[openfabrics-ewg] ipath patches for RC3
Bryan O'Sullivan
bos at pathscale.com
Tue Aug 29 23:23:56 PDT 2006
On Tue, 2006-08-29 at 16:17 -0700, Bryan O'Sullivan wrote:
> This is a set of patches for the ipath driver for OFED 1.1 RC3.
I forgot to include a broken-out description of the patches earlier;
here it is.
<b
-------------- next part --------------
allow-disable-sma.patch:
IB/ipath - allow SMA to be disabled
This is useful for testing purposes.
control-rx-polarity-inversion.patch:
IB/ipath - control receive polarity inversion
count-attached-qps-correctly.patch:
IB/ipath - account for attached QPs correctly
count-cqs-properly.patch:
IB/ipath - lock and count allocated CQs properly
diag-send.patch:
IB/ipath - add new minor device to allow sending of diag packets
drop-double-unregister.patch:
IB/ipath - don't unregister the IB device twice
fix-ipath_poll-return.patch:
IB/ipath - fix return value from ipath_poll
This stops the generic poll code from waiting for a timeout.
fix-kpiobufs-handling.patch:
IB/ipath - fix handling of kpiobufs
Change comment: no longer imply that user can set ipath_kpiobufs to zero.
Actually set ipath_kpiobufs from parameter. Previously only altered
per-device ipath_lastport_piobuf, which was over-written in chip init.
fix-port_rcvhdrq-memleak.patch:
IB/ipath - fix memory leak if allocation fails
If the second allocation failed, the first structure allocated in this
routine was not freed.
fix-sq_sig_all.patch:
IB/ipath - handle sq_sig_all field correctly
fix-unload-crash-if-cfgports-lt-portcnt.patch:
IB/ipath - fix for crash on module unload, if cfgports < portcnt
Allocate enough pointers for all possible ports, to avoid problems in
cleanup/unload.
forbid-zero-entry-cq.patch:
IB/ipath - do not allow use of CQ entries with invalid counts
freeze-with-serno.patch:
IB/ipath - add serial number to hardware freeze error message
Also added the word "Hardware" after "Fatal" to make it more obvious
that it's hardware, not software.
guid-0-cleanup.patch:
IB/ipath - clean up handling of GUID 0
Respond with an error to the SM if our GUID is 0, and don't allow the
user to set our GUID to 0.
limit-packets-sent-without-ack.patch:
IB/ipath - limit # of packets sent without an ACK received
The sender requests an ACK every 1/2 MB to avoid retransmit timeouts that
were causing MVAPICH mod_bw to fail after a predictable number of sends.
limit-qp-create-count.patch:
IB/ipath - put a limit on the number of QPs that can be created
lock-resource-limit-counters-correctly.patch:
IB/ipath - lock resource limit counters correctly
merge-core-verbs.patch:
IB/ipath - merge ipath_core and ib_ipath drivers
There is little point in keeping the two drivers separate, so we are
merging them.
merge-debug-code.patch:
IB/ipath - simplify debugging code after ipath_core and ib_ipath merger
mmap-cq-rq.patch:
IB/ipath: Performance improvements via mmap of queues
Improve performance of userspace post receive, post SRQ receive, and
poll CQ operations for ipath by allowing userspace to directly mmap()
receive queues and completion queues. This eliminates the copying
between userspace and the kernel in the data path.
simplify-layer.patch:
IB/ipath - simplify layering code
A lot of ipath layer code was only called in one place. Now that the
ipath_core and ib_ipath drivers are merged, it's more sensible to simply
inline the simple stuff that the layer code was doing.
trivial-cleanups.patch:
IB/ipath - trivial cleanups
unregister-from-openib-early.patch:
IB/ipath - unregister from IB core early
This gives upper-level protocols a chance to unregister while the device
is still usable.
user-data-to-modify-methods.patch:
IB/uverbs: Pass userspace data to modify_srq and modify_qp methods
Pass a struct ib_udata to the low-level driver's ->modify_srq() and
->modify_qp() methods, so that it can get to the device-specific data
passed in by the userspace driver.
user-master-slave.patch:
IB/ipath - driver support for userspace sharing of HW contexts
This allows multiple userspace processes to share a single hardware
context in a master/slave arrangement. It is backwards binary compatible
with existing userspace.
validate-path_mig_state-properly.patch:
IB/ipath - validate path_mig_state properly
warn-if-no-lid.patch:
IB/ipath - print warning if LID not acquired within one minute
backport/bits_per_byte-2.6.15.patch:
BACKPORT - define BITS_PER_BYTE for 2.6.15 and earlier
backport/class-2.6.12.patch:
BACKPORT - changes to class API for 2.6.12 and earlier
backport/class-2.6.9.patch:
BACKPORT - changes to class API for 2.6.9 and earlier
backport/class-2.6.9_U4.patch:
BACKPORT - class changes for RHEL4 U4, 2.6.9 with 2.6.15 bits
backport/dont-leak-info-to-userspace.patch:
BACKPORT - Don't leak kernel information to userspace
This is done using vmalloc_user in newer kernels.
backport/fs-get_sb-2.6.17.patch:
BACKPORT - change signature of ipath_get_sb for 2.6.17 and earlier
backport/header.patch:
BACKPORT - ipath_backport.h header
backport/idr_destroy-2.6.13.patch:
BACKPORT - idr_destroy didn't appear until 2.6.14
backport/iounmap-2.6.9.patch:
BACKPORT - backport to handle change in iounmap signature in 2.6.9
backport/iowrite32_copy.patch:
BACKPORT - backport iowrite32_copy to 2.6.15 and earlier
backport/mutex-2.6.15.patch:
BACKPORT - turn mutexes into semaphores for 2.6.15 and earlier
backport/page-hacks-2.6.14.patch:
BACKPORT - fix up page management for kernels prior to 2.6.15
backport/page-hacks-2.6.9.patch:
BACKPORT - use more insistent gfp flags for 2.6.9
backport/pci_find_next_capability-2.6.14.patch:
BACKPORT - backport pci_find_next_capability to 2.6.14 and earlier
backport/pcie-linkstate-2.6.9.patch:
BACKPORT - define PCI_EXP_LNKSTA for 2.6.9
backport/remap_page_range-2.6.9.patch:
BACKPORT - remap_pfn_range was remap_page_range in 2.6.9
backport/rlimit-2.6.9.patch:
BACKPORT - rlim moved from current to current->signal in 2.6.10
backport/sysfs-show-2.6.12.patch:
BACKPORT - handle change to sysfs show method's signature in 2.6.13
More information about the ewg
mailing list