When casting unsigned long userspace pointer for use with
userspace-accessing functions, need to use __user attribute to make
sparse happy.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Casts for get/put_user and copy_to/from_user should have __user
attribute for the userspace buffer address.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This is a user pointer, so that makes sparse a lot happier.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cannot use memcpy, but use copy_to/from_user instead
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Also update all callers.
This fixes a bunch of address space mismatch warnings from sparse.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
sysfs store methods provide us with a kernel buffer already, but
lprocfs_write_helper is expecting a user buffer.
Replace lprocfs_write_helper with kstrto[u]int() calls instead in
contention_seconds_store() and lockless_truncate_store()
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The buffer that lprocfs_write_frac_u64_helper is working on is
presumed to be a userspace one, so it's not suitable for use
on kernel buffers.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This also silents a sparse address space warning
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Also update all methods and calls everywhere.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For some crazy reason ll_obd_statfs decided to decode async flag
passed from userspace and then pass it via a userspace pointer
argument to lov_iocontrol.
This patch moves flags decoding to lov_iocontrol where it belongs.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This also happens to silence sparce warnings about different
address spaces.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
arg is a userspace pointer and marking it as such makes sparse happy.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Make them void __user * instead of char * (or char __user *),
void * removes the necessity of explicit casts to proper type
where people also need to remember __user qualifiers, so I think
it works better here.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Make them void __user * instead of char * (or char __user *),
void * removes the necessity of explicit casts to proper type
where people also need to remember __user qualifiers, so I think
it works better here.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Just make clear which pointers are from userspace and which are not in
libcfs_ioctl_getdata/libcfs_ioctl_popdata and their callers.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In lustre/obdecho/echo_client.c, remove handling of lov_stripe_md
passed from userspace (since userspace never passes it). Remove the
LOV specific code (ed_next_islov) from the echo client (since it
doesn't work).
Remove echo_get_stripe_off_id() and all calls to it since the stripe
count of the passed in lsm is always 0 and the funciton does nothing
in this case. Remove the then unused lsm parameters of
echo_client_page_debug_setup() and echo_client_page_debug_check().
In the OBD_IOC_GETATTR and OBD_IOC_SETATTR cases of
echo_client_iocontrol() do not set the oi_md member of struct obd_info
since only LOV OBD methods access it.
Signed-off-by: John L. Hammond <john.hammond@intel.com>
Reviewed-on: http://review.whamcloud.com/12446
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-5814
Reviewed-by: Bobi Jam <bobijam@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This has been unused for ages and could be safely removed now.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
IOC_LIBCFS_DEBUG_PEER was added back in the stone ages to print debug
statistics on a peer when peer timeout happens.
Redo it properly as a separate LNet API call,
also get rid of "ioctl" forwarding into the underlying LNDs,
since no current LNDs implement this function anymore.
Signed-off-by: Oleg Drokin <green@linuxhacker.ru>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This export_operations structure is never modified, so declare it as const.
Most other structures of this type are already const.
Done with the help of Coccinelle.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
These lu_device_operations and cl_device_operations structures are never
modified, so declare them as const. Other structures of these types are
already const.
Done with the help of Coccinelle.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
nrs_resource_put_safe() might hold a lock one one struct
while operating on the other.
There are 2 levels of structures.
Use nrs_policy_put(), which has locking baked in.
sparse gives the following warning:
drivers/staging/lustre//lustre/ptlrpc/nrs.c:498:39:
warning: context imbalance in 'nrs_resource_put_safe' -
different lock contexts for basic block
Signed-off-by: Joshua Clayton <stillcompiling@gmail.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Clean up another case of the compiler remininding the programmer they
are an idiot:
drivers/staging/lustre/lustre/ptlrpc/sec_bulk.c:308:34: warning: comparison of unsigned expression >= 0 is always true [-Wtype-limits]
LASSERT(page_pools.epp_waitqlen >= 0);
Just lose the assert, and save a page of compiler spew.
