mirror of
https://github.com/a52xq-dev/kernel.git
synced 2026-11-19 16:18:04 +07:00
| Message | SHA1 | |
|---|---|---|
x86/ioapic: Unbreak check_timer()
commit 86a82ae0b5095ea24c55898a3f025791e7958b21 upstream.
Several people reported in the kernel bugzilla that between v4.12 and v4.13
the magic which works around broken hardware and BIOSes to find the proper
timer interrupt delivery mode stopped working for some older affected
platforms which need to fall back to ExtINT delivery mode.
The reason is that the core code changed to keep track of the masked and
disabled state of an interrupt line more accurately to avoid the expensive
hardware operations.
That broke an assumption in i8259_make_irq() which invokes
disable_irq_nosync();
irq_set_chip_and_handler();
enable_irq();
Up to v4.12 this worked because enable_irq() unconditionally unmasked the
interrupt line, but after the state tracking improvements this is not
longer the case because the IO/APIC uses lazy disabling. So the line state
is unmasked which means that enable_irq() does not call into the new irq
chip to unmask it.
In principle this is a shortcoming of the core code, but it's more than
unclear whether the core code should try to reset state. At least this
cannot be done unconditionally as that would break other existing use cases
where the chip type is changed, e.g. when changing the trigger type, but
the callers expect the state to be preserved.
As the way how check_timer() is switching the delivery modes is truly
unique, the obvious fix is to simply unmask the i8259 manually after
changing the mode to ExtINT delivery and switching the irq chip to the
legacy PIC.
Note, that the fixes tag is not really precise, but identifies the commit
which broke the assumptions in the IO/APIC and i8259 code and that's the
kernel version to which this needs to be backported.
Fixes:
|
9fddc16ff6 | |
genirq/affinity: Make affinity setting if activated opt-in
commit f0c7baca180046824e07fc5f1326e83a8fd150c7 upstream.
John reported that on a RK3288 system the perf per CPU interrupts are all
affine to CPU0 and provided the analysis:
"It looks like what happens is that because the interrupts are not per-CPU
in the hardware, armpmu_request_irq() calls irq_force_affinity() while
the interrupt is deactivated and then request_irq() with IRQF_PERCPU |
IRQF_NOBALANCING.
Now when irq_startup() runs with IRQ_STARTUP_NORMAL, it calls
irq_setup_affinity() which returns early because IRQF_PERCPU and
IRQF_NOBALANCING are set, leaving the interrupt on its original CPU."
This was broken by the recent commit which blocked interrupt affinity
setting in hardware before activation of the interrupt. While this works in
general, it does not work for this particular case. As contrary to the
initial analysis not all interrupt chip drivers implement an activate
callback, the safe cure is to make the deferred interrupt affinity setting
at activation time opt-in.
Implement the necessary core logic and make the two irqchip implementations
for which this is required opt-in. In hindsight this would have been the
right thing to do, but ...
Fixes: baedb87d1b53 ("genirq/affinity: Handle affinity setting on inactive interrupts correctly")
Reported-by: John Keeping <john@metanate.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Tested-by: Marc Zyngier <maz@kernel.org>
Acked-by: Marc Zyngier <maz@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/87blk4tzgm.fsf@nanos.tec.linutronix.de
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
5c4d9eefd3 | |
irqdomain/treewide: Free firmware node after domain removal
commit ec0160891e387f4771f953b888b1fe951398e5d9 upstream.
Commit 711419e504eb ("irqdomain: Add the missing assignment of
domain->fwnode for named fwnode") unintentionally caused a dangling pointer
page fault issue on firmware nodes that were freed after IRQ domain
allocation. Commit e3beca48a45b fixed that dangling pointer issue by only
freeing the firmware node after an IRQ domain allocation failure. That fix
no longer frees the firmware node immediately, but leaves the firmware node
allocated after the domain is removed.
The firmware node must be kept around through irq_domain_remove, but should be
freed it afterwards.
Add the missing free operations after domain removal where where appropriate.
Fixes: e3beca48a45b ("irqdomain/treewide: Keep firmware node unconditionally allocated")
Signed-off-by: Jon Derrick <jonathan.derrick@intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Acked-by: Bjorn Helgaas <bhelgaas@google.com> # drivers/pci
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/1595363169-7157-1-git-send-email-jonathan.derrick@intel.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
8e22f6848f | |
irqdomain/treewide: Keep firmware node unconditionally allocated
[ Upstream commit e3beca48a45b5e0e6e6a4e0124276b8248dcc9bb ]
Quite some non OF/ACPI users of irqdomains allocate firmware nodes of type
IRQCHIP_FWNODE_NAMED or IRQCHIP_FWNODE_NAMED_ID and free them right after
creating the irqdomain. The only purpose of these FW nodes is to convey
name information. When this was introduced the core code did not store the
pointer to the node in the irqdomain. A recent change stored the firmware
node pointer in irqdomain for other reasons and missed to notice that the
usage sites which do the alloc_fwnode/create_domain/free_fwnode sequence
are broken by this. Storing a dangling pointer is dangerous itself, but in
case that the domain is destroyed later on this leads to a double free.
Remove the freeing of the firmware node after creating the irqdomain from
all affected call sites to cure this.
Fixes: 711419e504eb ("irqdomain: Add the missing assignment of domain->fwnode for named fwnode")
Reported-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Acked-by: Marc Zyngier <maz@kernel.org>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/873661qakd.fsf@nanos.tec.linutronix.de
Signed-off-by: Sasha Levin <sashal@kernel.org>
|
c0c489e543 | |
genirq/affinity: Handle affinity setting on inactive interrupts correctly
commit baedb87d1b53532f81b4bd0387f83b05d4f7eb9a upstream.
Setting interrupt affinity on inactive interrupts is inconsistent when
hierarchical irq domains are enabled. The core code should just store the
affinity and not call into the irq chip driver for inactive interrupts
because the chip drivers may not be in a state to handle such requests.
X86 has a hacky workaround for that but all other irq chips have not which
causes problems e.g. on GIC V3 ITS.
Instead of adding more ugly hacks all over the place, solve the problem in
the core code. If the affinity is set on an inactive interrupt then:
- Store it in the irq descriptors affinity mask
- Update the effective affinity to reflect that so user space has
a consistent view
- Don't call into the irq chip driver
This is the core equivalent of the X86 workaround and works correctly
because the affinity setting is established in the irq chip when the
interrupt is activated later on.
Note, that this is only effective when hierarchical irq domains are enabled
by the architecture. Doing it unconditionally would break legacy irq chip
implementations.
For hierarchial irq domains this works correctly as none of the drivers can
have a dependency on affinity setting in inactive state by design.
Remove the X86 workaround as it is not longer required.
Fixes:
|
2048e4375c | |
x86/apic: Make TSC deadline timer detection message visible
[ Upstream commit de308d1815c9e8fe602a958c5c76142ff6501d75 ]
The commit
c84cb3735fd5 ("x86/apic: Move TSC deadline timer debug printk")
removed the message which said that the deadline timer was enabled.
It added a pr_debug() message which is issued when deadline timer
validation succeeds.
