| Age | Commit message (Collapse) | Author |
|---|
|
When CACHE_SET_IO_DISABLE is set on cache set flags, bcache allocator
thread routine bch_allocator_thread() may stop the while-loops and
exit. Then it is possible to observe the following kernel oops message,
[ 631.068366] bcache: bch_btree_insert() error -5
[ 631.069115] bcache: cached_dev_detach_finish() Caching disabled for sdf
[ 631.070220] BUG: unable to handle kernel NULL pointer dereference at 0000000000000000
[ 631.070250] PGD 0 P4D 0
[ 631.070261] Oops: 0002 [#1] SMP PTI
[snipped]
[ 631.070578] Workqueue: events cache_set_flush [bcache]
[ 631.070597] RIP: 0010:exit_creds+0x1b/0x50
[ 631.070610] RSP: 0018:ffffc9000705fe08 EFLAGS: 00010246
[ 631.070626] RAX: 0000000000000001 RBX: ffff880a622ad300 RCX: 000000000000000b
[ 631.070645] RDX: 0000000000000601 RSI: 000000000000000c RDI: 0000000000000000
[ 631.070663] RBP: ffff880a622ad300 R08: ffffea00190c66e0 R09: 0000000000000200
[ 631.070682] R10: ffff880a48123000 R11: ffff880000000000 R12: 0000000000000000
[ 631.070700] R13: ffff880a4b160e40 R14: ffff880a4b160000 R15: 0ffff880667e2530
[ 631.070719] FS: 0000000000000000(0000) GS:ffff880667e00000(0000) knlGS:0000000000000000
[ 631.070740] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 631.070755] CR2: 0000000000000000 CR3: 000000000200a001 CR4: 00000000003606e0
[ 631.070774] DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
[ 631.070793] DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
[ 631.070811] Call Trace:
[ 631.070828] __put_task_struct+0x55/0x160
[ 631.070845] kthread_stop+0xee/0x100
[ 631.070863] cache_set_flush+0x11d/0x1a0 [bcache]
[ 631.070879] process_one_work+0x146/0x340
[ 631.070892] worker_thread+0x47/0x3e0
[ 631.070906] kthread+0xf5/0x130
[ 631.070917] ? max_active_store+0x60/0x60
[ 631.070930] ? kthread_bind+0x10/0x10
[ 631.070945] ret_from_fork+0x35/0x40
[snipped]
[ 631.071017] RIP: exit_creds+0x1b/0x50 RSP: ffffc9000705fe08
[ 631.071033] CR2: 0000000000000000
[ 631.071045] ---[ end trace 011c63a24b22c927 ]---
[ 631.071085] bcache: bcache_device_free() bcache0 stopped
The reason is when cache_set_flush() tries to call kthread_stop() to stop
allocator thread, but it exits already due to cache device I/O errors.
This patch adds wait_for_kthread_stop() at tail of bch_allocator_thread(),
to prevent the thread routine exiting directly. Then the allocator thread
can be blocked at wait_for_kthread_stop() and wait for cache_set_flush()
to stop it by calling kthread_stop().
changelog:
v3: add Reviewed-by from Hannnes.
v2: not directly return from allocator_wait(), move 'return 0' to tail of
bch_allocator_thread().
v1: initial version.
Fixes: 771f393e8ffc ("bcache: add CACHE_SET_IO_DISABLE to struct cache_set flags")
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Commit c7b7bd07404c5 ("bcache: add io_disable to struct cached_dev")
counts backing device I/O requets and set dc->io_disable to true if error
counters exceeds dc->io_error_limit. But it only counts I/O errors for
regular I/O request, neglects errors of write back I/Os when backing device
is offline.
This patch counts the errors of writeback I/Os, in dirty_endio() if
bio->bi_status is not 0, it means error happens when writing dirty keys
to backing device, then bch_count_backing_io_errors() is called.
By this fix, even there is no reqular I/O request coming, if writeback I/O
errors exceed dc->io_error_limit, the bcache device may still be stopped
for the broken backing device.
Fixes: c7b7bd07404c5 ("bcache: add io_disable to struct cached_dev")
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Commit c7b7bd07404c5 ("bcache: add io_disable to struct cached_dev") tries
to stop bcache device by calling bcache_device_stop() when too many I/O
errors happened on backing device. But if there is internal I/O happening
on cache device (writeback scan, garbage collection, etc), a regular I/O
request triggers the internal I/Os may still holds a refcount of dc->count,
and the refcount may only be dropped after the internal I/O stopped.
By this patch, bch_cached_dev_error() will check if the backing device is
attached to a cache set, if yes that CACHE_SET_IO_DISABLE will be set to
flags of this cache set. Then internal I/Os on cache device will be
rejected and stopped immediately, and the bcache device can be stopped.
