Age | Commit message (Collapse) | Author | Files | Lines |
|
nconf.h includes <curses.h> and "ncurses.h", but it does not need to
include both. Generally, it should fall back to curses.h only when
ncurses.h is not found. But, looks like it has never happened;
these includes have been here for many years since commit 692d97c380c6
("kconfig: new configuration interface (nconfig)"), and nobody has
complained about hard-coding of ncurses.h . Let's simply drop the
curses.h inclusion.
I replaced "ncurses.h" with <ncurses.h> since it is not a local file.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
Currently, the unmet dependency warnings end up with endlessly long
expressions, most of which are false positives.
Here is test code to demonstrate how it currently works.
[Test Case]
config DEP1
def_bool y
config DEP2
bool "DEP2"
config A
bool "A"
select E
config B
bool "B"
depends on DEP2
select E
config C
bool "C"
depends on DEP1 && DEP2
select E
config D
def_bool n
select E
config E
bool
depends on DEP1 && DEP2
[Result]
$ make config
scripts/kconfig/conf --oldaskconfig Kconfig
*
* Linux Kernel Configuration
*
DEP2 (DEP2) [N/y/?] (NEW) n
A (A) [N/y/?] (NEW) y
warning: (A && B && D) selects E which has unmet direct
dependencies (DEP1 && DEP2)
Here, I see some points to be improved.
First, '(A || B || D)' would make more sense than '(A && B && D)'.
I am not sure if this is intentional, but expr_simplify_unmet_dep()
turns OR expressions into AND, like follows:
case E_OR:
return expr_alloc_and(
Second, we see false positives. 'A' is a real unmet dependency.
'B' is false positive because 'DEP1' is fixed to 'y', and 'B' depends
on 'DEP2'. 'C' was correctly dropped by expr_simplify_unmet_dep().
'D' is also false positive because it has no chance to be enabled.
Current expr_simplify_unmet_dep() cannot avoid those false positives.
After all, I decided to use the same helpers as used for printing
reverse dependencies in the help.
With this commit, unreadable warnings (most of the reported symbols are
false positives) in the real world:
$ make ARCH=score allyesconfig
scripts/kconfig/conf --allyesconfig Kconfig
warning: (HWSPINLOCK_QCOM && AHCI_MTK && STMMAC_PLATFORM &&
DWMAC_IPQ806X && DWMAC_LPC18XX && DWMAC_OXNAS && DWMAC_ROCKCHIP &&
DWMAC_SOCFPGA && DWMAC_STI && TI_CPSW && PINCTRL_GEMINI &&
PINCTRL_OXNAS && PINCTRL_ROCKCHIP && PINCTRL_DOVE &&
PINCTRL_ARMADA_37XX && PINCTRL_STM32 && S3C2410_WATCHDOG &&
VIDEO_OMAP3 && VIDEO_S5P_FIMC && USB_XHCI_MTK && RTC_DRV_AT91SAM9 &&
LPC18XX_DMAMUX && VIDEO_OMAP4 && COMMON_CLK_GEMINI &&
COMMON_CLK_ASPEED && COMMON_CLK_NXP && COMMON_CLK_OXNAS &&
COMMON_CLK_BOSTON && QCOM_ADSP_PIL && QCOM_Q6V5_PIL && QCOM_GSBI &&
ATMEL_EBI && ST_IRQCHIP && RESET_IMX7 && PHY_HI6220_USB &&
PHY_RALINK_USB && PHY_ROCKCHIP_PCIE && PHY_DA8XX_USB) selects
MFD_SYSCON which has unmet direct dependencies (HAS_IOMEM)
warning: (PINCTRL_AT91 && PINCTRL_AT91PIO4 && PINCTRL_OXNAS &&
PINCTRL_PISTACHIO && PINCTRL_PIC32 && PINCTRL_MESON &&
PINCTRL_NOMADIK && PINCTRL_MTK && PINCTRL_MT7622 && GPIO_TB10X)
selects OF_GPIO which has unmet direct dependencies (GPIOLIB && OF &&
HAS_IOMEM)
warning: (FAULT_INJECTION_STACKTRACE_FILTER && LATENCYTOP && LOCKDEP)
selects FRAME_POINTER which has unmet direct dependencies
(DEBUG_KERNEL && (CRIS || M68K || FRV || UML || SUPERH || BLACKFIN ||
MN10300 || METAG) || ARCH_WANT_FRAME_POINTERS)
will be turned into:
$ make ARCH=score allyesconfig
scripts/kconfig/conf --allyesconfig Kconfig
WARNING: unmet direct dependencies detected for MFD_SYSCON
Depends on [n]: HAS_IOMEM [=n]
Selected by [y]:
- PINCTRL_STM32 [=y] && PINCTRL [=y] && (ARCH_STM32 ||
COMPILE_TEST [=y]) && OF [=y]
- RTC_DRV_AT91SAM9 [=y] && RTC_CLASS [=y] && (ARCH_AT91 ||
COMPILE_TEST [=y])
- RESET_IMX7 [=y] && RESET_CONTROLLER [=y]
- PHY_HI6220_USB [=y] && (ARCH_HISI && ARM64 ||
COMPILE_TEST [=y])
- PHY_RALINK_USB [=y] && (RALINK || COMPILE_TEST [=y])
- PHY_ROCKCHIP_PCIE [=y] && (ARCH_ROCKCHIP && OF [=y] ||
COMPILE_TEST [=y])
WARNING: unmet direct dependencies detected for OF_GPIO
Depends on [n]: GPIOLIB [=y] && OF [=y] && HAS_IOMEM [=n]
Selected by [y]:
- PINCTRL_MTK [=y] && PINCTRL [=y] && (ARCH_MEDIATEK ||
COMPILE_TEST [=y]) && OF [=y]
- PINCTRL_MT7622 [=y] && PINCTRL [=y] && (ARCH_MEDIATEK ||
COMPILE_TEST [=y]) && OF [=y] && (ARM64 || COMPILE_TEST [=y])
WARNING: unmet direct dependencies detected for FRAME_POINTER
Depends on [n]: DEBUG_KERNEL [=y] && (CRIS || M68K || FRV || UML ||
SUPERH || BLACKFIN || MN10300 || METAG) || ARCH_WANT_FRAME_POINTERS [=n]
Selected by [y]:
- LATENCYTOP [=y] && DEBUG_KERNEL [=y] && STACKTRACE_SUPPORT [=y] &&
PROC_FS [=y] && !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND &&
!ARC && !X86
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Petr Vorel <petr.vorel@gmail.com>
|
|
Commit 246cf9c26bf1 ("kbuild: Warn on selecting symbols with unmet
direct dependencies") forcibly promoted ->dir_dep.tri to yes from mod.
So, the unmet direct dependencies of tristate symbols are not reported.
[Test Case]
config MODULES
def_bool y
option modules
config A
def_bool y
select B
config B
tristate "B"
depends on m
This causes unmet dependency because 'B' is forced 'y' ignoring
'depends on m'. This should be warned.
On the other hand, the following case ('B' is bool) should not be
warned, so 'depends on m' for bool symbols should be naturally treated
as 'depends on y'.
