aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2015-04-17drivers/rtc/rtc-ds1685.c: remove .owner assignment from platform_driverJoshua Kinard1-1/+0
The rtc driver core now sets the platform_driver 'owner' property, so remove the assignment from the DS1685 driver. Fixes: aaaf5fbf56f1: "rtc: add driver for DS1685 family of real time clocks" Signed-off-by: Joshua Kinard <[email protected]> Reported-by: kbuild test robot <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17drivers/rtc/rtc-ab-b5ze-s3.c: constify struct regmap_configKrzysztof Kozlowski1-1/+1
The regmap_config struct may be const because it is not modified by the driver and regmap_init() accepts pointer to const. Signed-off-by: Krzysztof Kozlowski <[email protected]> Cc: Alessandro Zummo <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17drivers/rtc/rtc-s3c.c: delete duplicate clock controlChanwoo Choi1-124/+39
The current functions in s3c-rtc driver execute clk_enable/disable() to control clocks and some functions execute s3c_rtc_alarm_clk_enable() unnecessarily. So this patch deletes the duplicate clock control and spilts s3c_rtc_alarm_clk_enable() out as s3c_rtc_enable_clk()/s3c_rtc_disable_clk() to improve readability. Signed-off-by: Chanwoo Choi <[email protected]> Cc: Alessandro Zummo <[email protected]> Cc: Kukjin Kim <[email protected]> Cc: Inki Dae <[email protected]> Cc: Kyungmin Park <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17rtc: hctosys: do not treat lack of RTC device as errorFloris Bos1-1/+1
When using device trees on the ARM platform, it is not certain at compile time whether or not the system will have a RTC. If one enables CONFIG_HCTOSYS just in case the system booted has a RTC, and it turns out not to be, this will result in a big fat "unable to open rtc device" error being printed to console, even when "quiet" is set in the kernel cmdline. Fix this by outputting the message with loglevel info instead. Signed-off-by: Floris Bos <[email protected]> Cc: Alessandro Zummo <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17drivers/rtc/rtc-x1205.c: use sign_extend32() for sign extensionMartin Kepplinger1-2/+2
Despite its name, sign_extend32() is safe to use for 8 bit types too. (See https://lkml.org/lkml/2015/1/18/289). Signed-off-by: Martin Kepplinger <[email protected]> Cc: Alessandro Zummo <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17drivers/rtc/rtc-em3027.c: add device tree supportPeter Robinson1-0/+11
Set the of_match_table for this driver so that devices can be described in the device tree. This device is used in the Trimslice and is already defined in the Trimslice device tree. Signed-off-by: Peter Robinson <[email protected]> Cc: Mike Rapoport <[email protected]> Cc: Alessandro Zummo <[email protected]> Cc: Grant Likely <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17rtc: stmp3xxx: use optional crystal in low power statesUwe Kleine-König2-6/+65
The rtc's status register allows to determine if a 32k crystal is connected to keep the rtc running in low power states provided the corresponding fuse bits were blown correctly during production. (In case they were not, the right frequency can be stated in the device tree.) If there is no such crystal available force the 24 MHz XTAL clock to keep running to retain the right date and time. Otherwise use the crystal to save some power. It would be nice to only switch to the crystal when the XTAL clock is about to be disabled and keep the crystal off when unneeded because XTAL is always on while the chip is powered on. But as sudden power loss isn't detectable this is not save. Signed-off-by: Uwe Kleine-König <[email protected]> Cc: Alessandro Zummo <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17drivers/rtc/rtc-pcf8563.c: simplify return from functionRobert Kmiec1-6/+1
This commit does not change any logic here. It just makes the code easier to read. This is how it looked like: If err != 0 return err; else return 0; Signed-off-by: Robert Kmiec <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17kasan: Makefile: shut up warnings if CONFIG_COMPILE_TEST=yAndrey Ryabinin1-2/+6