Signed-off-by: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
CC [M] drivers/staging/lustre/lustre/libcfs/module.o
drivers/staging/lustre/lustre/libcfs/module.c: In function 'lustre_insert_debugfs':
drivers/staging/lustre/lustre/libcfs/module.c:670:17: warning: variable 'entry' set but not used [-Wunused-but-set-variable]
struct dentry *entry;
^
Just ignore the dentry returned, and add a comment that we *know*
we're not really leaking the dentry because something else will be able
to reap it via recursion.
Signed-off-by: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/lustre/lustre/fid/../include/lustre_cfg.h: In function 'lustre_cfg_free':
drivers/staging/lustre/lustre/fid/../include/lustre_cfg.h:253:6: warning: variable 'len' set but not used [-Wunused-but-set-variable]
int len;
Yep, we're just gonna call kfree, no need to calculate len. Bye-bye.
Signed-off-by: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/lustre/lustre/fid/lproc_fid.c: In function 'ldebugfs_fid_write_common':
drivers/staging/lustre/lustre/fid/lproc_fid.c:67:6: warning: variable 'rc' set but not used [-Wunused-but-set-variable]
int rc;
We fix it by *using* the return code to help bulletproof it. It says it's
test code - it should be *more* bulletproof than production, not less.
Signed-off-by: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Low-hanging fruit first:
CC [M] drivers/staging/lustre/lustre/fid/fid_request.o
In file included from drivers/staging/lustre/lustre/fid/../include/lustre_net.h:66:0,
from drivers/staging/lustre/lustre/fid/../include/lustre_lib.h:64,
from drivers/staging/lustre/lustre/fid/../include/obd.h:52,
from drivers/staging/lustre/lustre/fid/fid_request.c:48:
drivers/staging/lustre/lustre/fid/../include/lu_object.h:765:1: warning: 'inline' is not at beginning of declaration [-Wold-style-declaration]
static const inline struct lu_device_operations *
^
So we just swap inline and const. 272 warnings gone. :)
Signed-off-by: Valdis Kletnieks <Valdis.Kletnieks@vt.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cleanup the last white space issues in the kernel comm code.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Doing if (rc != 0) or if (rc == 0) is bad form. This patch corrects
kernelcomm.c to behavior according to kernel coding standards.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The functon libcfs_kkuc_msg_put() returns -ENOSYS which is not
correct. Return -ENXIO instead if the kuc header is corrupt.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add in missing braces for libcfs_kkuc_group_put();.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add in missing space arouund '+' in the kernel_comm code.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Fixup the comments to the linux kernel style for the source and
headers related to the kernel_comm work.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The group variable was converted to an unsigned int in
libcfs_kkuc_group_add() to avoid a potential overflow.
The variable group is used in other kernelcomm functions
so it makes sense to convert the rest of the group
variables to unsigned int.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
In struct kkuc_reg, the "kr_data" is difficult to be freed
outside of libcfs, then it's better to change it to be
inline data instead of the data pointer.
Signed-off-by: Hongchao Zhang <hongchao.zhang@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6485
Reviewed-on: http://review.whamcloud.com/14638
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: frank zago <fzago@cray.com>
Reviewed-by: Robert Read <robert.read@intel.com>
Reviewed-by: Henri Doreau <henri.doreau@cea.fr>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Split the kernel comm header in libcfs into two new headers
to handle both kernel space and user space for the lustre layer.
This is broken out of the original patch 14270. The part covered
by this change is as follows:
The original libcfs_kernelcomm.h header is split into three parts:
* lustre_kernelcomm.h, a new header for the kernel parts;
* uapi_kernelcomm.h, a new header for the data structures shared
between userspace and kernelspace;
* lustreapi_internal.h receives the private liblustreapi prototypes.
Various names and filenames have been harmonized to *kernelcomm*.
