drivers/media/dvb-frontends/helene.c: In function 'helene_write_regs':
drivers/media/dvb-frontends/helene.c:312:5: warning: format '%lu' expects argument of type 'long unsigned int', but argument 5 has type 'unsigned int' [-Wformat=]
"wr reg=%04x: len=%d vs %lu is too big!\n",
^
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
correct is CXD2841ER and CXD2854ER
incorrect was CXD2441ER and CXD2454ER
Signed-off-by: Abylay Ospan <aospan@netup.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
now when new tuning parameters specified demod should retune.
Also ISDB-T frequency offset calculation added
(cxd2841er_get_carrier_offset_i).
While here, fix re-tune for DVB-C Annex A, using the desired
bandwidth, instead of using 8MHz.
Signed-off-by: Abylay Ospan <aospan@netup.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Avoid error message:
cxd2841er_read_status_s(): invalid state 1
Always force demod to shutdown state before initializing
Signed-off-by: Abylay Ospan <aospan@netup.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
mistakenly membase8_io used instead of membase8_config
in this case we can't read/write CAM module memory (TUPLES)
Signed-off-by: Abylay Ospan <aospan@netup.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
CXD2854ER is identical to CXD2841ER except ISDB-T/S added.
New method 'cxd2841er_attach_i' is added
xtal frequency now configurable. Available options:
20.5MHz, 24MHz, 41MHz
Signed-off-by: Abylay Ospan <aospan@netup.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
This is Sony HELENE DVB-S/S2 DVB-T/T2 DVB-C/C2 ISDB-T/S tuner
driver (CXD2858ER).
Tuner is used on NetUP Dual Universal DVB CI card (hardware revision 1.4).
Use 'helene_attach_s' to attach tuner in 'satellite mode'.
Use 'helene_attach' for 'terrestrial mode'.
Satellite delivery systems supported:
DVB-S/S2, ISDB-S
Terrestrial delivery systems supported:
DVB-T/T2, ISDB-T
Cable delivery systems supported:
DVB-C/C2
Signed-off-by: Abylay Ospan <aospan@netup.ru>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
The formula used to calculate bytesperline only works for packed format.
So far, all planar format we support have their bytesperline equal to
the image width (stride of the Y plane or a line of Y for M420).
Signed-off-by: Nicolas Dufresne <nicolas.dufresne@collabora.com>
Signed-off-by: Laurent Pinchart <laurent.pinchart@ideasonboard.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Add a DT binding documentation of Video Processor Unit for the
MT8173 SoC from Mediatek.
Signed-off-by: Andrew-CT Chen <andrew-ct.chen@mediatek.com>
Signed-off-by: Tiffany Lin <tiffany.lin@mediatek.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
User-space applications can use the VIDIOC_REQBUFS ioctl to determine if a
memory mapped, user pointer or DMABUF based I/O is supported by the driver.
So a set of VIDIOC_REQBUFS ioctl calls will be made with count 0 and then
the real VIDIOC_REQBUFS call with count == n. But for count 0, the driver
not only frees the buffer but also closes the MFC instance and s5p_mfc_ctx
state is set to MFCINST_FREE.
The VIDIOC_REQBUFS handler for the output device checks if the s5p_mfc_ctx
state is set to MFCINST_INIT (which happens on an VIDIOC_S_FMT) and fails
otherwise. So after a VIDIOC_REQBUFS(n), future VIDIOC_REQBUFS(n) calls
will fails unless a VIDIOC_S_FMT ioctl calls happens before the reqbufs.
But applications may first set the format and then attempt to determine
the I/O methods supported by the driver (for example Gstramer does it) so
the state won't be set to MFCINST_INIT again and VIDIOC_REQBUFS will fail.
To avoid this issue, only free the buffers on VIDIOC_REQBUFS(0) but don't
close the MFC instance to allow future VIDIOC_REQBUFS(n) calls to succeed.