[Test Case2 (not unmet dependency)]
config MODULES
def_bool y
option modules
config A
def_bool y
select B
config B
bool "B"
depends on m
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
If recursive inclusion is detected, it should fail with error
messages. Test this.
This also tests the line numbers in the error message, fixed by
commit 5ae6fcc4bb82 ("kconfig: fix line number in recursive inclusion
error message").
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
Recursive dependency should be detected and warned. Test this.
This indirectly tests the line number increments.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
Commit 3b9a19e08960 ("kconfig: loop as long as we changed some symbols
in randconfig") fixed randconfig where a choice contains a sub-choice.
Prior to that commit, the sub-choice values were not set.
I am not sure whether this is an intended feature or just something
people discovered works, but it is used in the real world;
drivers/usb/gadget/legacy/Kconfig is source'd in a choice context,
then creates a sub-choice in it.
For the test case in this commit, there are 3 possible results.
Case 1:
CONFIG_A=y
# CONFIG_B is not set
Case 2:
# CONFIG_A is not set
CONFIG_B=y
CONFIG_C=y
# CONFIG_D is not set
Case 3:
# CONFIG_A is not set
CONFIG_B=y
# CONFIG_C is not set
CONFIG_D=y
CONFIG_E=y
So, this test iterates several times, and checks if the result is
either of the three.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
Commit fbe98bb9ed3d ("kconfig: Fix defconfig when one choice menu
selects options that another choice menu depends on") fixed defconfig
when two choices interact (i.e. calculating the visibility of a choice
requires to calculate another choice).
The test code in that commit log was based on the real world example,
and complicated. So, I shrunk it down to the following:
defconfig.choice:
---8<---
CONFIG_CHOICE_VAL0=y
---8<---
---8<---
config MODULES
def_bool y
option modules
choice
prompt "Choice"
config CHOICE_VAL0
tristate "Choice 0"
config CHOICE_VAL1
tristate "Choice 1"
endchoice
choice
prompt "Another choice"
depends on CHOICE_VAL0
config DUMMY
bool "dummy"
endchoice
---8<---
Prior to commit fbe98bb9ed3d,
$ scripts/kconfig/conf --defconfig=defconfig.choice Kconfig.choice
resulted in:
CONFIG_MODULES=y
CONFIG_CHOICE_VAL0=m
# CONFIG_CHOICE_VAL1 is not set
CONFIG_DUMMY=y
where the expected result would be:
CONFIG_MODULES=y
CONFIG_CHOICE_VAL0=y
# CONFIG_CHOICE_VAL1 is not set
CONFIG_DUMMY=y
Roughly, this weird behavior happened like this:
Symbols are calculated a couple of times. First, all symbols are
calculated in conf_read(). The first 'choice' is evaluated to 'y'
due to the SYMBOL_DEF_USER flag, but sym_calc_choice() clears it
unless all of its choice values are explicitly set by the user.
conf_set_all_new_symbols() clears all SYMBOL_VALID flags. Then, only
choices are calculated. Here, the SYMBOL_DEF_USER for the first choice
has been forgotten, so it is evaluated to 'm'. set_all_choice_values()
sets SYMBOL_DEF_USER again to choice symbols.
When calculating the second choice, due to 'depends on CHOICE_VAL0',
it triggers the calculation of CHOICE_VAL0. As a result, SYMBOL_VALID
is set for CHOICE_VAL0.
Symbols except choices get the final chance of re-calculation in
conf_write(). In a normal case, CHOICE_VAL0 would be re-calculated,
then the first choice would be indirectly re-calculated with the
SYMBOL_DEF_USER which has been recalled by set_all_choice_values(),
which would be evaluated to 'y'. But, in this case, CHOICE_VAL0 has
already been marked as SYMBOL_VALID, so this re-calculation does not
happen. Then, =m from the conf_set_all_new_symbols() phase is written
out to the .config file.
Add a unit test for this naive case.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
If tristate choice values depend on symbols set to 'm', they should be
hidden when the choice containing them is changed from 'm' to 'y'
(i.e. exclusive choice).
This issue was fixed by commit fa64e5f6a35e ("kconfig/symbol.c: handle
choice_values that depend on 'm' symbols").
Add a test case to avoid regression.
For the input in this unit test, there is a room for argument if
"# CONFIG_CHOICE1 is not set" should be written to the .config file.
After commit fa64e5f6a35e, this line was written to the .config file.
With commit cb67ab2cd2b8 ("kconfig: do not write choice values when
their dependency becomes n"), it is not written now.
In this test, "# CONFIG_CHOICE1 is not set" is don't care.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
Commit cb67ab2cd2b8 ("kconfig: do not write choice values when their
dependency becomes n") fixed a problem where "# CONFIG_... is not set"
for choice values are wrongly written into the .config file when they
are once visible, then become invisible later.
Add a test for this naive case.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
If new choice values are added with new dependency, and they become
visible during user configuration, oldconfig should recognize them
as (NEW), and ask the user for choice.
This issue was fixed by commit 5d09598d488f ("kconfig: fix new choices
being skipped upon config update").
This is a subtle corner case. Add a test case to avoid breakage.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
If a symbols has dependency on the preceding symbol, the menu entry
should become the submenu of the preceding one, and displayed with
deeper indentation.
This is done by restructuring the menu tree in menu_finalize().
It is a bit complicated computation, so let's add a test case.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
The calculation of 'choice' is a bit complicated part in Kconfig.
The behavior of 'y' choice is intuitive. If choice values are tristate,
the choice can be 'm' where each value can be enabled independently.
Also, if a choice is marked as 'optional', the whole choice can be
invisible.
Test basic functionality of choice.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
Many parts in Kconfig are so cryptic and need refactoring. However,
its complexity prevents us from moving forward. There are several
naive corner cases where it is difficult to notice breakage. If
those are covered by unit tests, we will be able to touch the code
with more confidence.
Here is a simple test framework based on pytest. The conftest.py
provides a fixture useful to run commands such as 'oldaskconfig' etc.
and to compare the resulted .config, stdout, stderr with expectations.
How to add test cases?
----------------------
For each test case, you should create a subdirectory under
scripts/kconfig/tests/ (so test cases are separated from each other).
Every test case directory should contain the following files:
- __init__.py: describes test functions
- Kconfig: the top level Kconfig file for the test
To do a useful job, test cases generally need additional data like
input .config and information about expected results.
How to run tests?
-----------------
You need python3 and pytest. Then, run "make testconfig". O= option
is supported. If V=1 is given, detailed logs captured during tests
are displayed.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
The variable 'PYTHON' allows users to specify a proper executable
name in case the default 'python' does not work. However, this does
not address the case where both Python 2.x and 3.x scripts are used
in one source tree.
PEP 394 (https://www.python.org/dev/peps/pep-0394/) provides a
convention for Python scripts portability. Here is a quotation:
In order to tolerate differences across platforms, all new code
that needs to invoke the Python interpreter should not specify
'python', but rather should specify either 'python2' or 'python3'.