For people who are not familiar with the interesting refcount dependance,
let me explain a bit more how the fix works. Example the writeback thread
will scan cache device for dirty data writeback purpose. Before it stopps,
it holds a refcount of dc->count. When CACHE_SET_IO_DISABLE bit is set,
the internal I/O will stopped and the while-loop in bch_writeback_thread()
quits and calls cached_dev_put() to drop dc->count. If this is the last
refcount to drop, then cached_dev_detach_finish() will be called. In this
call back function, in turn closure_put(dc->disk.cl) is called to drop a
refcount of closure dc->disk.cl. If this is the last refcount of this
closure to drop, then cached_dev_flush() will be called. Then the cached
device is freed. So if CACHE_SET_IO_DISABLE is not set, the bache device
can not be stopped until all inernal cache device I/O stopped. For large
size cache device, and writeback thread competes locks with gc thread,
there might be a quite long time to wait.
Fixes: c7b7bd07404c5 ("bcache: add io_disable to struct cached_dev")
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
Current code uses bdevname() or bio_devname() to reference gendisk
disk name when bcache needs to display the disk names in kernel message.
It was safe before bcache device failure handling patch set merged in,
because when devices are failed, there was deadlock to prevent bcache
printing error messages with gendisk disk name. But after the failure
handling patch set merged, the deadlock is fixed, so it is possible
that the gendisk structure bdev->hd_disk is released when bdevname() is
called to reference bdev->bd_disk->disk_name[]. This is why I receive
bug report of NULL pointers deference panic.
This patch stores gendisk disk name in a buffer inside struct cache and
struct cached_dev, then print out the offline device name won't reference
bdev->hd_disk anymore. And this patch also avoids extra function calls
of bdevname() and bio_devnmae().
Changelog:
v3, add Reviewed-by from Hannes.
v2, call bdevname() earlier in register_bdev()
v1, first version with segguestion from Junhui Tang.
Fixes: c7b7bd07404c5 ("bcache: add io_disable to struct cached_dev")
Fixes: 5138ac6748e38 ("bcache: fix misleading error message in bch_count_io_errors()")
Signed-off-by: Coly Li <colyli@suse.de>
Reviewed-by: Hannes Reinecke <hare@suse.com>
Signed-off-by: Jens Axboe <axboe@kernel.dk>
|
|
A misaligned address is only worth a warning, and not
stopping the while execution path with a BUG_ON().
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
It allows to flush more than 4GB of device TLBs. So the mask should be
64bit wide. UBSAN captured this fault as below.
[ 3.760024] ================================================================================
[ 3.768440] UBSAN: Undefined behaviour in drivers/iommu/dmar.c:1348:3
[ 3.774864] shift exponent 64 is too large for 32-bit type 'int'
[ 3.780853] CPU: 2 PID: 0 Comm: swapper/2 Tainted: G U 4.17.0-rc1+ #89
[ 3.788661] Hardware name: Dell Inc. OptiPlex 7040/0Y7WYT, BIOS 1.2.8 01/26/2016
[ 3.796034] Call Trace:
[ 3.798472] <IRQ>
[ 3.800479] dump_stack+0x90/0xfb
[ 3.803787] ubsan_epilogue+0x9/0x40
[ 3.807353] __ubsan_handle_shift_out_of_bounds+0x10e/0x170
[ 3.812916] ? qi_flush_dev_iotlb+0x124/0x180
[ 3.817261] qi_flush_dev_iotlb+0x124/0x180
[ 3.821437] iommu_flush_dev_iotlb+0x94/0xf0
[ 3.825698] iommu_flush_iova+0x10b/0x1c0
[ 3.829699] ? fq_ring_free+0x1d0/0x1d0
[ 3.833527] iova_domain_flush+0x25/0x40
[ 3.837448] fq_flush_timeout+0x55/0x160
[ 3.841368] ? fq_ring_free+0x1d0/0x1d0
[ 3.845200] ? fq_ring_free+0x1d0/0x1d0
[ 3.849034] call_timer_fn+0xbe/0x310
[ 3.852696] ? fq_ring_free+0x1d0/0x1d0
[ 3.856530] run_timer_softirq+0x223/0x6e0
[ 3.860625] ? sched_clock+0x5/0x10
[ 3.864108] ? sched_clock+0x5/0x10
[ 3.867594] __do_softirq+0x1b5/0x6f5
[ 3.871250] irq_exit+0xd4/0x130
[ 3.874470] smp_apic_timer_interrupt+0xb8/0x2f0
[ 3.879075] apic_timer_interrupt+0xf/0x20
[ 3.883159] </IRQ>
[ 3.885255] RIP: 0010:poll_idle+0x60/0xe7
[ 3.889252] RSP: 0018:ffffb1b201943e30 EFLAGS: 00000246 ORIG_RAX: ffffffffffffff13
[ 3.896802] RAX: 0000000080200000 RBX: 000000000000008e RCX: 000000000000001f
[ 3.903918] RDX: 0000000000000000 RSI: 000000002819aa06 RDI: 0000000000000000
[ 3.911031] RBP: ffff9e93c6b33280 R08: 00000010f717d567 R09: 000000000010d205
[ 3.918146] R10: ffffb1b201943df8 R11: 0000000000000001 R12: 00000000e01b169d
[ 3.925260] R13: 0000000000000000 R14: ffffffffb12aa400 R15: 0000000000000000
[ 3.932382] cpuidle_enter_state+0xb4/0x470
[ 3.936558] do_idle+0x222/0x310
[ 3.939779] cpu_startup_entry+0x78/0x90
[ 3.943693] start_secondary+0x205/0x2e0
[ 3.947607] secondary_startup_64+0xa5/0xb0
[ 3.951783] ================================================================================
Signed-off-by: Changbin Du <changbin.du@intel.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
This pretty much reverts commit 273df9635385 ("iommu/dma: Make PCI
window reservation generic") by moving the PCI window region
reservation back into the dma specific path so that these regions
doesn't get exposed via the IOMMU API interface. With this change,
the vfio interface will report only iommu specific reserved regions
to the user space.