It might be annoying to constantly see this: scripts/Makefile.kasan:16: Cannot use CONFIG_KASAN: -fsanitize=kernel-address is not supported by compiler while performing allmodconfig/allyesconfig build tests. Disable this warning if CONFIG_COMPILE_TEST=y. Signed-off-by: Andrey Ryabinin <[email protected]> Cc: Michal Marek <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17fs/binfmt_misc.c: simplify entry_status()Rasmus Villemoes1-21/+9
sprintf() reliably returns the number of characters printed, so we don't need to ask strlen() where we are. Also replace calling sprintf("%02x") in a loop with the much simpler bin2hex(). [[email protected]: it's odd to include kernel.h after everything else] Signed-off-by: Rasmus Villemoes <[email protected]> Cc: Al Viro <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: avoid "spaces required around that ':'" false positiveJoe Perches1-0/+8
Since commit 1f65f947a6a8 ("checkpatch: add checks for question mark and colon spacing") back in 2008, checkpatch has reported false positive for asm volatile uses of "::" checkpatch thinks colons should always have spaces around it. Add an exception for colons with colons on either side for this valid asm volatile (and c++) use. Signed-off-by: Joe Perches <[email protected]> Reported-by: Yehuda Yitschak <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: fix --fix use with a patch of multiple filesJoe Perches1-1/+1
If a patch touches multiple files, the --fix and --fix-inplace option doesn't keep the proper line count and makes the new patch file not able to be applied via bad offset line numbers when lines are added or deleted by the --fix option. Dunno how that extra backslash snuck in there. Signed-off-by: Joe Perches <[email protected]> Cc: Andy Whitcroft <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17errno.h: Improve ENOSYS's commentAndy Lutomirski1-1/+10
ENOSYS is the mechanism used by user code to detect whether the running kernel implements a given system call. It should not be returned by anything except an unimplemented system call. Unfortunately, it is rather frequently used in the kernel to indicate that various new functions of existing system calls are not implemented. This should be discouraged. Improve the comment in errno.h to help clarify ENOSYS's purpose. Signed-off-by: Andy Lutomirski <[email protected]> Cc: Pavel Machek <[email protected]> Cc: Michael Kerrisk <[email protected]> Cc: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch.pl: new instances of ENOSYS are errorsAndy Lutomirski1-0/+8
ENOSYS means that a nonexistent system call was called. We have a bad habit of using it for things like invalid operations on otherwise valid syscalls. We should avoid this in new code. Pervasive incorrect usage of ENOSYS came up at the kernel summit ABI review discussion. Let's see if checkpatch can help. I'll submit a separate patch for include/uapi/asm-generic/errno.h. Signed-off-by: Andy Lutomirski <[email protected]> Cc: Pavel Machek <[email protected]> Cc: Michael Kerrisk <[email protected]> Cc: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: add a test for const with __read_mostly usesJoe Perches1-0/+10
const objects shouldn't be __read_mostly. They are read-only. Marking these objects as __read_mostly causes section conflicts with LTO linking. So add a test to try to avoid this issue. Signed-off-by: Joe Perches <[email protected]> Cc: Andy Whitcroft <[email protected]> Cc: Andi Kleen <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: improve operator spacing checkSam Bobroff1-1/+1
Code such as: x = timercmp(&now, &end, <); Will currently trigger a checkpatch error. e.g. ERROR: spaces required around that '<' This is because the "Ignore operators passed as parameters" check looks only for a comma following the operator. Improve the check by also looking for a close parenthesis. Signed-off-by: Sam Bobroff <[email protected]> Cc: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: add 'Prefer ARRAY_SIZE" testJoe Perches1-0/+13
Add a test for sizeof(foo)/sizeof(foo[0]) that could be ARRAY_SIZE(foo). Signed-off-by: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: add uart_ops to normally const structsJoe Perches1-0/+1
