Pull vfs updates from Al Viro:
"All kinds of stuff this time around; some more notable parts:
- RCU'd vfsmounts handling
- new primitives for coredump handling
- files_lock is gone
- Bruce's delegations handling series
- exportfs fixes
plus misc stuff all over the place"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (101 commits)
ecryptfs: ->f_op is never NULL
locks: break delegations on any attribute modification
locks: break delegations on link
locks: break delegations on rename
locks: helper functions for delegation breaking
locks: break delegations on unlink
namei: minor vfs_unlink cleanup
locks: implement delegations
locks: introduce new FL_DELEG lock flag
vfs: take i_mutex on renamed file
vfs: rename I_MUTEX_QUOTA now that it's not used for quotas
vfs: don't use PARENT/CHILD lock classes for non-directories
vfs: pull ext4's double-i_mutex-locking into common code
exportfs: fix quadratic behavior in filehandle lookup
exportfs: better variable name
exportfs: move most of reconnect_path to helper function
exportfs: eliminate unused "noprogress" counter
exportfs: stop retrying once we race with rename/remove
exportfs: clear DISCONNECTED on all parents sooner
exportfs: more detailed comment for path_reconnect
...
The CONFIG_HPET_MMAP Kconfig option exposes the memory map of the HPET
registers to userspace. The Kconfig help points out that in some cases
this can be a security risk as some systems may erroneously configure the
map such that additional data is exposed to userspace.
This is a problem for distributions -- some users want the MMAP
functionality but it comes with a significant security risk. In an effort
to mitigate this risk, and due to the low number of users of the MMAP
functionality, I've introduced a kernel parameter, hpet_mmap_enable, that
is required in order to actually have the HPET MMAP exposed.
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Acked-by: Matt Wilson <msw@amazon.com>
Signed-off-by: Clemens Ladisch <clemens@ladisch.de>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: Tomas Winkler <tomas.winkler@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Pull devicetree updates from Rob Herring:
"DeviceTree updates for 3.13. This is a bit larger pull request than
usual for this cycle with lots of clean-up.
- Cross arch clean-up and consolidation of early DT scanning code.
- Clean-up and removal of arch prom.h headers. Makes arch specific
prom.h optional on all but Sparc.
- Addition of interrupts-extended property for devices connected to
multiple interrupt controllers.
- Refactoring of DT interrupt parsing code in preparation for
deferred probe of interrupts.
- ARM cpu and cpu topology bindings documentation.
- Various DT vendor binding documentation updates"
* tag 'devicetree-for-3.13' of git://git.kernel.org/pub/scm/linux/kernel/git/robh/linux: (82 commits)
powerpc: add missing explicit OF includes for ppc
dt/irq: add empty of_irq_count for !OF_IRQ
dt: disable self-tests for !OF_IRQ
of: irq: Fix interrupt-map entry matching
MIPS: Netlogic: replace early_init_devtree() call
of: Add Panasonic Corporation vendor prefix
of: Add Chunghwa Picture Tubes Ltd. vendor prefix
of: Add AU Optronics Corporation vendor prefix
of/irq: Fix potential buffer overflow
of/irq: Fix bug in interrupt parsing refactor.
of: set dma_mask to point to coherent_dma_mask
of: add vendor prefix for PHYTEC Messtechnik GmbH
DT: sort vendor-prefixes.txt
of: Add vendor prefix for Cadence
of: Add empty for_each_available_child_of_node() macro definition
arm/versatile: Fix versatile irq specifications.
of/irq: create interrupts-extended property
microblaze/pci: Drop PowerPC-ism from irq parsing
of/irq: Create of_irq_parse_and_map_pci() to consolidate arch code.
of/irq: Use irq_of_parse_and_map()
...
Pull powerpc updates from Benjamin Herrenschmidt:
"The bulk of this is LE updates. One should now be able to build an LE
kernel and even run some things in it.
I'm still sitting on a handful of patches to enable the new ABI that I
*might* still send this merge window around, but due to the
incertainty (they are pretty fresh) I want to keep them separate.
Other notable changes are some infrastructure bits to better handle
PCI pass-through under KVM, some bits and pieces added to the new
PowerNV platform support such as access to the CPU SCOM bus via sysfs,
and support for EEH error handling on PHB3 (Power8 PCIe).
We also grew arch_get_random_long() for both pseries and powernv when
running on P7+ and P8, exploiting the HW rng.
