Keep track of the connected procs in vader_add_procs().
Otherwise, the same rank will reconnect the same shmem
segment (rank 0+...) multiple times instead of the next
one as intended.
Signed-off-by: Austen Lauria <awlauria@us.ibm.com>
(cherry picked from commit f69c8d6819)
This PR removes the constant defining the max attachment address and
replaces it with the largest address that shows up in /proc/self/maps.
This should address issues found on AARCH64 where the max address
may differ based on the configuration.
Since the calculated max address may differ between processes the
max address is sent as part of the modex and stored in the endpoint
data.
Signed-off-by: Nathan Hjelm <hjelmn@google.com>
(cherry picked from commit 728d51f9f3)
This commit fixes an issue discovered in the XPMEM registration cache. It
was possible for a registration to be invalidated by multiple threads
leading to a double-free situation or re-use of an invalidated registration.
This commit fixes the issue by setting the INVALID flag on a registation
when it will be deleted. The flag is set while iterating over the tree
to take advantage of the fact that a registration can not be removed
from the VMA tree by a thread while another thread is traversing the VMA
tree.
References #6524
References #7030Closes#6534
Signed-off-by: Nathan Hjelm <hjelmn@google.com>
(cherry picked from commit f86f805be1)
Trying out to run processes via mpirun in Podman containers has shown
that the CMA btl_vader_single_copy_mechanism does not work when user
namespaces are involved.
Creating containers with Podman requires at least user namespaces to be
able to do unprivileged mounts in a container
Even if running the container with user namespace user ID mappings which
result in the same user ID on the inside and outside of all involved
containers, the check in the kernel to allow ptrace (and thus
process_vm_{read,write}v()), fails if the same IDs are not in the same
user namespace.
One workaround is to specify '--mca btl_vader_single_copy_mechanism none'
and this commit adds code to automatically skip CMA if user namespaces
are detected and fall back to MCA_BTL_VADER_EMUL.
Signed-off-by: Adrian Reber <areber@redhat.com>
(cherry picked from commit fc68d8a90f)
This commit changes how the single-copy emulation in the vader btl
operates. Before this change the BTL set its put and get limits
based on the max send size. After this change the limits are unset
and the put or get operation is fragmented internally.
References #6568
Signed-off-by: Nathan Hjelm <hjelmn@google.com>
(cherry picked from commit ae91b11de2)
free the component mpool in mca_btl_vader_component_close()
and after freeing soem objects that depend on it such as
mca_btl_vader_component.vader_frags_user
Thanks Christoph Niethammer for reporting this.
Refs. open-mpi/ompi#6524
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
(cherry picked from commit open-mpi/ompi@77060cad07)
This commit fixes a bug introduced in
f62d26ddbc. That commit changed how
vader allocates fragment memory from the shared memory
segment. Unfortunately, the values used for the fragment sizes did not
include space for the fragment header. This can cause an overrun of
data from one fragment to the header of the next fragment.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit updates btl/vader to use an mpool for handling all shared
memory allocations (frags, fboxes).
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit fixes a bug where add_procs can incorrectly return an
error when going through the dynamic add_procs path. This doesn't
happen normally, only when pml/ob1 is not in use.
References #6201
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
(cherry picked from commit 30b8336cb4)
This commit works around an Oracle C compiler bug in 5.15 (not sure
when it was introduced). The bug is triggered when we chain
assignments of atomic variables. Ex:
_Atomic intptr x, y;
intptr_t z = 0;
x = y = z;
Will produce a compiler error of the form:
operand cannot have void type: op "="
assignment type mismatch:
long "=" void
To work around the issue we are removing the chain assignment and
setting the head and tail on different lines.
Fixes#5814
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
(cherry picked from commit dfa8d3a81a)
On some platfoms reading a 64-bit value is non-atomic and it is
possible that the two 32-bit values are read in the wrong order. To
ensure the tag is always read first this commit reads the tag before
reading the full 64-bit value.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
(cherry picked from commit 66a7dc4c72)
To ensure fast box entries are complete when processed by the
receiving process the tag must be written last. This includes a zero
header for the next fast box entry (in some cases). This commit fixes
two instances where the tag was written too early. In one case, on
32-bit systems it is possible for the tag part of the header to be
written before the size. The second instance is an ordering issue. The
zero header was being written after the fastbox header.
Fixes#5375, #5638
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
(cherry picked from commit 850fbff441)
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The write memory barrier was intended to precede setting a fast-box
header but instead follows it. This commit moves the memory barrier to
the intended location.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
(cherry picked from commit dca3516765)
This commit fixes two bugs in the RMA/atomic emulation code:
1) Fix a fragment leak when using AMO emulation.
