29 Apr, 2008
40 commits
-
Signed-off-by: Harvey Harrison
Acked-by: OGAWA Hirofumi
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Cc: Petr Vandrovec
Cc: Antonino Daplas
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Acked-by: Greg Kroah-Hartman
Cc: Alan Stern
Cc: David Brownell
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Cc: Dominik Brodowski
Cc: Daniel Ritz
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Cc: John W. Linville
Cc: Michael Buesch
Cc: Daniel Drake
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Cc: Jeff Garzik
Cc: Auke Kok
Cc: John Ronciak
Cc: Jesse Brandeburg
Cc: Grant Grundler
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Cc: David Brownell
Cc: Tony Jones
Cc: Pierre Ossman
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Cc: Dmitry Torokhov
Cc: Jiri Kosina
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Remove unnecessary temp variable from_buf in snsc_event.c
Signed-off-by: Harvey Harrison
Cc: Jiri Slaby
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Acked-by: Jiri Kosina
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Harvey Harrison
Cc: Ed L. Cashin
Cc: Jens Axboe
Cc: Al Viro
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Unaligned access is ok for the following arches:
cris, m68k, mn10300, powerpc, s390, x86Arches that use the memmove implementation for native endian, and
the byteshifting for the opposite endianness.
h8300, m32r, xtensaPacked struct for native endian, byteshifting for other endian:
alpha, blackfin, ia64, parisc, sparc, sparc64, mips, shm86knommu is generic_be for Coldfire, otherwise unaligned access is ok.
frv, arm chooses endianness based on compiler settings, uses the byteshifting
versions. Remove the unaligned trap handler from frv as it is now unused.v850 is le, uses the byteshifting versions for both be and le.
Remove the now unused asm-generic implementation.
Signed-off-by: Harvey Harrison
Acked-by: David S. Miller
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Create a linux/unaligned directory similar in spirit to the linux/byteorder
folder to hold generic implementations collected from various arches.Currently there are five implementations:
1) packed_struct.h: C-struct based, from asm-generic/unaligned.h
2) le_byteshift.h: Open coded byte-swapping, heavily based on asm-arm
3) be_byteshift.h: Open coded byte-swapping, heavily based on asm-arm
4) memmove.h: taken from multiple implementations in tree
5) access_ok.h: taken from x86 and others, unaligned access is ok.All of the new implementations checks for sizes not equal to 1,2,4,8
and will fail to link.API additions:
get_unaligned_{le16|le32|le64|be16|be32|be64}(p) which is meant to replace
code of the form:
le16_to_cpu(get_unaligned((__le16 *)p));put_unaligned_{le16|le32|le64|be16|be32|be64}(val, pointer) which is meant to
replace code of the form:
put_unaligned(cpu_to_le16(val), (__le16 *)p);The headers that arches should include from their asm/unaligned.h:
access_ok.h : Wrappers of the byteswapping functions in asm/byteorder
Choose a particular implementation for little-endian access:
le_byteshift.h
le_memmove.h (arch must be LE)
le_struct.h (arch must be LE)Choose a particular implementation for big-endian access:
be_byteshift.h
be_memmove.h (arch must be BE)
be_struct.h (arch must be BE)After including as needed from the above, include unaligned/generic.h and
define your arch's get/put_unaligned as (for LE):Signed-off-by: Harvey Harrison
Cc:
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Since isn't exported to userspace, there is little
point checking that this is a GNU-compatible compiler.Signed-off-by: Robert P. J. Day
Acked-by: Christoph Hellwig
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
There is no "PNPACPI" driver interface as such. PNPACPI is an internal
backend of PNP, and drivers just use the generic PNP interface.The drivers should depend on CONFIG_PNP, not CONFIG_PNPACPI.