Well, issued only when CONFIG_DYNAMIC_DEBUG is enabled - otherwise
pr_debug() calls get optimized away if DEBUG is not defined in the
compilation unit.
Therefore, make the above message pr_info() so that it is visible in
dmesg.
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/20200525104218.27018-1-bp@alien8.de
Signed-off-by: Sasha Levin <sashal@kernel.org>
|
bad579b96b | |
x86/apic: Move TSC deadline timer debug printk
[ Upstream commit c84cb3735fd53c91101ccdb191f2e3331a9262cb ] Leon reported that the printk_once() in __setup_APIC_LVTT() triggers a lockdep splat due to a lock order violation between hrtimer_base::lock and console_sem, when the 'once' condition is reset via /sys/kernel/debug/clear_warn_once after boot. The initial printk cannot trigger this because that happens during boot when the local APIC timer is set up on the boot CPU. Prevent it by moving the printk to a place which is guaranteed to be only called once during boot. Mark the deadline timer check related functions and data __init while at it. Reported-by: Leon Romanovsky <leon@kernel.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Link: https://lkml.kernel.org/r/87y2qhoshi.fsf@nanos.tec.linutronix.de Signed-off-by: Sasha Levin <sashal@kernel.org> |
3c52564729 | |
x86/apic/msi: Plug non-maskable MSI affinity race
commit 6f1a4891a5928a5969c87fa5a584844c983ec823 upstream. Evan tracked down a subtle race between the update of the MSI message and the device raising an interrupt internally on PCI devices which do not support MSI masking. The update of the MSI message is non-atomic and consists of either 2 or 3 sequential 32bit wide writes to the PCI config space. - Write address low 32bits - Write address high 32bits (If supported by device) - Write data When an interrupt is migrated then both address and data might change, so the kernel attempts to mask the MSI interrupt first. But for MSI masking is optional, so there exist devices which do not provide it. That means that if the device raises an interrupt internally between the writes then a MSI message is sent built from half updated state. On x86 this can lead to spurious interrupts on the wrong interrupt vector when the affinity setting changes both address and data. As a consequence the device interrupt can be lost causing the device to become stuck or malfunctioning. Evan tried to handle that by disabling MSI accross an MSI message update. That's not feasible because disabling MSI has issues on its own: If MSI is disabled the PCI device is routing an interrupt to the legacy INTx mechanism. The INTx delivery can be disabled, but the disablement is not working on all devices. Some devices lose interrupts when both MSI and INTx delivery are disabled. Another way to solve this would be to enforce the allocation of the same vector on all CPUs in the system for this kind of screwed devices. That could be done, but it would bring back the vector space exhaustion problems which got solved a few years ago. Fortunately the high address (if supported by the device) is only relevant when X2APIC is enabled which implies interrupt remapping. In the interrupt remapping case the affinity setting is happening at the interrupt remapping unit and the PCI MSI message is programmed only once when the PCI device is initialized. That makes it possible to solve it with a two step update: 1) Target the MSI msg to the new vector on the current target CPU 2) Target the MSI msg to the new vector on the new target CPU In both cases writing the MSI message is only changing a single 32bit word which prevents the issue of inconsistency. After writing the final destination it is necessary to check whether the device issued an interrupt while the intermediate state #1 (new vector, current CPU) was in effect. This is possible because the affinity change is always happening on the current target CPU. The code runs with interrupts disabled, so the interrupt can be detected by checking the IRR of the local APIC. If the vector is pending in the IRR then the interrupt is retriggered on the new target CPU by sending an IPI for the associated vector on the target CPU. This can cause spurious interrupts on both the local and the new target CPU. 1) If the new vector is not in use on the local CPU and the device affected by the affinity change raised an interrupt during the transitional state (step #1 above) then interrupt entry code will ignore that spurious interrupt. The vector is marked so that the 'No irq handler for vector' warning is supressed once. 2) If the new vector is in use already on the local CPU then the IRR check might see an pending interrupt from the device which is using this vector. The IPI to the new target CPU will then invoke the handler of the device, which got the affinity change, even if that device did not issue an interrupt 3) If the new vector is in use already on the local CPU and the device affected by the affinity change raised an interrupt during the transitional state (step #1 above) then the handler of the device which uses that vector on the local CPU will be invoked. expose issues in device driver interrupt handlers which are not prepared to handle a spurious interrupt correctly. This not a regression, it's just exposing something which was already broken as spurious interrupts can happen for a lot of reasons and all driver handlers need to be able to deal with them. Reported-by: Evan Green <evgreen@chromium.org> Debugged-by: Evan Green <evgreen@chromium.org> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Evan Green <evgreen@chromium.org> Cc: stable@vger.kernel.org Link: https://lore.kernel.org/r/87imkr4s7n.fsf@nanos.tec.linutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
032a2bf978 | |
x86/ioapic: Prevent inconsistent state when moving an interrupt
[ Upstream commit df4393424af3fbdcd5c404077176082a8ce459c4 ]
There is an issue with threaded interrupts which are marked ONESHOT
and using the fasteoi handler:
if (IS_ONESHOT())
mask_irq();
....
cond_unmask_eoi_irq()
chip->irq_eoi();
if (setaffinity_pending) {
mask_ioapic();
...
move_affinity();
unmask_ioapic();
}
So if setaffinity is pending the interrupt will be moved and then
unconditionally unmasked at the ioapic level, which is wrong in two
aspects:
1) It should be kept masked up to the point where the threaded handler
finished.
2) The physical chip state and the software masked state are inconsistent
Guard both the mask and the unmask with a check for the software masked
state. If the line is marked masked then the ioapic line is also masked, so
both mask_ioapic() and unmask_ioapic() can be skipped safely.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Andy Shevchenko <andy.shevchenko@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Sebastian Siewior <bigeasy@linutronix.de>
Fixes:
|
00f0f29495 | |
x86/apic/32: Avoid bogus LDR warnings
commit fe6f85ca121e9c74e7490fe66b0c5aae38e332c3 upstream.
The removal of the LDR initialization in the bigsmp_32 APIC code unearthed
a problem in setup_local_APIC().
The code checks unconditionally for a mismatch of the logical APIC id by
comparing the early APIC id which was initialized in get_smp_config() with
the actual LDR value in the APIC.
Due to the removal of the bogus LDR initialization the check now can
trigger on bigsmp_32 APIC systems emitting a warning for every booting
CPU. This is of course a false positive because the APIC is not using
logical destination mode.
Restrict the check and the possibly resulting fixup to systems which are
actually using the APIC in logical destination mode.
[ tglx: Massaged changelog and added Cc stable ]
Fixes: bae3a8d3308 ("x86/apic: Do not initialize LDR and DFR for bigsmp")
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/666d8f91-b5a8-1afd-7add-821e72a35f03@suse.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
caddaf43b0 | |
x86/apic/x2apic: Fix a NULL pointer deref when handling a dying cpu
commit 7a22e03b0c02988e91003c505b34d752a51de344 upstream.