Cc: Joerg Roedel <joro@8bytes.org>
Signed-off-by: Shameer Kolothum <shameerali.kolothum.thodi@huawei.com>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Fixes: 273df9635385 ('iommu/dma: Make PCI window reservation generic')
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
iommu clocks are optional, so the driver should not fail if they are not
present. Instead just set the number of clocks to 0, which the clk-blk APIs
can handle just fine.
Fixes: f2e3a5f557ad ("iommu/rockchip: Control clocks needed to access the IOMMU")
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Reviewed-by: Robin Murphy <robin.murphy@arm.com>
Tested-by: Enric Balletbo i Serra <enric.balletbo@collabora.com>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
The newly introduced lock is only used when CONFIG_IRQ_REMAP is enabled:
drivers/iommu/amd_iommu.c:86:24: error: 'iommu_table_lock' defined but not used [-Werror=unused-variable]
static DEFINE_SPINLOCK(iommu_table_lock);
This moves the definition next to the user, within the #ifdef protected
section of the file.
Fixes: ea6166f4b83e ("iommu/amd: Split irq_lookup_table out of the amd_iommu_devtable_lock")
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
It was noticed that the IRTE configured for guest OS kernel
was over-written while the guest was running. As a result,
vt-d Posted Interrupts configured for the guest are not being
delivered directly, and instead bounces off the host. Every
interrupt delivery takes a VM Exit.
It was noticed that the following stack is doing the over-write:
[ 147.463177] modify_irte+0x171/0x1f0
[ 147.463405] intel_ir_set_affinity+0x5c/0x80
[ 147.463641] msi_domain_set_affinity+0x32/0x90
[ 147.463881] irq_do_set_affinity+0x37/0xd0
[ 147.464125] irq_set_affinity_locked+0x9d/0xb0
[ 147.464374] __irq_set_affinity+0x42/0x70
[ 147.464627] write_irq_affinity.isra.5+0xe1/0x110
[ 147.464895] proc_reg_write+0x38/0x70
[ 147.465150] __vfs_write+0x36/0x180
[ 147.465408] ? handle_mm_fault+0xdf/0x200
[ 147.465671] ? _cond_resched+0x15/0x30
[ 147.465936] vfs_write+0xad/0x1a0
[ 147.466204] SyS_write+0x52/0xc0
[ 147.466472] do_syscall_64+0x74/0x1a0
[ 147.466744] entry_SYSCALL_64_after_hwframe+0x3d/0xa2
reversing the sense of force check in intel_ir_reconfigure_irte()
restores proper posted interrupt functionality
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Fixes: d491bdff888e ('iommu/vt-d: Reevaluate vector configuration on activate()')
Signed-off-by: Joerg Roedel <jroedel@suse.de>
|
|
When speculation flaw mitigations are opt-in (via prctl), using seccomp
will automatically opt-in to these protections, since using seccomp
indicates at least some level of sandboxing is desired.
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
|
|
As done with seccomp and no_new_privs, also show speculation flaw
mitigation state in /proc/$pid/status.
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
|
|
Adjust arch_prctl_get/set_spec_ctrl() to operate on tasks other than
current.
This is needed both for /proc/$pid/status queries and for seccomp (since
thread-syncing can trigger seccomp in non-current threads).
Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
|
|
Add prctl based control for Speculative Store Bypass mitigation and make it
the default mitigation for Intel and AMD.
Andi Kleen provided the following rationale (slightly redacted):
There are multiple levels of impact of Speculative Store Bypass:
1) JITed sandbox.
It cannot invoke system calls, but can do PRIME+PROBE and may have call
interfaces to other code
2) Native code process.
No protection inside the process at this level.
3) Kernel.
4) Between processes.
The prctl tries to protect against case (1) doing attacks.
If the untrusted code can do random system calls then control is already
lost in a much worse way. So there needs to be system call protection in
some way (using a JIT not allowing them or seccomp). Or rather if the
process can subvert its environment somehow to do the prctl it can already
execute arbitrary code, which is much worse than SSB.
To put it differently, the point of the prctl is to not allow JITed code
to read data it shouldn't read from its JITed sandbox. If it already has
escaped its sandbox then it can already read everything it wants in its
address space, and do much worse.