And finally various embedded updates from freescale"
* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/benh/powerpc: (154 commits)
powerpc: Fix fatal SLB miss when restoring PPR
powerpc/powernv: Reserve the correct PE number
powerpc/powernv: Add PE to its own PELTV
powerpc/powernv: Add support for indirect XSCOM via debugfs
powerpc/scom: Improve debugfs interface
powerpc/scom: Enable 64-bit addresses
powerpc/boot: Properly handle the base "of" boot wrapper
powerpc/bpf: Support MOD operation
powerpc/bpf: Fix DIVWU instruction opcode
of: Move definition of of_find_next_cache_node into common code.
powerpc: Remove big endianness assumption in of_find_next_cache_node
powerpc/tm: Remove interrupt disable in __switch_to()
powerpc: word-at-a-time optimization for 64-bit Little Endian
powerpc/bpf: BPF JIT compiler for 64-bit Little Endian
powerpc: Only save/restore SDR1 if in hypervisor mode
powerpc/pmu: Fix ADB_PMU_LED_IDE dependencies
powerpc/nvram: Fix endian issue when using the partition length
powerpc/nvram: Fix endian issue when reading the NVRAM size
powerpc/nvram: Scan partitions only once
powerpc/mpc512x: remove unnecessary #if
...
The Tausworthe PRNG is initialized at late_initcall time. At that time the
entropy pool serving get_random_bytes is not filled sufficiently. This
patch adds an additional reseeding step as soon as the nonblocking pool
gets marked as initialized.
On some machines it might be possible that late_initcall gets called after
the pool has been initialized. In this situation we won't reseed again.
(A call to prandom_seed_late blocks later invocations of early reseed
attempts.)
Joint work with Daniel Borkmann.
Cc: Eric Dumazet <eric.dumazet@gmail.com>
Cc: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: Daniel Borkmann <dborkman@redhat.com>
Acked-by: "Theodore Ts'o" <tytso@mit.edu>
Signed-off-by: David S. Miller <davem@davemloft.net>
We only depend on the intel-gtt module for GTT frobbign on older gens.
The intel_agp module is optional, except for UMS and some old XvMC
userland on gen3. So make AGP support optional. As before, we will
fail the i915 init for UMS and gen3 KMS the same as before if
intel_agp isn't around.
intel-gtt.c is left with a somewhat ugly ifdef mess, but I'm going
to save that for a later cleaning.
At least my gen2 still works with the patch and CONFIG_AGP=n.
v2: Make i915 depend on X86 and PCI, and intel-gtt depend on PCI
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Since we initialize jiffies to wrap five minutes before boot (see
INITIAL_JIFFIES defined in include/linux/jiffies.h) it's important to
make sure the last_time field is initialized to INITIAL_JIFFIES.
Otherwise, the entropy estimator will overestimate the amount of
entropy resulting from the first call to add_timer_randomness(),
generally by about 8 bits.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The rand_initialize() function was being run fairly late in the kernel
boot sequence. This was unfortunate, since it zero'ed the entropy
counters, thus throwing away credit that was accumulated earlier in
the boot sequence, and it also meant that initcall functions run
before rand_initialize were using a minimally initialized pool.
To fix this, fix init_std_data() to no longer zap the entropy counter;
it wasn't necessary, and move rand_initialize() to be an early
initcall.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Print a notification to the console when the nonblocking pool is
initialized. Also printk a warning when a process tries reading from
/dev/urandom before it is fully initialized.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Change add_timer_randomness() so that it directs incoming entropy to
the nonblocking pool first if it hasn't been fully initialized yet.
This matches the strategy we use in add_interrupt_randomness(), which
allows us to push the randomness where we need it the most during when
the system is first booting up, so that get_random_bytes() and
/dev/urandom become safe to use as soon as possible.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
If virtqueue_get_buf() returns with a NULL pointer it should be verified
if the virtqueue is broken, in order to avoid loop calling cpu_relax().
Signed-off-by: Heinz Graalfs <graalfs@linux.vnet.ibm.com>
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This patch changes the displayed module name from
tpm_tis_i2c_infineon to its actual name tpm_i2c_infineon.
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
This is based on the work of Teddy Reed <teddy@prosauce.org> published
on GitHub:
https://github.com/theopolis/tpm-i2c-atmel.git
34894b988b67e0ae55088d6388e77b0dbf10c07d
That driver was never merged, I have taken it as a starting port,
forward ported, tested and revised the driver:
- Make it broadly textually similar to the Infineon and Nuvoton I2C
driver
- Place everything in a format suitable for mainline inclusion
- Use high level I2C functions i2c_master_send and
i2c_master_recv for data xfer
- Use the timeout system from the core code, by faking out a status
register
- Only I2C transfer the number of bytes in the reply, not a fixed
message size.
