navin patidar [Sat, 3 May 2014 11:45:26 +0000 (17:15 +0530)]
staging: rtl8188eu: Remove 'irp_pending' from struct recv_buf
irp_pending is initialized to false inside rtw_os_recvbuf_resource_alloc()
and value of irq_pending never changed after that, so 'if (!precvbuf->irp_pending)'
inside rtw_os_read_port() function will be always true.
staging/rtl8192e: Erroneous return codes (types and values)
This function returns a bool, that is supposed to be false when something
goes wrong. It's assumed this way by its lone calling function (which is
SetRFPowerState8190(), line 1445 of rtl8192e/rtl8192e/r8192E_phy.c)
Despite of this, this procedure returns non-null enumerations values or
negative codes instead. This patch fixes this.
Signed-off-by: Dominique van den Broeck <domdevlin@free.fr> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Ian Abbott [Fri, 2 May 2014 12:50:14 +0000 (13:50 +0100)]
staging: comedi: remove async parameter from resize_async_buffer()
The `struct comedi_async *async` parameter of `resize_async_buffer()` is
redundant as its value can be easily derived from the `struct
comedi_subdevice *s` parameter as `s->async`. Remove the parameter.
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 [Fri, 2 May 2014 12:50:13 +0000 (13:50 +0100)]
staging: comedi: remove redundant pointer dereference in comedi_poll()
`s->async->subdevice` in `comedi_poll()` points to the same `struct
comedi_subdevice` as `s`, so the double pointer reference is redundant.
Just use `s`.
(`s->async->subdevice` is initialized by
`__comedi_device_postconfig_async()` in
"drivers/staging/comedi/drivers.c" and doesn't change.)
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 [Thu, 1 May 2014 16:38:24 +0000 (17:38 +0100)]
staging: comedi: amplc_dio200_common: correct bound on counter mode
For the mode configured by the `INSN_CONFIG_SET_COUNTER_MODE` comedi
instruction for the counter subdevice channels supported by this module,
the upper bound should be `I8254_MODE5 | I8254_BCD` ((5 << 1) | 1)
rather than `I8254_MODE5 | I8254_BINARY` ((5 << 1) | 0). Fix it.
Reported-by: Hartley Sweeten <HartleyS@visionengravers.com> 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 [Thu, 1 May 2014 16:38:23 +0000 (17:38 +0100)]
staging: comedi: drivers: correct mode check for i8254_set_mode()
The upper bound check on the `mode` parameter of `i8254_set_mode()` and
`i8254_mm_set_mode()` is incorrect. The `mode` parameter value consists
of a mode number in the range 0 to 5 in bits 3..1 {represented by the
constants `I8254_MODE0` (0 << 1) through to `I8254_MODE5` (2 << 1)} ORed
with a BCD/binary flag in bit 0 {represented by the constants
`I8254_BINARY` (0) and `I8254_BCD` (1)}. The maximum allowed value
ought to be `I8254_MODE5 | I8254_BCD` ((5 << 1) | 1), but is currently
`I8254_MODE5 | I8254_BINARY` ((5 << 1) | 0). Fix it.
None of the comedi drivers use `I8254_BCD` but some of the low-level
drivers allow user-space to configure the counter mode, so all legal
values ought to be allowed. However, it's pretty unlikely anyone would
want to set the counters to count in BCD (binary coded decimal) so the
bug is not that significant.
Reported-by: Hartley Sweeten <HartleyS@visionengravers.com> 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>
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
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>
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
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>
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
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>
As pointed out by Ian Abbott, the i8254_load() function does not
use the I8254_MODE* values to specify the "mode".
The labpc_counter_load() function in this driver is passed an
I8254_MODE* value so we need to use the i8254_set_mode() and
i8254_write() helpers instead of i8254_load() to program the
timers.
The calls to labpc_counter_load() will not fail so change the
return to void and remove all the unnecessary error handling.
Similarly, change the return type of labpc_counter_set_mode() to
void and remove the unnecessary error handling.
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>
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
The das800_set_frequency() function will not fail. For aesthetics
change the return to a void and remove the error handling.
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>
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
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>
As pointed out by Ian Abbott, the i8254_load() function does not
use the I8254_MODE* values to specify the "mode".
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
Also, use the CR_CHAN() macro to make sure the only the channel is
used in the insh->chanspec to specify the counter number.
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>
As pointed out by Ian Abbott, the i8254_load() function does not
use the I8254_MODE* values to specify the "mode".
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
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>
As pointed out by Ian Abbott, the i8254_load() function does not
use the I8254_MODE* values to specify the "mode".
Refactor the 8254 timer programming to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows
us to use the I8254_MODE* defines to clarify the code.
For aesthetics, factor the timer programming out as a separate
function.
