summaryrefslogtreecommitdiffstats
path: root/arch/x86
AgeCommit message (Collapse)Author
2014-04-17uprobes/x86: Emulate unconditional relative jmp'sOleg Nesterov
Currently we always execute all insns out-of-line, including relative jmp's and call's. This assumes that even if regs->ip points to nowhere after the single-step, default_post_xol_op(UPROBE_FIX_IP) logic will update it correctly. However, this doesn't work if this regs->ip == xol_vaddr + insn_offset is not canonical. In this case CPU generates #GP and general_protection() kills the task which tries to execute this insn out-of-line. Now that we have uprobe_xol_ops we can teach uprobes to emulate these insns and solve the problem. This patch adds branch_xol_ops which has a single branch_emulate_op() hook, so far it can only handle rel8/32 relative jmp's. TODO: move ->fixup into the union along with rip_rela_target_address. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reported-by: Jonathan Lebon <jlebon@redhat.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com>
2014-04-17uprobes/x86: Introduce sizeof_long(), cleanup adjust_ret_addr() and ↵Oleg Nesterov
arch_uretprobe_hijack_return_addr() 1. Add the trivial sizeof_long() helper and change other callers of is_ia32_task() to use it. TODO: is_ia32_task() is not what we actually want, TS_COMPAT does not necessarily mean 32bit. Fortunately syscall-like insns can't be probed so it actually works, but it would be better to rename and use is_ia32_frame(). 2. As Jim pointed out "ncopied" in arch_uretprobe_hijack_return_addr() and adjust_ret_addr() should be named "nleft". And in fact only the last copy_to_user() in arch_uretprobe_hijack_return_addr() actually needs to inspect the non-zero error code. TODO: adjust_ret_addr() should die. We can always calculate the value we need to write into *regs->sp, just UPROBE_FIX_CALL should record insn->length. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com>
2014-04-17uprobes/x86: Teach arch_uprobe_post_xol() to restart if possibleOleg Nesterov
SIGILL after the failed arch_uprobe_post_xol() should only be used as a last resort, we should try to restart the probed insn if possible. Currently only adjust_ret_addr() can fail, and this can only happen if another thread unmapped our stack after we executed "call" out-of-line. Most probably the application if buggy, but even in this case it can have a handler for SIGSEGV/etc. And in theory it can be even correct and do something non-trivial with its memory. Of course we can't restart unconditionally, so arch_uprobe_post_xol() does this only if ->post_xol() returns -ERESTART even if currently this is the only possible error. default_post_xol_op(UPROBE_FIX_CALL) can always restart, but as Jim pointed out it should not forget to pop off the return address pushed by this insn executed out-of-line. Note: this is not "perfect", we do not want the extra handler_chain() after restart, but I think this is the best solution we can realistically do without too much uglifications. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com>
2014-04-17uprobes/x86: Send SIGILL if arch_uprobe_post_xol() failsOleg Nesterov
Currently the error from arch_uprobe_post_xol() is silently ignored. This doesn't look good and this can lead to the hard-to-debug problems. 1. Change handle_singlestep() to loudly complain and send SIGILL. Note: this only affects x86, ppc/arm can't fail. 2. Change arch_uprobe_post_xol() to call arch_uprobe_abort_xol() and avoid TF games if it is going to return an error. This can help to to analyze the problem, if nothing else we should not report ->ip = xol_slot in the core-file. Note: this means that handle_riprel_post_xol() can be called twice, but this is fine because it is idempotent. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com>
2014-04-17uprobes/x86: Conditionalize the usage of handle_riprel_insn()Oleg Nesterov
