| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
The DSI driver for Tegra requires the SOR power partition to be enabled.
Now that Tegra supports the generic PM domain framework we manage the
SOR power partition via this framework. However, the sequence for
gating/ungating the SOR power partition requires that the DSI reset is
asserted/de-asserted at the time the SOR power partition is
gated/ungated, respectively. Now that the reset control core assumes
that resets are exclusive, the Tegra generic PM domain code and the DSI
driver cannot request the same reset unless we mark the reset as shared.
Sharing resets will not work in this case because we cannot guarantee
that the reset will be asserted/de-asserted at the appropriate time.
Therefore, given that the Tegra generic PM domain code will handle the
resets, do not request the reset in the DSI driver if the DSI device has
a PM domain associated.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Provide a per-SoC mapping of lanes which can be used to configure the
XBAR.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
When running in HDMI mode, the sor1 IP block needs to use the sor1_src
as parent clock, and in turn configure the sor1_src to use pll_d2_out0
as its parent.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
The SOR clock can have various sources, with the most commonly used
being the sor_safe, pll_d2_out0, pll_dp and sor_brick clocks. These
are configured using a three level mux, of which the first 2 levels
can be treated as one. The direct parents of the SOR clock are the
sor_safe, sor_brick and sor_src clocks, whereas the pll_d2_out0 and
pll_dp clocks can be selected as parents of the sor_src clock via a
second mux.
Previous generations of Tegra have only supported eDP and LVDS with
the SOR, where LVDS was never used on publicly available hardware.
Clocking for this only ever required the first level mux (to select
between sor_safe and sor_brick).
Tegra210 has a new revision of the SOR that supports HDMI and hence
needs to support the second level mux to allow selecting pll_d2_out0
as the SOR clock's parent. This second mux is knows as sor_src, and
operating system software needs a reference to it in order to select
the proper parent.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
sor1_brick is a clock that can be used as a source for the sor1 clock.
The registers to control the clock output are part of the sor1 IP block
and hence the sor driver is the best place to implement it.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Use runtime PM to clock-(un)gate and (de)assert reset to the SOR
controller. This ties in nicely with atomic DPMS in that a runtime PM
reference is taken before a pipe is enabled and dropped after it has
been shut down.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Use runtime PM to clock-(un)gate and (de)assert reset to the HDMI
controller. This ties in nicely with atomic DPMS in that a runtime PM
reference is taken before a pipe is enabled and dropped after it has
been shut down.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Use runtime PM to clock-(un)gate, (de)assert reset and control power to
the DSI controller. This ties in nicely with atomic DPMS in that a
runtime PM reference is taken before a pipe is enabled and dropped after
it has been shut down.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Use runtime PM to clock-gate, assert reset and powergate the display
controller. This ties in nicely with atomic DPMS in that a runtime PM
reference is taken before a pipe is enabled and dropped after it has
been shut down.
To make sure this works, make sure to only ever update planes on active
CRTCs, otherwise register accesses to a clock-gated and reset CRTC will
hang the CPU.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
In order to use the HDA codec to forward audio data to the HDMI codec it
needs the ELD that is parsed from the monitor's EDID.
Also implement an interoperability mechanism between the HDA controller
and the HDMI codec. This uses vendor-defined scratch registers to pass
data from the HDMI codec driver to the HDMI driver (that implements the
receiving end of the HDMI codec). A custom format is used to pass audio
sample rate and channel count to the HDMI driver.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Current generations of Tegra do not support deep color modes, so force
8 bits per color even if the connected monitor or panel supports more.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
The code to set a video mode is common to all types of outputs that the
SOR can drive. Extract it into a separate function so that it can be
shared.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
This function is useful in both eDP and DP modes, so split it out in
anticipation of adding DP support.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Use a slightly more sensible name, tegra_sor_compute_config().
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Switching the SOR parent clock can glitch if done while the clock is
enabled. Extract a common function that can be used to disable the
module clock, switch the parent and reenable the module clock.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
The DPAUX pins are shared with an internal I2C controller. To allow
these pins to be muxed to the I2C controller, register a pinctrl device
for the DPAUX device.