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>
This private data member is set when an analog input async command is
started by apci3120_cyclic_ai() and cleared when the command finishes
or is canceled.
Rename the CamelCase member to 'ai_running' and convert it to a
bit-field flag.
Remove the unnecessary clearing of the flag before or after calling
apci3120_cancel().
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: addi_common.h: remove 'b_AiContinuous' from private data
This member of the private data can be determined by checking the cmd->stop_src.
Do that instead.
Fix the b_AiContinuous check in apci3120_ai_cmd(). The (*do_cmdtest) validates
that the cmd->stop_arg os >= 1 for TRIG_COUNT and 0 for TRIG_NONE. The correct
test to set this flag is (cmd->stop_src == TRIG_NONE).
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: addi_common.h: remove 'ui_AiNbrofScans' from private data
This member of the private data is a copy of the cmd->stop_arg. Use that
instead.
Fix the b_AiContinuous check in apci3120_ai_cmd(). The (*do_cmdtest) validates
that the cmd->stop_arg os >= 1 for TRIG_COUNT and 0 for TRIG_NONE. The correct
test to set this flag is (cmd->stop_src == TRIG_NONE).
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: addi_common.h: remove 'ui_AiScanLength' from private data
This member of the private data is a copy of the cmd->scan_end_arg. Use that
instead.
Use a local variable in apci3120_cyclic_ai() for the DMA 'scan_bytes', which
is the cmd->scan_end_arg * 2. Replace the open-coded '2' with sizeof(short).
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: adl_pci9111: tidy up (*do_cmdtest) Step 4
This step in the (*do_cmdtest) fixes the comedi_cmd arguments, usually
for TRIG_TIMER sources.
Tidy up this step to clarify the code and remove the extra local
variables. All the arguments are unsigned long so change the 'tmp'
local variable type and rename it for aesthetics.
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: 8253.h: remove the unused i8253_cascade_ns_to_timer_*()
None of the comedi drivers use the i8253_cascade_ns_to_timer_2div_old()
or i8253_cascade_ns_to_timer_power() helpers to calculate the cascaded
divisors. Remove them to avoid any confusion.
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>
All of the comedi drivers use the i8253_cascade_ns_to_timer macro
when calculating the divisors for the cascaded timers. This macro
just causes the i8253_cascade_ns_to_timer_2div() inline to be used.
Rename i8253_cascade_ns_to_timer_2div() and remove the macro.
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>
The analog input async command can use the pacer for the scan_begin_src
or the convert_src. The (*do_cmdtest) calculates the divisors when
validating the cmd argument.
There is no reason to recalc the divisors in the (*do_cmd). Just use the
values from the private data.
For aesthetics, rename the setup_counters() function so it has namespace
associated with the driver. Refactor the function to use the values from
the private data and absorb das1800_set_frequency() to clarify the code.
Refactor the function to use the i8254_set_mode() and i8254_write()
helpers instead of i8254_load(). This allows us to use the I8254_* defines
when setting the mode to clarify the code.
This function will not fail so change the return type to void to simplify
the (*do_cmd) a bit.
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>
The analog input async command can use the pacer for the convert_src. The
(*do_cmdtest) calculates the divisors when validating the cmd argument.
There is no reason to recalc the divisors in the (*do_cmd). Just use the
values from the private data.
Refactor das16m1_set_pacer() to use the i8254_set_mode() and i8254_write()
helpers instead of i8254_load(). This allows us to use the I8254_* defines
when setting the mode to clarify the code.
Tidy up das16m1_cmd_exec() a bit. The pacer only needs to be set when the
convert_src is TRIG_TIMER.
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: cb_pcidas: don't calc ao pacer divisors twice
The analog output async command can use the pacer for the scan_begin_src.
The (*do_cmdtest) calculates the divisors when validating the cmd argument.
There is no reason to recalc the divisors in the (*do_cmd). Just use the
values from the private data.
For aesthetics, factor out the code that loads the counters to match the
analog input. Refactor the code to use the i8254_set_mode() and i8254_write()
helpers instead of i8254_load(). This allows us to use the I8254_* defines
when setting the mode to clarify the code.
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: cb_pcidas: don't calc ai pacer divisors twice
The analog input async command can use the pacer for the scan_bagin_src
or the convert_src. The (*do_cmdtest) validates that only one of these
sources has the TRIG_TIMER selected and calculates the divisors when
validating the cmd argument.
There is no reason to recalc the divisors in the (*do_cmd). Just use
the values from the private data.
Also, refactor cb_pcidas_load_counters() to use the i8254_set_mode()
and i8254_write() helpers instead of i8254_load(). This allows us to
use the I8254_* defines when setting the mode to clarify the code.