- checkpatch cleanups
- Testing on ARM Kirkwood, with this device tree, using a
AT97SC3204T-X1A180
tpm@29 {
compatible = "atmel,at97sc3204t";
reg = <0x29>;
};
Signed-off-by: Teddy Reed <teddy@prosauce.org>
[jgg: revised and tested]
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
[phuewe: minor whitespace changes]
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
This chip is/was also branded as a Winbond WPCT301.
Originally written by Dan Morav <dmorav@nuvoton.com> and posted to LKML:
https://lkml.org/lkml/2011/9/7/206
The original posting was not merged, I have taken it as a
starting point, forward ported, tested and revised the driver:
- Rework interrupt handling to work properly with level triggered
interrupts. The old version just locked up.
- Synchronize various items with Peter Huewe's Infineon driver:
* Add durations/timeouts sysfs calls
* Remove I2C device auto-detection
* Don't fiddle with chip->release
* Call tpm_dev_vendor_release in the probe error path
* Use MODULE_DEVICE_TABLE for the I2C ids
* Provide OF compatible strings for DT support
* Use SIMPLE_DEV_PM_OPS
* Use module_i2c_driver
- checkpatch cleanups
- Testing on ARM Kirkwood with GPIO interrupts, with this device tree:
tpm@57 {
compatible = "nuvoton,npct501";
reg = <0x57>;
interrupt-parent = <&gpio1>;
interrupts = <6 IRQ_TYPE_LEVEL_LOW>;
};
Signed-off-by: Dan Morav <dmorav@nuvoton.com>
[jgg: revised and tested]
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
[phuewe: minor whitespace changes, fixed module name in kconfig]
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Now that we can have multiple .c files in the tpm module there is
no reason for tpm-bios.
tpm-bios exported several functions: tpm_bios_log_setup,
tpm_bios_log_teardown, tpm_add_ppi, and tpm_remove_ppi.
They are only used by tpm, and if tpm-bios is built then
tpm will unconditionally require them. Further, tpm-bios does
nothing on its own, it has no module_init function.
Thus we remove the exports and merge the modules to simplify things.
The Makefile conditions are changed slightly to match the code,
tpm_ppi is always required if CONFIG_ACPI is set.
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
This is preparation for making the tpm module multi-file. kbuild does
not like having a .c file with the same name as a module. We wish to
keep the tpm module name so that userspace doesn't see this change.
tpm-interface.c is chosen because the next several commits in the series
migrate items into tpm-sysfs.c, tpm-dev.c and tpm-class.c. All that will
be left is tpm command processing and interfacing code.
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
before we rename the file it might be a good idea to cleanup the long
persisting checkpatch warnings.
Since everything is really trivial, splitting the patch up would only
result in noise.
For the interested reader - here the checkpatch warnings:
(regrouped for easer readability)
ERROR: trailing whitespace
+ * Specifications at www.trustedcomputinggroup.org^I $
+ * $
+^I/* $
+^I parameters (RSA 12->bytes: keybit, #primes, expbit) $
WARNING: unnecessary whitespace before a quoted newline
+ "invalid count value %x %zx \n", count, bufsiz);
ERROR: do not use assignment in if condition
+ if ((rc = chip->vendor.send(chip, (u8 *) buf, count)) < 0) {
ERROR: space required after that ',' (ctx:VxV)
+ len = tpm_transmit(chip,(u8 *) cmd, len);
^
ERROR: "foo * bar" should be "foo *bar"
+ssize_t tpm_show_enabled(struct device * dev, struct device_attribute * attr,
+ssize_t tpm_show_enabled(struct device * dev, struct device_attribute * attr,
+ssize_t tpm_show_active(struct device * dev, struct device_attribute * attr,
+ssize_t tpm_show_active(struct device * dev, struct device_attribute * attr,
+ssize_t tpm_show_owned(struct device * dev, struct device_attribute * attr,
+ssize_t tpm_show_owned(struct device * dev, struct device_attribute * attr,
+ssize_t tpm_show_temp_deactivated(struct device * dev,
+ struct device_attribute * attr, char *buf)
WARNING: please, no space before tabs
+ * @chip_num: ^Itpm idx # or ANY$
+ * @res_buf: ^ITPM_PCR value$
+ * ^I^Isize of res_buf is 20 bytes (or NULL if you don't care)$
+ * @chip_num: ^Itpm idx # or AN&$
+ * @hash: ^Ihash value used to extend pcr value$
ERROR: code indent should use tabs where possible
+^I TPM_ORD_CONTINUE_SELFTEST);$
WARNING: line over 80 characters
+static bool wait_for_tpm_stat_cond(struct tpm_chip *chip, u8 mask, bool check_cancel,
ERROR: trailing whitespace
+ * Called from tpm_<specific>.c probe function only for devices $
total: 16 errors, 7 warnings, 1554 lines checked
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
The version of the TPM should not depend on the bus it is connected
through. 1.1, 1.2 and soon 2.0 TPMS will be all be able to use the
same bus interfaces.