This distinction should be made in shebangs, when invoking from a
shell script, when invoking via the system() call, or when invoking
in any other context.
One exception to this is scripts that are deliberately written to
be source compatible with both Python 2.x and 3.x. Such scripts may
continue to use python on their shebang line without affecting their
portability.
To meet this requirement, this commit adds new variables 'PYTHON2'
and 'PYTHON3'.
arch/ia64/scripts/unwcheck.py is the only script that has ever used
$(PYTHON). Recent commit bd5edbe67794 ("ia64: convert unwcheck.py to
python3") converted it to be compatible with both Python 2.x and 3.x,
so this is the exceptional case where the use of 'python' is allowed.
So, I did not touch arch/ia64/Makefile.
tools/perf/Makefile.config sets PYTHON and PYTHON2 by itself, so it
is not affected by this commit.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
The local{yes,mod}config targets currently have streamline_config.pl as
a prerequisite. This is redundant, because streamline_config.pl is a
checked-in file with no prerequisites.
Remove the prerequisite and reference streamline_config.pl directly in
the recipe of the rule instead.
Signed-off-by: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
As commit cedd55d49dee ("kconfig: Remove silentoldconfig from help
and docs; fix kconfig/conf's help") mentioned, 'silentoldconfig' is a
historical misnomer. That commit removed it from help and docs since
it is an internal interface. If so, it should be allowed to rename
it to something more intuitive. 'syncconfig' is the one I came up
with because it updates the .config if necessary, then synchronize
include/generated/autoconf.h and include/config/* with it.
You should not manually invoke 'silentoldcofig'. Display warning if
used in case existing scripts are doing wrong.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
The purpose of local{yes,mod}config is to arrange the .config file
based on actually loaded modules. It is unnecessary to update
include/generated/autoconf.h and include/config/* stuff here.
They will be updated as needed during the build.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
Historically, "make oldconfig" has changed its behavior several times,
quieter or louder. (I attached the history below.) Currently, it is
not as quiet as it should be. This commit addresses it.
Test Case
---------
---------------------------(Kconfig)----------------------------
menu "menu"
config FOO
bool "foo"
menu "sub menu"
config BAR
bool "bar"
endmenu
endmenu
menu "sibling menu"
config BAZ
bool "baz"
endmenu
----------------------------------------------------------------
---------------------------(.config)----------------------------
CONFIG_BAR=y
CONFIG_BAZ=y
----------------------------------------------------------------
With the Kconfig and .config above, "make silentoldconfig" and
"make oldconfig" work differently, like follows:
$ make silentoldconfig
scripts/kconfig/conf --silentoldconfig Kconfig
*
* Restart config...
*
*
* menu
*
foo (FOO) [N/y/?] (NEW) y
#
# configuration written to .config
#
$ make oldconfig
scripts/kconfig/conf --oldconfig Kconfig
*
* Restart config...
*
*
* menu
*
foo (FOO) [N/y/?] (NEW) y
*
* sub menu
*
bar (BAR) [Y/n/?] y
#
# configuration written to .config
#
Both hide "sibling node" since it is irrelevant. The difference is
that silentoldconfig hides "sub menu" whereas oldconfig does not.
The behavior of silentoldconfig is preferred since the "sub menu"
does not contain any new symbol.
The root cause is in conf(). There are three input modes that can
call conf(); oldaskconfig, oldconfig, and silentoldconfig.
Everytime conf() encounters a menu entry, it calls check_conf() to
check if it contains new symbols. If no new symbol is found, the
menu is just skipped.
Currently, this happens only when input_mode == silentoldconfig.
The oldaskconfig enters into the check_conf() loop as silentoldconfig,
so oldaskconfig works likewise for the second loop or later, but it
never happens for oldconfig. So, irrelevant sub-menus are shown for
oldconfig.
Change the test condition to "input_mode != oldaskconfig". This is
false only for the first loop of oldaskconfig; it must ask the user
all symbols, so no need to call check_conf().
History of oldconfig
--------------------
[0] Originally, "make oldconfig" was as loud as "make config" (It
showed the entire .config file)
[1] Commit cd9140e1e73a ("kconfig: make oldconfig is now less chatty")
made oldconfig quieter, but it was still less quieter than
silentoldconfig. (oldconfig did not hide sub-menus)
[2] Commit 204c96f60904 ("kconfig: fix silentoldconfig") changed
the input_mode of oldconfig to "ask_silent" from "ask_new".
So, oldconfig really became as quiet as silentoldconfig.
(oldconfig hided irrelevant sub-menus)
[3] Commit 4062f1a4c030 ("kconfig: use long options in conf") made
oldconfig as loud as [0] due to misconversion.
[4] Commit 14828349719a ("kconfig: fix make oldconfig") addressed
the misconversion of [3], but it made oldconfig quieter only to
the same level as [1], not [2].
This commit is restoring the behavior of [2].
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
check_conf() never increments conf_cnt for listnewconfig, so conf_cnt
is always zero.
In other words, conf_cnt is not zero, "input_mode != listnewconfig"
is met.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
conf() is never called for listnewconfig / olddefconfig.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
check_conf() traverses the menu tree, but it is completely no-op for
olddefconfig because the following if-else block does nothing.
if (input_mode == listnewconfig) {
...
} else if (input_mode != olddefconfig) {
...
}
As the help message says, olddefconfig automatically sets new symbols
to their default value. There is no room for manual intervention.
So, calling check_conf() for olddefconfig is odd in the first place.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
|
|
=== Background ===
- Visible n-valued bool/tristate symbols generate a
'# CONFIG_FOO is not set' line in the .config file. The idea is to
remember the user selection without having to set a Makefile
variable. Having n correspond to the variable being undefined in the
Makefiles makes for easy CONFIG_* tests.
- Invisible n-valued bool/tristate symbols normally do not generate a
'# CONFIG_FOO is not set' line, because user values from .config
files have no effect on invisible symbols anyway.
Currently, there is one exception to this rule: Any bool/tristate symbol
that gets the value n through a 'default' property generates a
'# CONFIG_FOO is not set' line, even if the symbol is invisible.
Note that this only applies to explicitly given defaults, and not when
the symbol implicitly defaults to n (like bool/tristate symbols without
'default' properties do).
This is inconsistent, and seems redundant:
- As mentioned, the '# CONFIG_FOO is not set' won't affect the symbol
once the .config is read back in.
- Even if the symbol is invisible at first but becomes visible later,
there shouldn't be any harm in recalculating the default value
rather than viewing the '# CONFIG_FOO is not set' as a previous
user value of n.
=== Changes ===
Change sym_calc_value() to only set SYMBOL_WRITE (write to .config) for
non-n-valued 'default' properties.
Note that SYMBOL_WRITE is always set for visible symbols regardless of whether
they have 'default' properties or not, so this change only affects invisible
symbols.
This reduces the size of the x86 .config on my system by about 1% (due
to removed '# CONFIG_FOO is not set' entries).
One side effect of (and the main motivation for) this change is making
the following two definitions behave exactly the same:
config FOO
bool
config FOO
bool
default n
With this change, neither of these will generate a
'# CONFIG_FOO is not set' line (assuming FOO isn't selected/implied).