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>
This define enables code that checks for analog input channel dropout
when reading samples. The define is enabled so we might as well always
enable the code and remove the define.
Factor out the common channel dropout detect code as a helper function
and cleanup the code.
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: adv_pci1710: remove local var in pci171x_ai_cmd()
The local variable 'mode' is not necessary. We can determine the mode
by checking the cmd->convert_src and cmd->start_src. Do this instead
to clarify the code.
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: adv_pci1710: only calc the pacer divisors once
When the cmd->convert_src == TRIG_TIMER the divisors needed to
generate the pacer time are calculated in the (*do_cmdtest) to
validate the cmd->convert_arg. The core always does the (*do_cmdtest)
before the (*do_cmd) so there is no reason to recalc the divisors.
Save the calculated divisors in the private data as 'next_divisor[12]'.
The (*do_cmd) then transfers them to the private data 'divisor[12]' so
that they can be used to set the timer for the command immediately or
later when the cmd->start_src is TRIG_EXT (mode 2 in this driver).
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: adv_pci1710: don't check the chanlist twice
The chanlist is checked in Step 5 of the (*do_cmdtest) there is no
reason to check it again in the (*do_cmd). The only reason its done
again is to get the actual 'seglen', the non-repeating length of the
chanlist.
Save the 'seglen' found by pci171x_ai_check_chanlist() in the private
data and use that in the (*do_cmd). Rename the private data member to
clarify it. Also, remove the unused 'act_chanlist_pos' member from the
private data.
Refactor the error handling in pci171x_ai_check_chanlist() so it returns
and errno for failure and 0 for success.
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>
Chase Southwood [Sat, 3 May 2014 07:49:15 +0000 (02:49 -0500)]
staging: comedi: addi_apci_1564: remove use of devpriv->s_EeParameters
This driver no longer reads the eeprom to find the board specific data,
all the necessary data is in the boardinfo. Use the boardinfo directly
instead of passing through devpriv->s_EeParameters.
Signed-off-by: Chase Southwood <chase.southwood@gmail.com> Cc: Ian Abbott <abbotti@mev.co.uk> Reviewed-by: H Hartley Sweeten <hsweeten@visionengravers.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging: comedi: addi_apci_1564: remove unnecessary info from boardinfo
The i_IorangeBase1, i_PCIEeprom, and pc_EepromChip data in the boardinfo
was only needed to work out the usage of the PCI bars. Now that that is
squared away, this info is no longer needed and can be removed.
Signed-off-by: Chase Southwood <chase.southwood@gmail.com> Reviewed-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: addi_apci_1564: board has 32 digital inputs
This board always has 32 digital inputs. Remove the test when
initializing the subdevice.
Also, since this board is the only one supported by this driver,
remove the boardinfo about the digital inputs and just use the
data directly in the subdevice init.
Signed-off-by: Chase Southwood <chase.southwood@gmail.com> Reviewed-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: remove duplicate pointer assignments in attach functions
Some board pointer are assigned twice via comedi_board() in the comedi low
level driver attach functions. Remove the duplicate assignment from the
variable definition where the pointer is not used anyway until assigned later
in the function when dev->board_ptr, that comedi_board() relies on, is setup
correctly.
Signed-off-by: Christian Engelmayer <cengelma@gmx.at> Reviewed-by: Ian Abbott <abbotti@mev.co.uk> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Carpenter [Wed, 30 Apr 2014 10:06:03 +0000 (13:06 +0300)]
staging: rtl8192u: cleanup a switch statement
The white space was all wrong here. The case statements were indented
too far. The if else blocks weren't indented at all. There was a break
statement aligned with the else block and it confused my static checker
because "were curly braces intended" so that the break statement was
only on the else side? Also I removed some commented out code.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Dan Carpenter [Fri, 4 Apr 2014 21:18:55 +0000 (00:18 +0300)]
staging: r8188eu: cleanup by using "len" consistently
"*(p + 1)" and "len" are the same thing. For reviewers who don't know
that, then this code is worrying because we cap "len", but pass
"*(p + 1)" to memcpy().