Add another struct to the list of normally const struct types Signed-off-by: Joe Perches <[email protected]> Cc: Andy Whitcroft <[email protected]> Cc: Maxime Coquelin <[email protected]> Cc: Greg Kroah-Hartman <[email protected]> Cc: Jiri Slaby <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: add #define foo "string" long line exceptionJoe Perches1-2/+3
There are #defines with long string constants like: #define foo "some really long string > 80 columns" Add a long line exception for them. Miscellanea: Use the $String variable for slightly better readability Signed-off-by: Joe Perches <[email protected]> Reported-by: Madalin-Cristian Bucur <[email protected]> Cc: Andy Whitcroft <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch, SubmittingPatches: suggest line wrapping commit messages at 75 ↵Joe Perches2-2/+11
columns Commit messages lines are sometimes overly long. Suggest line wrapping at 75 columns so the default git commit log indentation of 4 plus the commit message text still fits on an 80 column screen. Add a checkpatch test for long commit messages lines too. Signed-off-by: Joe Perches <[email protected]> Cc: David Miller <[email protected]> Cc: Jonathan Corbet <[email protected]> Cc: Ian Morris <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: don't ask for asm/file.h to linux/file.h unconditionallyFabian Frederick1-7/+11
Currently checkpatch warns when asm/file.h is included and linux/file.h exists. That conversion can be made when linux/file.h includes asm/file.h which is not always the case.(See signal.h) Signed-off-by: Fabian Frederick <[email protected]> Cc: Andy Whitcroft <[email protected]> Acked-by: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: add test for repeated const usesJoe Perches1-0/+18
Using 'const <type> const *' is generally meant to be written 'const <type> * const'. Add a test for the miswritten form. Signed-off-by: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: improve return negative errno checkJoe Perches1-3/+3
Add a few conditions to the test to find return (ERRNO); Make the output message a bit less cryptic too. Signed-off-by: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: match more world writable permissionsJoe Perches1-2/+10
Currently checkpatch will fuss if one uses world writable settings in debugfs files and DEVICE_ATTR uses by testing S_IWUGO but not testing S_IWOTH, S_IRWXUGO or S_IALLUGO. Extend the check to catch all cases exporting world writable permissions including octal values. [[email protected]: remove stray $] Signed-off-by: Joe Perches <[email protected]> Original-patch-by: Nicholas Mc Guire <[email protected]> Cc: Guenter Roeck <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: add optional --codespell dictionary to find more typosJoe Perches1-4/+34
If a codespell dictionary exists, use it if desired. default is off, maybe it could be turned on later. codespell's dictionary format allows multiple possible corrections, ignore that for now and only use the first suggestion. Also add \b to spelling test so that consecutive misspelled words are found properly. Signed-off-by: Joe Perches <[email protected]> Cc: Andy Whitcroft <[email protected]> Cc: Kees Cook <[email protected]> Cc: Masanari Iida <[email protected]> Cc: Lucas De Marchi <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: spell check reudceJani Nikula1-0/+1
References: http://mid.gmane.org/[email protected] Signed-off-by: Jani Nikula <[email protected]> Cc: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: add spell checking of email subject lineJoe Perches1-1/+2
Only commit log and patch additions are checked for typos and spelling errors currently. Add a check of the email subject line too. Signed-off-by: Joe Perches <[email protected]> Suggested-by: Jani Nikula <[email protected]> Tested-by: Jani Nikula <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17checkpatch: improve "no space is necessary after a cast" testJoe Perches1-2/+9