That might make it clearer to people that a bare 'default n' is
redundant.
This change only affects generated .config files and not autoconf.h:
autoconf.h only includes #defines for non-n bool/tristate symbols.
=== Testing ===
The following testing was done with the x86 Kconfigs:
- .config files generated before and after the change were compared to
verify that the only difference is some '# CONFIG_FOO is not set'
entries disappearing. A couple of these were inspected manually, and
most turned out to be from redundant 'default n/def_bool n'
properties.
- The generated include/generated/autoconf.h was compared before and
after the change and verified to be identical.
- As a sanity check, the same modification was done to Kconfiglib.
The Kconfiglib test suite was then run to check for any mismatches
against the output of the C implementation.
Signed-off-by: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
Surprisingly or not, disabling a CONFIG option (which is assumed to
be unneeded) may be not so trivial. Especially it is not trivial, when
this CONFIG option is selected by a dozen of other configs. Before the
moment commit 1ccb27143360 ("kconfig: make "Selected by:" and
"Implied by:" readable") popped up in v4.16-rc1, it was an absolute pain
to break down the "Selected by" reverse dependency expression in order
to identify all those configs which select (IOW *do not allow
disabling*) a certain feature (assumed to be not needed).
This patch tries to make one step further by putting at users'
fingertips the revdep top level OR sub-expressions grouped/clustered by
the tristate value they evaluate to. This should allow the users to
directly concentrate on and tackle the _active_ reverse dependencies.
To give some numbers and quantify the complexity of certain reverse
dependencies, assuming commit 617aebe6a97e ("Merge tag
'usercopy-v4.16-rc1' of
git://git.kernel.org/pub/scm/linux/kernel/git/kees/linux"), ARCH=arm64
and vanilla arm64 defconfig, here is the top 10 CONFIG options with
the highest amount of top level "||" sub-expressions/tokens that make
up the final "Selected by" reverse dependency expression.
| Config | All revdep | Active revdep |
|-------------------|------------|---------------|
| REGMAP_I2C | 212 | 9 |
| CRC32 | 167 | 25 |
| FW_LOADER | 128 | 5 |
| MFD_CORE | 124 | 9 |
| FB_CFB_IMAGEBLIT | 114 | 2 |
| FB_CFB_COPYAREA | 111 | 2 |
| FB_CFB_FILLRECT | 110 | 2 |
| SND_PCM | 103 | 2 |
| CRYPTO_HASH | 87 | 19 |
| WATCHDOG_CORE | 86 | 6 |
The story behind the above is that users need to visually
review/evaluate 212 expressions which *potentially* select REGMAP_I2C
in order to identify the expressions which *actually* select REGMAP_I2C,
for a particular ARCH and for a particular defconfig used.
To make this experience smoother, change the way reverse dependencies
are displayed to the user from [1] to [2].
[1] Old representation of DMA_ENGINE_RAID:
Selected by:
- AMCC_PPC440SPE_ADMA [=n] && DMADEVICES [=y] && (440SPe || 440SP)
- BCM_SBA_RAID [=m] && DMADEVICES [=y] && (ARM64 [=y] || ...
- FSL_RAID [=n] && DMADEVICES [=y] && FSL_SOC && ...
- INTEL_IOATDMA [=n] && DMADEVICES [=y] && PCI [=y] && X86_64
- MV_XOR [=n] && DMADEVICES [=y] && (PLAT_ORION || ARCH_MVEBU [=y] ...
- MV_XOR_V2 [=y] && DMADEVICES [=y] && ARM64 [=y]
- XGENE_DMA [=n] && DMADEVICES [=y] && (ARCH_XGENE [=y] || ...
- DMATEST [=n] && DMADEVICES [=y] && DMA_ENGINE [=y]
[2] New representation of DMA_ENGINE_RAID:
Selected by [y]:
- MV_XOR_V2 [=y] && DMADEVICES [=y] && ARM64 [=y]
Selected by [m]:
- BCM_SBA_RAID [=m] && DMADEVICES [=y] && (ARM64 [=y] || ...
Selected by [n]:
- AMCC_PPC440SPE_ADMA [=n] && DMADEVICES [=y] && (440SPe || ...
- FSL_RAID [=n] && DMADEVICES [=y] && FSL_SOC && ...
- INTEL_IOATDMA [=n] && DMADEVICES [=y] && PCI [=y] && X86_64
- MV_XOR [=n] && DMADEVICES [=y] && (PLAT_ORION || ARCH_MVEBU [=y] ...
- XGENE_DMA [=n] && DMADEVICES [=y] && (ARCH_XGENE [=y] || ...
- DMATEST [=n] && DMADEVICES [=y] && DMA_ENGINE [=y]
Suggested-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Eugeniu Rosca <erosca@de.adit-jv.com>
Reviewed-by: Petr Vorel <pvorel@suse.cz>
Reviewed-by: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
This commit splits out the special E_OR handling ('-' instead of '||')
into a dedicated helper expr_print_revdev().
Restore the original expr_print() prior to commit 1ccb27143360
("kconfig: make "Selected by:" and "Implied by:" readable").
This makes sense because:
- We need to chop those expressions only when printing the reverse
dependency, and only when E_OR is encountered
- Otherwise, it should be printed as before, so fall back to
expr_print()
This also improves the behavior; for a single line, it was previously
displayed in the same line as "Selected by", like this:
Selected by: A [=n] && B [=n]
This will be displayed in a new line, consistently:
Selected by:
- A [=n] && B [=n]
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Reviewed-by: Petr Vorel <pvorel@suse.cz>
|
|
IMO, we should discourage '---help---' for new help texts, even in cases
where it would be consistent with other help texts in the file. This
will help if we ever want to get rid of '---help---' in the future.
Also simplify the code to only check for exactly '---help---'. Since
commit c2264564df3d ("kconfig: warn of unhandled characters in Kconfig
commands"), '---help---' is a proper keyword and can only appear in that
form. Prior to that commit, '---help---' working was more of a syntactic
quirk.
Signed-off-by: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
Currently, only Kconfig symbols are checked for a missing or short help
text, and are only checked if they are defined with the 'config'
keyword.
To make the check more general, extend it to also check help texts for
choices and for symbols defined with the 'menuconfig' keyword.
This increases the accuracy of the check for symbols that would already
have been checked as well, since e.g. a 'menuconfig' symbol after a help
text will be recognized as ending the preceding symbol/choice
definition.
To increase the accuracy of the check further, also recognize 'if',
'endif', 'menu', 'endmenu', 'endchoice', and 'source' as ending a
symbol/choice definition.
Signed-off-by: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
The check for a missing or short help text only considers symbols with a
prompt, but doesn't recognize any of the following as a prompt:
bool 'foo'
tristate 'foo'
prompt "foo"
prompt 'foo'
Make the check recognize those too.
Signed-off-by: Ulf Magnusson <ulfalizer@gmail.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
|
|
|
|
On the CP110 components which are present on the Armada 7K/8K SoC we need
to explicitly enable the clock for the registers. However it is not
needed for the AP8xx component, that's why this clock is optional.