I have changed the code to use "len" throughout.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
staging/lustre/llite: Fix unused variable warning if !CONFIG_FS_POSIX_ACL
If CONFIG_FS_POSIX_ACL=n:
drivers/staging/lustre/lustre/llite/xattr.c: In function 'll_setxattr_common':
drivers/staging/lustre/lustre/llite/xattr.c:115:27: warning: unused variable 'rce' [-Wunused-variable]
drivers/staging/lustre/lustre/llite/rw26.c: In function 'll_direct_IO_26':
drivers/staging/lustre/lustre/llite/rw26.c:383:2: warning: format '%lu' expects argument of type 'long unsigned int', but argument 6 has type 'ssize_t' [-Wformat]
drivers/staging/lustre/lustre/llite/rw26.c:383:2: warning: format '%lu' expects argument of type 'long unsigned int', but argument 10 has type 'ssize_t' [-Wformat]
Join the quoted string split across lines to fix a checkpatch warning while
we're at it.
staging/lustre/libcfs: Fix build errors for !CONFIG_SMP
When CONFIG_SMP=n:
drivers/staging/lustre/include/linux/libcfs/linux/linux-mem.h:58:31: fatal error: libcfs/libcfs_cpu.h: No such file or directory
drivers/staging/lustre/lustre/libcfs/libcfs_cpu.c:78:1: error: redefinition of 'cfs_cpt_table_print'
drivers/staging/lustre/include/linux/libcfs/libcfs_cpu.h:109:1: note: previous definition of 'cfs_cpt_table_print' was here
Dan Carpenter [Mon, 28 Apr 2014 10:58:58 +0000 (13:58 +0300)]
staging: lustre: integer overflow in obd_ioctl_is_invalid()
The obd_ioctl_getdata() function caps "data->ioc_len" at
OBD_MAX_IOCTL_BUFFER and then calls this obd_ioctl_is_invalid() to check
that the other values inside data are valid.
There are several lengths inside data but when they are added together
they must not be larger than "data->ioc_len". The checks against
"(data->ioc_inllen1 > (1<<30))" are supposed to ensure that the addition
does not have an integer overflow. But "(1<<30) * 4" actually can
overflow 32 bits, so the checks are insufficient.
I have changed it to "> OBD_MAX_IOCTL_BUFFER" instead.
Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Pointer 'ni' checked for NULL at line 1569 may be passed to
function and may be dereferenced there by passing argument 1 to
function 'lnet_ni_notify_locked' at line 1621.
found by Klocwork Insight tool
staging/lustre: Replace jobid acquiring with per node setting
Insted of meddling directly in process environment variables
(which is also not possible on certain platforms due to not exported
symbols), create jobid_name proc file to represent this info
(to be filled by job scheduler epilogue).
Will Deacon [Wed, 23 Apr 2014 16:52:52 +0000 (17:52 +0100)]
word-at-a-time: avoid undefined behaviour in zero_bytemask macro
The asm-generic, big-endian version of zero_bytemask creates a mask of
bytes preceding the first zero-byte by left shifting ~0ul based on the
position of the first zero byte.
Unfortunately, if the first (top) byte is zero, the output of
prep_zero_mask has only the top bit set, resulting in undefined C
behaviour as we shift left by an amount equal to the width of the type.
As it happens, GCC doesn't manage to spot this through the call to fls(),
but the issue remains if architectures choose to implement their shift
instructions differently.
An example would be arch/arm/ (AArch32), where LSL Rd, Rn, #32 results
in Rd == 0x0, whilst on arch/arm64 (AArch64) LSL Xd, Xn, #64 results in
Xd == Xn.
Rather than check explicitly for the problematic shift, this patch adds
an extra shift by 1, replacing fls with __fls. Since zero_bytemask is
never called with a zero argument (has_zero() is used to check the data
first), we don't need to worry about calling __fls(0), which is
undefined.
Cc: <stable@vger.kernel.org> Cc: Victor Kamensky <victor.kamensky@linaro.org> Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
This merges the patch to fix possible loss of dirty bit on munmap() or
madvice(DONTNEED). If there are concurrent writers on other CPU's that
have the unmapped/unneeded page in their TLBs, their writes to the page
could possibly get lost if a third CPU raced with the TLB flush and did
a page_mkclean() before the page was fully written.
Admittedly, if you unmap() or madvice(DONTNEED) an area _while_ another
thread is still busy writing to it, you deserve all the lost writes you
could get. But we kernel people hold ourselves to higher quality
standards than "crazy people deserve to lose", because, well, we've seen
people do all kinds of crazy things.
So let's get it right, just because we can, and we don't have to worry
about it.
* safe-dirty-tlb-flush:
mm: split 'tlb_flush_mmu()' into tlb flushing and memory freeing parts
Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs
Pull btrfs fixes from Chris Mason.
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/linux-btrfs:
Btrfs: limit the path size in send to PATH_MAX
Btrfs: correctly set profile flags on seqlock retry
Btrfs: use correct key when repeating search for extent item
Btrfs: fix inode caching vs tree log
Btrfs: fix possible memory leaks in open_ctree()
Btrfs: avoid triggering bug_on() when we fail to start inode caching task
Btrfs: move btrfs_{set,clear}_and_info() to ctree.h
btrfs: replace error code from btrfs_drop_extents
btrfs: Change the hole range to a more accurate value.
btrfs: fix use-after-free in mount_subvol()