Impact: bogus error codes (+other?) on x86-64
The rdmsr_safe/wrmsr_safe routines have macros for the handling of the
edx:eax arguments. Those macros take a variable number of assembly
arguments. This is rather inherently incompatible with using
%digit-style escapes in the inline assembly; replace those with
%[name]-style escapes.
This fixes miscompilation on x86-64, which at the very least caused
bogus return values. It is possible that this could also corrupt the
return value; I am not sure.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
Propagate error (-ENXIO) from smp_call_function_single(). These
errors can happen when a CPU is unplugged while the MSR driver is
open.
Signed-off-by: H. Peter Anvin <hpa@zytor.com>
The header file for use by (in-kernel) PCMCIA sockets deserved a major
cleanup. Some stuff only used by the pcmcia core modules was moved to
cs_internal.h
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
* 'x86-fixes-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip:
x86: add X86_FEATURE_XMM4_2 definitions
x86: fix cpufreq + sched_clock() regression
x86: fix HPET regression in 2.6.26 versus 2.6.25, check hpet against BAR, v3
x86: do not enable TSC notifier if we don't need it
x86 MCE: Fix CPU hotplug problem with multiple multicore AMD CPUs
x86: fix: make PCI ECS for AMD CPUs hotplug capable
x86: fix: do not run code in amd_bus.c on non-AMD CPUs
* 'kvm-updates-2.6.27' of git://git.kernel.org/pub/scm/linux/kernel/git/avi/kvm:
KVM: fix userspace ABI breakage
KVM: MMU: Fix torn shadow pte
KVM: Use .fixup instead of .text.fixup on __kvm_handle_fault_on_reboot
Added Intel processor SSE4.2 feature flag.
No in-tree user at the moment, but makes the tree-merging life easier
for the crypto tree.
Signed-off-by: Austin Zhang <austin.zhang@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
The following part of commit 9ef621d3be
(KVM: Support mixed endian machines) changed on the size of a struct
that is exported to userspace:
include/linux/kvm.h:
@@ -318,14 +318,14 @@ struct kvm_trace_rec {
__u32 vcpu_id;
union {
struct {
- __u32 cycle_lo, cycle_hi;
+ __u64 cycle_u64;
__u32 extra_u32[KVM_TRC_EXTRA_MAX];
} cycle;
struct {
__u32 extra_u32[KVM_TRC_EXTRA_MAX];
} nocycle;
} u;
-};
+} __attribute__((packed));
Packing a struct was the correct idea, but it packed the wrong struct.
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Avi Kivity <avi@qumranet.com>
vmlinux.lds expects the fixup code to be on a section named .fixup. The
.text.fixup section is not mentioned on vmlinux.lds, and is included on
the resulting vmlinux (just after .text) only because of ld heuristics on
placing orphan sections.
However, placing .text.fixup outside .text breaks the definition of
_etext, making it exclude the .text.fixup contents. That makes .text.fixup
be ignored by the kernel initialization code that needs to know about
section locations, such as the code setting page protection bits.
Signed-off-by: Eduardo Habkost <ehabkost@redhat.com>
Signed-off-by: Avi Kivity <avi@qumranet.com>
Bug Description:
a customer reported under IRQ stress, running applications may
wrongly trigger an ICPLB miss and be killed. after playing a
bit more, here's a test case that may be triggering the same bug.
Fixing:
After modifying page protections, only modify the active CPLBs if the
protection change was made for the active mm.
Signed-off-by: Bernd Schmidt <bernds_cb1@t-online.de>
Signed-off-by: Bryan Wu <cooloney@kernel.org>
It allows paravirt implementations of cpu_disable to share the
cpu_disable_common code, without having to take on board APIC
writes, which may not be appropriate.
Signed-off-by: Alex Nixon <alex.nixon@citrix.com>
Acked-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Add the new play_dead into smpboot.c, as it fits more cleanly in there
alongside other CONFIG_HOTPLUG functions.