Check that the per-cpu cluster mask pointer has been set prior to
clearing a dying cpu's bit. The per-cpu pointer is not set until the
target cpu reaches smp_callin() during CPUHP_BRINGUP_CPU, whereas the
teardown function, x2apic_dead_cpu(), is associated with the earlier
CPUHP_X2APIC_PREPARE. If an error occurs before the cpu is awakened,
e.g. if do_boot_cpu() itself fails, x2apic_dead_cpu() will dereference
the NULL pointer and cause a panic.
smpboot: do_boot_cpu failed(-22) to wakeup CPU#1
BUG: kernel NULL pointer dereference, address: 0000000000000008
RIP: 0010:x2apic_dead_cpu+0x1a/0x30
Call Trace:
cpuhp_invoke_callback+0x9a/0x580
_cpu_up+0x10d/0x140
do_cpu_up+0x69/0xb0
smp_init+0x63/0xa9
kernel_init_freeable+0xd7/0x229
? rest_init+0xa0/0xa0
kernel_init+0xa/0x100
ret_from_fork+0x35/0x40
Fixes:
|
4dedaa73c4 | |
x86/apic/vector: Warn when vector space exhaustion breaks affinity
[ Upstream commit 743dac494d61d991967ebcfab92e4f80dc7583b3 ] On x86, CPUs are limited in the number of interrupts they can have affined to them as they only support 256 interrupt vectors per CPU. 32 vectors are reserved for the CPU and the kernel reserves another 22 for internal purposes. That leaves 202 vectors for assignement to devices. When an interrupt is set up or the affinity is changed by the kernel or the administrator, the vector assignment code attempts to honor the requested affinity mask. If the vector space on the CPUs in that affinity mask is exhausted the code falls back to a wider set of CPUs and assigns a vector on a CPU outside of the requested affinity mask silently. While the effective affinity is reflected in the corresponding /proc/irq/$N/effective_affinity* files the silent breakage of the requested affinity can lead to unexpected behaviour for administrators. Add a pr_warn() when this happens so that adminstrators get at least informed about it in the syslog. [ tglx: Massaged changelog and made the pr_warn() more informative ] Reported-by: djuran@redhat.com Signed-off-by: Neil Horman <nhorman@tuxdriver.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: djuran@redhat.com Link: https://lkml.kernel.org/r/20190822143421.9535-1-nhorman@tuxdriver.com Signed-off-by: Sasha Levin <sashal@kernel.org> |
b619496540 | |
x86/apic: Soft disable APIC before initializing it
[ Upstream commit 2640da4cccf5cc613bf26f0998b9e340f4b5f69c ] If the APIC was already enabled on entry of setup_local_APIC() then disabling it soft via the SPIV register makes a lot of sense. That masks all LVT entries and brings it into a well defined state. Otherwise previously enabled LVTs which are not touched in the setup function stay unmasked and might surprise the just booting kernel. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org> Link: https://lkml.kernel.org/r/20190722105219.068290579@linutronix.de Signed-off-by: Sasha Levin <sashal@kernel.org> |
b40c15c20e | |
x86/apic: Make apic_pending_intr_clear() more robust
[ Upstream commit cc8bf191378c1da8ad2b99cf470ee70193ace84e ]
In course of developing shorthand based IPI support issues with the
function which tries to clear eventually pending ISR bits in the local APIC
were observed.
1) O-day testing triggered the WARN_ON() in apic_pending_intr_clear().
This warning is emitted when the function fails to clear pending ISR
bits or observes pending IRR bits which are not delivered to the CPU
after the stale ISR bit(s) are ACK'ed.
Unfortunately the function only emits a WARN_ON() and fails to dump
the IRR/ISR content. That's useless for debugging.
Feng added spot on debug printk's which revealed that the stale IRR
bit belonged to the APIC timer interrupt vector, but adding ad hoc
debug code does not help with sporadic failures in the field.
Rework the loop so the full IRR/ISR contents are saved and on failure
dumped.
2) The loop termination logic is interesting at best.
If the machine has no TSC or cpu_khz is not known yet it tries 1
million times to ack stale IRR/ISR bits. What?
With TSC it uses the TSC to calculate the loop termination. It takes a
timestamp at entry and terminates the loop when:
(rdtsc() - start_timestamp) >= (cpu_hkz << 10)
That's roughly one second.
Both methods are problematic. The APIC has 256 vectors, which means
that in theory max. 256 IRR/ISR bits can be set. In practice this is
impossible and the chance that more than a few bits are set is close
to zero.
With the pure loop based approach the 1 million retries are complete
overkill.
With TSC this can terminate too early in a guest which is running on a
heavily loaded host even with only a couple of IRR/ISR bits set. The
reason is that after acknowledging the highest priority ISR bit,
pending IRRs must get serviced first before the next round of
acknowledge can take place as the APIC (real and virtualized) does not
honour EOI without a preceeding interrupt on the CPU. And every APIC
read/write takes a VMEXIT if the APIC is virtualized. While trying to
reproduce the issue 0-day reported it was observed that the guest was
scheduled out long enough under heavy load that it terminated after 8
iterations.
Make the loop terminate after 512 iterations. That's plenty enough
in any case and does not take endless time to complete.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20190722105219.158847694@linutronix.de
Signed-off-by: Sasha Levin <sashal@kernel.org>
|
d29c7b8be5 | |
x86/apic: Fix arch_dynirq_lower_bound() bug for DT enabled machines
[ Upstream commit 3e5bedc2c258341702ddffbd7688c5e6eb01eafa ] Rahul Tanwar reported the following bug on DT systems: > 'ioapic_dynirq_base' contains the virtual IRQ base number. Presently, it is > updated to the end of hardware IRQ numbers but this is done only when IOAPIC > configuration type is IOAPIC_DOMAIN_LEGACY or IOAPIC_DOMAIN_STRICT. There is > a third type IOAPIC_DOMAIN_DYNAMIC which applies when IOAPIC configuration > comes from devicetree. > > See dtb_add_ioapic() in arch/x86/kernel/devicetree.c > > In case of IOAPIC_DOMAIN_DYNAMIC (DT/OF based system), 'ioapic_dynirq_base' > remains to zero initialized value. This means that for OF based systems, > virtual IRQ base will get set to zero. Such systems will very likely not even boot. For DT enabled machines ioapic_dynirq_base is irrelevant and not updated, so simply map the IRQ base 1:1 instead. Reported-by: Rahul Tanwar <rahul.tanwar@linux.intel.com> Tested-by: Rahul Tanwar <rahul.tanwar@linux.intel.com> Tested-by: Andy Shevchenko <andriy.shevchenko@intel.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: alan@linux.intel.com Cc: bp@alien8.de Cc: cheol.yong.kim@intel.com Cc: qi-ming.wu@intel.com Cc: rahul.tanwar@intel.com Cc: rppt@linux.ibm.com Cc: tony.luck@intel.com Link: http://lkml.kernel.org/r/20190821081330.1187-1-rahul.tanwar@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org> Signed-off-by: Sasha Levin <sashal@kernel.org> |
e997c07305 | |
Revert "x86/apic: Include the LDR when clearing out APIC registers"
[ Upstream commit 950b07c14e8c59444e2359f15fd70ed5112e11a0 ]
This reverts commit 558682b5291937a70748d36fd9ba757fb25b99ae.