The "no space is necessary after a cast" sizeof exclusion doesn't work properly. The test reports a false positive for code like: BUILD_BUG_ON(sizeof(struct batadv_bla_claim_dst) != 6); Make it work, simplify the exclusions, and add some comments. Signed-off-by: Joe Perches <[email protected]> Reported-by: Marek Lindner <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17firmware/ihex2fw.c: restore missing default in switch statementNicolas Iooss1-0/+1
Commit 2473238eac95 ("ihex: add support for CS:IP/EIP records") removes the "default:" statement in the switch block, making the "return usage();" line dead code and ihex2fw silently ignoring unknown options. Restore this statement. This bug was found by building with HOSTCC=clang and adding -Wunreachable-code-return to HOSTCFLAGS. Fixes: 2473238eac95 ("ihex: add support for CS:IP/EIP records") Signed-off-by: Nicolas Iooss <[email protected]> Cc: Mark Brown <[email protected]> Cc: David Woodhouse <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17lib/bitmap.c: bitmap_[empty,full]: remove code duplicationYury Norov2-34/+4
bitmap_empty() has its own implementation. But it's clearly as simple as: find_first_bit(src, nbits) == nbits The same is true for 'bitmap_full'. Signed-off-by: Yury Norov <[email protected]> Cc: George Spelvin <[email protected]> Cc: Alexey Klimov <[email protected]> Cc: Rasmus Villemoes <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17ASoC: pcm512x: use DIV_ROUND_CLOSEST_ULL() from kernel.hJavi Merino1-2/+1
Now that the kernel provides DIV_ROUND_CLOSEST_ULL(), drop the internal implementation and use the kernel one. Signed-off-by: Javi Merino <[email protected]> Reported-by: kbuild test robot <[email protected]> Cc: Peter Rosin <[email protected]> Acked-by: Mark Brown <[email protected]> Cc: Liam Girdwood <[email protected]> Cc: Jaroslav Kysela <[email protected]> Cc: Takashi Iwai <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17media: cxd2820r: use DIV_ROUND_CLOSEST_ULL()Javi Merino5-11/+3
Now that the kernel provides DIV_ROUND_CLOSEST_ULL(), drop the internal implementation and use the kernel one. Signed-off-by: Javi Merino <[email protected]> Cc: Mauro Carvalho Chehab <[email protected]> Acked-by: Antti Palosaari <[email protected]> Reviewed-by: Antti Palosaari <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17cpuidle: menu: use DIV_ROUND_CLOSEST_ULL()Javi Merino1-7/+1
Now that the kernel provides DIV_ROUND_CLOSEST_ULL(), drop the internal implementation and use the kernel one. Signed-off-by: Javi Merino <[email protected]> Acked-by: Rafael J. Wysocki <[email protected]> Cc: Mel Gorman <[email protected]> Cc: Stephen Hemminger <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17clk: bcm/kona: use DIV_ROUND_CLOSEST_ULL()Javi Merino2-22/+7
Now that the kernel provides DIV_ROUND_CLOSEST_ULL(), drop the internal implementation and use the kernel one. Signed-off-by: Javi Merino <[email protected]> Cc: Mike Turquette <[email protected]> Cc: Stephen Boyd <[email protected]> Acked-by: Alex Elder <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17kernel.h: implement DIV_ROUND_CLOSEST_ULLJavi Merino3-3/+13
We have grown a number of different implementations of DIV_ROUND_CLOSEST_ULL throughout the kernel. Move the i915 one to kernel.h so that it can be reused. Signed-off-by: Javi Merino <[email protected]> Reviewed-by: Jeff Epler <[email protected]> Cc: Jani Nikula <[email protected]> Cc: David Airlie <[email protected]> Cc: Guenter Roeck <[email protected]> Acked-by: Daniel Vetter <[email protected]> Cc: "Rafael J. Wysocki" <[email protected]> Cc: Alex Elder <[email protected]> Cc: Antti Palosaari <[email protected]> Cc: Javi Merino <[email protected]> Cc: Mauro Carvalho Chehab <[email protected]> Cc: Mel Gorman <[email protected]> Cc: Mike Turquette <[email protected]> Cc: Stephen Boyd <[email protected]> Cc: Stephen Hemminger <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17lib/vsprintf.c: improve put_dec_trunc8 slightlyRasmus Villemoes1-6/+4