The ability to control Speculative Store Bypass allows to enable the
protection selectively without affecting overall system performance.
Based on an initial patch from Tim Chen. Completely rewritten.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
|
|
The Speculative Store Bypass vulnerability can be mitigated with the
Reduced Data Speculation (RDS) feature. To allow finer grained control of
this eventually expensive mitigation a per task mitigation control is
required.
Add a new TIF_RDS flag and put it into the group of TIF flags which are
evaluated for mismatch in switch_to(). If these bits differ in the previous
and the next task, then the slow path function __switch_to_xtra() is
invoked. Implement the TIF_RDS dependent mitigation control in the slow
path.
If the prctl for controlling Speculative Store Bypass is disabled or no
task uses the prctl then there is no overhead in the switch_to() fast
path.
Update the KVM related speculation control functions to take TID_RDS into
account as well.
Based on a patch from Tim Chen. Completely rewritten.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
|
|
Add two new prctls to control aspects of speculation related vulnerabilites
and their mitigations to provide finer grained control over performance
impacting mitigations.
PR_GET_SPECULATION_CTRL returns the state of the speculation misfeature
which is selected with arg2 of prctl(2). The return value uses bit 0-2 with
the following meaning:
Bit Define Description
0 PR_SPEC_PRCTL Mitigation can be controlled per task by
PR_SET_SPECULATION_CTRL
1 PR_SPEC_ENABLE The speculation feature is enabled, mitigation is
disabled
2 PR_SPEC_DISABLE The speculation feature is disabled, mitigation is
enabled
If all bits are 0 the CPU is not affected by the speculation misfeature.
If PR_SPEC_PRCTL is set, then the per task control of the mitigation is
available. If not set, prctl(PR_SET_SPECULATION_CTRL) for the speculation
misfeature will fail.
PR_SET_SPECULATION_CTRL allows to control the speculation misfeature, which
is selected by arg2 of prctl(2) per task. arg3 is used to hand in the
control value, i.e. either PR_SPEC_ENABLE or PR_SPEC_DISABLE.
The common return values are:
EINVAL prctl is not implemented by the architecture or the unused prctl()
arguments are not 0
ENODEV arg2 is selecting a not supported speculation misfeature
PR_SET_SPECULATION_CTRL has these additional return values:
ERANGE arg3 is incorrect, i.e. it's not either PR_SPEC_ENABLE or PR_SPEC_DISABLE
ENXIO prctl control of the selected speculation misfeature is disabled
The first supported controlable speculation misfeature is
PR_SPEC_STORE_BYPASS. Add the define so this can be shared between
architectures.
Based on an initial patch from Tim Chen and mostly rewritten.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
|
|
Having everything in nospec-branch.h creates a hell of dependencies when
adding the prctl based switching mechanism. Move everything which is not
required in nospec-branch.h to spec-ctrl.h and fix up the includes in the
relevant files.
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Expose the CPUID.7.EDX[31] bit to the guest, and also guard against various
combinations of SPEC_CTRL MSR values.
The handling of the MSR (to take into account the host value of SPEC_CTRL
Bit(2)) is taken care of in patch:
KVM/SVM/VMX/x86/spectre_v2: Support the combination of guest and host IBRS
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
AMD does not need the Speculative Store Bypass mitigation to be enabled.
The parameters for this are already available and can be done via MSR
C001_1020. Each family uses a different bit in that MSR for this.
[ tglx: Expose the bit mask via a variable and move the actual MSR fiddling
into the bugs code as that's the right thing to do and also required
to prepare for dynamic enable/disable ]
Suggested-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Intel and AMD SPEC_CTRL (0x48) MSR semantics may differ in the
future (or in fact use different MSRs for the same functionality).
As such a run-time mechanism is required to whitelist the appropriate MSR
values.
[ tglx: Made the variable __ro_after_init ]
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Intel CPUs expose methods to:
- Detect whether RDS capability is available via CPUID.7.0.EDX[31],
- The SPEC_CTRL MSR(0x48), bit 2 set to enable RDS.
- MSR_IA32_ARCH_CAPABILITIES, Bit(4) no need to enable RRS.
With that in mind if spec_store_bypass_disable=[auto,on] is selected set at
boot-time the SPEC_CTRL MSR to enable RDS if the platform requires it.
Note that this does not fix the KVM case where the SPEC_CTRL is exposed to
guests which can muck with it, see patch titled :
KVM/SVM/VMX/x86/spectre_v2: Support the combination of guest and host IBRS.
And for the firmware (IBRS to be set), see patch titled:
x86/spectre_v2: Read SPEC_CTRL MSR during boot and re-use reserved bits
[ tglx: Distangled it from the intel implementation and kept the call order ]
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Contemporary high performance processors use a common industry-wide
optimization known as "Speculative Store Bypass" in which loads from
addresses to which a recent store has occurred may (speculatively) see an
older value. Intel refers to this feature as "Memory Disambiguation" which
is part of their "Smart Memory Access" capability.