2) Always initialize the single-copy emulation code. This is required
to use the AMO support.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds support for atomic operations as well as rdma for
systems without rdma support. This support is implemented using an
internal send tag.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Important note :
According to the man page
"On success, process_vm_readv() returns the number of bytes read and
process_vm_writev() returns the number of bytes written. This return
value may be less than the total number of requested bytes, if a
partial read/write occurred. (Partial transfers apply at the
granularity of iovec elements. These system calls won't perform a
partial transfer that splits a single iovec element.)"
So since we use a single iovec element, the returned size should either
be 0 or size, and the do loop should not be needed here.
We tried on various Linux kernels with size > 2 GB, and surprisingly,
the returned value is always 0x7ffff000 (fwiw, it happens to be the size
of the larger number of pages that fits a signed 32 bits integer).
We do not know whether this is a bug from the kernel, the libc or even
the man page, but for the time being, we do as is process_vm_readv() could
return any value.
Thanks Heiko Bauke for the bug report.
Refs. open-mpi/ompi#4829
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
This commit replaces the current VMA tree implementation with one that
uses the new opal_interval_tree_t class. Since the VMA tree lock is no
longer used this commit also updates rcache/grdma and btl/vader to
take better care when searching for existing registrations.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Since open-mpi/ompi@47fd2313ab
the backing file is now in /dev/shm by default. As a consequence,
the backing file name has to include the jobid so more than one job
can run at a time.
Signed-off-by: Gilles Gouaillardet <gilles@rist.or.jp>
Resolve a race condition between registering for a file to be removed upon termination and actual creation of that file by providing attributes that identify whether the path is a file or directory. This removes the need for PMIx to detect the difference.
Refs #4686
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
This commit moves the backing files to /dev/shm to avoid limitations
that may be set on /tmp. The files are registered with pmix to ensure
they are cleaned up after an erroneous exit.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
(cherry picked from commit 48101278160672317ade352365592f56ef3b8977)
If available, have apps use registration capability to cleanup their session directories. Setup capability for vader to register its shared memory file location - let someone familiar with that code do so.
Final cleanup to track uid/gid, update the opal/pmix API to pass flags for ignore and leave top directory alone
Signed-off-by: Ralph Castain <rhc@open-mpi.org>
There were multiple paths that could lead to a fast box
allocation. One of them made little sense (in-place send) so it has
been removed to allow a rework of the fast-box send function. This
should fix a number of issues with hanging/crashing when using the
vader btl.
References #4260
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds support for fetch-and-op atomics. This is needed
because and and or are irreversible operations so there needs to be a
way to get the old value atomically. These are also the only semantics
supported by C11 (there is not atomic_op_fetch, just
atomic_fetch_op). The old op-and-fetch atomics have been defined in
terms of fetch-and-op.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit renames the arithmetic atomic operations in opal to
indicate that they return the new value not the old value. This naming
differentiates these routines from new functions that return the old
value.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit eliminates the old opal_atomic_bool_cmpset functions. They
have been replaced by the opal_atomic_compare_exchange_strong
functions.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit renames the atomic compare-and-swap functions to indicate
the return value. This is in preperation for adding support for a
compare-and-swap that returns the old value. At the same time the
return type has been changed to bool.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
* Resolves#3705
* Components should link against the project level library to better
support `dlopen` with `RTLD_LOCAL`.
* Extend the `mca_FRAMEWORK_COMPONENT_la_LIBADD` in the `Makefile.am`
with the appropriate project level library:
```
MCA components in ompi/
$(top_builddir)/ompi/lib@OMPI_LIBMPI_NAME@.la
MCA components in orte/
$(top_builddir)/orte/lib@ORTE_LIB_PREFIX@open-rte.la
MCA components in opal/
$(top_builddir)/opal/lib@OPAL_LIB_PREFIX@open-pal.la
MCA components in oshmem/
$(top_builddir)/oshmem/liboshmem.la"
```
Note: The changes in this commit were automated by the script in
the commit that proceeds it with the `libadd_mca_comp_update.py`
script. Some components were not included in this change because
they are statically built only.
Signed-off-by: Joshua Hursey <jhursey@us.ibm.com>
This commit ensures that the pml callback is always made when
sending fragments. This is needed to avoid #3845. Once that is
fixed the #if 0'd code can be restored.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The vader btl kept a per-peer registration cache to keep track of
attachments. This is not really a problem with small numbers of local
ranks but can be a problem with large SMP machines. To reduce the
footprint there is now one registration cache for all xpmem
attachments. This will probably increase the lookup time for large
transfers but is a worthwhile trade-off.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
The original VADER_MAX_ADDRESS was tunned for x86_64 platforms only.
For non x86_64 platforms we can use XPMEM_MAXADDR_SIZE.