This is based upon work by Thierry Reding <treding@nvidia.com>.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
On Tegra124, Tegra132 and Tegra210 devices the pads used by the Display
Port Auxiliary (DPAUX) channel are multiplexed such that they can also
be used by one of the internal I2C controllers. Note that this is
different from I2C-over-AUX supported by the DPAUX controller. The
register that configures these pads is part of the DPAUX controllers
register set and so a pinctrl driver is being added for the DPAUX device
to share these pads. Add the device-tree binding documentation for the
DPAUX pad controller.
Although there is only one group of pads associated with the DPAUX that
can be multiplexed, the group still needs to be described by the binding.
If the 'groups' property is not present in the binding, then the pads
will not be allocated by the pinctrl core for a client and this would
allow another client to re-configure the same pads that may already be
in-use.
Please note that although the "off" function for the DPAUX pads is not
technically a pin-mux setting but more of a pin-conf setting it is
simpler to expose these as a function so that the user can simply select
either "aux", "i2c" or "off" as the current function/mode.
Update the main DPAUX binding documentation to reference the DPAUX pad
controller binding document and add the 'i2c-bus' subnode. The 'i2c-bus'
subnode is used for populating I2C slaves for the DPAUX device so that
the I2C driver core does not attempt to add the DPAUX pad controller
nodes as I2C slaves.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
To utilise the DPAUX on Tegra, the SOR power partition must be enabled.
Now that Tegra supports the generic PM domain framework we manage the
SOR power partition via this framework for DPAUX. However, the sequence
for gating/ungating the SOR power partition requires that the DPAUX
reset is asserted/de-asserted at the time the SOR power partition is
gated/ungated, respectively. Now that the reset control core assumes
that resets are exclusive, the Tegra generic PM domain code and the
DPAUX driver cannot request the same reset unless we mark the resets as
shared. Sharing resets will not work in this case because we cannot
guarantee that the reset will be asserted/de-asserted at the appropriate
time. Therefore, given that the Tegra generic PM domain code will handle
the DPAUX reset, do not request the reset in the DPAUX driver if the
DPAUX device has a PM domain associated.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Update the DPAUX compatibility string information for Tegra124, Tegra132
and Tegra210.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
In preparation for adding pinctrl support for the DPAUX pads, add a
couple of helpers functions to configure the pads and control their
power.
Please note that although a simple if-statement could be used instead
of a case statement for configuring the pads as there are only two
possible modes, a case statement is used because when integrating with
the pinctrl framework, we need to be able to handle invalid modes that
could be passed.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
If the probing of the DPAUX fails, then clocks are left enabled and the
DPAUX reset de-asserted. Add code to perform the necessary clean-up on
probe failure by disabling clocks and asserting the reset.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
The local 'val' variable is used to store a value and immediately return
it to its caller, and hence serves no purpose. Just drop it and directly
return the value.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
This array never needs to be modified and therefore can be read-only
data.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
There's no need to wrap the BIT() macro into an extra set of parentheses
because it's already implemented to use its own set.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Insert a number of blank lines in places where they increase readability
of the code. Also collapse various variable declarations to shorten some
functions and finally rewrite some code for readability.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Fix a couple of occurrences where no blank line was used to separate
variable declarations from code or where block comments were wrongly
formatted.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Use kcalloc() to allocate arrays rather than passing the product of the
size per element by the number of elements to kzalloc().
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
The local 'pos' variable doesn't serve any purpose other than being a
shortcut for pb->pos, but the result doesn't remove much, so simply drop
the local variable.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
find_first_zero_bit() returns an unsigned long, so make the local
variable that stores the result the same type for consistency.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
IDs can never be negative so use unsigned int. In some instances an
explicitly sized type (such as u32) was used for no particular reason,
so turn those into unsigned int as well for consistency.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
The number of channels, syncpoints, bases and mlocks can never be
negative, so use unsigned int instead of int. Also make loop variables
the same type for consistency.
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
System workqueues have been able to handle high level of concurrency
for a long time now and there's no reason to use dedicated workqueues
just to gain concurrency. Since the workqueue host->intr_wq is involved
in sync point interrupts, and sync point wait and is not being used on
a memory reclaim path, dedicated host->intr_wq has been replaced with the
use of system_wq.
