still want to get a clean solution for the "free XRC domain with other objects still associated because multiple contexts share XRCDs" problem.

Here all the patches I already have in my for-next branch:

Chien Tung (2):
RDMA/nes: Add loopback check to make_cm_node() RDMA/nes: Cleanup warnings

Dave Olson (4):
IB/ipath: Don't count IB symbol and link errors unless link is UP IB/ipath: Only do 1X workaround on rev1 chips
IB/ipath: Fix spi_pioindex value
IB/ipath: Add locking for interrupt use of ipath_pd contexts vs free

David Disseldorp (1):
IB/iser: Avoid recv buffer exhaustion caused by unexpected PDUs

Faisal Latif (6):
RDMA/nes: Cleanup cqp_request list usage
RDMA/nes: Lock down connected_nodes list while processing it RDMA/nes: Avoid race between MPA request and reset event to rdma_cm RDMA/nes: Forward packets for a new connection with stale APBVT entry RDMA/nes: Fix TCP compliance test failures
RDMA/nes: Check cqp_avail_reqs is empty after locking the list

Joachim Fenkes (1):
IB/ehca: Fix locking for shca_list_lock

Julia Lawall (1):
IB/ehca: Remove redundant test of vpage

Michael Ellerman (1):
IB/ipath: Fix pointer-to-pointer thinko in ipath_fs.c

Ralph Campbell (3):
IB/ipath: Improve UD loopback performance by allocating temp array only once IB/ipath: Fix PSN of send WQEs after an RDMA read resend IB/ipath: Check return value of dma_map_single()

Roland Dreier (1):
mlx4_core: Delete incorrect comment

Stefan Roscher (1):
IB/ehca: Replace modulus operations in flush error completion path

Yevgeny Petrilin (1):
mlx4_core: Add support for multiple completion event vectors

_______________________________________________
general mailing list
[email protected]
http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general

To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general

#2 lore raw
Quoted 1 line

Any chance create_qp_flags will get in?

Nope, since that's stuck behind XRC.

#3 lore raw

I know it was stuck there,
that's why I've sent a new version that doesn't rely on it. I've implemented the more_ops (in libibvers) in my patches - that's the only thing I needed from XRC.

Ron

Quoted 4 lines · Roland Dreier
Quoted 1 line

Any chance create_qp_flags will get in?

Nope, since that's stuck behind XRC.

#4 lore raw
In reply to Roland Dreier
Quoted 4 lines · Roland Dreier

HW specific:
- A bunch of ipath fixes.
- A bunch of nes cleanups and fixes.
- A few ehca fixes and cleanups.

What about Jack's raw QP support/fixes patch set which he reposted on December 15th?

http://lists.openfabrics.org/pipermail/general/2008-December/055954.html
http://lists.openfabrics.org/pipermail/general/2008-December/055953.html
http://lists.openfabrics.org/pipermail/general/2008-December/055955.html

Or.

#5 lore raw
Quoted 2 lines

What about Jack's raw QP support/fixes patch set which he reposted on December 15th?

didn't make this window. How are you using raw QPs?

#6 lore raw
Quoted 2 lines · Roland Dreier
Quoted 1 line

What about Jack's raw QP support/fixes patch set which he reposted on December 15th?

didn't make this window. How are you using raw QPs?

soon want to play/experience with sniffer prototype, any chance they would eventually make it for this window?

Or.