Signed-off-by: Pavel Shamis (Pasha) <pasharesearch@gmail.com>
This commit fixes several bugs in the registration cache code:
- Fix a programming error in the grdma invalidation function that can
cause an infinite loop if more than 100 registrations are
associated with a munmapped region. This happens because the
mca_rcache_base_vma_find_all function returns the same 100
registrations on each call. This has been fixed by adding an
iterate function to the vma tree interface.
- Always obtain the vma lock when needed. This is required because
there may be other threads in the system even if
opal_using_threads() is false. Additionally, since it is safe to do
so (the vma lock is recursive) the vma interface has been made
thread safe.
- Avoid calling free() while holding a lock. This avoids race
conditions with locks held outside the Open MPI code.
Fixesopen-mpi/ompi#1654.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit fixes a compile/link issue caused by vader. The vader btl
was using OPAL_THREAD_ADD64 to increment a counter which may not be
available on 32-bit systems. Changed to use OPAL_THREAD_ADD_SIZE_T
which will be 64-bit or 32-bit depending on the system.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
These macros should really be named OPAL_SUMMARY_*; they're used in
all projects, and therefore should be in the lowest later project (OPAL).
Signed-off-by: Jeff Squyres <jsquyres@cisco.com>
This commit rewrites both the mpool and rcache frameworks. Summary of
changes:
- Before this change a significant portion of the rcache
functionality lived in mpool components. This meant that it was
impossible to add a new memory pool to use with rdma networks
(ugni, openib, etc) without duplicating the functionality of an
existing mpool component. All the registration functionality has
been removed from the mpool and placed in the rcache framework.
- All registration cache mpools components (udreg, grdma, gpusm,
rgpusm) have been changed to rcache components. rcaches are
allocated and released in the same way mpool components were.
- It is now valid to pass NULL as the resources argument when
creating an rcache. At this time the gpusm and rgpusm components
support this. All other rcache components require non-NULL
resources.
- A new mpool component has been added: hugepage. This component
supports huge page allocations on linux.
- Memory pools are now allocated using "hints". Each mpool component
is queried with the hints and returns a priority. The current hints
supported are NULL (uses posix_memalign/malloc), page_size=x (huge
page mpool), and mpool=x.
- The sm mpool has been moved to common/sm. This reflects that the sm
mpool is specialized and not meant for any general
allocations. This mpool may be moved back into the mpool framework
if there is any objection.
- The opal_free_list_init arguments have been updated. The unused0
argument is not used to pass in the registration cache module. The
mpool registration flags are now rcache registration flags.
- All components have been updated to make use of the new framework
interfaces.
As this commit makes significant changes to both the mpool and rcache
frameworks both versions have been bumped to 3.0.0.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds two m4 macros: OPAL_SUMMARY_ADD, OPAL_SUMMARY_PRINT.
OPAL_SUMMARY_ADD adds an item to a section in the summary. For example
OPAL_SUMMARY_ADD([[Transports]],[[Foo]],...,[yes]) will add the
following to the summary:
Transports
-----------------------
Foo: yes
With this commit two sections are added: Transports, Resource Managers.
The OPAL_SUMMARY_PRINT macro is called after AC_OUTPUT and prints out
some information about the build (version, projects, etc) and then
the summarys sections. It will additionally print a warning if
internal debugging is enabled.
Example output:
Open MPI configuration:
-----------------------
Version: 3.0.0 a1
Build Open Platform Abstration project: yes
Build Open Runtime project: yes
Build Open MPI project: yes
Build Open SHMEM project: no
MPI C++ bindings (deprecated): no
MPI Fortran bindings: mpif.h, use mpi, use mpi_f08
Debug build: yes
Transports
-----------------------
Cray uGNI (Gemini/Aries): no
Intel Omnipath (PSM2): no
KNEM Shared Memory: no
Linux CMA IPC: no
Mellanox MXM: no
Open UCX: no
OpenFabrics libfabric: no
OpenFabrics Verbs: no
portals4: no
QLogic Infinipath (PSM): no
tcp: yes
XPMEM Shared Memory: no
Resource Managers
-----------------------
Cray Alps: no
Grid Engine: no
LSF: no
Slurm: yes
Torque: yes
INTERNAL DEBUGGING IS ENABLED. DO NOT USE THIS BUILD FOR PERFORMANCE MEASUREMENTS!
Signed-off-by: Nathan Hjelm <hjelmn@me.com>
This commit fixes several threading bugs:
- Add an additional lock to the btl_base_endpoint_t structure to lock
the list of pending frags. This allows the progress function to
attempt to send pending frags without needing to drop/reaquire the
lock. This should provide a small improvement in performance and
fixes a potential race between adding an removing items from the
pending list.
- Ensure fast boxes are only set up once by updating the send count
using atomics when needed and do not set the fast box buffer
pointer until the fast box is set up.
Closesopen-mpi/ompi#1408
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