Chris Wilson reports that it breaks his CPU hotplug test scripts. In
particular, it breaks offlining and then re-onlining the boot CPU, which
we treat specially (and the BIOS does too).
The symptoms are that we can offline the CPU, but it then does not come
back online again:
smpboot: CPU 0 is now offline
smpboot: Booting Node 0 Processor 0 APIC 0x0
smpboot: do_boot_cpu failed(-1) to wakeup CPU#0
Thomas says he knows why it's broken (my personal suspicion: our magic
handling of the "cpu0_logical_apicid" thing), but for 5.3 the right fix
is to just revert it, since we've never touched the LDR bits before, and
it's not worth the risk to do anything else at this stage.
[ Hotpluging of the boot CPU is special anyway, and should be off by
default. See the "BOOTPARAM_HOTPLUG_CPU0" config option and the
cpu0_hotplug kernel parameter.
In general you should not do it, and it has various known limitations
(hibernate and suspend require the boot CPU, for example).
But it should work, even if the boot CPU is special and needs careful
treatment - Linus ]
Link: https://lore.kernel.org/lkml/156785100521.13300.14461504732265570003@skylake-alporthouse-com/
Reported-by: Chris Wilson <chris@chris-wilson.co.uk>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Bandan Das <bsd@redhat.com>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
|
991467a47c | |
x86/apic: Include the LDR when clearing out APIC registers
commit 558682b5291937a70748d36fd9ba757fb25b99ae upstream. Although APIC initialization will typically clear out the LDR before setting it, the APIC cleanup code should reset the LDR. This was discovered with a 32-bit KVM guest jumping into a kdump kernel. The stale bits in the LDR triggered a bug in the KVM APIC implementation which caused the destination mapping for VCPUs to be corrupted. Note that this isn't intended to paper over the KVM APIC bug. The kernel has to clear the LDR when resetting the APIC registers except when X2APIC is enabled. This lacks a Fixes tag because missing to clear LDR goes way back into pre git history. [ tglx: Made x2apic_enabled a function call as required ] Signed-off-by: Bandan Das <bsd@redhat.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/20190826101513.5080-3-bsd@redhat.com Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
edc454cd5a | |
x86/apic: Do not initialize LDR and DFR for bigsmp
commit bae3a8d3308ee69a7dbdf145911b18dfda8ade0d upstream.
Legacy apic init uses bigsmp for smp systems with 8 and more CPUs. The
bigsmp APIC implementation uses physical destination mode, but it
nevertheless initializes LDR and DFR. The LDR even ends up incorrectly with
multiple bit being set.
This does not cause a functional problem because LDR and DFR are ignored
when physical destination mode is active, but it triggered a problem on a
32-bit KVM guest which jumps into a kdump kernel.
The multiple bits set unearthed a bug in the KVM APIC implementation. The
code which creates the logical destination map for VCPUs ignores the
disabled state of the APIC and ends up overwriting an existing valid entry
and as a result, APIC calibration hangs in the guest during kdump
initialization.
Remove the bogus LDR/DFR initialization.
This is not intended to work around the KVM APIC bug. The LDR/DFR
ininitalization is wrong on its own.
The issue goes back into the pre git history. The fixes tag is the commit
in the bitkeeper import which introduced bigsmp support in 2003.
git://git.kernel.org/pub/scm/linux/kernel/git/tglx/history.git
Fixes: db7b9e9f26b8 ("[PATCH] Clustered APIC setup for >8 CPU systems")
Suggested-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Bandan Das <bsd@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: stable@vger.kernel.org
Link: https://lkml.kernel.org/r/20190826101513.5080-2-bsd@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
|
959832657c | |
x86/apic: Handle missing global clockevent gracefully
commit f897e60a12f0b9146357780d317879bce2a877dc upstream. Some newer machines do not advertise legacy timers. The kernel can handle that situation if the TSC and the CPU frequency are enumerated by CPUID or MSRs and the CPU supports TSC deadline timer. If the CPU does not support TSC deadline timer the local APIC timer frequency has to be known as well. Some Ryzens machines do not advertize legacy timers, but there is no reliable way to determine the bus frequency which feeds the local APIC timer when the machine allows overclocking of that frequency. As there is no legacy timer the local APIC timer calibration crashes due to a NULL pointer dereference when accessing the not installed global clock event device. Switch the calibration loop to a non interrupt based one, which polls either TSC (if frequency is known) or jiffies. The latter requires a global clockevent. As the machines which do not have a global clockevent installed have a known TSC frequency this is a non issue. For older machines where TSC frequency is not known, there is no known case where the legacy timers do not exist as that would have been reported long ago. Reported-by: Daniel Drake <drake@endlessm.com> Reported-by: Jiri Slaby <jslaby@suse.cz> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Daniel Drake <drake@endlessm.com> Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/r/alpine.DEB.2.21.1908091443030.21433@nanos.tec.linutronix.de Link: http://bugzilla.opensuse.org/show_bug.cgi?id=1142926#c12 Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
685e598e44 | |
x86/apic: Silence -Wtype-limits compiler warnings
[ Upstream commit ec6335586953b0df32f83ef696002063090c7aef ]
There are many compiler warnings like this,
In file included from ./arch/x86/include/asm/smp.h:13,
from ./arch/x86/include/asm/mmzone_64.h:11,
from ./arch/x86/include/asm/mmzone.h:5,
from ./include/linux/mmzone.h:969,
from ./include/linux/gfp.h:6,
from ./include/linux/mm.h:10,
from arch/x86/kernel/apic/io_apic.c:34:
arch/x86/kernel/apic/io_apic.c: In function 'check_timer':
./arch/x86/include/asm/apic.h:37:11: warning: comparison of unsigned
expression >= 0 is always true [-Wtype-limits]
if ((v) <= apic_verbosity) \
^~
arch/x86/kernel/apic/io_apic.c:2160:2: note: in expansion of macro
'apic_printk'
apic_printk(APIC_QUIET, KERN_INFO "..TIMER: vector=0x%02X "
^~~~~~~~~~~
./arch/x86/include/asm/apic.h:37:11: warning: comparison of unsigned
expression >= 0 is always true [-Wtype-limits]
if ((v) <= apic_verbosity) \
^~
arch/x86/kernel/apic/io_apic.c:2207:4: note: in expansion of macro
'apic_printk'
apic_printk(APIC_QUIET, KERN_ERR "..MP-BIOS bug: "
^~~~~~~~~~~
APIC_QUIET is 0, so silence them by making apic_verbosity type int.
Signed-off-by: Qian Cai <cai@lca.pw>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Link: https://lkml.kernel.org/r/1562621805-24789-1-git-send-email-cai@lca.pw
Signed-off-by: Sasha Levin <sashal@kernel.org>
|
242666b2b7 | |
x86/irq: Seperate unused system vectors from spurious entry again
commit f8a8fe61fec8006575699559ead88b0b833d5cad upstream
Quite some time ago the interrupt entry stubs for unused vectors in the
system vector range got removed and directly mapped to the spurious
interrupt vector entry point.
Sounds reasonable, but it's subtly broken. The spurious interrupt vector
entry point pushes vector number 0xFF on the stack which makes the whole
logic in __smp_spurious_interrupt() pointless.