tpm_nsc.c doesn't use PNP at all, so we can just remove the dependency
completely. It probably *should* use PNP to discover the device, but until it
does, there's no point in depending on PNP.Signed-off-by: Bjorn Helgaas
Cc: Kylene Jo Hall
Cc: Marcel Selhorst
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Fix following warning:
WARNING: vmlinux.o(.init.text+0x32804): Section mismatch in reference from the function init_nsc() to the function .devexit.text:tpm_nsc_remove()The function tpm_nsc_remove() are used outside __exit, so remove the __exit
annotation to make sure the function is always avilable.Note: Trying to compare this module with other users of platform_device gve me
the impression that this driver needs some work to match other users.Signed-off-by: Sam Ravnborg
Cc: Kylene Hall
Cc: Marcel Selhorst
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Wrap PNP probe code in #ifdef CONFIG_PNP. We already do the same for
CONFIG_PCI.Without this change, we'll have unresolved references to pnp_get_resource()
function when CONFIG_PNP=n. (This is a new interface that's not in mainline
yet.)Signed-off-by: Bjorn Helgaas
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add support for the CB.ProbeUuid cache manager RPC op. This allows a modern
OpenAFS server to quickly ask if the client has been rebooted.Signed-off-by: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The AFS RxRPC op CBGetCapabilities is actually CBTellMeAboutYourself.
Signed-off-by: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert P. J. Day
Acked-by: David Howells
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
I implemented opstate_init() as a inline function in linux/edac.h.
added calling opstate_init() to:
i82443bxgx_edac.c
i82860_edac.c
i82875p_edac.c
i82975x_edac.cI wrote a fixed patch of
edac-fix-module-initialization-on-several-modules.patch,
and tested building 2.6.25-rc7 with applying this. It was succeed.
I think the patch is now correct.Cc: Alan Cox
Signed-off-by: Hitoshi Mitake
Signed-off-by: Doug Thompson
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Collection of patches, merged into one, from Adrian that do the following:
1) This patch makes the following needlessly global functions static:
- edac_pci_get_log_pe()
- edac_pci_get_log_npe()
- edac_pci_get_panic_on_pe()
- edac_pci_unregister_sysfs_instance_kobj()
- edac_pci_main_kobj_setup()2) Remove unneeded function edac_device_find()
3) Added #if 0 around function edac_pci_find()
4) make the needlessly global edac_pci_generic_check() static
5) Removed function edac_check_mc_devices()
Doug Thompson modified Adrian's patches, to bettern represent
the direction of EDAC, and make them one patch.Cc: Alan Cox
Signed-off-by: Adrian Bunk
Signed-off-by: Doug Thompson
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Robert P. J. Day
Acked-by: Doug Thompson
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add a module parameter "sysbus_parity" to allow forcing system bus parity
error checking on or off. Also add support to automatically disable system
bus parity errors for processors which do not support it.If the sysbus_parity parameter is specified, sysbus parity detection will be
forced on or off. If it is not specified, the driver will attempt to look at
the CPU identifier string and determine if the CPU supports system bus parity.
A blacklist was used instead of a whitelist so that system bus parity would
be enabled by default and to minimize the chances of breaking things for those
people already using the driver which for some reason have a processor that
does not have a valid CPU identifier string.[akpm@linux-foundation.org: coding-style fixes]
Cc: Alan Cox
Signed-off-by: Peter Tyser
Signed-off-by: Doug Thompson
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add Intel 3100 chipset support to e752x EDAC driver.
Cc: Alan Cox
Signed-off-by: Andrei Konovalov
Signed-off-by: Doug Thompson
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Avoid a possible kmem_cache_create() failure by creating idr_layer_cache
unconditionary at boot time rather than creating it on-demand when idr_init()
is called the first time.This change also enables us to eliminate the check every time idr_init() is
called.[akpm@linux-foundation.org: rename init_id_cache() to idr_init_cache()]
[akpm@linux-foundation.org: fix alpha build]
Signed-off-by: Akinobu Mita
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Some drivers have duplicated unlikely() macros. IS_ERR() already has
unlikely() in itself.This patch cleans up such pointless code.