[javier: Rewrote changelog to explain the problem more detailed]
Signed-off-by: ayaka <ayaka@soulik.info>
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Acked-by: Nicolas Dufresne <nicolas@collabora.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Failing to get the struct s5p_mfc_pm .clock is a non-fatal error so the
clock field can have a errno pointer value. But s5p_mfc_final_pm() only
checks if .clock is not NULL before attempting to unprepare and put it.
This leads to the following warning in clk_put() due s5p_mfc_final_pm():
WARNING: CPU: 3 PID: 1023 at drivers/clk/clk.c:2814 s5p_mfc_final_pm+0x48/0x74 [s5p_mfc]
CPU: 3 PID: 1023 Comm: rmmod Tainted: G W 4.6.0-rc6-next-20160502-00005-g5a15a49106bc #9
Hardware name: SAMSUNG EXYNOS (Flattened Device Tree)
[<c010e1bc>] (unwind_backtrace) from [<c010af28>] (show_stack+0x10/0x14)
[<c010af28>] (show_stack) from [<c032485c>] (dump_stack+0x88/0x9c)
[<c032485c>] (dump_stack) from [<c011b8e8>] (__warn+0xe8/0x100)
[<c011b8e8>] (__warn) from [<c011b9b0>] (warn_slowpath_null+0x20/0x28)
[<c011b9b0>] (warn_slowpath_null) from [<bf16004c>] (s5p_mfc_final_pm+0x48/0x74 [s5p_mfc])
[<bf16004c>] (s5p_mfc_final_pm [s5p_mfc]) from [<bf157414>] (s5p_mfc_remove+0x8c/0x94 [s5p_mfc])
[<bf157414>] (s5p_mfc_remove [s5p_mfc]) from [<c03fe1f8>] (platform_drv_remove+0x24/0x3c)
[<c03fe1f8>] (platform_drv_remove) from [<c03fcc70>] (__device_release_driver+0x84/0x110)
[<c03fcc70>] (__device_release_driver) from [<c03fcdd8>] (driver_detach+0xac/0xb0)
[<c03fcdd8>] (driver_detach) from [<c03fbff8>] (bus_remove_driver+0x4c/0xa0)
[<c03fbff8>] (bus_remove_driver) from [<c01886a8>] (SyS_delete_module+0x174/0x1b8)
[<c01886a8>] (SyS_delete_module) from [<c01078c0>] (ret_fast_syscall+0x0/0x3c)
Assign the pointer to NULL in case of a lookup failure to fix the issue.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Acked-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
This patch fixes build break caused by lack of dma-iommu API on ARM64
(this API is specific to ARM 32bit architecture).
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
* 'for-v4.8/media/exynos-mfc' of git://linuxtv.org/snawrocki/samsung:
media: s5p-mfc: add iommu support
media: s5p-mfc: replace custom reserved memory handling code with generic one
media: s5p-mfc: use generic reserved memory bindings
of: reserved_mem: add support for using more than one region for given device
media: set proper max seg size for devices on Exynos SoCs
media: vb2-dma-contig: add helper for setting dma max seg size
s5p-mfc: Fix race between s5p_mfc_probe() and s5p_mfc_open()
s5p-mfc: Add release callback for memory region devs
s5p-mfc: Set device name for reserved memory region devs
This patch adds support for IOMMU to s5p-mfc device driver. MFC firmware
is limited and it cannot use the default configuration. If IOMMU is
available, the patch disables the default DMA address space
configuration and creates a new address space of size limited to 256M
and base address set to 0x20000000.
For now the same address space is shared by both 'left' and 'right'
memory channels, because the DMA/IOMMU frameworks do not support
configuring them separately. This is not optimal, but besides limiting
total address space available has no other drawbacks (MFC firmware
supports 256M of address space per each channel).