With this patch both clock have now a name, but in order to be backward
compatible, the name of the first clock is not used. It allows to still
use this clock with a device tree using the old binding.
Signed-off-by: Gregory CLEMENT <gregory.clement@bootlin.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Signed-off-by: Vinod Koul <vinod.koul@intel.com>
|
|
As reported by Dan the parentheses is in the wrong place, and since
unlikely() call returns either 0 or 1 it's never less than zero. The
second issue is that signed integer overflows like "INT_MAX + 1" are
undefined behavior.
Since num_test_devs represents the number of devices, we want to stop
prior to hitting the max, and not rely on the wrap arround at all. So
just cap at num_test_devs + 1, prior to assigning a new device.
Link: http://lkml.kernel.org/r/20180224030046.24238-1-mcgrof@kernel.org
Fixes: d9c6a72d6fa2 ("kmod: add test driver to stress test the module loader")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Luis R. Rodriguez <mcgrof@kernel.org>
Acked-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Fix userfaultfd_hugetlb on hosts which have more than 64 cpus.
---------------------------
running userfaultfd_hugetlb
---------------------------
invalid MiB
Usage: <MiB> <bounces>
[FAIL]
Via userfaultfd.c we can know, hugetlb_size needs to meet hugetlb_size
>= nr_cpus * hugepage_size. hugepage_size is often 2M, so when host
cpus > 64, it requires more than 128M.
[zhijianx.li@intel.com: update changelog/comments and variable name]
Link: http://lkml.kernel.org/r/20180302024356.83359-1-zhijianx.li@intel.com
Link: http://lkml.kernel.org/r/20180303125027.81638-1-zhijianx.li@intel.com
Link: http://lkml.kernel.org/r/20180302024356.83359-1-zhijianx.li@intel.com
Signed-off-by: Li Zhijian <zhijianx.li@intel.com>
Cc: Shuah Khan <shuah@kernel.org>
Cc: SeongJae Park <sj38.park@gmail.com>
Cc: Philippe Ombredanne <pombredanne@nexb.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Commit b92df1de5d28 ("mm: page_alloc: skip over regions of invalid pfns
where possible") introduced a bug where move_freepages() triggers a
VM_BUG_ON() on uninitialized page structure due to pageblock alignment.
To fix this, simply align the skipped pfns in memmap_init_zone() the
same way as in move_freepages_block().
Seen in one of the RHEL reports:
crash> log | grep -e BUG -e RIP -e Call.Trace -e move_freepages_block -e rmqueue -e freelist -A1
kernel BUG at mm/page_alloc.c:1389!
invalid opcode: 0000 [#1] SMP
--
RIP: 0010:[<ffffffff8118833e>] [<ffffffff8118833e>] move_freepages+0x15e/0x160
RSP: 0018:ffff88054d727688 EFLAGS: 00010087
--
Call Trace:
[<ffffffff811883b3>] move_freepages_block+0x73/0x80
[<ffffffff81189e63>] __rmqueue+0x263/0x460
[<ffffffff8118c781>] get_page_from_freelist+0x7e1/0x9e0
[<ffffffff8118caf6>] __alloc_pages_nodemask+0x176/0x420
--
RIP [<ffffffff8118833e>] move_freepages+0x15e/0x160
RSP <ffff88054d727688>
crash> page_init_bug -v | grep RAM
<struct resource 0xffff88067fffd2f8> 1000 - 9bfff System RAM (620.00 KiB)
<struct resource 0xffff88067fffd3a0> 100000 - 430bffff System RAM ( 1.05 GiB = 1071.75 MiB = 1097472.00 KiB)
<struct resource 0xffff88067fffd410> 4b0c8000 - 4bf9cfff System RAM ( 14.83 MiB = 15188.00 KiB)
<struct resource 0xffff88067fffd480> 4bfac000 - 646b1fff System RAM (391.02 MiB = 400408.00 KiB)
<struct resource 0xffff88067fffd560> 7b788000 - 7b7fffff System RAM (480.00 KiB)
<struct resource 0xffff88067fffd640> 100000000 - 67fffffff System RAM ( 22.00 GiB)
crash> page_init_bug | head -6
<struct resource 0xffff88067fffd560> 7b788000 - 7b7fffff System RAM (480.00 KiB)
<struct page 0xffffea0001ede200> 1fffff00000000 0 <struct pglist_data 0xffff88047ffd9000> 1 <struct zone 0xffff88047ffd9800> DMA32 4096 1048575
<struct page 0xffffea0001ede200> 505736 505344 <struct page 0xffffea0001ed8000> 505855 <struct page 0xffffea0001edffc0>
<struct page 0xffffea0001ed8000> 0 0 <struct pglist_data 0xffff88047ffd9000> 0 <struct zone 0xffff88047ffd9000> DMA 1 4095
<struct page 0xffffea0001edffc0> 1fffff00000400 0 <struct pglist_data 0xffff88047ffd9000> 1 <struct zone 0xffff88047ffd9800> DMA32 4096 1048575
BUG, zones differ!
Note that this range follows two not populated sections
68000000-77ffffff in this zone. 7b788000-7b7fffff is the first one
after a gap. This makes memmap_init_zone() skip all the pfns up to the
beginning of this range. But this range is not pageblock (2M) aligned.
In fact no range has to be.
crash> kmem -p 77fff000 78000000 7b5ff000 7b600000 7b787000 7b788000
PAGE PHYSICAL MAPPING INDEX CNT FLAGS
ffffea0001e00000 78000000 0 0 0 0
ffffea0001ed7fc0 7b5ff000 0 0 0 0
ffffea0001ed8000 7b600000 0 0 0 0 <<<<
ffffea0001ede1c0 7b787000 0 0 0 0
ffffea0001ede200 7b788000 0 0 1 1fffff00000000
Top part of page flags should contain nodeid and zonenr, which is not
the case for page ffffea0001ed8000 here (<<<<).
crash> log | grep -o fffea0001ed[^\ ]* | sort -u
fffea0001ed8000
fffea0001eded20
fffea0001edffc0
crash> bt -r | grep -o fffea0001ed[^\ ]* | sort -u
fffea0001ed8000
fffea0001eded00
fffea0001eded20
fffea0001edffc0
Initialization of the whole beginning of the section is skipped up to
the start of the range due to the commit b92df1de5d28. Now any code
calling move_freepages_block() (like reusing the page from a freelist as
in this example) with a page from the beginning of the range will get
the page rounded down to start_page ffffea0001ed8000 and passed to
move_freepages() which crashes on assertion getting wrong zonenr.
> VM_BUG_ON(page_zone(start_page) != page_zone(end_page));
Note, page_zone() derives the zone from page flags here.