As a consequence any spurious interrupt which comes from a vector != 0xFF
is treated as a real spurious interrupt (vector 0xFF) and not
acknowledged. That subsequently stalls all interrupt vectors of equal and
lower priority, which brings the system to a grinding halt.
This can happen because even on 64-bit the system vector space is not
guaranteed to be fully populated. A full compile time handling of the
unused vectors is not possible because quite some of them are conditonally
populated at runtime.
Bring the entry stubs back, which wastes 160 bytes if all stubs are unused,
but gains the proper handling back. There is no point to selectively spare
some of the stubs which are known at compile time as the required code in
the IDT management would be way larger and convoluted.
Do not route the spurious entries through common_interrupt and do_IRQ() as
the original code did. Route it to smp_spurious_interrupt() which evaluates
the vector number and acts accordingly now that the real vector numbers are
handed in.
Fixup the pr_warn so the actual spurious vector (0xff) is clearly
distiguished from the other vectors and also note for the vectored case
whether it was pending in the ISR or not.
"Spurious APIC interrupt (vector 0xFF) on CPU#0, should never happen."
"Spurious interrupt vector 0xed on CPU#1. Acked."
"Spurious interrupt vector 0xee on CPU#1. Not pending!."
Fixes:
|
fc6975ee93 | |
x86/irq: Handle spurious interrupt after shutdown gracefully
commit b7107a67f0d125459fe41f86e8079afd1a5e0b15 upstream
Since the rework of the vector management, warnings about spurious
interrupts have been reported. Robert provided some more information and
did an initial analysis. The following situation leads to these warnings:
CPU 0 CPU 1 IO_APIC
interrupt is raised
sent to CPU1
Unable to handle
immediately
(interrupts off,
deep idle delay)
mask()
...
free()
shutdown()
synchronize_irq()
clear_vector()
do_IRQ()
-> vector is clear
Before the rework the vector entries of legacy interrupts were statically
assigned and occupied precious vector space while most of them were
unused. Due to that the above situation was handled silently because the
vector was handled and the core handler of the assigned interrupt
descriptor noticed that it is shut down and returned.
While this has been usually observed with legacy interrupts, this situation
is not limited to them. Any other interrupt source, e.g. MSI, can cause the
same issue.
After adding proper synchronization for level triggered interrupts, this
can only happen for edge triggered interrupts where the IO-APIC obviously
cannot provide information about interrupts in flight.
While the spurious warning is actually harmless in this case it worries
users and driver developers.
Handle it gracefully by marking the vector entry as VECTOR_SHUTDOWN instead
of VECTOR_UNUSED when the vector is freed up.
If that above late handling happens the spurious detector will not complain
and switch the entry to VECTOR_UNUSED. Any subsequent spurious interrupt on
that line will trigger the spurious warning as before.
Fixes:
|
9494cd3928 | |
x86/ioapic: Implement irq_get_irqchip_state() callback
commit dfe0cf8b51b07e56ded571e3de0a4a9382517231 upstream
When an interrupt is shut down in free_irq() there might be an inflight
interrupt pending in the IO-APIC remote IRR which is not yet serviced. That
means the interrupt has been sent to the target CPUs local APIC, but the
target CPU is in a state which delays the servicing.
So free_irq() would proceed to free resources and to clear the vector
because synchronize_hardirq() does not see an interrupt handler in
progress.
That can trigger a spurious interrupt warning, which is harmless and just
confuses users, but it also can leave the remote IRR in a stale state
because once the handler is invoked the interrupt resources might be freed
already and therefore acknowledgement is not possible anymore.
Implement the irq_get_irqchip_state() callback for the IO-APIC irq chip. The
callback is invoked from free_irq() via __synchronize_hardirq(). Check the
remote IRR bit of the interrupt and return 'in flight' if it is set and the
interrupt is configured in level mode. For edge mode the remote IRR has no
meaning.
As this is only meaningful for level triggered interrupts this won't cure
the potential spurious interrupt warning for edge triggered interrupts, but
the edge trigger case does not result in stale hardware state. This has to
be addressed at the vector/interrupt entry level seperately.
Fixes:
|
7897f5a443 | |
x86/apic: Fix integer overflow on 10 bit left shift of cpu_khz
[ Upstream commit ea136a112d89bade596314a1ae49f748902f4727 ]
The left shift of unsigned int cpu_khz will overflow for large values of
cpu_khz, so cast it to a long long before shifting it to avoid overvlow.
For example, this can happen when cpu_khz is 4194305, i.e. ~4.2 GHz.
Addresses-Coverity: ("Unintentional integer overflow")
Fixes:
|
2a6ee36917 | |
irq/matrix: Spread managed interrupts on allocation
[ Upstream commit 76f99ae5b54d48430d1f0c5512a84da0ff9761e0 ] Linux spreads out the non managed interrupt across the possible target CPUs to avoid vector space exhaustion. Managed interrupts are treated differently, as for them the vectors are reserved (with guarantee) when the interrupt descriptors are initialized. When the interrupt is requested a real vector is assigned. The assignment logic uses the first CPU in the affinity mask for assignment. If the interrupt has more than one CPU in the affinity mask, which happens when a multi queue device has less queues than CPUs, then doing the same search as for non managed interrupts makes sense as it puts the interrupt on the least interrupt plagued CPU. For single CPU affine vectors that's obviously a NOOP. Restructre the matrix allocation code so it does the 'best CPU' search, add the sanity check for an empty affinity mask and adapt the call site in the x86 vector management code. [ tglx: Added the empty mask check to the core and improved change log ] Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: hpa@zytor.com Link: https://lkml.kernel.org/r/20180908175838.14450-2-dou_liyang@163.com Signed-off-by: Sasha Levin <sashal@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org> |
8cae7757e8 | |
x86/apic/vector: Make error return value negative
activate_managed() returns EINVAL instead of -EINVAL in case of
error. While this is unlikely to happen, the positive return value would
cause further malfunction at the call site.
Fixes:
|
47b7360ce5 | |
x86/smp: fix non-SMP broken build due to redefinition of apic_id_is_primary_thread
The function has an inline "return false;" definition with CONFIG_SMP=n
but the "real" definition is also visible leading to "redefinition of
‘apic_id_is_primary_thread’" compiler error.
Guard it with #ifdef CONFIG_SMP
Signed-off-by: Vlastimil Babka <vbabka@suse.cz>
Fixes:
|
d0055f351e | |
Merge branch 'l1tf-final' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Merge L1 Terminal Fault fixes from Thomas Gleixner:
"L1TF, aka L1 Terminal Fault, is yet another speculative hardware
engineering trainwreck. It's a hardware vulnerability which allows
unprivileged speculative access to data which is available in the
Level 1 Data Cache when the page table entry controlling the virtual
address, which is used for the access, has the Present bit cleared or
other reserved bits set.
If an instruction accesses a virtual address for which the relevant
page table entry (PTE) has the Present bit cleared or other reserved
bits set, then speculative execution ignores the invalid PTE and loads
the referenced data if it is present in the Level 1 Data Cache, as if
the page referenced by the address bits in the PTE was still present
and accessible.