arch_uprobe_analyze_insn() calls handle_riprel_insn() at the start, but only "0xff" and "default" cases need the UPROBE_FIX_RIP_ logic. Move the callsite into "default" case and change the "0xff" case to fall-through. We are going to add the various hooks to handle the rip-relative jmp/call instructions (and more), we need this change to enforce the fact that the new code can not conflict with is_riprel_insn() logic which, after this change, can only be used by default_xol_ops. Note: arch_uprobe_abort_xol() still calls handle_riprel_post_xol() directly. This is fine unless another _xol_ops we may add later will need to reuse "UPROBE_FIX_RIP_AX|UPROBE_FIX_RIP_CX" bits in ->fixup. In this case we can add uprobe_xol_ops->abort() hook, which (perhaps) we will need anyway in the long term. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
2014-04-17uprobes/x86: Introduce uprobe_xol_ops and arch_uprobe->opsOleg Nesterov
Introduce arch_uprobe->ops pointing to the "struct uprobe_xol_ops", move the current UPROBE_FIX_{RIP*,IP,CALL} code into the default set of methods and change arch_uprobe_pre/post_xol() accordingly. This way we can add the new uprobe_xol_ops's to handle the insns which need the special processing (rip-relative jmp/call at least). Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com>
2014-04-17uprobes/x86: move the UPROBE_FIX_{RIP,IP,CALL} code at the end of pre/post hooksOleg Nesterov
No functional changes. Preparation to simplify the review of the next change. Just reorder the code in arch_uprobe_pre/post_xol() functions so that UPROBE_FIX_{RIP_*,IP,CALL} logic goes to the end. Also change arch_uprobe_pre_xol() to use utask instead of autask, to make the code more symmetrical with arch_uprobe_post_xol(). Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
2014-04-17uprobes/x86: Gather "riprel" functions togetherOleg Nesterov
Cosmetic. Move pre_xol_rip_insn() and handle_riprel_post_xol() up to the closely related handle_riprel_insn(). This way it is simpler to read and understand this code, and this lessens the number of ifdef's. While at it, update the comment in handle_riprel_post_xol() as Jim suggested. TODO: rename them somehow to make the naming consistent. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com>
2014-04-17uprobes/x86: Kill the "ia32_compat" check in handle_riprel_insn(), remove ↵Oleg Nesterov
"mm" arg Kill the "mm->context.ia32_compat" check in handle_riprel_insn(), if it is true insn_rip_relative() must return false. validate_insn_bits() passed "ia32_compat" as !x86_64 to insn_init(), and insn_rip_relative() checks insn->x86_64. Also, remove the no longer needed "struct mm_struct *mm" argument and the unnecessary "return" at the end. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
2014-04-17uprobes/x86: Fold prepare_fixups() into arch_uprobe_analyze_insn()Oleg Nesterov
No functional changes, preparation. Shift the code from prepare_fixups() to arch_uprobe_analyze_insn() with the following modifications: - Do not call insn_get_opcode() again, it was already called by validate_insn_bits(). - Move "case 0xea" up. This way "case 0xff" can fall through to default case. - change "case 0xff" to use the nested "switch (MODRM_REG)", this way the code looks a bit simpler. - Make the comments look consistent. While at it, kill the initialization of rip_rela_target_address and ->fixups, we can rely on kzalloc(). We will add the new members into arch_uprobe, it would be better to assume that everything is zero by default. TODO: cleanup/fix the mess in validate_insn_bits() paths: - validate_insn_64bits() and validate_insn_32bits() should be unified. - "ifdef" is not used consistently; if good_insns_64 depends on CONFIG_X86_64, then probably good_insns_32 should depend on CONFIG_X86_32/EMULATION - the usage of mm->context.ia32_compat looks wrong if the task is TIF_X32. Signed-off-by: Oleg Nesterov <oleg@redhat.com> Reviewed-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Reviewed-by: Jim Keniston <jkenisto@us.ibm.com> Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
2014-04-17Merge tag 'stable/for-linus-3.15-rc1-tag' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip Pull Xen fixes from David Vrabel: "Xen regression and bug fixes for 3.15-rc1: - fix completely broken 32-bit PV guests caused by x86 refactoring 32-bit thread_info. - only enable ticketlock slow path on Xen (not bare metal) - fix two bugs with PV guests not shutting down when requested - fix a minor memory leak in xen-pciback error path" * tag 'stable/for-linus-3.15-rc1-tag' of git://git.kernel.org/pub/scm/linux/kernel/git/xen/tip: xen/manage: Poweroff forcefully if user-space is not yet up. xen/xenbus: Avoid synchronous wait on XenBus stalling shutdown/restart. xen/spinlock: Don't enable them unconditionally. xen-pciback: silence an unwanted debug printk xen: fix memory leak in __xen_pcibk_add_pci_dev() x86/xen: Fix 32-bit PV guests's usage of kernel_stack