From similar machine before commit b92df1de5d28:
crash> kmem -p 77fff000 78000000 7b5ff000 7b600000 7b7fe000 7b7ff000
PAGE PHYSICAL MAPPING INDEX CNT FLAGS
fffff73941e00000 78000000 0 0 1 1fffff00000000
fffff73941ed7fc0 7b5ff000 0 0 1 1fffff00000000
fffff73941ed8000 7b600000 0 0 1 1fffff00000000
fffff73941edff80 7b7fe000 0 0 1 1fffff00000000
fffff73941edffc0 7b7ff000 ffff8e67e04d3ae0 ad84 1 1fffff00020068 uptodate,lru,active,mappedtodisk
All the pages since the beginning of the section are initialized.
move_freepages()' not gonna blow up.
The same machine with this fix applied:
crash> kmem -p 77fff000 78000000 7b5ff000 7b600000 7b7fe000 7b7ff000
PAGE PHYSICAL MAPPING INDEX CNT FLAGS
ffffea0001e00000 78000000 0 0 0 0
ffffea0001e00000 7b5ff000 0 0 0 0
ffffea0001ed8000 7b600000 0 0 1 1fffff00000000
ffffea0001edff80 7b7fe000 0 0 1 1fffff00000000
ffffea0001edffc0 7b7ff000 ffff88017fb13720 8 2 1fffff00020068 uptodate,lru,active,mappedtodisk
At least the bare minimum of pages is initialized preventing the crash
as well.
Customers started to report this as soon as 7.4 (where b92df1de5d28 was
merged in RHEL) was released. I remember reports from
September/October-ish times. It's not easily reproduced and happens on
a handful of machines only. I guess that's why. But that does not make
it less serious, I think.
Though there actually is a report here:
https://bugzilla.kernel.org/show_bug.cgi?id=196443
And there are reports for Fedora from July:
https://bugzilla.redhat.com/show_bug.cgi?id=1473242
and CentOS:
https://bugs.centos.org/view.php?id=13964
and we internally track several dozens reports for RHEL bug
https://bugzilla.redhat.com/show_bug.cgi?id=1525121
Link: http://lkml.kernel.org/r/0485727b2e82da7efbce5f6ba42524b429d0391a.1520011945.git.neelx@redhat.com
Fixes: b92df1de5d28 ("mm: page_alloc: skip over regions of invalid pfns where possible")
Signed-off-by: Daniel Vacek <neelx@redhat.com>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
This is just a cleanup. It aids handling the special end case in the
next commit.
[akpm@linux-foundation.org: make it work against current -linus, not against -mm]
[akpm@linux-foundation.org: make it work against current -linus, not against -mm some more]
Link: http://lkml.kernel.org/r/1ca478d4269125a99bcfb1ca04d7b88ac1aee924.1520011944.git.neelx@redhat.com
Signed-off-by: Daniel Vacek <neelx@redhat.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Vlastimil Babka <vbabka@suse.cz>
Cc: Mel Gorman <mgorman@techsingularity.net>
Cc: Pavel Tatashin <pasha.tatashin@oracle.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
KVM is hanging during postcopy live migration with userfaultfd because
get_user_pages_unlocked is not capable to handle FOLL_NOWAIT.
Earlier FOLL_NOWAIT was only ever passed to get_user_pages.
Specifically faultin_page (the callee of get_user_pages_unlocked caller)
doesn't know that if FAULT_FLAG_RETRY_NOWAIT was set in the page fault
flags, when VM_FAULT_RETRY is returned, the mmap_sem wasn't actually
released (even if nonblocking is not NULL). So it sets *nonblocking to
zero and the caller won't release the mmap_sem thinking it was already
released, but it wasn't because of FOLL_NOWAIT.
Link: http://lkml.kernel.org/r/20180302174343.5421-2-aarcange@redhat.com
Fixes: ce53053ce378c ("kvm: switch get_user_page_nowait() to get_user_pages_unlocked()")
Signed-off-by: Andrea Arcangeli <aarcange@redhat.com>
Reported-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Tested-by: Dr. David Alan Gilbert <dgilbert@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Commit b8347c219649 ("x86/debug: Handle warnings before the notifier
chain, to fix KGDB crash") changed the ordering of fixups, and did not
take into account the case of x86 processing non-WARN() and non-BUG()
exceptions. This would lead to output of a false BUG line with no other
information.
In the case of a refcount exception, it would be immediately followed by
the refcount WARN(), producing very strange double-"cut here":
lkdtm: attempting bad refcount_inc() overflow
------------[ cut here ]------------
Kernel BUG at 0000000065f29de5 [verbose debug info unavailable]
------------[ cut here ]------------
refcount_t overflow at lkdtm_REFCOUNT_INC_OVERFLOW+0x6b/0x90 in cat[3065], uid/euid: 0/0
WARNING: CPU: 0 PID: 3065 at kernel/panic.c:657 refcount_error_report+0x9a/0xa4
...
In the prior ordering, exceptions were searched first:
do_trap_no_signal(struct task_struct *tsk, int trapnr, char *str,
...
if (fixup_exception(regs, trapnr))
return 0;
- if (fixup_bug(regs, trapnr))
- return 0;
-
As a result, fixup_bugs()'s is_valid_bugaddr() didn't take into account
needing to search the exception list first, since that had already
happened.
So, instead of searching the exception list twice (once in
is_valid_bugaddr() and then again in fixup_exception()), just add a
simple sanity check to report_bug() that will immediately bail out if a
BUG() (or WARN()) entry is not found.
Link: http://lkml.kernel.org/r/20180301225934.GA34350@beast
Fixes: b8347c219649 ("x86/debug: Handle warnings before the notifier chain, to fix KGDB crash")
Signed-off-by: Kees Cook <keescook@chromium.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Richard Weinberger <richard.weinberger@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
The BUG and stack protector reports were still using a raw %p. This
changes it to %pB for more meaningful output.
Link: http://lkml.kernel.org/r/20180301225704.GA34198@beast
Fixes: ad67b74d2469 ("printk: hash addresses printed with %p")
Signed-off-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Andrew Morton <akpm@linux-foundation.org>
Cc: Ingo Molnar <mingo@kernel.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Richard Weinberger <richard.weinberger@gmail.com>,
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Dan Rue has noticed that libhugetlbfs test suite fails counter test:
# mount_point="/mnt/hugetlb/"
# echo 200 > /proc/sys/vm/nr_hugepages
# mkdir -p "${mount_point}"
# mount -t hugetlbfs hugetlbfs "${mount_point}"
# export LD_LIBRARY_PATH=/root/libhugetlbfs/libhugetlbfs-2.20/obj64
# /root/libhugetlbfs/libhugetlbfs-2.20/tests/obj64/counters
Starting testcase "/root/libhugetlbfs/libhugetlbfs-2.20/tests/obj64/counters", pid 3319
Base pool size: 0
Clean...
FAIL Line 326: Bad HugePages_Total: expected 0, actual 1
The bug was bisected to 0c397daea1d4 ("mm, hugetlb: further simplify
hugetlb allocation API").
The reason is that alloc_surplus_huge_page() misaccounts per node
surplus pages. We should increase surplus_huge_pages_node rather than
nr_huge_pages_node which is already handled by alloc_fresh_huge_page.