Make tpm_show_caps try the 1.2 capability first. If that fails then
fall back to the 1.1 capability. This effectively auto-detects what
interface the TPM supports at run-time.
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Joel Schopp <jschopp@linux.vnet.ibm.com>
Reviewed-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
For some reason this driver thinks that chip->data_buffer needs
to be set before it can call tpm_pm_*. This is not true. data_buffer
is used only by /dev/tpmX, which is why it is managed exclusively
by the fops functions.
Cc: Mathias Leblanc <mathias.leblanc@st.com>
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Joel Schopp <jschopp@linux.vnet.ibm.com>
TPM drivers should not call dev_set_drvdata (or aliases), only the core
code is allowed to call dev_set_drvdata, and it does it during
tpm_register_hardware.
These extra sets are harmless, but are an anti-pattern that many drivers
have copied.
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Joel Schopp <jschopp@linux.vnet.ibm.com>
Reviewed-by: Peter Huewe <peterhuewe@gmx.de>
Acked-by: Ashley Lai <adlai@linux.vnet.ibm.com>
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
misc_open sets the file->private_date to the misc_dev when calling
open. We can use container_of to go from the misc_dev back to the
tpm_chip.
Future clean ups will move tpm_open into a new file and this change
means we do not have to export the tpm_chip list.
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Joel Schopp <jschopp@linux.vnet.ibm.com>
Reviewed-by: Peter Huewe <peterhuewe@gmx.de>
Acked-by: Ashley Lai <adlai@linux.vnet.ibm.com>
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
Commit e0dd03caf2 ("tpm: return chip from
tpm_register_hardware") changed the code path here so that
ateml_get_base_addr no longer directly altered the tpm_vendor_specific
structure, and instead placed the base address on the stack.
The commit missed updating the request_region call, which would have
resulted in request_region being called with 0 as the base address.
I don't know if request_region(0, ..) will fail, if so the
driver has been broken since 2006 and we should remove it
from the tree as it has no users.
Signed-off-by: Jason Gunthorpe <jgunthorpe@obsidianresearch.com>
Reviewed-by: Joel Schopp <jschopp@linux.vnet.ibm.com>
Reviewed-by: Peter Huewe <peterhuewe@gmx.de>
Signed-off-by: Peter Huewe <peterhuewe@gmx.de>
This lets the transport do endian conversion if necessary, and insulates
the drivers from the difference.
Most drivers can use the simple helpers virtio_cread() and virtio_cwrite().
Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
This driver provides kernel-side support for the Random Number
Generator hardware found on OMAP34xx processors.
This driver comes from Maemo 2.6.28 kernel and was tested on Nokia RX-51.
It is platform device because it needs board specific function for smc calls.
Signed-off-by: Pali Rohár <pali.rohar@gmail.com>
Signed-off-by: Juha Yrjola <juha.yrjola@solidboot.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Pull Xen fixes from Stefano Stabellini:
"A small fix for Xen on x86_32 and a build fix for xen-tpmfront on
arm64"
* tag 'stable/for-linus-3.12-rc4-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip:
xen: Fix possible user space selector corruption
tpm: xen-tpmfront: fix missing declaration of xen_domain
Add a driver for the hwrng found in power7+ systems, based on the
existing code for the arch_get_random_long() hook.
We only register a single instance of the driver, not one per device,
because we use the existing per_cpu array of devices in the arch code.
This means we always read from the "closest" device, avoiding inter-chip
memory traffic.
Signed-off-by: Guo Chao <yan@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
We don't expect to get errors from the hypervisor when reading the rng,
but if we do we should pass the error up to the hwrng driver. Otherwise
the hwrng driver will continue calling us forever.
Signed-off-by: Michael Ellerman <michael@ellerman.id.au>
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Pull /dev/random changes from Ted Ts'o:
"These patches are designed to enable improvements to /dev/random for
non-x86 platforms, in particular MIPS and ARM"
* tag 'random_for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tytso/random:
random: allow architectures to optionally define random_get_entropy()
random: run random_int_secret_init() run after all late_initcalls
Instead of using the random driver's ad-hoc DEBUG_ENT() mechanism, use
tracepoints instead. This allows for a much more fine-grained control
of which debugging mechanism which a developer might need, and unifies
the debugging messages with all of the existing tracepoints.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
As the input pool gets filled, start transfering entropy to the output
pools until they get filled. This allows us to use the output pools
to store more system entropy. Waste not, want not....
