Because of the removal of the linux memory component it is no longer
necessary to initialize the memory component in opal_init(). This
commit moves the initialization to the creation of the first rcache
component.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit fixes bugs that can cause crashes and memory corruption
when the mremap hook is called. The problem occurs because of the
ellipses (...) in the mremap intercept function. The ellipses cover
the optional new_addr argument on Linux. This commit removes the
ellipses and adds an explicit 5th argument.
This commit also adds a hook for shmdt. The code only works on Linux
at the moment as it needs to read /proc/self/maps to determine the
size of the shared memory segment.
Additionally, this commit removes the mmap hook. There is no
apparent benefit for detecting mmap(..., PROT_NONE, ...) and it
seems to cause problems when threads are in use.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit removes the ptmalloc2 memory hooks. This is necessary in
order to support lazy registration of memory hooks. A feature that is
not supported by the ptmalloc hooks but is supported by the new
patcher hooks.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds a framework to abstract runtime code patching.
Components in the new framework can provide functions for either
patching a named function or a function pointer. The later
functionality is not being used but may provide a way to allow memory
hooks when dlopen functionality is disabled.
This commit adds two different flavors of code patching. The first is
provided by the overwrite component. This component overwrites the
first several instructions of the target function with code to jump to
the provided hook function. The hook is expected to provide the full
functionality of the hooked function.
The linux patcher component is based on the memory hooks in ucx. It
only works on linux and operates by overwriting function pointers in
the symbol table. In this case the hook is free to call the original
function using the function pointer returned by dlsym.
Both components restore the original functions when the patcher
framework closes.
Changes had to be made to support Power/PowerPC with the Linux
dynamic loader patcher. Some of the changes:
- Move code necessary for powerpc/power support to the patcher
base. The code is needed by both the overwrite and linux
components.
- Move patch structure down to base and move the patch list to
mca_patcher_base_module_t. The structure has been modified to
include a function pointer to the function that will unapply the
patch. This allows the mixing of multiple different types of
patches in the patch_list.
- Update linux patching code to keep track of the matching between
got entry and original (unpatched) address. This allows us to
completely clean up the patch on finalize.
All patchers keep track of the changes they made so that they can be
reversed when the patcher framework is closed.
At this time there are bugs in the Linux dynamic loader patcher so
its priority is lower than the overwrite patcher.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit makes it possible to set relative priorities for
components. Before the addition of the patched component there was
only one component that would run on any system but that is no longer
the case. When determining which component to open each component's
query function is called and the one that returns the highest priority
is opened. The default priority of the patcher component is set
slightly higher than the old ptmalloc2/ummunotify component.
This commit fixes a long-standing break in the abstration of the
memory components. ompi_mpi_init.c was referencing the linux malloc
hook initilize function to ensure the hooks are initialized for
libmpi.so. The abstraction break has been fixed by adding a memory
base function that calls the open memory component's malloc hook init
function if it has one. The code is not yet complete but is intended
to support ptmalloc in 2.0.0. In that case the base function will
always call the ptmalloc hook init if exists.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit adds support for runtime binary patching. The support is
broken down into two parts: util/opal_patcher.[ch] which contains the
functionality for runtime patching of symbols, and mca/memory/patcher
which patches the various symbols needed to provide support for memory
hooks. This work is preliminary and is based off work donated by IBM.
The patcher code is disabled if dlopen is disabled.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
https://github.com/pmix/master/pull/71
Have OMPI's current version of pmix120 nicely fail in case of
too long sun_path (longer than 108 or in case of OSX 103 chars).
And have OMPI return proper error messages with hints how to
amend.
This comit adds code that sets the fragment reg_cnt to 1 when sending
the completion message for an eager get. Without this the btl will
either hang or abort.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
Added mca parameter to turn progress thread on/off
Add a flag to check if we have btl progress thread.
Added macro for ob1 matching lock.
Update the AUTHORS file.
All BTL-only operations (basically all data movements
with the exception of the matching operation) can now
be handled for the TCP BTL by a progress thread.
The reference counting was broken which led PMIx_Finalize
to release resources early. This fixes the "use after free" scenarios
that I encountered.
(based on commit pmix/master@abfaa4c)
This reverts commit open-mpi/ompi@f7257a8310.
Ensure that we properly cleanup the session directory tree. Prior code had issues with symlinks, especially if the file that the link points to was already removed as we traverse the tree. Also found that the dirent checks for directory type weren't fully portable, and so fall back to the stat-based approach which is known to be portable.
Fix singularity singletons by detecting we are in a container and properly setting the pmix selection to pick the isolated component. Remove a stale restriction blocking use of the sm btl
This commit brings the scif btl up to date with changes made on master
to rework the mpool and rcache frameworks.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
This commit uses the new flag "enumerator" to support comma-delimited
lists of flags for both the btl and btl atomic flags. After this
commit is is valid to specify something like -mca btl_foo_flags
self,put,get,in-place. All non-deprecated flags are supported by the
enumerator.
Signed-off-by: Nathan Hjelm <hjelmn@lanl.gov>