While this is a purely speculative mechanism and the instruction will
raise a page fault when it is retired eventually, the pure act of
loading the data and making it available to other speculative
instructions opens up the opportunity for side channel attacks to
unprivileged malicious code, similar to the Meltdown attack.
While Meltdown breaks the user space to kernel space protection, L1TF
allows to attack any physical memory address in the system and the
attack works across all protection domains. It allows an attack of SGX
and also works from inside virtual machines because the speculation
bypasses the extended page table (EPT) protection mechanism.
The assoicated CVEs are: CVE-2018-3615, CVE-2018-3620, CVE-2018-3646
The mitigations provided by this pull request include:
- Host side protection by inverting the upper address bits of a non
present page table entry so the entry points to uncacheable memory.
- Hypervisor protection by flushing L1 Data Cache on VMENTER.
- SMT (HyperThreading) control knobs, which allow to 'turn off' SMT
by offlining the sibling CPU threads. The knobs are available on
the kernel command line and at runtime via sysfs
- Control knobs for the hypervisor mitigation, related to L1D flush
and SMT control. The knobs are available on the kernel command line
and at runtime via sysfs
- Extensive documentation about L1TF including various degrees of
mitigations.
Thanks to all people who have contributed to this in various ways -
patches, review, testing, backporting - and the fruitful, sometimes
heated, but at the end constructive discussions.
There is work in progress to provide other forms of mitigations, which
might be less horrible performance wise for a particular kind of
workloads, but this is not yet ready for consumption due to their
complexity and limitations"
* 'l1tf-final' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: (75 commits)
x86/microcode: Allow late microcode loading with SMT disabled
tools headers: Synchronise x86 cpufeatures.h for L1TF additions
x86/mm/kmmio: Make the tracer robust against L1TF
x86/mm/pat: Make set_memory_np() L1TF safe
x86/speculation/l1tf: Make pmd/pud_mknotpresent() invert
x86/speculation/l1tf: Invert all not present mappings
cpu/hotplug: Fix SMT supported evaluation
KVM: VMX: Tell the nested hypervisor to skip L1D flush on vmentry
x86/speculation: Use ARCH_CAPABILITIES to skip L1D flush on vmentry
x86/speculation: Simplify sysfs report of VMX L1TF vulnerability
Documentation/l1tf: Remove Yonah processors from not vulnerable list
x86/KVM/VMX: Don't set l1tf_flush_l1d from vmx_handle_external_intr()
x86/irq: Let interrupt handlers set kvm_cpu_l1tf_flush_l1d
x86: Don't include linux/irq.h from asm/hardirq.h
x86/KVM/VMX: Introduce per-host-cpu analogue of l1tf_flush_l1d
x86/irq: Demote irq_cpustat_t::__softirq_pending to u16
x86/KVM/VMX: Move the l1tf_flush_l1d test to vmx_l1d_flush()
x86/KVM/VMX: Replace 'vmx_l1d_flush_always' with 'vmx_l1d_flush_cond'
x86/KVM/VMX: Don't set l1tf_flush_l1d to true from vmx_l1d_flush()
cpu/hotplug: detect SMT disabled by BIOS
...
|
958f338e96 | |
Merge branch 'x86-platform-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 platform updates from Thomas Gleixner: "Trivial cleanups and improvements" * 'x86-platform-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86/platform/UV: Remove redundant check of p == q x86/platform/olpc: Use PTR_ERR_OR_ZERO() x86/platform/UV: Mark memblock related init code and data correctly |
7edcf0d314 | |
Merge branch 'x86-apic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 apic update from Thomas Gleixner: "Trivial cleanups of the APIC related code" * 'x86-apic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86/apic: Trivial coding style fixes x86/vector: Merge allocate_vector() into assign_vector_locked() |
66e22087bd | |
Merge 4.18-rc7 into master to pick up the KVM dependcy
Signed-off-by: Thomas Gleixner <tglx@linutronix.de> |
f2701b77bb | |
x86: Don't include linux/irq.h from asm/hardirq.h
The next patch in this series will have to make the definition of
irq_cpustat_t available to entering_irq().
Inclusion of asm/hardirq.h into asm/apic.h would cause circular header
dependencies like
asm/smp.h
asm/apic.h
asm/hardirq.h
linux/irq.h
linux/topology.h
linux/smp.h
asm/smp.h
or
linux/gfp.h
linux/mmzone.h
asm/mmzone.h
asm/mmzone_64.h
asm/smp.h
asm/apic.h
asm/hardirq.h
linux/irq.h
linux/irqdesc.h
linux/kobject.h
linux/sysfs.h
linux/kernfs.h
linux/idr.h
linux/gfp.h
and others.
This causes compilation errors because of the header guards becoming
effective in the second inclusion: symbols/macros that had been defined
before wouldn't be available to intermediate headers in the #include chain
anymore.
A possible workaround would be to move the definition of irq_cpustat_t
into its own header and include that from both, asm/hardirq.h and
asm/apic.h.
However, this wouldn't solve the real problem, namely asm/harirq.h
unnecessarily pulling in all the linux/irq.h cruft: nothing in
asm/hardirq.h itself requires it. Also, note that there are some other
archs, like e.g. arm64, which don't have that #include in their
asm/hardirq.h.
Remove the linux/irq.h #include from x86' asm/hardirq.h.
Fix resulting compilation errors by adding appropriate #includes to *.c
files as needed.
Note that some of these *.c files could be cleaned up a bit wrt. to their
set of #includes, but that should better be done from separate patches, if
at all.
Signed-off-by: Nicolai Stange <nstange@suse.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
|
447ae31667 | |
x86/apic: Trivial coding style fixes
There is inconsistent indenting in calibrate_APIC_clock() and activate_managed(). Remove the surplus TAB. Signed-off-by: Yi Wang <wang.yi59@zte.com.cn> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Jiang Biao <jiang.biao2@zte.com.cn> Acked-by: Steven Rostedt (VMware) <rostedt@goodmis.org> Cc: hpa@zytor.com Cc: douly.fnst@cn.fujitsu.com Cc: jgross@suse.com Cc: ville.syrjala@linux.intel.com Cc: len.brown@intel.com Cc: gregkh@linuxfoundation.org Cc: zhong.weidong@zte.com.cn Link: https://lkml.kernel.org/r/1532672103-32250-1-git-send-email-wang.yi59@zte.com.cn |
843c408905 | |
x86/platform/UV: Mark memblock related init code and data correctly
parse_mem_block_size() and mem_block_size are only used during init. Mark
them accordingly.
Fixes:
|
24cfd8ca1d | |
x86/apic: Future-proof the TSC_DEADLINE quirk for SKX
All SKX with stepping higher than 4 support the TSC_DEADLINE,
no matter the microcode version.
Without this patch, upcoming SKX steppings will not be able to use
their TSC_DEADLINE timer.
Signed-off-by: Len Brown <len.brown@intel.com>
Cc: <stable@kernel.org> # v4.14+
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes:
|
d9e6dbcf28 | |
Revert "x86/apic: Ignore secondary threads if nosmt=force"
Dave Hansen reported, that it's outright dangerous to keep SMT siblings
disabled completely so they are stuck in the BIOS and wait for SIPI.