The unused symbol KUC_FL_BLOCK has been removed.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/14270
Reviewed-by: Nathan Rutman <nathan.rutman@seagate.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Move the kernel portion from libcfs to obdclass. This code is
only used by lustre. This is broken out of the original patch
14270. The part covered by this change is as follows:
The original code in kernel_user_comm.c is split into two parts:
* obdclass/kernelcomm.c for the kernel part. filp_user_write() was
moved there, and linux-fs.c deleted;
* liblustreapi_kernelconn.c for the user part. The calls to CDEBUG
have been removed, and calls to CERROR have been transformed to
llapi_err_noerrno. The type lustre_kernelcomm has been removed and
replace by struct lustre_kernelcomm.
Various names and filenames have been harmonized to *kernelcomm*.
Signed-off-by: frank zago <fzago@cray.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-6245
Reviewed-on: http://review.whamcloud.com/14270
Reviewed-by: Nathan Rutman <nathan.rutman@seagate.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Associate copytool registration to a given MDC import so that
multiple mounts of the same filesystem do not lead to having the
copytool registered multiple time.
Signed-off-by: Henri Doreau <henri.doreau@cea.fr>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3882
Reviewed-on: http://review.whamcloud.com/7612
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Jinshan Xiong <jinshan.xiong@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
During the removal of the cfs wrappers the kg_sem semaphore
was handled incorrectly. We need to take a write lock when
writing data to the kkuc_groups. The libcfs_kkuc_group_foreach
needs to only take a read lock. This makes use match the
OpenSFS development branch.
Signed-off-by: James Simmons <jsimmons@infradead.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cleanup the a style issues for the lu_object_header_flags
enum by adding a comma for the last field. This is
broken out of patch http://review.whamcloud.com/6321.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2914
Reviewed-on: http://review.whamcloud.com/6321
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove libcfs function cfs_strncasecmp() since the kernel
already has its own strncasecmp(). Lastly remove from libcfs.h
cfs_get_blocked_sigs() since this function no longer exist.
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-3963
Reviewed-on: http://review.whamcloud.com/13070
Reviewed-by: John L. Hammond <john.hammond@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Patrick Farrell <paf@cray.com>
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Add the ability to debug LFSCK to libcfs. This is
broken out of patch http://review.whamcloud.com/6321.
Signed-off-by: Fan Yong <fan.yong@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-2914
Reviewed-on: http://review.whamcloud.com/6321
Reviewed-by: Alex Zhuravlev <alexey.zhuravlev@intel.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
For UMP and SMP machines the struct cfs_cpt_table are
defined differently. In the case handled by this patch
nodemask is defined as a integer for the UMP case and
as a pointer for the SMP case. This will cause a problem
for ost_setup which reads the nodemask directly. Instead
we create a UMP version of cfs_cpt_nodemask and use that
in ost_setup.
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4199
Reviewed-on: http://review.whamcloud.com/9219
Starting in 3.14 kernels nodemask_t was changed from a
a unsigned long to a linux bitmap so more than 32 cores
could be supported. Using set_bit in cfs_cpt_table_alloc
no longer compiles so this patch backports bits of the
node management function that use a linux bitmap back
end. Cleaned up libcfs bitmap.h to use the libcfs layers
memory allocation function. This was pulling in lustre
related code that was not defined.
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-4993
Reviewed-on: http://review.whamcloud.com/10332
Signed-off-by: James Simmons <uja.ornl@gmail.com>
Reviewed-by: Liang Zhen <liang.zhen@intel.com>
Reviewed-by: Li Xi <pkuelelixi@gmail.com>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Reviewed-by: Bob Glossman <bob.glossman@intel.com>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Remove IOC_LIBCFS_MEMHOG ioctl functionality as it is no longer needed thereby
making functions like - kportal_memhog_alloc(), kportal_memhog_free() and type -
struct libcfs_device_userstate unused.
Signed-off-by: Niranjan Dighe <ndighe@visteon.com>
Acked-by: James Simmons <jsimmons@infradead.org>
Reviewed-by: Andreas Dilger <andreas.dilger@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