Signed-off-by: Hirofumi Nakagawa
Acked-by: David S. Miller
Acked-by: Jeff Garzik
Cc: Paul Clements
Cc: Richard Purdie
Cc: Alessandro Zummo
Cc: David Brownell
Cc: James Bottomley
Cc: Michael Halcrow
Cc: Anton Altaparmakov
Cc: Al Viro
Cc: Carsten Otte
Cc: Patrick McHardy
Cc: Paul Mundt
Cc: Jaroslav Kysela
Cc: Takashi Iwai
Acked-by: Mike Frysinger
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Add async notification support to /dev/random.
A little test case is below. Without this patch, you get:
$ ./async-random
Drained the pool
Found more randomnessWith it, you get:
$ ./async-random
Drained the pool
SIGIO
Found more randomness#include
#include
#include
#include
#includestatic void handler(int sig)
{
printf("SIGIO\n");
}int main(int argc, char **argv)
{
int fd, n, err, flags;if(signal(SIGIO, handler) < 0){
perror("setting SIGIO handler");
exit(1);
}fd = open("/dev/random", O_RDONLY);
if(fd < 0){
perror("open");
exit(1);
}flags = fcntl(fd, F_GETFL);
if (flags < 0){
perror("getting flags");
exit(1);
}flags |= O_NONBLOCK;
if (fcntl(fd, F_SETFL, flags) < 0){
perror("setting flags");
exit(1);
}while((err = read(fd, &n, sizeof(n))) > 0) ;
if(err == 0){
printf("random returned 0\n");
exit(1);
}
else if(errno != EAGAIN){
perror("read");
exit(1);
}flags |= O_ASYNC;
if (fcntl(fd, F_SETFL, flags) < 0){
perror("setting flags");
exit(1);
}if (fcntl(fd, F_SETOWN, getpid()) < 0) {
perror("Setting SIGIO");
exit(1);
}printf("Drained the pool\n");
read(fd, &n, sizeof(n));
printf("Found more randomness\n");return(0);
}Signed-off-by: Jeff Dike
Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
- emphasize bits in the name
- make zero bits lock-free
- simplify logicSigned-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Switch add_entropy_words to a byte-oriented interface, eliminating numerous
casts and byte/word size rounding issues. This also reduces the overall
bit/byte/word confusion in this code.We now mix a byte at a time into the word-based pool. This takes four times
as many iterations, but should be negligible compared to hashing overhead.
This also increases our pool churn, which adds some depth against some
theoretical failure modes.The function name is changed to emphasize pool mixing and deemphasize entropy
(the samples mixed in may not contain any). extract is added to the core
function to make it clear that it extracts from the pool.Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The add_ptr variable wasn't used in a sensible way, use only i instead.
i got reused later for a different purpose, use j instead.While we're here, put tap0 first in the tap list and add a comment.
Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
The urandom output pool (ie the fast path) fits in one cacheline, so
this is pretty unnecessary. Further, the output path has already
fetched the entire pool to hash it before calling in here.(This was the only user of prefetch_range in the kernel, and it passed
in words rather than bytes!)Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
- eliminate new_rotate
- move input_rotate masking
- simplify input_rotate update
- move input_rotate update to end of inner loop for readabilitySigned-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Earlier changes greatly reduce the number of times we grab the lock
per output byte, so we shouldn't need this particular hack any more.Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
At each extraction, we change (poolbits / 16) + 32 bits in the pool,
or 96 bits in the case of the secondary pools. Thus, a brute-force
backtracking attack on the pool state is less difficult than breaking
the hash. In certain cases, this difficulty may be is reduced to 2^64
iterations.Instead, hash the entire pool in one go, then feedback the whole hash
(160 bits) in one go. This will make backtracking at least as hard as
inverting the hash.Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
- split the SHA variables apart into hash and workspace
- rename data to extract
- wipe extract and workspace after hashingSigned-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
No locking actually needed.
Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds -
Signed-off-by: Matt Mackall
Cc: Theodore Ts'o
Signed-off-by: Andrew Morton
Signed-off-by: Linus Torvalds