Paul Gortmaker [Mon, 27 Apr 2015 05:25:40 +0000 (01:25 -0400)]
rtl8192u: delete another embedded instance of generic reason codes
We have global copies of all these reason codes. We don't need local
copies. Worse is that these seem totally unused; a grep for some of
the fields comes up empty, and it still compiles after its complete
removal.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Paul Gortmaker [Mon, 27 Apr 2015 05:25:39 +0000 (01:25 -0400)]
rtl8192u: don't duplicate ieee80211 constants for status/auth/reason
These are all defined as a part of the standard and should not be
duplicated on a per-driver basis. Use the global ones and delete the
local ones.
In switching to <linux/ieee80211.h> we have to delete a local copy of
an identical struct that we prepped earlier to have identical field
names, and we add explicit include <...> where needed in order to
preserve getting a successful compile.
This isn't the totality of duplicated data removed, but it is a start.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Paul Gortmaker [Mon, 27 Apr 2015 05:25:38 +0000 (01:25 -0400)]
rtl8192u: align local ieee80211_wmm_ac_param struct fields with global
The <linux/ieee80211.h> and this local file both have a struct of the
same name. They also have the same field sizes and generally the
same fields, as can be seen here:
Here we just align the local field names with the main system one. Then
we can add an include of the system one and delete the local copy in one
smooth step in a follow-on commit.
Not that the replacement:
for i in `find . -name '*.[ch]'` ; do sed -i 's/ac_aci_acm_aifsn/aci_aifsn/g' $i ; done
for i in `find . -name '*.[ch]'` ; do sed -i 's/ac_ecwmin_ecwmax/cw/g' $i ; done
for i in `find . -name '*.[ch]'` ; do sed -i 's/ac_txop_limit/txop_limit/g' $i ; done
implicitly shows that only one of the three fields is currently used.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Paul Gortmaker [Mon, 27 Apr 2015 05:25:37 +0000 (01:25 -0400)]
rtl8192u: promote auth_mode to a full 8 bits
Currently LEAP is defined to two locally but the identically named
global constant is 128 in <linux/ieee80211.h>. In order for us to
switch over to using the global value, we need to adjust the local
storage which is currently not enough to hold the larger value.
This is now consistent with the similar struct used in
drivers/net/wireless/ipw2x00/libipw.h and other drivers.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Paul Gortmaker [Mon, 27 Apr 2015 05:25:36 +0000 (01:25 -0400)]
rtl8192u: don't trample on <linux/ieee80211.h> struct namespace
In order to start reducing the duplicated code/constants/macros in this
driver, we need to include <linux/ieee80211.h> to provide the defacto
versions. However this driver has structs with the same name as the
ones in the main include, so namespace collision prevents us from doing
step #1.
Since the structs actually differ in their respective fields, we can't
simply delete the local ones without impacting the runtime; a conversion
to use the global ones can be considered at a later date if desired.
Rename the ones here with a vendor specific prefix so that we won't have
the namespace collision, and hence can continue on with the cleanup.
Automated conversion done with:
for i in `find . -name '*.[ch]'` ; do \
sed -i 's/struct ieee80211_hdr/struct rtl_80211_hdr/g' $i ; \
done
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Paul Gortmaker [Mon, 27 Apr 2015 05:25:35 +0000 (01:25 -0400)]
rtl8712: don't duplicate ieee80211 constants for status/reason
These are all defined as a part of the standard and should not be
duplicated on a per-driver basis. Use the global ones and delete the
local ones.
It seems that ieee80211 was already included everywhere it was needed,
since no explicit include <...> were needed to be added in order to
preserve getting a successful compile.
This isn't the totality of duplicated data removed, but it is a start.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Paul Gortmaker [Mon, 27 Apr 2015 05:25:34 +0000 (01:25 -0400)]
rtl8188eu: don't duplicate ieee80211 constants for status/reason
These are all defined as a part of the standard and should not be
duplicated on a per-driver basis. Use the global ones and delete the
local ones.
Note that a couple of them had slight wording differences, things like
INVALID vs. NOT_VALID or similar, so they are aligned with the global
naming conventions here, as dictated by compile testing.
This isn't the totality of duplicated data removed, but it is a start.