2014-04-17KVM: VMX: speed up wildcard MMIO EVENTFDMichael S. Tsirkin
With KVM, MMIO is much slower than PIO, due to the need to do page walk and emulation. But with EPT, it does not have to be: we know the address from the VMCS so if the address is unique, we can look up the eventfd directly, bypassing emulation. Unfortunately, this only works if userspace does not need to match on access length and data. The implementation adds a separate FAST_MMIO bus internally. This serves two purposes: - minimize overhead for old userspace that does not use eventfd with lengtth = 0 - minimize disruption in other code (since we don't know the length, devices on the MMIO bus only get a valid address in write, this way we don't need to touch all devices to teach them to handle an invalid length) At the moment, this optimization only has effect for EPT on x86. It will be possible to speed up MMIO for NPT and MMU using the same idea in the future. With this patch applied, on VMX MMIO EVENTFD is essentially as fast as PIO. I was unable to detect any measureable slowdown to non-eventfd MMIO. Making MMIO faster is important for the upcoming virtio 1.0 which includes an MMIO signalling capability. The idea was suggested by Peter Anvin. Lots of thanks to Gleb for pre-review and suggestions. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-17KVM: support any-length wildcard ioeventfdMichael S. Tsirkin
It is sometimes benefitial to ignore IO size, and only match on address. In hindsight this would have been a better default than matching length when KVM_IOEVENTFD_FLAG_DATAMATCH is not set, In particular, this kind of access can be optimized on VMX: there no need to do page lookups. This can currently be done with many ioeventfds but in a suboptimal way. However we can't change kernel/userspace ABI without risk of breaking some applications. Use len = 0 to mean "ignore length for matching" in a more optimal way. Signed-off-by: Michael S. Tsirkin <mst@redhat.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-17x86/efi: Save and restore FPU context around efi_calls (i386)Ricardo Neri
Do a complete FPU context save/restore around the EFI calls. This required as runtime EFI firmware may potentially use the FPU. This change covers only the i386 configuration. Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com> Cc: Borislav Petkov <bp@suse.de> Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2014-04-17x86/efi: Save and restore FPU context around efi_calls (x86_64)Ricardo Neri
Do a complete FPU context save/restore around the EFI calls. This required as runtime EFI firmware may potentially use the FPU. This change covers only the x86_64 configuration. Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com> Cc: Borislav Petkov <bp@suse.de> Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2014-04-17x86/efi: Implement a __efi_call_virt macroRicardo Neri
For i386, all the EFI system runtime services functions return efi_status_t except efi_reset_system_system. Therefore, not all functions can be covered by the same macro in case the macro needs to do more than calling the function (i.e., return a value). The purpose of the __efi_call_virt macro is to be used when no return value is expected. For x86_64, this macro would not be needed as all the runtime services return u64. However, the same code is used for both x86_64 and i386. Thus, the macro __efi_call_virt is also defined to not break compilation. Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com> Cc: Borislav Petkov <bp@suse.de> Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2014-04-17x86, fpu: Extend the use of static_cpu_has_safeMatt Fleming
It may be necessary to save and restore the FPU context during EFI runtime system services calls. However, this may happen during boot and before alternatives have run. Thus, we need to use static_cpu_has_safe instead. The rationale behind the use of static_cpu_has_safe is the same as in commit 5f8c4218148822fde6ee ("x86, fpu: Use static_cpu_has_safe before alternatives") by Borislav Petkov. Signed-off-by: Matt Fleming <matt.fleming@intel.com> Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com> Cc: Borislav Petkov <bp@suse.de>