Unlike a dedicated per-cpu workqueue created with create_workqueue(),
system_wq allows multiple work items to overlap executions even on
the same CPU; however, a per-cpu workqueue doesn't have any CPU
locality or global ordering guarantees unless the target CPU is
explicitly specified and thus the increase of local concurrency
shouldn't make any difference.
cancel_work_sync() has been used in _host1x_free_syncpt_irq() to ensure
that no work is pending by the time exit path runs.
Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com>
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/linusw/linux-pinctrl into drm/tegra/for-next
|
|
The pinconf-generic.h file exposes functions for creating generic mappings
but it does not expose a function for freeing the mappings. Add a function
for freeing generic mappings.
Signed-off-by: Jon Hunter <jonathanh@nvidia.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
|
|
|
|
The self-test was updated to cover zero-length strings; the function
needs to be updated, too.
Reported-by: Geert Uytterhoeven <geert@linux-m68k.org>
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Fixes: fcfd2fbf22d2 ("fs/namei.c: Add hashlen_string() function")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
The original name was simply hash_string(), but that conflicted with a
function with that name in drivers/base/power/trace.c, and I decided
that calling it "hashlen_" was better anyway.
But you have to do it in two places.
[ This caused build errors for architectures that don't define
CONFIG_DCACHE_WORD_ACCESS - Linus ]
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Fixes: fcfd2fbf22d2 ("fs/namei.c: Add hashlen_string() function")
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
The HPFS filesystem used generic_show_options to produce string that is
displayed in /proc/mounts. However, there is a problem that the options
may disappear after remount. If we mount the filesystem with option1
and then remount it with option2, /proc/mounts should show both option1
and option2, however it only shows option2 because the whole option
string is replaced with replace_mount_options in hpfs_remount_fs.
To fix this bug, implement the hpfs_show_options function that prints
options that are currently selected.
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Commit c8f33d0bec99 ("affs: kstrdup() memory handling") checks if the
kstrdup function returns NULL due to out-of-memory condition.
However, if we are remounting a filesystem with no change to
filesystem-specific options, the parameter data is NULL. In this case,
kstrdup returns NULL (because it was passed NULL parameter), although no
out of memory condition exists. The mount syscall then fails with
ENOMEM.
This patch fixes the bug. We fail with ENOMEM only if data is non-NULL.
The patch also changes the call to replace_mount_options - if we didn't
pass any filesystem-specific options, we don't call
replace_mount_options (thus we don't erase existing reported options).
Fixes: c8f33d0bec99 ("affs: kstrdup() memory handling")
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org # v4.1+
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Commit ce657611baf9 ("hpfs: kstrdup() out of memory handling") checks if
the kstrdup function returns NULL due to out-of-memory condition.
However, if we are remounting a filesystem with no change to
filesystem-specific options, the parameter data is NULL. In this case,
kstrdup returns NULL (because it was passed NULL parameter), although no
out of memory condition exists. The mount syscall then fails with
ENOMEM.
This patch fixes the bug. We fail with ENOMEM only if data is non-NULL.
The patch also changes the call to replace_mount_options - if we didn't
pass any filesystem-specific options, we don't call
replace_mount_options (thus we don't erase existing reported options).
Fixes: ce657611baf9 ("hpfs: kstrdup() out of memory handling")
Signed-off-by: Mikulas Patocka <mpatocka@redhat.com>
Cc: stable@vger.kernel.org
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Pull more MIPS updates from Ralf Baechle:
"This is the secondnd batch of MIPS patches for 4.7. Summary:
CPS:
- Copy EVA configuration when starting secondary VPs.
EIC:
- Clear Status IPL.
Lasat:
- Fix a few off by one bugs.
lib:
- Mark intrinsics notrace. Not only are the intrinsics
uninteresting, it would cause infinite recursion.
MAINTAINERS:
- Add file patterns for MIPS BRCM device tree bindings.
- Add file patterns for mips device tree bindings.
MT7628:
- Fix MT7628 pinmux typos.