Signed-off-by: Paul Gortmaker <paul.gortmaker@windriver.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: sm7xxfb: use framebuffer_alloc and release
use the standard framebuffer_alloc() and framebuffer_release() instead
of custom defined function. for making that change we had to change a
member of the private structure from a variable to pointer and had to
touch almost all places of the file. since fb was changed into a pointer
so all instance of "sfb->fb." has been changed into "sfb->fb->".
now we will get build warning about smtc_alloc_fb_info() and
smtc_free_fb_info() to be unused which will be removed in the next
patch of the series.
staging: fsl-mc: Remove redundant initalization of the .owner field
This patch removes the redundant static initialization of the .owner field from
this driver as it is being overidden by the call from the platform driver register
Signed-off-by: Hari Prasath <hgujulan@visteon.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
This patch fixes the following sparse warnings in r8192U_core.c:
CHECK drivers/staging/rtl8192u/r8192U_core.c
drivers/staging/rtl8192u/r8192U_core.c:3212:6: warning: symbol
'rtl819x_watchdog_wqcallback' was not declared. Should it be static?
drivers/staging/rtl8192u/r8192U_core.c:3276:6: warning: symbol
'watch_dog_timer_callback' was not declared. Should it be static?
drivers/staging/rtl8192u/r8192U_core.c:3282:5: warning: symbol
'_rtl8192_up' was not declared. Should it be static?
drivers/staging/rtl8192u/r8192U_core.c:3333:5: warning: symbol
'rtl8192_close' was not declared. Should it be static?
drivers/staging/rtl8192u/r8192U_core.c:3406:6: warning: symbol
'rtl8192_restart' was not declared. Should it be static?
drivers/staging/rtl8192u/r8192U_core.c:4618:6: warning: symbol
'rtl8192_irq_rx_tasklet' was not declared. Should it be static?
drivers/staging/rtl8192u/r8192U_core.c:4736:6: warning: symbol
'rtl8192_cancel_deferred_work' was not declared. Should it be static?
Ian Abbott [Fri, 1 May 2015 17:03:59 +0000 (18:03 +0100)]
staging: comedi: mite.h: remove "../comedi_pci.h" and make self-reliant
"mite.h" relies on a lot of things declared including "../comedi_pci.h",
but doesn't need anything in "../comedi_pci.h" itself. None of the
Comedi modules that include "mite.h" rely on it to include
"../comedi_pci.h" on their behalf.
Remove the inclusion of "../comedi_pci.h" from "mite.h", and add enough
other includes and (incomplete) `struct` declarations to make it compile
without warnings or errors even when included first in a .c file.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The Comedi driver modules build fine if the inclusion of <linux/slab.h>
is removed from "mite.h", so remove it. However, since "mite.c" calls
`kzalloc()`, include <linux/slab.h> there.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: ion: chunk_heap: use pr_debug for heap creation print
We're currently printing to the kernel log at `info' level when we
successfully create the chunk heap, but success messages should be done
at `debug' level. Fix this.
staging: dgnc: clean up allocation of ->channels[i]
Check if kzalloc fails in dgnc_tty_init() and if it does then free all
previously allocated ->channels[i] and set them to NULL. This makes the code
less error/bug prone because instead of needing programmers attention to add
checks everywhere we do that in one place. Also, remove a bogus comment and
check in the same loop because ->channels[i] isn't allocated anywhere else.
Finally, remove a unnecessary check if ->channels[i] is NULL in the next loop
because it can't be.
Tolga Ceylan [Sat, 2 May 2015 06:40:35 +0000 (23:40 -0700)]
drivers: staging: fbtft: fbtft-bus.c: Fix different address space warning on I/O mem
To fix sparse warning of incorrect type in assignment
(different address space), added annotation __iomem to
vmem8 and modified direct reads with ioread8().
Julia Lawall [Sun, 3 May 2015 13:21:44 +0000 (15:21 +0200)]
staging/lustre/ptlrpc: Replace OBD_CPT_ALLOC etc by kzalloc_node
Replace OBD_CPT_ALLOC, OBD_CPT_ALLOC_PTR, and OBD_CPT_ALLOC_GFP by
corresponding calls to kzalloc_node. The semantic patch for the
OBD_CPT_ALLOC case is as follows: (http://coccinelle.lip6.fr/).