2014-04-17x86/efi: Delete most of the efi_call* macrosMatt Fleming
We really only need one phys and one virt function call, and then only one assembly function to make firmware calls. Since we are not using the C type system anyway, we're not really losing much by deleting the macros apart from no longer having a check that we are passing the correct number of parameters. The lack of duplicated code seems like a worthwhile trade-off. Cc: Ricardo Neri <ricardo.neri-calderon@linux.intel.com> Cc: Borislav Petkov <bp@suse.de> Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2014-04-17efi: x86: Handle arbitrary Unicode charactersH. Peter Anvin
Instead of truncating UTF-16 assuming all characters is ASCII, properly convert it to UTF-8. Signed-off-by: H. Peter Anvin <hpa@linux.intel.com> [ Bug and style fixes. ] Signed-off-by: Roy Franz <roy.franz@linaro.org> Signed-off-by: Leif Lindholm <leif.lindholm@linaro.org> Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2014-04-17kprobes/x86: Fix page-fault handling logicMasami Hiramatsu
Current kprobes in-kernel page fault handler doesn't expect that its single-stepping can be interrupted by an NMI handler which may cause a page fault(e.g. perf with callback tracing). In that case, the page-fault handled by kprobes and it misunderstands the page-fault has been caused by the single-stepping code and tries to recover IP address to probed address. But the truth is the page-fault has been caused by the NMI handler, and do_page_fault failes to handle real page fault because the IP address is modified and causes Kernel BUGs like below. ---- [ 2264.726905] BUG: unable to handle kernel NULL pointer dereference at 0000000000000020 [ 2264.727190] IP: [<ffffffff813c46e0>] copy_user_generic_string+0x0/0x40 To handle this correctly, I fixed the kprobes fault handler to ensure the faulted ip address is its own single-step buffer instead of checking current kprobe state. Signed-off-by: Masami Hiramatsu <masami.hiramatsu.pt@hitachi.com> Cc: Andi Kleen <andi@firstfloor.org> Cc: Ananth N Mavinakayanahalli <ananth@in.ibm.com> Cc: Sandeepa Prabhu <sandeepa.prabhu@linaro.org> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: fche@redhat.com Cc: systemtap@sourceware.org Link: http://lkml.kernel.org/r/20140417081644.26341.52351.stgit@ltc230.yrl.intra.hitachi.co.jp Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-17x86/mce: Fix CMCI preemption bugsIngo Molnar
The following commit: 27f6c573e0f7 ("x86, CMCI: Add proper detection of end of CMCI storms") Added two preemption bugs: - machine_check_poll() does a get_cpu_var() without a matching put_cpu_var(), which causes preemption imbalance and crashes upon bootup. - it does percpu ops without disabling preemption. Preemption is not disabled due to the mistaken use of a raw spinlock. To fix these bugs fix the imbalance and change cmci_discover_lock to a regular spinlock. Reported-by: Owen Kibel <qmewlo@gmail.com> Reported-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Ingo Molnar <mingo@kernel.org> Cc: Chen, Gong <gong.chen@linux.intel.com> Cc: Josh Boyer <jwboyer@fedoraproject.org> Cc: Tony Luck <tony.luck@intel.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Alexander Todorov <atodorov@redhat.com> Cc: Borislav Petkov <bp@alien8.de> Link: http://lkml.kernel.org/n/tip-jtjptvgigpfkpvtQxpEk1at2@git.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org> -- arch/x86/kernel/cpu/mcheck/mce.c | 4 +--- arch/x86/kernel/cpu/mcheck/mce_intel.c | 18 +++++++++--------- 2 files changed, 10 insertions(+), 12 deletions(-)
2014-04-16Merge branch 'x86-urgent-for-linus' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip Pull x86 fixes from Ingo Molnar: "Various fixes: - reboot regression fix - build message spam fix - GPU quirk fix - 'make kvmconfig' fix plus the wire-up of the renameat2() system call on i386" * 'x86-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86: Remove the PCI reboot method from the default chain x86/build: Supress "Nothing to be done for ..." messages x86/gpu: Fix sign extension issue in Intel graphics stolen memory quirks x86/platform: Fix "make O=dir kvmconfig" i386: Wire up the renameat2() syscall