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
This patch removes custom code for initialization and handling of
reserved memory regions in s5p-mfc driver and replaces it with generic
reserved memory regions api.
s5p-mfc driver now handles two reserved memory regions defined by
generic reserved memory bindings. Support for non-dt platform has been
removed, because all supported platforms have been already converted to
device tree.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Use generic reserved memory bindings and mark old, custom properties
as obsoleted.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Reviewed-by: Javier Martinez Canillas <javier@osg.samsung.com>
Acked-by: Rob Herring <robh@kernel.org>
Acked-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
This patch allows device drivers to initialize more than one reserved
memory region assigned to given device. When driver needs to use more
than one reserved memory region, it should allocate child devices and
initialize regions by index for each of its child devices.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
All multimedia devices found on Exynos SoCs support only contiguous
buffers, so set DMA max segment size to DMA_BIT_MASK(32) to let memory
allocator to correctly create contiguous memory mappings.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
Add a helper function for device drivers to set DMA's max_seg_size.
Setting it to largest possible value lets DMA-mapping API always create
contiguous mappings in DMA address space. This is essential for all
devices, which use dma-contig videobuf2 memory allocator and shared
buffers.
Till now, the only case when vb2-dma-contig really 'worked' was a case
where userspace provided USERPTR buffer, which was in fact mmaped
contiguous buffer from the other v4l2/drm device. Also DMABUF made of
contiguous buffer worked only when its exporter did not split it into
several chunks in the scatter-list. Any other buffer failed, regardless
of the arch/platform used and the presence of the IOMMU of the device bus.
This patch provides interface to fix this issue.
Signed-off-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
The s5p_mfc_probe() function registers the video devices before all the
resources needed by s5p_mfc_open() are correctly initalized.
So if s5p_mfc_open() function is called before s5p_mfc_probe() finishes
(since the video dev is already registered), a NULL pointer dereference
will happen due s5p_mfc_open() accessing uninitialized vars such as the
struct s5p_mfc_dev .watchdog_timer and .mfc_ops fields.
An example is following BUG caused by add_timer() getting a NULL pointer:
[ 45.765374] kernel BUG at kernel/time/timer.c:790!
[ 45.765381] Internal error: Oops - BUG: 0 [#1] PREEMPT SMP ARM
...
[ 45.766149] [<c016fdf4>] (mod_timer) from [<bf181d18>] (s5p_mfc_open+0x274/0x4d4 [s5p_mfc])
[ 45.766416] [<bf181d18>] (s5p_mfc_open [s5p_mfc]) from [<bf0214a0>] (v4l2_open+0x9c/0x100 [videodev])
[ 45.766547] [<bf0214a0>] (v4l2_open [videodev]) from [<c01e355c>] (chrdev_open+0x9c/0x178)
[ 45.766575] [<c01e355c>] (chrdev_open) from [<c01dceb4>] (do_dentry_open+0x1e0/0x300)
[ 45.766595] [<c01dceb4>] (do_dentry_open) from [<c01ec2f0>] (path_openat+0x800/0x10d4)
[ 45.766610] [<c01ec2f0>] (path_openat) from [<c01ed8b8>] (do_filp_open+0x5c/0xc0)
[ 45.766624] [<c01ed8b8>] (do_filp_open) from [<c01de218>] (do_sys_open+0x10c/0x1bc)
[ 45.766642] [<c01de218>] (do_sys_open) from [<c01078c0>] (ret_fast_syscall+0x0/0x3c)
[ 45.766655] Code: eaffffe3 e3a00001 e28dd008 e8bd81f0 (e7f001f2)
Fix it by registering the video devs as the last step in s5p_mfc_probe().
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
When s5p_mfc_remove() calls put_device() for the reserved memory region
devs, the driver core warns that the dev doesn't have a release callback:
WARNING: CPU: 0 PID: 591 at drivers/base/core.c:251 device_release+0x8c/0x90
Device 's5p-mfc-l' does not have a release() function, it is broken and must be fixed.