I hadn't had enough coffee when I wrote this. Currently, the final increment of buf depends on the value loaded from the table, and causes gcc to emit a cmov immediately before the return. It is smarter to let it depend on r, since the increment can then be computed in parallel with the final load/store pair. It also shaves 16 bytes of .text. Signed-off-by: Rasmus Villemoes <[email protected]> Cc: Tejun Heo <[email protected]> Cc: Peter Zijlstra <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17hwmon: (w83795) use find_closest_descending() in pwm_freq_to_reg()Bartosz Golaszewski1-5/+3
Replace the loop iterating over pwm_freq_cksel0 with a call to find_closest_descending(). Signed-off-by: Bartosz Golaszewski <[email protected]> Cc: Guenter Roeck <[email protected]> Cc: Steven Rostedt <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17hwmon: (lm85) use find_closest() in x_TO_REG() functionsBartosz Golaszewski1-18/+8
Replace RANGE_TO_REG() and FREQ_TO_REG() implementations with calls to find_closest(). Signed-off-by: Bartosz Golaszewski <[email protected]> Cc: Guenter Roeck <[email protected]> Cc: Steven Rostedt <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17hwmon: (ina2xx) replace ina226_avg_bits() with find_closest()Bartosz Golaszewski1-14/+3
Use find_closest() to locate the closest average in ina226_avg_tab. Signed-off-by: Bartosz Golaszewski <[email protected]> Cc: Guenter Roeck <[email protected]> Cc: Steven Rostedt <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17documentation: update CodingStyle on local variables naming in macrosBartosz Golaszewski1-0/+13
Describe proper naming convention for local variables in macros resembling functions. Signed-off-by: Bartosz Golaszewski <[email protected]> Cc: Guenter Roeck <[email protected]> Cc: Steven Rostedt <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17util_macros.h: add find_closest() macroBartosz Golaszewski1-0/+40
This series unduplicates the code used to find the member in an array closest to 'x'. The first patch adds a macro implementing the algorithm in two flavors - for arrays sorted in ascending and descending order. The second updates Documentation/CodingStyle on the naming convention for local variables in macros resembling functions. Other three patches replace duplicated code with calls to one of these macros in some hwmon drivers. This patch (of 5): Searching for the member of an array closest to 'x' is duplicated in several places. Add a new include - util_macros.h - and two macros that implement this algorithm for arrays sorted both in ascending and descending order. Uses linear search. Signed-off-by: Bartosz Golaszewski <[email protected]> Cc: Guenter Roeck <[email protected]> Cc: Steven Rostedt <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17lib/dma-debug: fix bucket_find_contain()Sebastian Ott1-1/+1
bucket_find_contain() will search the bucket list for a dma_debug_entry. When the entry isn't found it needs to search other buckets too, since only the start address of a dma range is hashed (which might be in a different bucket). A copy of the dma_debug_entry is used to get the previous hash bucket but when its list is searched the original dma_debug_entry is to be used not its modified copy. This fixes false "device driver tries to sync DMA memory it has not allocated" warnings. Signed-off-by: Sebastian Ott <[email protected]> Cc: Florian Fainelli <[email protected]> Cc: Horia Geanta <[email protected]> Cc: Jiri Kosina <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17lib/vsprintf.c: even faster binary to decimal conversionRasmus Villemoes1-118/+128
The most expensive part of decimal conversion is the divisions by 10 (albeit done using reciprocal multiplication with appropriately chosen constants). I decided to see if one could eliminate around half of these multiplications by emitting two digits at a time, at the cost of a 200 byte lookup table, and it does indeed seem like there is something to be gained, especially on 64 bits. Microbenchmarking shows improvements ranging from -50% (for numbers uniformly distributed in [0, 2^64-1]) to -25% (for numbers heavily biased toward the smaller end, a more realistic distribution). On a larger scale, perf shows that top, one of the big consumers of /proc data, uses 0.5-1.0% fewer cpu