Separate out the common code into its own function.
Signed-off-by: Alex Nixon <alex.nixon@citrix.com>
Acked-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
The removal of the CPU from the various maps was redundant as it already
happened in cpu_disable.
After cleaning this up, cpu_uninit only resets the tlb state, so rename
it and create a noop version for the X86_64 case (so the two play_deads
can be unified later).
Signed-off-by: Alex Nixon <alex.nixon@citrix.com>
Acked-by: Jeremy Fitzhardinge <jeremy@goop.org>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
- more register naming work
- finally figured out that weird CR register stuff
(and did I mention that I hate _really_ undecipherable open-coded values?)
- fix handling of IRQ sharing in interrupt handler
(hopefully properly, otherwise I'd be grateful to hear your
pedantic comments ;)
- add handy SPECS_PAGE references wherever useful
- comments, cleanup
- add me as module author
Signed-off-by: Andreas Mohr <andi@lisas.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Most hardwares have limited buffer-descriptor table length. This
also restricts the max buffer size of the sound driver.
For example, snd-hda-intel has 1MB buffer size limit, and this is
because it can have at most 256 BDL entries. For supporting larger
buffers, we need to allocate larger pages even for sg-buffers.
This patch changes the sgbuf allocation code to try to allocate
larger pages first. At each head of the allocated pages, the
number of allocated pages is stored in the lowest bits of the
corresponding entry of the table addr field. This change isn't
visible as long as the driver uses snd_sgbuf_get_addr() helper.
Also, the patch adds a new function, snd_pcm_sgbuf_get_chunk_size().
This returns the size of the chunk on continuous pages starting at
the given position offset. If the chunk reaches to a non-continuous
page, it returns the size to the boundary.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Clean up SG-buffer helper functions and macros. Helpers take substream
as arguments now.
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
As pointed out during review d_add_ci argument order should match d_add,
so switch the dentry and inode arguments.
Signed-off-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-2.6:
ipv6: protocol for address routes
icmp: icmp_sk() should not use smp_processor_id() in preemptible code
pkt_sched: Fix qdisc list locking
pkt_sched: Fix qdisc_watchdog() vs. dev_deactivate() race
sctp: fix potential panics in the SCTP-AUTH API.
This patch lets the files using linux/version.h match the files that
#include it.
Signed-off-by: Adrian Bunk <bunk@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
if get_rtc_time() is _ever_ called with IRQs off, we deadlock badly
in it, waiting for jiffies to increment.
So make the code more robust by doing an explicit mdelay(20).
This solves a very hard to reproduce/debug hard lockup reported
by Mikael Pettersson.
Reported-by: Mikael Pettersson <mikpe@it.uu.se>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
During CPU hot-remove the sysfs directory created by
threshold_create_bank(), defined in
arch/x86/kernel/cpu/mcheck/mce_amd_64.c, has to be removed before
its parent directory, created by mce_create_device(), defined in
arch/x86/kernel/cpu/mcheck/mce_64.c . Moreover, when the CPU in
question is hotplugged again, obviously the latter has to be created
before the former. At present, the right ordering is not enforced,
because all of these operations are carried out by CPU hotplug
notifiers which are not appropriately ordered with respect to each
other. This leads to serious problems on systems with two or more
multicore AMD CPUs, among other things during suspend and hibernation.
Fix the problem by placing threshold bank CPU hotplug callbacks in
mce_cpu_callback(), so that they are invoked at the right places,
if defined. Additionally, use kobject_del() to remove the sysfs
directory associated with the kobject created by
kobject_create_and_add() in threshold_create_bank(), to prevent the
kernel from crashing during CPU hotplug operations on systems with
two or more multicore AMD CPUs.
This patch fixes bug #11337.
Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Andi Kleen <andi@firstfloor.org>
Tested-by: Mark Langsdorf <mark.langsdorf@amd.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
cpuid leaf 0xb provides extended topology enumeration. This interface provides
the 32-bit x2APIC id of the logical processor and it also provides a new
mechanism to detect SMT and core siblings (which provides increased
addressability).