Memory Disambiguation can expose a cache side-channel attack against such
speculatively read values. An attacker can create exploit code that allows
them to read memory outside of a sandbox environment (for example,
malicious JavaScript in a web page), or to perform more complex attacks
against code running within the same privilege level, e.g. via the stack.
As a first step to mitigate against such attacks, provide two boot command
line control knobs:
nospec_store_bypass_disable
spec_store_bypass_disable=[off,auto,on]
By default affected x86 processors will power on with Speculative
Store Bypass enabled. Hence the provided kernel parameters are written
from the point of view of whether to enable a mitigation or not.
The parameters are as follows:
- auto - Kernel detects whether your CPU model contains an implementation
of Speculative Store Bypass and picks the most appropriate
mitigation.
- on - disable Speculative Store Bypass
- off - enable Speculative Store Bypass
[ tglx: Reordered the checks so that the whole evaluation is not done
when the CPU does not support RDS ]
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Add the CPU feature bit CPUID.7.0.EDX[31] which indicates whether the CPU
supports Reduced Data Speculation.
[ tglx: Split it out from a later patch ]
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Add the sysfs file for the new vulerability. It does not do much except
show the words 'Vulnerable' for recent x86 cores.
Intel cores prior to family 6 are known not to be vulnerable, and so are
some Atoms and some Xeon Phi.
It assumes that older Cyrix, Centaur, etc. cores are immune.
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
A guest may modify the SPEC_CTRL MSR from the value used by the
kernel. Since the kernel doesn't use IBRS, this means a value of zero is
what is needed in the host.
But the 336996-Speculative-Execution-Side-Channel-Mitigations.pdf refers to
the other bits as reserved so the kernel should respect the boot time
SPEC_CTRL value and use that.
This allows to deal with future extensions to the SPEC_CTRL interface if
any at all.
Note: This uses wrmsrl() instead of native_wrmsl(). I does not make any
difference as paravirt will over-write the callq *0xfff.. with the wrmsrl
assembler code.
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
The 336996-Speculative-Execution-Side-Channel-Mitigations.pdf refers to all
the other bits as reserved. The Intel SDM glossary defines reserved as
implementation specific - aka unknown.
As such at bootup this must be taken it into account and proper masking for
the bits in use applied.
A copy of this document is available at
https://bugzilla.kernel.org/show_bug.cgi?id=199511
[ tglx: Made x86_spec_ctrl_base __ro_after_init ]
Suggested-by: Jon Masters <jcm@redhat.com>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Those SysFS functions have a similar preamble, as such make common
code to handle them.
Suggested-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Combine the various logic which goes through all those
x86_cpu_id matching structures in one function.
Suggested-by: Borislav Petkov <bp@suse.de>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Borislav Petkov <bp@suse.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
The macro is not type safe and I did look for why that "g" constraint for
the asm doesn't work: it's because the asm is more fundamentally wrong.
It does
movl %[val], %%eax
but "val" isn't a 32-bit value, so then gcc will pass it in a register,
and generate code like
movl %rsi, %eax
and gas will complain about a nonsensical 'mov' instruction (it's moving a
64-bit register to a 32-bit one).
Passing it through memory will just hide the real bug - gcc still thinks
the memory location is 64-bit, but the "movl" will only load the first 32
bits and it all happens to work because x86 is little-endian.
Convert it to a type safe inline function with a little trick which hands
the feature into the ALTERNATIVE macro.
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
|
|
Need to configure PHY interrupt as active low for P3310 Tegra186
platform otherwise it results in spurious interrupts.
This issue wasn't seen before because the generic PHY driver without
interrupt support was used.
Signed-off-by: Bhadram Varka <vbhadram@nvidia.com>
Signed-off-by: Thierry Reding <treding@nvidia.com>
|
|
Even with the wait-loop fixed, there is a further issue with
kthread_parkme(). Upon hotplug, when we do takedown_cpu(),
smpboot_park_threads() can return before all those threads are in fact
blocked, due to the placement of the complete() in __kthread_parkme().
When that happens, sched_cpu_dying() -> migrate_tasks() can end up
migrating such a still runnable task onto another CPU.
Normally the task will have hit schedule() and gone to sleep by the
time we do kthread_unpark(), which will then do __kthread_bind() to
re-bind the task to the correct CPU.
However, when we loose the initial TASK_PARKED store to the concurrent
wakeup issue described previously, do the complete(), get migrated, it
is possible to either:
- observe kthread_unpark()'s clearing of SHOULD_PARK and terminate
the park and set TASK_RUNNING, or
- __kthread_bind()'s wait_task_inactive() to observe the competing
TASK_RUNNING store.
Either way the WARN() in __kthread_bind() will trigger and fail to
correctly set the CPU affinity.
Fix this by only issuing the complete() when the kthread has scheduled
out. This does away with all the icky 'still running' nonsense.