cycles. I had to jump through some hoops to get the 32 bit code to compile and run on my 64 bit machine, so I'm not sure how relevant these numbers are, but just for comparison the microbenchmark showed improvements between -30% and -10%. The bloat-o-meter costs are around 150 bytes (the generated code is a little smaller, so it's not the full 200 bytes) on both 32 and 64 bit. I'm aware that extra cache misses won't show up in a microbenchmark as used above, but on the other hand decimal conversions often happen in bulk (for example in the case of top). I have of course tested that the new code generates the same output as the old, for both the first and last 1e10 numbers in [0,2^64-1] and 4e9 'random' numbers in-between. Test and verification code on github: https://github.com/Villemoes/dec. Signed-off-by: Rasmus Villemoes <[email protected]> Tested-by: Jeff Epler <[email protected]> Cc: "Peter Zijlstra (Intel)" <[email protected]> Cc: Tejun Heo <[email protected]> Cc: Joe Perches <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17lib: rename lib/find_next_bit.c to lib/find_bit.cYury Norov2-1/+1
This file contains implementation for all find_*_bit{,_le} So giving it more generic name looks reasonable. Signed-off-by: Yury Norov <[email protected]> Reviewed-by: Rasmus Villemoes <[email protected]> Reviewed-by: George Spelvin <[email protected]> Cc: Alexey Klimov <[email protected]> Cc: David S. Miller <[email protected]> Cc: Daniel Borkmann <[email protected]> Cc: Hannes Frederic Sowa <[email protected]> Cc: Lai Jiangshan <[email protected]> Cc: Mark Salter <[email protected]> Cc: AKASHI Takahiro <[email protected]> Cc: Thomas Graf <[email protected]> Cc: Valentin Rothberg <[email protected]> Cc: Chris Wilson <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17lib: move find_last_bit to lib/find_next_bit.cYury Norov2-2/+27
Currently all 'find_*_bit' family is located in lib/find_next_bit.c, except 'find_last_bit', which is in lib/find_last_bit.c. It seems, there's no major benefit to have it separated. Signed-off-by: Yury Norov <[email protected]> Reviewed-by: Rasmus Villemoes <[email protected]> Reviewed-by: George Spelvin <[email protected]> Cc: Alexey Klimov <[email protected]> Cc: David S. Miller <[email protected]> Cc: Daniel Borkmann <[email protected]> Cc: Hannes Frederic Sowa <[email protected]> Cc: Lai Jiangshan <[email protected]> Cc: Mark Salter <[email protected]> Cc: AKASHI Takahiro <[email protected]> Cc: Thomas Graf <[email protected]> Cc: Valentin Rothberg <[email protected]> Cc: Chris Wilson <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17lib: find_*_bit reimplementationYury Norov3-216/+91
This patchset does rework to find_bit function family to achieve better performance, and decrease size of text. All rework is done in patch 1. Patches 2 and 3 are about code moving and renaming. It was boot-tested on x86_64 and MIPS (big-endian) machines. Performance tests were ran on userspace with code like this: /* addr[] is filled from /dev/urandom */ start = clock(); while (ret < nbits) ret = find_next_bit(addr, nbits, ret + 1); end = clock(); printf("%ld\t", (unsigned long) end - start); On Intel(R) Core(TM) i7-3770 CPU @ 3.40GHz measurements are: (for find_next_bit, nbits is 8M, for find_first_bit - 80K) find_next_bit: find_first_bit: new current new current 26932 43151 14777 14925 26947 43182 14521 15423 26507 43824 15053 14705 27329 43759 14473 14777 26895 43367 14847 15023 26990 43693 15103 15163 26775 43299 15067 15232 27282 42752 14544 15121 27504 43088 14644 14858 26761 43856 14699 15193 26692 43075 14781 14681 27137 42969 14451 15061 ... ... find_next_bit performance gain is 35-40%; find_first_bit - no measurable difference. On ARM machine, there is arch-specific implementation for find_bit. Thanks a lot to George Spelvin and Rasmus Villemoes for hints and helpful discussions. This patch (of 3): New implementations takes less space in source file (see diffstat) and in object. For me it's 710 vs 453 bytes of text. It also shows better performance. find_last_bit description fixed due to obvious typo. [[email protected]: include