- wled_an pinmux gpio.
- EPHY LEDs pinmux support.
Pistachio:
- Enable KASLR
VDSO:
- Build microMIPS VDSO for microMIPS kernels.
- Fix aliasing warning by building with `-fno-strict-aliasing' for
debugging but also tracing them might result in recursion.
Misc:
- Add missing FROZEN hotplug notifier transitions.
- Fix clk binding example for varioius PIC32 devices.
- Fix cpu interrupt controller node-names in the DT files.
- Fix XPA CPU feature separation.
- Fix write_gc0_* macros when writing zero.
- Add inline asm encoding helpers.
- Add missing VZ accessor microMIPS encodings.
- Fix little endian microMIPS MSA encodings.
- Add 64-bit HTW fields and fix its configuration.
- Fix sigreturn via VDSO on microMIPS kernel.
- Lots of typo fixes.
- Add definitions of SegCtl registers and use them"
* 'upstream' of git://git.linux-mips.org/pub/scm/ralf/upstream-linus: (49 commits)
MIPS: Add missing FROZEN hotplug notifier transitions
MIPS: Build microMIPS VDSO for microMIPS kernels
MIPS: Fix sigreturn via VDSO on microMIPS kernel
MIPS: devicetree: fix cpu interrupt controller node-names
MIPS: VDSO: Build with `-fno-strict-aliasing'
MIPS: Pistachio: Enable KASLR
MIPS: lib: Mark intrinsics notrace
MIPS: Fix 64-bit HTW configuration
MIPS: Add 64-bit HTW fields
MAINTAINERS: Add file patterns for mips device tree bindings
MAINTAINERS: Add file patterns for mips brcm device tree bindings
MIPS: Simplify DSP instruction encoding macros
MIPS: Add missing tlbinvf/XPA microMIPS encodings
MIPS: Fix little endian microMIPS MSA encodings
MIPS: Add missing VZ accessor microMIPS encodings
MIPS: Add inline asm encoding helpers
MIPS: Spelling fix lets -> let's
MIPS: VR41xx: Fix typo
MIPS: oprofile: Fix typo
MIPS: math-emu: Fix typo
...
|
|
Various builds (such as i386:allmodconfig) fail with
fs/binfmt_aout.c:133:2: error: expected identifier or '(' before 'return'
fs/binfmt_aout.c:134:1: error: expected identifier or '(' before '}' token
[ Oops. My bad, I had stupidly thought that "allmodconfig" covered this
on x86-64 too, but it obviously doesn't. Egg on my face. - Linus ]
Fixes: 5d22fc25d4fc ("mm: remove more IS_ERR_VALUE abuses")
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
|
|
Pull string hash improvements from George Spelvin:
"This series does several related things:
- Makes the dcache hash (fs/namei.c) useful for general kernel use.
(Thanks to Bruce for noticing the zero-length corner case)
- Converts the string hashes in <linux/sunrpc/svcauth.h> to use the
above.
- Avoids 64-bit multiplies in hash_64() on 32-bit platforms. Two
32-bit multiplies will do well enough.
- Rids the world of the bad hash multipliers in hash_32.
This finishes the job started in commit 689de1d6ca95 ("Minimal
fix-up of bad hashing behavior of hash_64()")
The vast majority of Linux architectures have hardware support for
32x32-bit multiply and so derive no benefit from "simplified"
multipliers.
The few processors that do not (68000, h8/300 and some models of
Microblaze) have arch-specific implementations added. Those
patches are last in the series.
- Overhauls the dcache hash mixing.
The patch in commit 0fed3ac866ea ("namei: Improve hash mixing if
CONFIG_DCACHE_WORD_ACCESS") was an off-the-cuff suggestion.
Replaced with a much more careful design that's simultaneously
faster and better. (My own invention, as there was noting suitable
in the literature I could find. Comments welcome!)
- Modify the hash_name() loop to skip the initial HASH_MIX(). This
would let us salt the hash if we ever wanted to.
- Sort out partial_name_hash().