Also, the declared DMA memory using dma_declare_coherent_memory() isn't
relased so add a dev .release that calls dma_release_declared_memory().
Cc: <stable@vger.kernel.org>
Fixes: 6e83e6e25e ("[media] s5p-mfc: Fix kernel warning on memory init")
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
The devices don't have a name set, so makes dev_name() returns NULL which
makes harder to identify the devices that are causing issues, for example:
WARNING: CPU: 2 PID: 616 at drivers/base/core.c:251 device_release+0x8c/0x90
Device '(null)' does not have a release() function, it is broken and must be fixed.
And after setting the device name:
WARNING: CPU: 0 PID: 591 at drivers/base/core.c:251 device_release+0x8c/0x90
Device 's5p-mfc-l' does not have a release() function, it is broken and must be fixed.
Cc: <stable@vger.kernel.org>
Fixes: 6e83e6e25e ("[media] s5p-mfc: Fix kernel warning on memory init")
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Sylwester Nawrocki <s.nawrocki@samsung.com>
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1
iQEcBAABAgAGBQJXSxkJAAoJEHm+PkMAQRiGwfcH/iK8kqycIUhLc/Wn2TnY0TcO
9FPW6EREzc7cRu38rXAgLyU15ysCK56QtV8anqvxkalHgY0ovmE3bwR3BKyUtW3W
IdevRChiGCH0jwG+KjlGrguBI1PCejnY9/El2F1mzUvlkMv7OYYMcUHPUm96HN15
TLjOYgUQyK+YBsj/LBWj7qUmpWEa+DuQyY/AoohtGQR9KHRAxxAr0SsvFrC+/phJ
xkozkqhy5/snqiqbVgDcqxqFBqHO23IYAbWjaCWG3M2/9HPOjw7tJwM7P2d6LQuz
mQi3+FY3sIKgjGS6T9vKYpt3PE3KmziVkzL13iU0yX/iuatjrhUKapifwPZfiKc=
=Ihgt
-----END PGP SIGNATURE-----
Merge tag 'v4.7-rc1' into patchwork
Linux 4.7-rc1
* tag 'v4.7-rc1': (10534 commits)
Linux 4.7-rc1
hash_string: Fix zero-length case for !DCACHE_WORD_ACCESS
Rename other copy of hash_string to hashlen_string
hpfs: implement the show_options method
affs: fix remount failure when there are no options changed
hpfs: fix remount failure when there are no options changed
fs: fix binfmt_aout.c build error
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>
Revert "platform/chrome: chromeos_laptop: Add Leon Touch"
i2c: dev: use after free in detach
MIPS: Add missing FROZEN hotplug notifier transitions
...
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: fcfd2fbf22 ("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: fcfd2fbf22 ("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 c8f33d0bec ("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: c8f33d0bec ("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 ce657611ba ("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: ce657611ba ("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>
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: 5d22fc25d4 ("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 689de1d6ca ("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 0fed3ac866 ("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>
Finally, the first use of previous two patches: eliminate the
separate ad-hoc string hash functions in the sunrpc code.
Now hash_str() is a wrapper around hash_string(), and hash_mem() is
likewise a wrapper around full_name_hash().
Note that sunrpc code *does* call hash_mem() with a zero length, which
is why the previous patch needed to handle that in full_name_hash().
(Thanks, Bruce, for finding that!)
This also eliminates the only caller of hash_long which asks for
more than 32 bits of output.
The comment about the quality of hashlen_string() and full_name_hash()
is jumping the gun by a few patches; they aren't very impressive now,
but will be improved greatly later in the series.
Signed-off-by: George Spelvin <linux@sciencehorizons.net>
Tested-by: J. Bruce Fields <bfields@redhat.com>
Acked-by: J. Bruce Fields <bfields@redhat.com>
Cc: Jeff Layton <jlayton@poochiereds.net>
Cc: linux-nfs@vger.kernel.org