Link: http://lkml.kernel.org/r/20180221191439.GM2231@dhcp22.suse.cz
Fixes: 0c397daea1d4 ("mm, hugetlb: further simplify hugetlb allocation API")
Signed-off-by: Michal Hocko <mhocko@suse.com>
Reported-by: Dan Rue <dan.rue@linaro.org>
Tested-by: Dan Rue <dan.rue@linaro.org>
Reviewed-by: Mike Kravetz <mike.kravetz@oracle.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
The user can provide very large cqe_size which will cause to integer
overflow as it can be seen in the following UBSAN warning:
=======================================================================
UBSAN: Undefined behaviour in drivers/infiniband/hw/mlx5/cq.c:1192:53
signed integer overflow:
64870 * 65536 cannot be represented in type 'int'
CPU: 0 PID: 267 Comm: syzkaller605279 Not tainted 4.15.0+ #90 Hardware
name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS
rel-1.7.5-0-ge51488c-20140602_164612-nilsson.home.kraxel.org 04/01/2014
Call Trace:
dump_stack+0xde/0x164
? dma_virt_map_sg+0x22c/0x22c
ubsan_epilogue+0xe/0x81
handle_overflow+0x1f3/0x251
? __ubsan_handle_negate_overflow+0x19b/0x19b
? lock_acquire+0x440/0x440
mlx5_ib_resize_cq+0x17e7/0x1e40
? cyc2ns_read_end+0x10/0x10
? native_read_msr_safe+0x6c/0x9b
? cyc2ns_read_end+0x10/0x10
? mlx5_ib_modify_cq+0x220/0x220
? sched_clock_cpu+0x18/0x200
? lookup_get_idr_uobject+0x200/0x200
? rdma_lookup_get_uobject+0x145/0x2f0
ib_uverbs_resize_cq+0x207/0x3e0
? ib_uverbs_ex_create_cq+0x250/0x250
ib_uverbs_write+0x7f9/0xef0
? cyc2ns_read_end+0x10/0x10
? print_irqtrace_events+0x280/0x280
? ib_uverbs_ex_create_cq+0x250/0x250
? uverbs_devnode+0x110/0x110
? sched_clock_cpu+0x18/0x200
? do_raw_spin_trylock+0x100/0x100
? __lru_cache_add+0x16e/0x290
__vfs_write+0x10d/0x700
? uverbs_devnode+0x110/0x110
? kernel_read+0x170/0x170
? sched_clock_cpu+0x18/0x200
? security_file_permission+0x93/0x260
vfs_write+0x1b0/0x550
SyS_write+0xc7/0x1a0
? SyS_read+0x1a0/0x1a0
? trace_hardirqs_on_thunk+0x1a/0x1c
entry_SYSCALL_64_fastpath+0x1e/0x8b
RIP: 0033:0x433549
RSP: 002b:00007ffe63bd1ea8 EFLAGS: 00000217
=======================================================================
Cc: syzkaller <syzkaller@googlegroups.com>
Cc: <stable@vger.kernel.org> # 3.13
Fixes: bde51583f49b ("IB/mlx5: Add support for resize CQ")
Reported-by: Noa Osherovich <noaos@mellanox.com>
Reviewed-by: Yishai Hadas <yishaih@mellanox.com>
Signed-off-by: Leon Romanovsky <leonro@mellanox.com>
Signed-off-by: Doug Ledford <dledford@redhat.com>
|
|
The original commit of this patch has a munged log message that is
missing several of the tags the original author intended to be on the
patch. This was due to patchworks misinterpreting a cut-n-paste
separator line as an end of message line and munging the mbox that was
used to import the patch:
https://patchwork.kernel.org/patch/10264089/
The original patch will be reapplied with a fixed commit message so the
proper tags are applied.
This reverts commit aa0de36a40f446f5a21a7c1e677b98206e242edb.
Signed-off-by: Doug Ledford <dledford@redhat.com>
|
|
A recent update to the ARM SMCCC ARCH_WORKAROUND_1 specification
allows firmware to return a non zero, positive value to describe
that although the mitigation is implemented at the higher exception
level, the CPU on which the call is made is not affected.
Let's relax the check on the return value from ARCH_WORKAROUND_1
so that we only error out if the returned value is negative.
Fixes: b092201e0020 ("arm64: Add ARM_SMCCC_ARCH_WORKAROUND_1 BP hardening support")
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
|
|
Sphinx 1.7 removed sphinx.util.compat.Directive so people
who have upgraded cannot build the documentation. Switch to
docutils.parsers.rst.Directive which has been available since
docutils 0.5 released in 2009.
Bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=1083694
Co-developed-by: Takashi Iwai <tiwai@suse.de>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Cc: stable@vger.kernel.org
Signed-off-by: Matthew Wilcox <mawilcox@microsoft.com>
Signed-off-by: Jonathan Corbet <corbet@lwn.net>
|
|
When the DELL_SMBIOS_SMM backend is enabled, the DELL_SMBIOS symbol
depends on DELL_DCDBAS, and we must avoid the situation where
DELL_SMBIOS=y and DCDBAS=m.
Adding the conditional dependency to DELL_SMBIOS such as:
depends !DELL_SMBIOS_SMM || (DCDBAS || DCDBAS=n)
results in the Kconfig tooling complaining about a circular dependency,
although it appears to work in practice.
Avoid the errors by simplifying the dependency and forcing DELL_SMBIOS
to be <= DCDBAS if DCDBAS is enabled (thanks to Greg KH for the
suggestion).
Cc: Mario.Limonciello@dell.com
Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org>
|
|
Avoid accidental configurations by setting default y for DELL_SMBIOS
backends. Avoid this impacting the default build size, by making them
dependent on DELL_SMBIOS, so they only appear when DELL_SMBIOS is
manually selected, or by DELL_LAPTOP or DELL_WMI.
While DELL_SMBIOS does have a prompt, it does not have any dependencies.
Keeping DELL_SMBIOS visible, despite being "select"ed by DELL_LAPTOP and
DELL_WMI, is a deliberate choice to provide context for the WMI and SMM
backends, which would otherwise appear to float without context within
the menu.
Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org>
|
|
Some race conditions were raised due to dell-smbios and its backends
not being ready by the time that a consumer would call one of the
exported methods.
To avoid this problem, guarantee that all initialization has been
done by linking them all together and running init for them all.
As part of this change the Kconfig needs to be adjusted so that
CONFIG_DELL_SMBIOS_SMM and CONFIG_DELL_SMBIOS_WMI are boolean
rather than modules.
CONFIG_DELL_SMBIOS is a visually selectable option again and both
CONFIG_DELL_SMBIOS_WMI and CONFIG_DELL_SMBIOS_SMM are optional.
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
[dvhart: Update prompt and help text for DELL_SMBIOS_* backends]
Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org>
|
|
This is being done to faciliate a later change to link all the dell-smbios
drivers together.