Note that the original OBD macros would check if the cptab argument was
NULL and fall back on kzalloc in that case. Oleg Drokin argues that this
test is not needed because the code containing these calls is only invoked
after initialization has been completed, in which case the possible cptab
arguments are not NULL.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Julia Lawall [Sun, 3 May 2015 13:21:43 +0000 (15:21 +0200)]
drivers: staging: lustre: lustre: Replace OBD_CPT_ALLOC etc by kzalloc_node
Replace OBD_CPT_ALLOC, OBD_CPT_ALLOC_PTR, and OBD_CPT_ALLOC_GFP by
corresponding calls to kzalloc_node. The semantic patch for the
OBD_CPT_ALLOC case is as follows: (http://coccinelle.lip6.fr/).
Note that the original OBD macros would check if the cptab argument was
NULL and fall back on kzalloc in that case. Oleg Drokin argues that this
test is not needed because the code containing these calls is only invoked
after initialization has been completed, in which case the possible cptab
arguments are not NULL.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Julia Lawall [Sun, 3 May 2015 13:21:42 +0000 (15:21 +0200)]
staging: lustre: ptlrpc: Replace OBD_CPT_ALLOC etc by kzalloc_node
Replace OBD_CPT_ALLOC, OBD_CPT_ALLOC_PTR, and OBD_CPT_ALLOC_GFP by
corresponding calls to kzalloc_node. The semantic patch for the
OBD_CPT_ALLOC case is as follows: (http://coccinelle.lip6.fr/).
Note that the original OBD macros would check if the cptab argument was
NULL and fall back on kzalloc in that case. Oleg Drokin argues that this
test is not needed because the code containing these calls is only invoked
after initialization has been completed, in which case the possible cptab
arguments are not NULL.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Julia Lawall [Fri, 1 May 2015 19:37:53 +0000 (21:37 +0200)]
staging: lustre: obdclass: obd_config: remove unneeded null test before free
Kfree can cope with a null argument, so drop null tests.
The semantic patch that identifies this issue is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@@ expression ptr; @@
- if (ptr != NULL)
kfree(ptr);
// </smpl>
The first part of the patch introduces new labels to avoid unnecessary
calls to kfree. In addition, lprof->lp_md is always null in the cleanup
code at the end of the function, so that kfree is just dropped.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Julia Lawall [Fri, 1 May 2015 19:38:01 +0000 (21:38 +0200)]
Staging: lustre: llite: file: remove unneeded null test before free
Kfree can cope with a null argument, so drop null tests.
The semantic patch that finds this issue is as follows:
(http://coccinelle.lip6.fr/)
// <smpl>
@@ expression ptr; @@
- if (ptr != NULL)
kfree(ptr);
// </smpl>
In the first case, llss can never be null at the point of the kfree anyway.
In the second set of changes, the gotos are either eliminated, when no
freeing is needed (hss failure case), or rewritten so that there is no call
to free on data that has not yet been allocated (free_hss goto case).
There is no goto at which attr is not null, so the out label is simply
dropped.
Signed-off-by: Julia Lawall <Julia.Lawall@lip6.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/lustre/lustre/llite/dir.c:1440:3-8: WARNING: NULL check before freeing functions like kfree, debugfs_remove, debugfs_remove_recursive or usb_free_urb is not needed. Maybe consider reorganizing relevant code to avoid passing NULL values.
NULL check before some freeing functions is not needed.
Based on checkpatch warning
"kfree(NULL) is safe this check is probably not required"
and kfreeaddr.cocci by Julia Lawall.
drivers/staging/lustre/lustre/llite/llite_lib.c:989:2-7: WARNING: NULL check before freeing functions like kfree, debugfs_remove, debugfs_remove_recursive or usb_free_urb is not needed. Maybe consider reorganizing relevant code to avoid passing NULL values.
drivers/staging/lustre/lustre/llite/llite_lib.c:991:2-7: WARNING: NULL check before freeing functions like kfree, debugfs_remove, debugfs_remove_recursive or usb_free_urb is not needed. Maybe consider reorganizing relevant code to avoid passing NULL values.
