With this patch applied I get the following failure 100% of the time,
I'd prefer to debug it and keep moving forward but I do not have the
time right now so I'm reverting the patch to the version which worked.
Ricardo please fix.
(gdb) bt
0 ztest_dmu_write_parallel (za=0x2aaaac898960) at
../../cmd/ztest/ztest.c:2566
1 0x0000000000405a79 in ztest_thread (arg=<value optimized out>)
at ../../cmd/ztest/ztest.c:3862
2 0x00002b2e6a7a841d in zk_thread_helper (arg=<value optimized out>)
at ../../lib/libzpool/kernel.c:131
3 0x000000379be06367 in start_thread (arg=<value optimized out>)
at pthread_create.c:297
4 0x000000379b2d30ad in clone () from /lib64/libc.so.6
This resolves previous scalabily concerns about the cost of calling
curthread which previously required a list walk. The kthread address
is now tracked as thread specific data which can be quickly returned.
Signed-off-by: Brian Behlendorf <behlendorf1@llnl.gov>
The 2.6.30 kernel build systems sets -Wframe-larger-than=2048 which causes
a warning to be generated when an individual stack frame exceeds 2048.
This caught the spa_history_log() and dmu_objset_snapshot() functions
which declared a data structure on the stack which contained a char
array of MAXPATHLEN. This in defined to be 4096 in the linux kernel
and I imagine it is quite large under Solaris as well. Regardless, the
offending data structures were moved to the heap to correctly keep the
stack depth to a minimum. We might consider setting this value even
lower to catch additional offenders because we are expecting deep stacks.
The intent here is to fully remove the previous Solaris thread
implementation so we don't need to simulate both Solaris kernel
and user space thread APIs. The few user space consumers of the
thread API have been updated to use the kthread API. In order
to support this we needed to more fully support the kthread API
and that means not doing crazy things like casting a thread id
to a pointer and using that as was done before. This first
implementation is not effecient but it does provide all the
corrent semantics. If/when performance becomes and issue we
can and should just natively adopt pthreads which is portable.
Let me finish by saying I'm not proud of any of this and I would
love to see it improved. However, this slow implementation does
at least provide all the correct kthread API semantics whereas
the previous method of casting the thread ID to a pointer was
dodgy at best.
gcc-unused and gcc-uninit topic branches at the same time and
then ran 'tg update'. I'll need to keep that sort of thing
in mind when updating multiple topic branches between updates.
within an ASSERT with the ASSERTV macro which will ensure it will
be removed when the ASSERTs are commented out. This makes gcc much
happier, makes the variables usage explicit, and removes the need
for the compiler to detect it is unused and do the right thing.
These changes bring the zfs-0.4.4 tree in to compliance with
the spl-0.4.4 packaging changes. The bottom line is 2 source
rpms and 4 binary rpms will now be generated when creating
packages there will be:
zfs-<version>.src.rpm
- Fully rebuildable source rpm for libzfs and utils.
zfs-modules-<version>.src.rpm
- Fully rebuildable source rpm for kernel modules.
zfs-<version>.<arch>.rpm
- Binary rpm for libzfs and utils. The utils in this package are
compatible with all zfs-module rpms of the same version.
zfs-devel-<version>.<arch>.rpm
- Binary rpm containing headers for building against libzfs libraries.
zfs-modules-<verion>-<kernel>.arch.rpm
- Binary rpm containing the kernel modules for a specific kernel build.
The package name contains the kernel version and you should have one
of these packages installed to match every kernel on your system.
zfs-modules-devel-<verion>-<kernel>.arch.rpm
- Binary rpm containing development header and module symbols needed
for building additional kernel modules which are dependent on the
zfs module stack.
Expect minor interations on these changes as I validate they work
properly on CHAOS, RHEL, Fedora, and SLES style distros.
The extra call to the constructor was there to reinitialize the non-
trivial primatives in the dnode (lists, mutexs, condvars, avl tree, etc).
This was safe, although not exactly clean, on Solaris because none of
the primitives allocate memory. In the Linux port this is not true.
To keep stack usage to a minimum several of the primatives dynamically
allocate memory thus initializing them twice results in a memory leak.
This patch resolves this problem for Solaris and Linux by ensuring all
*_inits are called in the constructor, and all *_destroys are called
in the destructor. Additionally we ensure that all dnode objects are
properly deconstructed before being freed to the slab, and when the
objects are allocated from the slab all required data members are
explicity initialized to correct values.
Brian Behlendorf
Ricardo M. Correia