2014-04-16Merge branch 'perf-urgent-for-linus' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip Pull perf fixes from Ingo Molnar: "Tooling fixes, plus a simple hardware-enablement patch for the Intel RAPL PMU (energy use measurement) on Haswell CPUs, which I hope is still fine at this stage" * 'perf-urgent-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: perf tools: Instead of redirecting flex output, use -o perf tools: Fix double free in perf test 21 (code-reading.c) perf stat: Initialize statistics correctly perf bench: Set more defaults in the 'numa' suite perf bench: Fix segfault at the end of an 'all' execution perf bench: Update manpage to mention numa and futex perf probe: Use dwarf_getcfi_elf() instead of dwarf_getcfi() perf probe: Fix to handle errors in line_range searching perf probe: Fix --line option behavior perf tools: Pick up libdw without explicit LIBDW_DIR MAINTAINERS: Change e-mail to kernel.org one perf callchains: Disable unwind libraries when libelf isn't found tools lib traceevent: Do not call warning() directly tools lib traceevent: Print event name when show warning if possible perf top: Fix documentation of invalid -s option perf/x86: Enable DRAM RAPL support on Intel Haswell
2014-04-16KVM: x86: Fix page-tables reserved bitsNadav Amit
KVM does not handle the reserved bits of x86 page tables correctly: In PAE, bits 5:8 are reserved in the PDPTE. In IA-32e, bit 8 is not reserved. Signed-off-by: Nadav Amit <namit@cs.technion.ac.il> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-16x86/build: Supress realmode.bin is up to date messagePeter Foley
Supress this unnecessary message during kernel re-build: make[3]: 'arch/x86/realmode/rm/realmode.bin' is up to date. Signed-off-by: Peter Foley <pefoley2@pefoley.com> Link: http://lkml.kernel.org/r/1397584693-15902-1-git-send-email-pefoley2@pefoley.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-16Merge commit 'b13b1d2d8692' into x86/mmIngo Molnar
It got into x86/urgent but isn't really urgent material. Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-16x86/irq: Fix fixup_irqs() error handlingPrarit Bhargava
Several patches to fix cpu hotplug and the down'd cpu's irq relocations have been submitted in the past month or so. The patches should resolve the problems with cpu hotplug and irq relocation, however, there is always a possibility that a bug still exists. The big problem with debugging these irq reassignments is that the cpu down completes and then we get random stack traces from drivers for which irqs have not been properly assigned to a new cpu. The stack traces are a mix of storage, network, and other kernel subsystem (I once saw the serial port stop working ...) warnings and failures. The problem with these failures is that they are difficult to diagnose. There is no warning in the cpu hotplug down path to indicate that an IRQ has failed to be assigned to a new cpu, and all we are left with is a stack trace from a driver, or a non-functional device. If we had some information on the console debugging these situations would be much easier; after all we can map an IRQ to a device by simply using lspci or /proc/interrupts. The current code, fixup_irqs(), which migrates IRQs from the down'd cpu and is called close to the end of the cpu down path, calls chip->set_irq_affinity which eventually calls __assign_irq_vector(). Errors are not propogated back from this function call and this results in silent irq relocation failures. This patch fixes this issue by returning the error codes up the call stack and prints out a warning if there is a relocation failure. Signed-off-by: Prarit Bhargava <prarit@redhat.com> Acked-by: Thomas Gleixner <tglx@linutronix.de> Cc: Rui Wang <rui.y.wang@intel.com> Cc: Liu Ping Fan <kernelfans@gmail.com> Cc: Bjorn Helgaas <bhelgaas@google.com> Cc: Yoshihiro YUNOMAE <yoshihiro.yunomae.ez@hitachi.com> Cc: Lv Zheng <lv.zheng@intel.com> Cc: Seiji Aguchi <seiji.aguchi@hds.com> Cc: Yang Zhang <yang.z.zhang@intel.com> Cc: Andi Kleen <ak@linux.intel.com> Cc: Steven Rostedt (Red Hat) <rostedt@goodmis.org> Cc: Li Fei <fei.li@intel.com> Cc: gong.chen@linux.intel.com Link: http://lkml.kernel.org/r/1396440673-18286-1-git-send-email-prarit@redhat.com [ Made small cleanliness tweaks. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-16Merge branch 'x86/apic' into x86/irq, to consolidate branches.Ingo Molnar