The hash function is declared as using a long state, even though
it's truncated to 32 bits at the end and the extra internal state
contributes nothing to the result. And some callers do odd things:
- fs/hfs/string.c only allocates 32 bits of state
- fs/hfsplus/unicode.c uses it to hash 16-bit unicode symbols not bytes
- Modify bytemask_from_count to handle inputs of 1..sizeof(long)
rather than 0..sizeof(long)-1. This would simplify users other
than full_name_hash"
Special thanks to Bruce Fields for testing and finding bugs in v1. (I
learned some humbling lessons about "obviously correct" code.)
On the arch-specific front, the m68k assembly has been tested in a
standalone test harness, I've been in contact with the Microblaze
maintainers who mostly don't care, as the hardware multiplier is never
omitted in real-world applications, and I haven't heard anything from
the H8/300 world"
* 'hash' of git://ftp.sciencehorizons.net/linux:
h8300: Add <asm/hash.h>
microblaze: Add <asm/hash.h>
m68k: Add <asm/hash.h>
<linux/hash.h>: Add support for architecture-specific functions
fs/namei.c: Improve dcache hash function
Eliminate bad hash multipliers from hash_32() and hash_64()
Change hash_64() return value to 32 bits
<linux/sunrpc/svcauth.h>: Define hash_str() in terms of hashlen_string()
fs/namei.c: Add hashlen_string() function
Pull out string hash to <linux/stringhash.h>
|
|
This will improve the performance of hash_32() and hash_64(), but due
to complete lack of multi-bit shift instructions on H8, performance will
still be bad in surrounding code.
Designing H8-specific hash algorithms to work around that is a separate
project. (But if the maintainers would like to get in touch...)
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: uclinux-h8-devel@lists.sourceforge.jp
|
|
Microblaze is an FPGA soft core that can be configured various ways.
If it is configured without a multiplier, the standard __hash_32()
will require a call to __mulsi3, which is a slow software loop.
Instead, use a shift-and-add sequence for the constant multiply.
GCC knows how to do this, but it's not as clever as some.
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Alistair Francis <alistair.francis@xilinx.com>
Cc: Michal Simek <michal.simek@xilinx.com>
|
|
This provides a multiply by constant GOLDEN_RATIO_32 = 0x61C88647
for the original mc68000, which lacks a 32x32-bit multiply instruction.
Yes, the amount of optimization effort put in is excessive. :-)
Shift-add chain found by Yevgen Voronenko's Hcub algorithm at
http://spiral.ece.cmu.edu/mcm/gen.html
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Greg Ungerer <gerg@linux-m68k.org>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Cc: Philippe De Muyter <phdm@macq.eu>
Cc: linux-m68k@lists.linux-m68k.org
|
|
This is just the infrastructure; there are no users yet.
This is modelled on CONFIG_ARCH_RANDOM; a CONFIG_ symbol declares
the existence of <asm/hash.h>.
That file may define its own versions of various functions, and define
HAVE_* symbols (no CONFIG_ prefix!) to suppress the generic ones.
Included is a self-test (in lib/test_hash.c) that verifies the basics.
It is NOT in general required that the arch-specific functions compute
the same thing as the generic, but if a HAVE_* symbol is defined with
the value 1, then equality is tested.
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Greg Ungerer <gerg@linux-m68k.org>
Cc: Andreas Schwab <schwab@linux-m68k.org>
Cc: Philippe De Muyter <phdm@macq.eu>
Cc: linux-m68k@lists.linux-m68k.org
Cc: Alistair Francis <alistai@xilinx.com>
Cc: Michal Simek <michal.simek@xilinx.com>
Cc: Yoshinori Sato <ysato@users.sourceforge.jp>
Cc: uclinux-h8-devel@lists.sourceforge.jp
|
|
Patch 0fed3ac866 improved the hash mixing, but the function is slower
than necessary; there's a 7-instruction dependency chain (10 on x86)
each loop iteration.
Word-at-a-time access is a very tight loop (which is good, because
link_path_walk() is one of the hottest code paths in the entire kernel),
and the hash mixing function must not have a longer latency to avoid
slowing it down.
There do not appear to be any published fast hash functions that:
1) Operate on the input a word at a time, and
2) Don't need to know the length of the input beforehand, and
3) Have a single iterated mixing function, not needing conditional
branches or unrolling to distinguish different loop iterations.