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The add_timer_randomness() used to drop into trickle mode when entropy
pool was estimated to be 87.5% full. This was important when
add_timer_randomness() was used to sample interrupts. It's not used
for this any more --- add_interrupt_randomness() now uses fast_mix()
instead. By elimitating trickle mode, it allows us to fully utilize
entropy provided by add_input_randomness() and add_disk_randomness()
even when the input pool is above the old trickle threshold of 87.5%.
This helps to answer the criticism in [1] in their hypothetical
scenario where our entropy estimator was inaccurate, even though the
measurements in [2] seem to indicate that our entropy estimator given
real-life entropy collection is actually pretty good, albeit on the
conservative side (which was as it was designed).
[1] http://eprint.iacr.org/2013/338.pdf
[2] http://eprint.iacr.org/2012/251.pdf
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Our mixing functions were analyzed by Lacharme, Roeck, Strubel, and
Videau in their paper, "The Linux Pseudorandom Number Generator
Revisited" (see: http://eprint.iacr.org/2012/251.pdf).
They suggested a slight change to improve our mixing functions
slightly. I also adjusted the comments to better explain what is
going on, and to document why the polynomials were changed.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
By mixing the entropy in chunks of 32-bit words instead of byte by
byte, we can speed up the fast_mix function significantly. Since it
is called on every single interrupt, on systems with a very heavy
interrupt load, this can make a noticeable difference.
Also fix a compilation warning in add_interrupt_randomness() and avoid
xor'ing cycles and jiffies together just in case we have an
architecture which tries to define random_get_entropy() by returning
jiffies.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Reported-by: Jörn Engel <joern@logfs.org>
In order to avoid draining the input pool of its entropy at too high
of a rate, enforce a minimum time interval between reseedings of the
urandom pool. This is set to 60 seconds by default.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
The add_device_randomness() function calls mix_pool_bytes() twice for
the input pool and the non-blocking pool, for a total of four times.
By using _mix_pool_byte() and taking the spinlock in
add_device_randomness(), we can halve the number of times we need
take each pool's spinlock.
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
Fix a problem where get_random_bytes_arch() was calling the tracepoint
get_random_bytes(). So add a new tracepoint for
get_random_bytes_arch(), and make get_random_bytes() and
get_random_bytes_arch() call their correct tracepoint.
Also, add a new tracepoint for add_device_randomness()
Signed-off-by: "Theodore Ts'o" <tytso@mit.edu>
When we write entropy into a non-empty pool, we currently don't
account at all for the fact that we will probabilistically overwrite
some of the entropy in that pool. This means that unless the pool is
fully empty, we are currently *guaranteed* to overestimate the amount
of entropy in the pool!
Assuming Shannon entropy with zero correlations we end up with an
exponentally decaying value of new entropy added:
entropy <- entropy + (pool_size - entropy) *
(1 - exp(-add_entropy/pool_size))
However, calculations involving fractional exponentials are not
practical in the kernel, so apply a piecewise linearization:
For add_entropy <= pool_size/2 then
(1 - exp(-add_entropy/pool_size)) >= (add_entropy/pool_size)*0.7869...
... so we can approximate the exponential with
3/4*add_entropy/pool_size and still be on the
safe side by adding at most pool_size/2 at a time.
In order for the loop not to take arbitrary amounts of time if a bad
ioctl is received, terminate if we are within one bit of full. This
way the loop is guaranteed to terminate after no more than
log2(poolsize) iterations, no matter what the input value is. The
vast majority of the time the loop will be executed exactly once.
The piecewise linearization is very conservative, approaching 3/4 of
the usable input value for small inputs, however, our entropy
estimation is pretty weak at best, especially for small values; we
have no handle on correlation; and the Shannon entropy measure (Rényi
entropy of order 1) is not the correct one to use in the first place,
but rather the correct entropy measure is the min-entropy, the Rényi
entropy of infinite order.
As such, this conservatism seems more than justified.
This does introduce fractional bit values. I have left it to have 3
bits of fraction, so that with a pool of 2^12 bits the multiply in
credit_entropy_bits() can still fit into an int, as 2*(3+12) < 31. It
is definitely possible to allow for more fractional accounting, but
that multiply then would have to be turned into a 32*32 -> 64 multiply.
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Cc: DJ Johnston <dj.johnston@intel.com>