NULL check before some freeing functions is not needed.
Based on checkpatch warning
"kfree(NULL) is safe this check is probably not required"
and kfreeaddr.cocci by Julia Lawall.
drivers/staging/lustre/lustre/lov/lov_request.c:74:3-8: WARNING: NULL check before freeing functions like kfree, debugfs_remove, debugfs_remove_recursive or usb_free_urb is not needed. Maybe consider reorganizing relevant code to avoid passing NULL values.
NULL check before some freeing functions is not needed.
Based on checkpatch warning
"kfree(NULL) is safe this check is probably not required"
and kfreeaddr.cocci by Julia Lawall.
drivers/staging/lustre/lustre/lov/lov_obd.c:574:3-8: WARNING: NULL check before freeing functions like kfree, debugfs_remove, debugfs_remove_recursive or usb_free_urb is not needed. Maybe consider reorganizing relevant code to avoid passing NULL values.
NULL check before some freeing functions is not needed.
Based on checkpatch warning
"kfree(NULL) is safe this check is probably not required"
and kfreeaddr.cocci by Julia Lawall.
Ian Abbott [Tue, 21 Apr 2015 12:18:11 +0000 (13:18 +0100)]
staging: comedi: move COMEDI_SRF_... macros to "comedi_fops.c"
The `COMEDI_SRF_...` macros define flag combinations in the `runflags`
member of `struct comedi_subdevice`. They are only used directly in
"comedi_fops.c", so move them to there.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ian Abbott [Tue, 21 Apr 2015 12:18:10 +0000 (13:18 +0100)]
staging: comedi: wrap COMEDI_SRF_FREE_SPRIV usage
The `COMEDI_SRF_FREE_SPRIV` flag in the `runflags` member of `struct
comedi_subdevice` indicates that the memory pointed to by the `private`
member can be automatically freed by the comedi core on subdevice
clean-up (when the low-level comedi device is being "detached"). the
flag doesn't really belong in `runflags`, but it was somewhere
convenient to keep it without having to add a new member to the
structure.
Rather than access the `COMEDI_SRF_FREE_SPRIV` flag directly, use some
new wrapper functions:
* comedi_can_auto_free_spriv(s) - checks whether the subdevice's
`s->private` points to memory that can be freed automatically.
* comedi_set_spriv_auto_free(s) - marks the subdevice as having a
`s->private` that points to memory that can be freed automatically.
Export `comedi_set_spriv_auto_free()` for use by the low-level comedi
driver modules, in particular the "amplc_dio200_common" module.
Signed-off-by: Ian Abbott <abbotti@mev.co.uk> Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
serial2002_setup_subdevices() initializes each subdevice based on the
config read from the attached serial device. Part of this initialization
is to setup the subdevice range_table_list for the non digital subdevices.
The range_table_list is allocated only when a 'range' is passed to the
functions. Each channel of the subdevice then has it's 'range' initialized
and that range is added to the range_table_list.
The logic of this function works but causes Coverity complain about an
Explicit null dereference of the allocated 'range_table_list'. Add a check
for the 'range_table_list' to quiet the Coverity issue.
Reported-by: coverity (CID 1011632) Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
ni_write_caldac() checks the boardinfo 'caldac' array to determine what
caldac is used for a given 'addr'. It then calculates the 'bitstring' and
number of 'bits' used to write a value to that caldac address.
After checking the caldac array, if the number of bits is 0 there is no
caldac associated with the address. If this happens we shouldn't try
writing to the non-existing caldac.
Reported-by: coverity (CID 1192116) Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: comedi_bond: fix 'b_mask' calc in bonding_dio_insn_bits()
'b_chans' may be a valud up to 32. 'b_mask' is an unsigned int and a left shift of
more than 31 bits has undefined behavior. Fix the calc so it works correctly with
a 'b_chans' of 32..
Reported-by: coverity (CID 1192244) Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
both the misc_deregister(), parport_release() and
parport_unregister_device() is there in the module_exit function also.
detach is called from parport_unregister_driver() and by the time
detach executes misc_deregister(), parport_release() and
parport_unregister_device() has already executed marking
keypad_initialized and lcd.initialized as false. so this part of the
code will never execute.
projects / linux-beck.git / log