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-16x86/mm: In the PTE swapout page reclaim case clear the accessed bit instead ↵Shaohua Li
of flushing the TLB We use the accessed bit to age a page at page reclaim time, and currently we also flush the TLB when doing so. But in some workloads TLB flush overhead is very heavy. In my simple multithreaded app with a lot of swap to several pcie SSDs, removing the tlb flush gives about 20% ~ 30% swapout speedup. Fortunately just removing the TLB flush is a valid optimization: on x86 CPUs, clearing the accessed bit without a TLB flush doesn't cause data corruption. It could cause incorrect page aging and the (mistaken) reclaim of hot pages, but the chance of that should be relatively low. So as a performance optimization don't flush the TLB when clearing the accessed bit, it will eventually be flushed by a context switch or a VM operation anyway. [ In the rare event of it not getting flushed for a long time the delay shouldn't really matter because there's no real memory pressure for swapout to react to. ] Suggested-by: Linus Torvalds <torvalds@linux-foundation.org> Signed-off-by: Shaohua Li <shli@fusionio.com> Acked-by: Rik van Riel <riel@redhat.com> Acked-by: Mel Gorman <mgorman@suse.de> Acked-by: Hugh Dickins <hughd@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: linux-mm@kvack.org Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Link: http://lkml.kernel.org/r/20140408075809.GA1764@kernel.org [ Rewrote the changelog and the code comments. ] Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-16x86: Remove the PCI reboot method from the default chainIngo Molnar
Steve reported a reboot hang and bisected it back to this commit: a4f1987e4c54 x86, reboot: Add EFI and CF9 reboot methods into the default list He heroically tested all reboot methods and found the following: reboot=t # triple fault ok reboot=k # keyboard ctrl FAIL reboot=b # BIOS ok reboot=a # ACPI FAIL reboot=e # EFI FAIL [system has no EFI] reboot=p # PCI 0xcf9 FAIL And I think it's pretty obvious that we should only try PCI 0xcf9 as a last resort - if at all. The other observation is that (on this box) we should never try the PCI reboot method, but close with either the 'triple fault' or the 'BIOS' (terminal!) reboot methods. Thirdly, CF9_COND is a total misnomer - it should be something like CF9_SAFE or CF9_CAREFUL, and 'CF9' should be 'CF9_FORCE' ... So this patch fixes the worst problems: - it orders the actual reboot logic to follow the reboot ordering pattern - it was in a pretty random order before for no good reason. - it fixes the CF9 misnomers and uses BOOT_CF9_FORCE and BOOT_CF9_SAFE flags to make the code more obvious. - it tries the BIOS reboot method before the PCI reboot method. (Since 'BIOS' is a terminal reboot method resulting in a hang if it does not work, this is essentially equivalent to removing the PCI reboot method from the default reboot chain.) - just for the miraculous possibility of terminal (resulting in hang) reboot methods of triple fault or BIOS returning without having done their job, there's an ordering between them as well. Reported-and-bisected-and-tested-by: Steven Rostedt <rostedt@goodmis.org> Cc: Li Aubrey <aubrey.li@linux.intel.com> Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Matthew Garrett <mjg59@srcf.ucam.org> Link: http://lkml.kernel.org/r/20140404064120.GB11877@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-15xen/spinlock: Don't enable them unconditionally.Konrad Rzeszutek Wilk
The git commit a945928ea2709bc0e8e8165d33aed855a0110279 ('xen: Do not enable spinlocks before jump_label_init() has executed') was added to deal with the jump machinery. Earlier the code that turned on the jump label was only called by Xen specific functions. But now that it had been moved to the initcall machinery it gets called on Xen, KVM, and baremetal - ouch!. And the detection machinery to only call it on Xen wasn't remembered in the heat of merge window excitement. This means that the slowpath is enabled on baremetal while it should not be. Reported-by: Waiman Long <waiman.long@hp.com> Acked-by: Steven Rostedt <rostedt@goodmis.org> CC: stable@vger.kernel.org CC: Boris Ostrovsky <boris.ostrovsky@oracle.com> Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