linux/bitmap.h, per Rasmus] Signed-off-by: Yury Norov <[email protected]> Reviewed-by: Rasmus Villemoes <[email protected]> Reviewed-by: George Spelvin <[email protected]> Cc: Alexey Klimov <[email protected]> Cc: David S. Miller <[email protected]> Cc: Daniel Borkmann <[email protected]> Cc: Hannes Frederic Sowa <[email protected]> Cc: Lai Jiangshan <[email protected]> Cc: Mark Salter <[email protected]> Cc: AKASHI Takahiro <[email protected]> Cc: Thomas Graf <[email protected]> Cc: Valentin Rothberg <[email protected]> Cc: Chris Wilson <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17alpha: forward declare struct pt_regs in processor.hRichard Weinberger1-0/+1
Removal of exec domains uncovered this new warning. processor.h re-used struct pt_regs from personality.h which is now gone. ./arch/alpha/include/asm/processor.h:47:33: warning: 'struct pt_regs' declared inside parameter list [enabled by default] Signed-off-by: Richard Weinberger <[email protected]> Cc: Richard Henderson <[email protected]> Cc: Ivan Kokshaysky <[email protected]> Cc: Matt Turner <[email protected]> Signed-off-by: Andrew Morton <[email protected]> Signed-off-by: Linus Torvalds <[email protected]>
2015-04-17Revert "mmc: core: Convert mmc_driver to device_driver"Ulf Hansson4-38/+69
This reverts commit 6685ac62b2f0 ("mmc: core: Convert mmc_driver to device_driver") The reverted commit went too far in simplifing the device driver parts for mmc. Let's restore the old mmc_driver to enable driver core to sooner or later to remove the ->probe(), ->remove() and ->shutdown() callbacks from the struct device_driver. Note that, the old ->suspend|resume() callbacks in the struct mmc_driver don't need to be restored, since the mmc block layer has converted to the modern system PM ops. Fixes: 6685ac62b2f0 ("mmc: core: Convert mmc_driver to device_driver") Signed-off-by: Ulf Hansson <[email protected]> Acked-by: Jaehoon Chung <[email protected]>
2015-04-17mmc: pwrseq: Fix error code propagation in mmc_pwrseq_simple_alloc()Javier Martinez Canillas1-1/+1
If the struct mmc_pwrseq_match .alloc function used to allocate a struct mmc_pwrseq fails, the error is propagated to mmc_of_parse(). But instead of returning the error code in pwrseq, host->pwrseq is returned which will always be 0. So mmc_of_parse() succeeds even if the pwrseq .alloc function failed and host->pwrseq is NULL. This makes the SDIO device to not be powered if the power sequencing .alloc functions wants to be deferred due a missing resource because the mmc controller driver probe did wrongly succeed. Fixes: 0f12a0ce4ce4a ("mmc: pwrseq: simplify alloc/free hooks") Signed-off-by: Javier Martinez Canillas <[email protected]> Reviewed-by: Alexandre Courbot <[email protected]> Signed-off-by: Ulf Hansson <[email protected]>
2015-04-17x86/fpu: Load xsave pointer *after* initializationBorislav Petkov1-2/+6
So I was playing with gdb today and did this simple thing: gdb /bin/ls ... (gdb) run Box exploded with this splat: BUG: unable to handle kernel NULL pointer dereference at 00000000000001d0 IP: [<ffffffff8100fe5a>] xstateregs_get+0x7a/0x120 [...] Call Trace: ptrace_regset ptrace_request ? wait_task_inactive ? preempt_count_sub arch_ptrace ? ptrace_get_task_struct SyS_ptrace system_call_fastpath ... because we do cache &target->thread.fpu.state->xsave into the local variable xsave but that pointer is NULL at that time and it gets initialized later, in init_fpu(), see: e7f180dcd8ab ("x86/fpu: Change xstateregs_get()/set() to use ->xsave.i387 rather than ->fxsave") The fix is simple: load xsave *after* init_fpu() has run. Also do the same in xstateregs_set(), as suggested by Oleg Nesterov. Signed-off-by: Borislav Petkov <[email protected]> Acked-by: Oleg Nesterov <[email protected]> Cc: Andy Lutomirski <[email protected]> Cc: H. Peter Anvin <[email protected]> Cc: Linus Torvalds <[email protected]> Cc: Rik van Riel <[email protected]> Cc: Tavis Ormandy <[email protected]> Cc: Thomas Gleixner <[email protected]> Link: http://lkml.kernel.org/r/[email protected] Signed-off-by: Ingo Molnar <[email protected]>
Powered by cgit, Gruvboxed by Blaster4385.