One of the algorithms which comes closest is Yann Collet's xxHash, but
that's two dependent multiplies per word, which is too much.
The key insights in this design are:
1) Barring expensive ops like multiplies, to diffuse one input bit
across 64 bits of hash state takes at least log2(64) = 6 sequentially
dependent instructions. That is more cycles than we'd like.
2) An operation like "hash ^= hash << 13" requires a second temporary
register anyway, and on a 2-operand machine like x86, it's three
instructions.
3) A better use of a second register is to hold a two-word hash state.
With careful design, no temporaries are needed at all, so it doesn't
increase register pressure. And this gets rid of register copying
on 2-operand machines, so the code is smaller and faster.
4) Using two words of state weakens the requirement for one-round mixing;
we now have two rounds of mixing before cancellation is possible.
5) A two-word hash state also allows operations on both halves to be
done in parallel, so on a superscalar processor we get more mixing
in fewer cycles.
I ended up using a mixing function inspired by the ChaCha and Speck
round functions. It is 6 simple instructions and 3 cycles per iteration
(assuming multiply by 9 can be done by an "lea" instruction):
x ^= *input++;
y ^= x; x = ROL(x, K1);
x += y; y = ROL(y, K2);
y *= 9;
Not only is this reversible, two consecutive rounds are reversible:
if you are given the initial and final states, but not the intermediate
state, it is possible to compute both input words. This means that at
least 3 words of input are required to create a collision.
(It also has the property, used by hash_name() to avoid a branch, that
it hashes all-zero to all-zero.)
The rotate constants K1 and K2 were found by experiment. The search took
a sample of random initial states (I used 1023) and considered the effect
of flipping each of the 64 input bits on each of the 128 output bits two
rounds later. Each of the 8192 pairs can be considered a biased coin, and
adding up the Shannon entropy of all of them produces a score.
The best-scoring shifts also did well in other tests (flipping bits in y,
trying 3 or 4 rounds of mixing, flipping all 64*63/2 pairs of input bits),
so the choice was made with the additional constraint that the sum of the
shifts is odd and not too close to the word size.
The final state is then folded into a 32-bit hash value by a less carefully
optimized multiply-based scheme. This also has to be fast, as pathname
components tend to be short (the most common case is one iteration!), but
there's some room for latency, as there is a fair bit of intervening logic
before the hash value is used for anything.
(Performance verified with "bonnie++ -s 0 -n 1536:-2" on tmpfs. I need
a better benchmark; the numbers seem to show a slight dip in performance
between 4.6.0 and this patch, but they're too noisy to quote.)
Special thanks to Bruce fields for diligent testing which uncovered a
nasty fencepost error in an earlier version of this patch.
[checkpatch.pl formatting complaints noted and respectfully disagreed with.]
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Tested-by: J. Bruce Fields <bfields@redhat.com>
|
|
The "simplified" prime multipliers made very bad hash functions, so get rid
of them. This completes the work of 689de1d6ca.
To avoid the inefficiency which was the motivation for the "simplified"
multipliers, hash_64() on 32-bit systems is changed to use a different
algorithm. It makes two calls to hash_32() instead.
drivers/media/usb/dvb-usb-v2/af9015.c uses the old GOLDEN_RATIO_PRIME_32
for some horrible reason, so it inherits a copy of the old definition.
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Cc: Antti Palosaari <crope@iki.fi>
Cc: Mauro Carvalho Chehab <m.chehab@samsung.com>
|
|
That's all that's ever asked for, and it makes the return
type of hash_long() consistent.
It also allows (upcoming patch) an optimized implementation
of hash_64 on 32-bit machines.
I tried adding a BUILD_BUG_ON to ensure the number of bits requested
was never more than 32 (most callers use a compile-time constant), but
adding <linux/bug.h> to <linux/hash.h> breaks the tools/perf compiler
unless tools/perf/MANIFEST is updated, and understanding that code base
well enough to update it is too much trouble. I did the rest of an
allyesconfig build with such a check, and nothing tripped.
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
|