The alternative is to promote TASK_PARKED to a special state, this
guarantees wait_task_inactive() cannot observe a 'stale' TASK_RUNNING
and we'll end up doing the right thing, but this preserves the whole
icky business of potentially migating the still runnable thing.
Reported-by: Gaurav Kohli <gkohli@codeaurora.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Gaurav reported a problem with __kthread_parkme() where a concurrent
try_to_wake_up() could result in competing stores to ->state which,
when the TASK_PARKED store got lost bad things would happen.
The comment near set_current_state() actually mentions this competing
store, but only mentions the case against TASK_RUNNING. This same
store, with different timing, can happen against a subsequent !RUNNING
store.
This normally is not a problem, because as per that same comment, the
!RUNNING state store is inside a condition based wait-loop:
for (;;) {
set_current_state(TASK_UNINTERRUPTIBLE);
if (!need_sleep)
break;
schedule();
}
__set_current_state(TASK_RUNNING);
If we loose the (first) TASK_UNINTERRUPTIBLE store to a previous
(concurrent) wakeup, the schedule() will NO-OP and we'll go around the
loop once more.
The problem here is that the TASK_PARKED store is not inside the
KTHREAD_SHOULD_PARK condition wait-loop.
There is a genuine issue with sleeps that do not have a condition;
this is addressed in a subsequent patch.
Reported-by: Gaurav Kohli <gkohli@codeaurora.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Oleg Nesterov <oleg@redhat.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
With commit:
31e77c93e432 ("sched/fair: Update blocked load when newly idle")
... we release the rq->lock when updating blocked load of idle CPUs.
This opens a time window during which another CPU can add a task to this
CPU's cfs_rq.
The check for newly added task of idle_balance() is not in the common path.
Move the out label to include this check.
Reported-by: Heiner Kallweit <hkallweit1@gmail.com>
Tested-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Paul E. McKenney <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 31e77c93e432 ("sched/fair: Update blocked load when newly idle")
Link: http://lkml.kernel.org/r/20180426103133.GA6953@linaro.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
Matt reported the following deadlock:
CPU0 CPU1
schedule(.prev=migrate/0) <fault>
pick_next_task() ...
idle_balance() migrate_swap()
active_balance() stop_two_cpus()
spin_lock(stopper0->lock)
spin_lock(stopper1->lock)
ttwu(migrate/0)
smp_cond_load_acquire() -- waits for schedule()
stop_one_cpu(1)
spin_lock(stopper1->lock) -- waits for stopper lock
Fix this deadlock by taking the wakeups out from under stopper->lock.
This allows the active_balance() to queue the stop work and finish the
context switch, which in turn allows the wakeup from migrate_swap() to
observe the context and complete the wakeup.
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reported-by: Matt Fleming <matt@codeblueprint.co.uk>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Matt Fleming <matt@codeblueprint.co.uk>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mike Galbraith <umgwanakikbuti@gmail.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/20180420095005.GH4064@hirez.programming.kicks-ass.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace
Pull tracing fixes from Steven Rostedt:
"Various fixes in tracing:
- Tracepoints should not give warning on OOM failures
- Use special field for function pointer in trace event
- Fix igrab issues in uprobes
- Fixes to the new histogram triggers"
* tag 'trace-v4.17-rc1-2' of git://git.kernel.org/pub/scm/linux/kernel/git/rostedt/linux-trace:
tracepoint: Do not warn on ENOMEM
tracing: Add field modifier parsing hist error for hist triggers
tracing: Add field parsing hist error for hist triggers
tracing: Restore proper field flag printing when displaying triggers
tracing: initcall: Ordered comparison of function pointers
tracing: Remove igrab() iput() call from uprobes.c
tracing: Fix bad use of igrab in trace_uprobe.c
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input
Pull input updates from Dmitry Torokhov:
"Just a few driver fixes"
* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/dtor/input:
Input: atmel_mxt_ts - add missing compatible strings to OF device table
Input: atmel_mxt_ts - fix the firmware update
Input: atmel_mxt_ts - add touchpad button mapping for Samsung Chromebook Pro
MAINTAINERS: Rakesh Iyer can't be reached anymore
Input: hideep_ts - fix a typo in Kconfig
Input: alps - fix reporting pressure of v3 trackstick
Input: leds - fix out of bound access
Input: synaptics-rmi4 - fix an unchecked out of memory error path
|
|
Allow some non-cached routes to use non-expired fnhe:
1. ip_del_fnhe: moved above and now called by find_exception.
The 4.5+ commit deed49df7390 expires fnhe only when caching
routes. Change that to:
1.1. use fnhe for non-cached local output routes, with the help
from (2)
1.2. allow __mkroute_input to detect expired fnhe (outdated
fnhe_gw, for example) when do_cache is false, eg. when itag!=0
for unicast destinations.
2. __mkroute_output: keep fi to allow local routes with orig_oif != 0
to use fnhe info even when the new route will not be cached into fnhe.