The reason is that Machine Check Exceptions are broadcasted to siblings and
the soft disabled sibling has CR4.MCE = 0. If a MCE is delivered to a
logical core with CR4.MCE = 0, it asserts IERR#, which shuts down or
reboots the machine. The MCE chapter in the SDM contains the following
blurb:
Because the logical processors within a physical package are tightly
coupled with respect to shared hardware resources, both logical
processors are notified of machine check errors that occur within a
given physical processor. If machine-check exceptions are enabled when
a fatal error is reported, all the logical processors within a physical
package are dispatched to the machine-check exception handler. If
machine-check exceptions are disabled, the logical processors enter the
shutdown state and assert the IERR# signal. When enabling machine-check
exceptions, the MCE flag in control register CR4 should be set for each
logical processor.
Reverting the commit which ignores siblings at enumeration time solves only
half of the problem. The core cpuhotplug logic needs to be adjusted as
well.
This thoughtful engineered mechanism also turns the boot process on all
Intel HT enabled systems into a MCE lottery. MCE is enabled on the boot CPU
before the secondary CPUs are brought up. Depending on the number of
physical cores the window in which this situation can happen is smaller or
larger. On a HSW-EX it's about 750ms:
MCE is enabled on the boot CPU:
[ 0.244017] mce: CPU supports 22 MCE banks
The corresponding sibling #72 boots:
[ 1.008005] .... node #0, CPUs: #72
That means if an MCE hits on physical core 0 (logical CPUs 0 and 72)
between these two points the machine is going to shutdown. At least it's a
known safe state.
It's obvious that the early boot can be hit by an MCE as well and then runs
into the same situation because MCEs are not yet enabled on the boot CPU.
But after enabling them on the boot CPU, it does not make any sense to
prevent the kernel from recovering.
Adjust the nosmt kernel parameter documentation as well.
Reverts:
|
506a66f374 | |
x86/platform/UV: Add kernel parameter to set memory block size
Add a kernel parameter that allows setting UV memory block size. This is to provide an adjustment for new forms of PMEM and other DIMM memory that might require alignment restrictions other than scanning the global address table for the required minimum alignment. The value set will be further adjusted by both the GAM range table scan as well as restrictions imposed by set_memory_block_size_order(). Signed-off-by: Mike Travis <mike.travis@hpe.com> Reviewed-by: Andrew Banman <andrew.banman@hpe.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Dimitri Sivanich <dimitri.sivanich@hpe.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Russ Anderson <russ.anderson@hpe.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dan.j.williams@intel.com Cc: jgross@suse.com Cc: kirill.shutemov@linux.intel.com Cc: mhocko@suse.com Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/lkml/20180524201711.854849120@stormcage.americas.sgi.com Signed-off-by: Ingo Molnar <mingo@kernel.org> |
d7609f4210 | |
x86/platform/UV: Use new set memory block size function
Add a call to the new function to "adjust" the current fixed UV memory block size of 2GB so it can be changed to a different physical boundary. This accommodates changes in the Intel BIOS, and therefore UV BIOS, which now can align boundaries different than the previous UV standard of 2GB. It also flags any UV Global Address boundaries from BIOS that cause a change in the mem block size (boundary). The current boundary of 2GB has been used on UV since the first system release in 2009 with Linux 2.6 and has worked fine. But the new NVDIMM persistent memory modules (PMEM), along with the Intel BIOS changes to support these modules caused the memory block size boundary to be set to a lower limit. Intel only guarantees that this minimum boundary at 64MB though the current Linux limit is 128MB. Note that the default remains 2GB if no changes occur. Signed-off-by: Mike Travis <mike.travis@hpe.com> Reviewed-by: Andrew Banman <andrew.banman@hpe.com> Cc: Andrew Morton <akpm@linux-foundation.org> Cc: Dimitri Sivanich <dimitri.sivanich@hpe.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Russ Anderson <russ.anderson@hpe.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: dan.j.williams@intel.com Cc: jgross@suse.com Cc: kirill.shutemov@linux.intel.com Cc: mhocko@suse.com Cc: stable@vger.kernel.org Link: https://lkml.kernel.org/lkml/20180524201711.732785782@stormcage.americas.sgi.com Signed-off-by: Ingo Molnar <mingo@kernel.org> |
bbbd2b51a2 | |
x86/apic: Ignore secondary threads if nosmt=force
nosmt on the kernel command line merely prevents the onlining of the
secondary SMT siblings.
nosmt=force makes the APIC detection code ignore the secondary SMT siblings
completely, so they even do not show up as possible CPUs. That reduces the
amount of memory allocations for per cpu variables and saves other
resources from being allocated too large.
This is not fully equivalent to disabling SMT in the BIOS because the low
level SMT enabling in the BIOS can result in partitioning of resources
between the siblings, which is not undone by just ignoring them. Some CPUs
can use the full resources when their sibling is not onlined, but this is
depending on the CPU family and model and it's not well documented whether
this applies to all partitioned resources. That means depending on the
workload disabling SMT in the BIOS might result in better performance.
Linus analysis of the Intel manual:
The intel optimization manual is not very clear on what the partitioning
rules are.
I find:
"In general, the buffers for staging instructions between major pipe
stages are partitioned. These buffers include µop queues after the
execution trace cache, the queues after the register rename stage, the
reorder buffer which stages instructions for retirement, and the load
and store buffers.
In the case of load and store buffers, partitioning also provided an
easier implementation to maintain memory ordering for each logical
processor and detect memory ordering violations"
but some of that partitioning may be relaxed if the HT thread is "not
active":
"In Intel microarchitecture code name Sandy Bridge, the micro-op queue
is statically partitioned to provide 28 entries for each logical
processor, irrespective of software executing in single thread or
multiple threads. If one logical processor is not active in Intel
microarchitecture code name Ivy Bridge, then a single thread executing
on that processor core can use the 56 entries in the micro-op queue"
but I do not know what "not active" means, and how dynamic it is. Some of
that partitioning may be entirely static and depend on the early BIOS
disabling of HT, and even if we park the cores, the resources will just be
wasted.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Acked-by: Ingo Molnar <mingo@kernel.org>
|
2207def700 | |
x86/smp: Provide topology_is_primary_thread()
If the CPU is supporting SMT then the primary thread can be found by checking the lower APIC ID bits for zero. smp_num_siblings is used to build the mask for the APIC ID bits which need to be taken into account. This uses the MPTABLE or ACPI/MADT supplied APIC ID, which can be different than the initial APIC ID in CPUID. But according to AMD the lower bits have to be consistent. Intel gave a tentative confirmation as well. Preparatory patch to support disabling SMT at boot/runtime. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Acked-by: Ingo Molnar <mingo@kernel.org> |
6a4d2657e0 | |
x86/apic/vector: Print APIC control bits in debugfs
Extend the debugability of the vector management by adding the state bits to the debugfs output. Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Tested-by: Song Liu <songliubraving@fb.com> Cc: Joerg Roedel <jroedel@suse.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Song Liu <liu.song.a23@gmail.com> Cc: Dmitry Safonov <0x7f454c46@gmail.com> Cc: Mike Travis <mike.travis@hpe.com> Cc: Borislav Petkov <bp@alien8.de> Cc: Tariq Toukan <tariqt@mellanox.com> Link: https://lkml.kernel.org/r/20180604162224.908136099@linutronix.de |
a07771ac6a | |
x86/ioapic: Use apic_ack_irq()
To address the EBUSY fail of interrupt affinity settings in case that the
previous setting has not been cleaned up yet, use the new apic_ack_irq()
function instead of directly invoking ack_APIC_irq().
