Skip to contentSkip to conversation

Please pull RDMA subsystem changes

GIT PULL Merged Jason Gunthorpe 3 participants 4 messages May 6, 2022 InfiniBand

lore
Share & tools
lore

Jason Gunthorpe

author maintainer
#0 lore raw

Hi Linus,

Small update, still some rxe bugs that need fixing, but that is normal for rxe.

Also, I replaced my Yubikey with a newer one that can do the smaller edwards curves and so have a generated a new pgp subkey. I signed the tag with the new key and this message with the old. It should be on the keyservers now, but if you have trouble it is in the kernel.org repo:

https://git.kernel.org/pub/scm/docs/kernel/pgpkeys.git/tree/keys/A5F46BDD553C74FA.asc

Thanks,
Jason

The following changes since commit ce522ba9ef7e2d9fb22a39eb3371c0c64e2a433e:

Linux 5.18-rc2 (2022-04-10 14:21:36 -1000)

are available in the Git repository at:

git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma.git tags/for-linus

for you to fetch changes up to bfdc0edd11f9501b891a069b5bbd3b16731941e1:

RDMA/rxe: Change mcg_lock to a _bh lock (2022-05-04 21:29:25 -0300)

----------------------------------------------------------------
v5.18 second rc pull request

A few recent regressions in rxe's multicast code, and some old driver bugs:

- Error case unwind bug in rxe for rkeys

- Dot not call netdev functions under a spinlock in rxe multicast code

- Use the proper BH lock type in rxe multicast code

- Fix idrma deadlock and crash

- Add a missing flush to drain irdma QPs when in error

- Fix high userspace latency in irdma during destroy due to synchronize_rcu()

- Rare race in siw MPA processing

----------------------------------------------------------------
Bob Pearson (4):
RDMA/rxe: Fix "Replace mr by rkey in responder resources" RDMA/rxe: Recheck the MR in when generating a READ reply RDMA/rxe: Do not call dev_mc_add/del() under a spinlock RDMA/rxe: Change mcg_lock to a _bh lock

Cheng Xu (1):
RDMA/siw: Fix a condition race issue in MPA request processing

Duoming Zhou (1):
RDMA/irdma: Fix deadlock in irdma_cleanup_cm_core()

Mustafa Ismail (1):
RDMA/irdma: Fix possible crash due to NULL netdev in notifier

Shiraz Saleem (1):
RDMA/irdma: Reduce iWARP QP destroy time

Tatyana Nikolova (1):
RDMA/irdma: Flush iWARP QP if modified to ERR from RTR state

 drivers/infiniband/hw/irdma/cm.c      | 33 ++++--------
 drivers/infiniband/hw/irdma/utils.c   | 21 ++++----
 drivers/infiniband/hw/irdma/verbs.c   |  4 +-
 drivers/infiniband/sw/rxe/rxe_mcast.c | 95 ++++++++++++++++-------------------
 drivers/infiniband/sw/rxe/rxe_resp.c  | 35 +++++++++----
 drivers/infiniband/sw/siw/siw_cm.c    |  7 +--
 6 files changed, 92 insertions(+), 103 deletions(-)
#1 lore raw
Quoted 2 lines · Jason Gunthorpe
 6 files changed, 92 insertions(+), 103 deletions(-)

I couldn't for the life of me understand how you got to that diffstat, since everything I tried gave me

 6 files changed, 85 insertions(+), 96 deletions(-)

instead.

Until I realized that you must be using the '--histogram' flag to git diff. It seems to give quite different results for rxe/rxe_mcast.c.

Better? Worse? I dunno. But it threw me.

Linus

merge bot Merged
#3 lore raw
In reply to Jason Gunthorpe
The pull request you sent on Fri, 6 May 2022 13:01:51 -0300:

The pull request you sent on Fri, 6 May 2022 13:01:51 -0300:

Quoted 1 line

git://git.kernel.org/pub/scm/linux/kernel/git/rdma/rdma.git tags/for-linus

has been merged into torvalds/linux.git:
https://git.kernel.org/torvalds/c/4df22ca85d3d73f9822b1a354bb56dd1872180cd

Thank you!

Signature · 2 lines
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/prtracker.html
author maintainer
#2 lore raw
In reply to Linus Torvalds
Quoted 7 lines · Linus Torvalds
Quoted 2 lines · Jason Gunthorpe
 6 files changed, 92 insertions(+), 103 deletions(-)

I couldn't for the life of me understand how you got to that diffstat, since everything I tried gave me

Quoted 6 lines
 6 files changed, 85 insertions(+), 96 deletions(-)

instead.

Until I realized that you must be using the '--histogram' flag to git diff. It seems to give quite different results for rxe/rxe_mcast.c.

Huh. That is surprising - I can't say why anymore, but yes this git tree did have the histrogram setting in the git config.

I took it away

Thanks,
Jason