After commit 839da4d98960 ("net: ipv4: set orig_oif based on fib
result for local traffic") it means all local routes will be affected
because they are not cached. This change is used to solve a PMTU
problem with IPVS (and probably Netfilter DNAT) setups that redirect
local clients from target local IP (local route to Virtual IP)
to new remote IP target, eg. IPVS TUN real server. Loopback has
64K MTU and we need to create fnhe on the local route that will
keep the reduced PMTU for the Virtual IP. Without this change
fnhe_pmtu is updated from ICMP but never exposed to non-cached
local routes. This includes routes with flowi4_oif!=0 for 4.6+ and
with flowi4_oif=any for 4.14+).
3. update_or_create_fnhe: make sure fnhe_expires is not 0 for
new entries
Fixes: 839da4d98960 ("net: ipv4: set orig_oif based on fib result for local traffic")
Fixes: d6d5e999e5df ("route: do not cache fib route info on local routes with oif")
Fixes: deed49df7390 ("route: check and remove route cache when we get route")
Cc: David Ahern <dsahern@gmail.com>
Cc: Xin Long <lucien.xin@gmail.com>
Signed-off-by: Julian Anastasov <ja@ssi.bg>
Acked-by: David Ahern <dsahern@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi
Pull SCSI fixes from James Bottomley:
"Three small bug fixes: an illegally overlapping memcmp in target code,
a potential infinite loop in isci under certain rare phy conditions
and an ATA queue depth (performance) correction for storvsc"
* tag 'scsi-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/jejb/scsi:
scsi: target: Fix fortify_panic kernel exception
scsi: isci: Fix infinite loop in while loop
scsi: storvsc: Set up correct queue depth values for IDE devices
|
|
git://anongit.freedesktop.org/drm/drm-misc into drm-fixes
vc4: Fix bo refcounts during async commits (Boris)
vga-dac: Fix edid memory leak (Sean)
Cc: Boris Brezillon <boris.brezillon@bootlin.com>
Cc: Sean Paul <seanpaul@chromium.org>
* tag 'drm-misc-fixes-2018-05-02' of git://anongit.freedesktop.org/drm/drm-misc:
drm/bridge: vga-dac: Fix edid memory leak
drm/vc4: Make sure vc4_bo_{inc,dec}_usecnt() calls are balanced
|
|
git://anongit.freedesktop.org/drm/drm-intel into drm-fixes
Add DMC firmware for Geminilake.
* tag 'drm-intel-fixes-2018-05-02' of git://anongit.freedesktop.org/drm/drm-intel:
drm/i915/glk: Add MODULE_FIRMWARE for Geminilake
|
|
Daniel Borkmann says:
====================
pull-request: bpf 2018-05-03
The following pull-request contains BPF updates for your *net* tree.
The main changes are:
1) Several BPF sockmap fixes mostly related to bugs in error path
handling, that is, a bug in updating the scatterlist length /
offset accounting, a missing sk_mem_uncharge() in redirect
error handling, and a bug where the outstanding bytes counter
sg_size was not zeroed, from John.
2) Fix two memory leaks in the x86-64 BPF JIT, one in an error
path where we still don't converge after image was allocated
and another one where BPF calls are used and JIT passes don't
converge, from Daniel.
3) Minor fix in BPF selftests where in test_stacktrace_build_id()
we drop useless args in urandom_read and we need to add a missing
newline in a CHECK() error message, from Song.
====================
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
John Fastabend says:
====================
When I added the test_sockmap to selftests I mistakenly changed the
test logic a bit. The result of this was on redirect cases we ended up
choosing the wrong sock from the BPF program and ended up sending to a
socket that had no receive handler. The result was the actual receive
handler, running on a different socket, is timing out and closing the
socket. This results in errors (-EPIPE to be specific) on the sending
side. Typically happening if the sender does not complete the send
before the receive side times out. So depending on timing and the size
of the send we may get errors. This exposed some bugs in the sockmap
error path handling.
This series fixes the errors. The primary issue is we did not do proper
memory accounting in these cases which resulted in missing a
sk_mem_uncharge(). This happened in the redirect path and in one case
on the normal send path. See the three patches for the details.
The other take-away from this is we need to fix the test_sockmap and
also add more negative test cases. That will happen in bpf-next.
Finally, I tested this using the existing test_sockmap program, the
older sockmap sample test script, and a few real use cases with
Cilium. All of these seem to be in working correctly.
v2: fix compiler warning, drop iterator variable 'i' that is no longer
used in patch 3.
====================
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When a redirect failure happens we release the buffers in-flight
without calling a sk_mem_uncharge(), the uncharge is called before
dropping the sock lock for the redirecte, however we missed updating
the ring start index. When no apply actions are in progress this
is OK because we uncharge the entire buffer before the redirect.
But, when we have apply logic running its possible that only a
portion of the buffer is being redirected. In this case we only
do memory accounting for the buffer slice being redirected and
expect to be able to loop over the BPF program again and/or if
a sock is closed uncharge the memory at sock destruct time.