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org>
|
|
WARNING: function definition argument 'struct calling_interface_buffer *'
should also have an identifier name
+ int (*call_fn)(struct calling_interface_buffer *);
WARNING: Block comments use * on subsequent lines
+ /* 4 bytes of table header, plus 7 bytes of Dell header,
plus at least
+ 6 bytes of entry */
WARNING: Block comments use a trailing */ on a separate line
+ 6 bytes of entry */
Signed-off-by: Mario Limonciello <mario.limonciello@dell.com>
Signed-off-by: Darren Hart (VMware) <dvhart@infradead.org>
|
|
The following commit:
commit aa4d86163e4e ("block: loop: switch to VFS ITER_BVEC")
replaced __do_lo_send_write(), which used ITER_KVEC iterators, with
lo_write_bvec() which uses ITER_BVEC iterators. In this change, though,
the WRITE flag was lost:
- iov_iter_kvec(&from, ITER_KVEC | WRITE, &kvec, 1, len);
+ iov_iter_bvec(&i, ITER_BVEC, bvec, 1, bvec->bv_len);
This flag is necessary for the DAX case because we make decisions based on
whether or not the iterator is a READ or a WRITE in dax_iomap_actor() and
in dax_iomap_rw().
We end up going through this path in configurations where we combine a PMEM
device with 4k sectors, a loopback device and DAX. The consequence of this
missed flag is that what we intend as a write actually turns into a read in
the DAX code, so no data is ever written.
The very simplest test case is to create a loopback device and try and
write a small string to it, then hexdump a few bytes of the device to see
if the write took. Without this patch you read back all zeros, with this
you read back the string you wrote.
For XFS this causes us to fail or panic during the following xfstests:
xfs/074 xfs/078 xfs/216 xfs/217 xfs/250
For ext4 we have a similar issue where writes never happen, but we don't
currently have any xfstests that use loopback and show this issue.
Fix this by restoring the WRITE flag argument to iov_iter_bvec(). This
causes the xfstests to all pass.
Cc: Al Viro <viro@zeniv.linux.org.uk>
Cc: stable@vger.kernel.org
Fixes: commit aa4d86163e4e ("block: loop: switch to VFS ITER_BVEC")
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Ross Zwisler <ross.zwisler@linux.intel.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
The ATMEL_ST config selects MFD_SYSCON, but does not depend on HAS_IOMEM.
Compile testing on architecture without HAS_IOMEM causes "unmet direct
dependencies" in Kconfig phase. Detected by "make ARCH=score allyesconfig".
Add the proper dependency to the ATMEL_ST config.
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Arnd Bergmann <arnd@arndb.de>
Link: https://lkml.kernel.org/r/1520335233-11277-1-git-send-email-yamada.masahiro@socionext.com
|
|
When running rcutorture with TREE03 config, CONFIG_PROVE_LOCKING=y, and
kernel cmdline argument "rcutorture.gp_exp=1", lockdep reports a
HARDIRQ-safe->HARDIRQ-unsafe deadlock:
================================
WARNING: inconsistent lock state
4.16.0-rc4+ #1 Not tainted
--------------------------------
inconsistent {IN-HARDIRQ-W} -> {HARDIRQ-ON-W} usage.
takes:
__schedule+0xbe/0xaf0
{IN-HARDIRQ-W} state was registered at:
_raw_spin_lock+0x2a/0x40
scheduler_tick+0x47/0xf0
...
other info that might help us debug this:
Possible unsafe locking scenario:
CPU0
----
lock(&rq->lock);
<Interrupt>
lock(&rq->lock);
*** DEADLOCK ***
1 lock held by rcu_torture_rea/724:
rcu_torture_read_lock+0x0/0x70
stack backtrace:
CPU: 2 PID: 724 Comm: rcu_torture_rea Not tainted 4.16.0-rc4+ #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.11.0-20171110_100015-anatol 04/01/2014
Call Trace:
lock_acquire+0x90/0x200
? __schedule+0xbe/0xaf0
_raw_spin_lock+0x2a/0x40
? __schedule+0xbe/0xaf0
__schedule+0xbe/0xaf0
preempt_schedule_irq+0x2f/0x60
retint_kernel+0x1b/0x2d
RIP: 0010:rcu_read_unlock_special+0x0/0x680
? rcu_torture_read_unlock+0x60/0x60
__rcu_read_unlock+0x64/0x70
rcu_torture_read_unlock+0x17/0x60
rcu_torture_reader+0x275/0x450
? rcutorture_booster_init+0x110/0x110
? rcu_torture_stall+0x230/0x230
? kthread+0x10e/0x130
kthread+0x10e/0x130
? kthread_create_worker_on_cpu+0x70/0x70
? call_usermodehelper_exec_async+0x11a/0x150
ret_from_fork+0x3a/0x50
This happens with the following even sequence:
preempt_schedule_irq();
local_irq_enable();
__schedule():
local_irq_disable(); // irq off
...
rcu_note_context_switch():
rcu_note_preempt_context_switch():
rcu_read_unlock_special():
local_irq_save(flags);
...
raw_spin_unlock_irqrestore(...,flags); // irq remains off
rt_mutex_futex_unlock():
raw_spin_lock_irq();
...
raw_spin_unlock_irq(); // accidentally set irq on
<return to __schedule()>
rq_lock():
raw_spin_lock(); // acquiring rq->lock with irq on
which means rq->lock becomes a HARDIRQ-unsafe lock, which can cause
deadlocks in scheduler code.
This problem was introduced by commit 02a7c234e540 ("rcu: Suppress
lockdep false-positive ->boost_mtx complaints"). That brought the user
of rt_mutex_futex_unlock() with irq off.
To fix this, replace the *lock_irq() in rt_mutex_futex_unlock() with
*lock_irq{save,restore}() to make it safe to call rt_mutex_futex_unlock()
with irq off.
Fixes: 02a7c234e540 ("rcu: Suppress lockdep false-positive ->boost_mtx complaints")
Signed-off-by: Boqun Feng <boqun.feng@gmail.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Lai Jiangshan <jiangshanlai@gmail.com>
Cc: Steven Rostedt <rostedt@goodmis.org>
Cc: Josh Triplett <josh@joshtriplett.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com>
Link: https://lkml.kernel.org/r/20180309065630.8283-1-boqun.feng@gmail.com
|
|
Disable the kprobe probing of the entry trampoline:
.entry_trampoline is a code area that is used to ensure page table
isolation between userspace and kernelspace.
At the beginning of the execution of the trampoline, we load the
kernel's CR3 register. This has the effect of enabling the translation
of the kernel virtual addresses to physical addresses. Before this
happens most kernel addresses can not be translated because the running
process' CR3 is still used.
If a kprobe is placed on the trampoline code before that change of the
CR3 register happens the kernel crashes because int3 handling pages are
not accessible.
To fix this, add the .entry_trampoline section to the kprobe blacklist
to prohibit the probing of code before all the kernel pages are
accessible.
Signed-off-by: Francis Deslauriers <francis.deslauriers@efficios.com>
Reviewed-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Brian Gerst <brgerst@gmail.com>
Cc: Denys Vlasenko <dvlasenk@redhat.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Josh Poimboeuf <jpoimboe@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: mathieu.desnoyers@efficios.com
Cc: mhiramat@kernel.org
Link: http://lkml.kernel.org/r/1520565492-4637-2-git-send-email-francis.deslauriers@efficios.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|