Preparatory change for the real fix
Fixes:
|
2b04e46d8d | |
x86/apic: Provide apic_ack_irq()
apic_ack_edge() is explicitely for handling interrupt affinity cleanup when
interrupt remapping is not available or disable.
Remapped interrupts and also some of the platform specific special
interrupts, e.g. UV, invoke ack_APIC_irq() directly.
To address the issue of failing an affinity update with -EBUSY the delayed
affinity mechanism can be reused, but ack_APIC_irq() does not handle
that. Adding this to ack_APIC_irq() is not possible, because that function
is also used for exceptions and directly handled interrupts like IPIs.
Create a new function, which just contains the conditional invocation of
irq_move_irq() and the final ack_APIC_irq().
Reuse the new function in apic_ack_edge().
Preparatory change for the real fix.
Fixes:
|
c0255770cc | |
x86/apic/vector: Prevent hlist corruption and leaks
Several people observed the WARN_ON() in irq_matrix_free() which triggers
when the caller tries to free an vector which is not in the allocation
range. Song provided the trace information which allowed to decode the root
cause.
The rework of the vector allocation mechanism failed to preserve a sanity
check, which prevents setting a new target vector/CPU when the previous
affinity change has not fully completed.
As a result a half finished affinity change can be overwritten, which can
cause the leak of a irq descriptor pointer on the previous target CPU and
double enqueue of the hlist head into the cleanup lists of two or more
CPUs. After one CPU cleaned up its vector the next CPU will invoke the
cleanup handler with vector 0, which triggers the out of range warning in
the matrix allocator.
Prevent this by checking the apic_data of the interrupt whether the
move_in_progress flag is false and the hlist node is not hashed. Return
-EBUSY if not.
This prevents the damage and restores the behaviour before the vector
allocation rework, but due to other changes in that area it also widens the
chance that user space can observe -EBUSY. In theory this should be fine,
but actually not all user space tools handle -EBUSY correctly. Addressing
that is not part of this fix, but will be addressed in follow up patches.
Fixes:
|
80ae7b1a91 | |
x86/vector: Merge allocate_vector() into assign_vector_locked()
assign_vector_locked() calls allocate_vector() to get a real vector for an IRQ. If the current target CPU is online and in the new requested affinity mask, allocate_vector() will return 0 and nothing should be done. But, assign_vector_locked() calls apic_update_irq_cfg() even in that case which is pointless. allocate_vector() is not called from anything else, so the functions can be merged and in case of no change the apic_update_irq_cfg() can be avoided. [ tglx: Massaged changelog ] Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: hpa@zytor.com Link: https://lkml.kernel.org/r/20180511080956.6316-1-douly.fnst@cn.fujitsu.com |
2773397171 | |
x86/apic/x2apic: Initialize cluster ID properly
Rick bisected a regression on large systems which use the x2apic cluster
mode for interrupt delivery to the commit wich reworked the cluster
management.
The problem is caused by a missing initialization of the clusterid field
in the shared cluster data structures. So all structures end up with
cluster ID 0 which only allows sharing between all CPUs which belong to
cluster 0. All other CPUs with a cluster ID > 0 cannot share the data
structure because they cannot find existing data with their cluster
ID. This causes malfunction with IPIs because IPIs are sent to the wrong
cluster and the caller waits for ever that the target CPU handles the IPI.
Add the missing initialization when a upcoming CPU is the first in a
cluster so that the later booting CPUs can find the data and share it for
proper operation.
Fixes:
|
fed71f7d98 | |
Merge branch 'WIP.x86/asm' into x86/urgent, because the topic is ready
Signed-off-by: Ingo Molnar <mingo@kernel.org> |
ef389b7346 | |
x86/apic: Fix signedness bug in APIC ID validity checks
The APIC ID as parsed from ACPI MADT is validity checked with the apic->apic_id_valid() callback, which depends on the selected APIC type. For non X2APIC types APIC IDs >= 0xFF are invalid, but values > 0x7FFFFFFF are detected as valid. This happens because the 'apicid' argument of the apic_id_valid() callback is type 'int'. So the resulting comparison apicid < 0xFF evaluates to true for all unsigned int values > 0x7FFFFFFF which are handed to default_apic_id_valid(). As a consequence, invalid APIC IDs in !X2APIC mode are considered valid and accounted as possible CPUs. Change the apicid argument type of the apic_id_valid() callback to u32 so the evaluation is unsigned and returns the correct result. [ tglx: Massaged changelog ] Signed-off-by: Li RongQing <lirongqing@baidu.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: stable@vger.kernel.org Cc: jgross@suse.com Cc: Dou Liyang <douly.fnst@cn.fujitsu.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: hpa@zytor.com Link: https://lkml.kernel.org/r/1523322966-10296-1-git-send-email-lirongqing@baidu.com |
a774635db5 | |
Merge branch 'x86-apic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip
Pull x86 apic updates from Ingo Molnar:
"The main x86 APIC/IOAPIC changes in this cycle were:
- Robustify kexec support to more carefully restore IRQ hardware
state before calling into kexec/kdump kernels. (Baoquan He)
- Clean up the local APIC code a bit (Dou Liyang)
- Remove unused callbacks (David Rientjes)"
* 'x86-apic-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
x86/apic: Finish removing unused callbacks
x86/apic: Drop logical_smp_processor_id() inline
x86/apic: Modernize the pending interrupt code
x86/apic: Move pending interrupt check code into it's own function
x86/apic: Set up through-local-APIC mode on the boot CPU if 'noapic' specified
x86/apic: Rename variables and functions related to x86_io_apic_ops
x86/apic: Remove the (now) unused disable_IO_APIC() function
x86/apic: Fix restoring boot IRQ mode in reboot and kexec/kdump
x86/apic: Split disable_IO_APIC() into two functions to fix CONFIG_KEXEC_JUMP=y
x86/apic: Split out restore_boot_irq_mode() from disable_IO_APIC()
x86/apic: Make setup_local_APIC() static
x86/apic: Simplify init_bsp_APIC() usage
x86/x2apic: Mark set_x2apic_phys_mode() as __init
|
2451d1e59d | |
x86/apic: Drop logical_smp_processor_id() inline
The logical_smp_processor_id() inline which is only called in setup_local_APIC() on x86_32 systems has no real value. Drop it and directly use GET_APIC_LOGICAL_ID() at the call site and use a more suitable variable name for readability Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com> Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Cc: andy.shevchenko@gmail.com Cc: bhe@redhat.com Cc: ebiederm@xmission.com Link: https://lkml.kernel.org/r/20180301055930.2396-4-douly.fnst@cn.fujitsu.com |
8f1561680f |