With an invalid start index however the program logic looks at
the start pointer index, checks the length, and when seeing the
length is zero (from the initial release and failure to update
the pointer) aborts without uncharging/releasing the remaining
memory.
The fix for this is simply to update the start index. To avoid
fixing this error in two locations we do a small refactor and
remove one case where it is open-coded. Then fix it in the
single function.
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When an error occurs during a redirect we have two cases that need
to be handled (i) we have a cork'ed buffer (ii) we have a normal
sendmsg buffer.
In the cork'ed buffer case we don't currently support recovering from
errors in a redirect action. So the buffer is released and the error
should _not_ be pushed back to the caller of sendmsg/sendpage. The
rationale here is the user will get an error that relates to old
data that may have been sent by some arbitrary thread on that sock.
Instead we simple consume the data and tell the user that the data
has been consumed. We may add proper error recovery in the future.
However, this patch fixes a bug where the bytes outstanding counter
sg_size was not zeroed. This could result in a case where if the user
has both a cork'ed action and apply action in progress we may
incorrectly call into the BPF program when the user expected an
old verdict to be applied via the apply action. I don't have a use
case where using apply and cork at the same time is valid but we
never explicitly reject it because it should work fine. This patch
ensures the sg_size is zeroed so we don't have this case.
In the normal sendmsg buffer case (no cork data) we also do not
zero sg_size. Again this can confuse the apply logic when the logic
calls into the BPF program when the BPF programmer expected the old
verdict to remain. So ensure we set sg_size to zero here as well. And
additionally to keep the psock state in-sync with the sk_msg_buff
release all the memory as well. Previously we did this before
returning to the user but this left a gap where psock and sk_msg_buff
states were out of sync which seems fragile. No additional overhead
is taken here except for a call to check the length and realize its
already been freed. This is in the error path as well so in my
opinion lets have robust code over optimized error paths.
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
When the call to do_tcp_sendpage() fails to send the complete block
requested we either retry if only a partial send was completed or
abort if we receive a error less than or equal to zero. Before
returning though we must update the scatterlist length/offset to
account for any partial send completed.
Before this patch we did this at the end of the retry loop, but
this was buggy when used while applying a verdict to fewer bytes
than in the scatterlist. When the scatterlist length was being set
we forgot to account for the apply logic reducing the size variable.
So the result was we chopped off some bytes in the scatterlist without
doing proper cleanup on them. This results in a WARNING when the
sock is tore down because the bytes have previously been charged to
the socket but are never uncharged.
The simple fix is to simply do the accounting inside the retry loop
subtracting from the absolute scatterlist values rather than trying
to accumulate the totals and subtract at the end.
Reported-by: Alexei Starovoitov <ast@kernel.org>
Signed-off-by: John Fastabend <john.fastabend@gmail.com>
Signed-off-by: Alexei Starovoitov <ast@kernel.org>
|
|
Normally, a socket can not be freed/reused unless all its TX packets
left qdisc and were TX-completed. However connect(AF_UNSPEC) allows
this to happen.
With commit fc59d5bdf1e3 ("pkt_sched: fq: clear time_next_packet for
reused flows") we cleared f->time_next_packet but took no special
action if the flow was still in the throttled rb-tree.
Since f->time_next_packet is the key used in the rb-tree searches,
blindly clearing it might break rb-tree integrity. We need to make
sure the flow is no longer in the rb-tree to avoid this problem.
Fixes: fc59d5bdf1e3 ("pkt_sched: fq: clear time_next_packet for reused flows")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
This reverts commit edd7ceb78296 ("ipv6: Allow non-gateway ECMP for
IPv6").
Eric reported a division by zero in rt6_multipath_rebalance() which is
caused by above commit that considers identical local routes to be
siblings. The division by zero happens because a nexthop weight is not
set for local routes.
Revert the commit as it does not fix a bug and has side effects.
To reproduce:
# ip -6 address add 2001:db8::1/64 dev dummy0
# ip -6 address add 2001:db8::1/64 dev dummy1
Fixes: edd7ceb78296 ("ipv6: Allow non-gateway ECMP for IPv6")
Signed-off-by: Ido Schimmel <idosch@mellanox.com>
Reported-by: Eric Dumazet <eric.dumazet@gmail.com>
Tested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
|
|
Fix three section mismatches:
1) Section mismatch in reference from the function ioread8() to the
function .init.text:pcibios_init_bridge()
2) Section mismatch in reference from the function free_initmem() to the
function .init.text:map_pages()
3) Section mismatch in reference from the function ccio_ioc_init() to
the function .init.text:count_parisc_driver()
Signed-off-by: Helge Deller <deller@gmx.de>
|
|
Fix two section mismatches in drivers.c:
1) Section mismatch in reference from the function alloc_tree_node() to
the function .init.text:create_tree_node().
2) Section mismatch in reference from the function walk_native_bus() to
the function .init.text:alloc_pa_dev().
Signed-off-by: Helge Deller <deller@gmx.de>
|
|
Make n signed to avoid leaking the pages array if __pipe_get_pages()
fails to allocate any pages.
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