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Signed-off-by: Ingo Molnar <mingo@elte.hu>
Use dev_printk() in cs_error()-based error reporting. While this
function-turned-macro will hopefully go away soon, using dev_printk
simplifies the code much.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Remoe the CS_ error codes. Drivers are expected to use default
error definitions (errno.h etc.).
Only one CS_ error code -- CS_IN_USE -- is used internally by the
(deprecated) pcmcia-cs userspace package. Therefore, keep this one
around so that it still compiles.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
CS_BAD_TUPLE was used to denote a bad tuple being passed to the parse
function. Therefore, replace it with -EINVAL and a verbose message.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
CS_BAD_ARGS mean a badly written driver or invalid userspace ioctl access,
so translate that to -EINVAL.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
These four error values mostly mean a badly written driver, so ds_dbg()
output and -EINVAL seems to be enough.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
CS_BAD_TYPE was only used in cs.c and already properly annotated by error
messages. CS_BAD_ATTRIBUTE and CS_BAD_PAGE mean a badly written driver, so
ds_dbg() output and -EINVAL seems to be enough.
(includes bugfix from and
Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
)
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
CS_NO_MORE_ITEMS is returned by the CIS tuple reading and parsing code if
the end of a tuple chain is reached. As at least one PCMCIA driver relies
on matching this return value, replace it with -ENOSPC which is now
uniquely used for this purpose within the in-kernel pcmcia subsystem.
CC: Russell King <rmk+kernel@arm.linux.org.uk>
CC: linux-serial@vger.kernel.org
CC: Michael Buesch <mb@bu3sch.de>
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
If a resource is already in use, mark it with -EBUSY. Same for cards already
asleep.
(includes a fix for a bug found by Larry Finger -- thanks!)
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
This error code meant that trying to change the configuration after the
initialization phase is forbidden.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
CS_BAD_HANDLE means that something went badly wrong: no parameter was passed,
or the paramater passed wasn't the correct one. Therefore, replace it with
-EINVAL.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
They were either used to report that changing voltage is not allowed, or that
changing voltage failed.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
CS_UNSUPPORTED_MODE and CS_UNSUPPORTED_FUNCTION were mostly used to denote
trying to use PCMCIA functions on CardBus cards.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
CS_OUT_OF_RESOURCE was almost only used to note -ENOMEM situations.
Therefore, use -ENOMEM explicitely, and also print out warnings.
CC: netdev@vger.kernel.org
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
Deprecate unused CS_ error codes by replacing their definitions with
generic error messages, and removing them from the error_t lookup table.
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
With the PCMCIA ioctl being the only remaining user of
_get_configuration_info, move the function to pcmcia_ioctl.c
Signed-off-by: Dominik Brodowski <linux@dominikbrodowski.net>
* pcmcia-config-loop:
pcmcia: pcmcia_config_loop() improvement by passing vcc
pcmcia: pcmcia_config_loop() default CIS entry handling
pcmcia: pcmcia_config_loop() ConfigIndex unification
pcmcia: use pcmcia_loop_config in misc pcmcia drivers
pcmcia: use pcmcia_loop_config in net pcmcia drivers
pcmcia: use pcmcia_loop_config in ISDN pcmcia drivers
pcmcia: use pcmcia_loop_config in scsi pcmcia drivers
pcmcia: use pcmcia_loop_config in bluetooth drivers
pcmcia: use pcmcia_loop_config in pata and ide drivers
pcmcia: add pcmcia_loop_config() helper
* pcmcia-printk:
pcmcia: don't add extra DEBUG cflag
pcmcia: remove unused cs_socket_name() definition
pcmcia: use dev_printk in module rsrc_nonstatic
pcmcia: use dev_printk in module pcmcia
pcmcia: use dev_printk in module pcmcia_core
pcmcia: use dev_printk and dev_dbg in yenta_socket