2014-04-15x86/xen: Fix 32-bit PV guests's usage of kernel_stackBoris Ostrovsky
Commit 198d208df4371734ac4728f69cb585c284d20a15 ("x86: Keep thread_info on thread stack in x86_32") made 32-bit kernels use kernel_stack to point to thread_info. That change missed a couple of updates needed by Xen's 32-bit PV guests: 1. kernel_stack needs to be initialized for secondary CPUs 2. GET_THREAD_INFO() now uses %fs register which may not be the kernel's version when executing xen_iret(). With respect to the second issue, we don't need GET_THREAD_INFO() anymore: we used it as an intermediate step to get to per_cpu xen_vcpu and avoid referencing %fs. Now that we are going to use %fs anyway we may as well go directly to xen_vcpu. Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com> Signed-off-by: David Vrabel <david.vrabel@citrix.com>
2014-04-14Merge git://git.kernel.org/pub/scm/virt/kvm/kvmLinus Torvalds
Pull KVM fixes from Marcelo Tosatti: - Fix for guest triggerable BUG_ON (CVE-2014-0155) - CR4.SMAP support - Spurious WARN_ON() fix * git://git.kernel.org/pub/scm/virt/kvm/kvm: KVM: x86: remove WARN_ON from get_kernel_ns() KVM: Rename variable smep to cr4_smep KVM: expose SMAP feature to guest KVM: Disable SMAP for guests in EPT realmode and EPT unpaging mode KVM: Add SMAP support when setting CR4 KVM: Remove SMAP bit from CR4_RESERVED_BITS KVM: ioapic: try to recover if pending_eoi goes out of range KVM: ioapic: fix assignment of ioapic->rtc_status.pending_eoi (CVE-2014-0155)
2014-04-14KVM: x86: remove WARN_ON from get_kernel_ns()Marcelo Tosatti
Function and callers can be preempted. https://bugzilla.kernel.org/show_bug.cgi?id=73721 Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com> Reviewed-by: Paolo Bonzini <pbonzini@redhat.com>
2014-04-14KVM: Rename variable smep to cr4_smepFeng Wu
Rename variable smep to cr4_smep, which can better reflect the meaning of the variable. Signed-off-by: Feng Wu <feng.wu@intel.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-14KVM: expose SMAP feature to guestFeng Wu
This patch exposes SMAP feature to guest Signed-off-by: Feng Wu <feng.wu@intel.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-14KVM: Disable SMAP for guests in EPT realmode and EPT unpaging modeFeng Wu
SMAP is disabled if CPU is in non-paging mode in hardware. However KVM always uses paging mode to emulate guest non-paging mode with TDP. To emulate this behavior, SMAP needs to be manually disabled when guest switches to non-paging mode. Signed-off-by: Feng Wu <feng.wu@intel.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-14KVM: Add SMAP support when setting CR4Feng Wu
This patch adds SMAP handling logic when setting CR4 for guests Thanks a lot to Paolo Bonzini for his suggestion to use the branchless way to detect SMAP violation. Signed-off-by: Feng Wu <feng.wu@intel.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-14KVM: Remove SMAP bit from CR4_RESERVED_BITSFeng Wu
This patch removes SMAP bit from CR4_RESERVED_BITS. Signed-off-by: Feng Wu <feng.wu@intel.com> Signed-off-by: Marcelo Tosatti <mtosatti@redhat.com>
2014-04-14x86, irq, pic: Probe for legacy PIC and set legacy_pic appropriatelyK. Y. Srinivasan
The legacy PIC may or may not be available and we need a mechanism to detect the existence of the legacy PIC that is applicable for all hardware (both physical as well as virtual) currently supported by Linux. On Hyper-V, when our legacy firmware presented to the guests, emulates the legacy PIC while when our EFI based firmware is presented we do not emulate the PIC. To support Hyper-V EFI firmware, we had to set the legacy_pic to the null_legacy_pic since we had to bypass PIC based calibration in the early boot code. While, on the EFI firmware, we know we don't emulate the legacy PIC, we need a generic mechanism to detect the presence of the legacy PIC that is not based on boot time state - this became apparent when we tried to get kexec to work on Hyper-V EFI firmware. This patch implements the proposal put forth by H. Peter Anvin <hpa@linux.intel.com>: Write a known value to the PIC data port and read it back. If the value read is the value written, we do have the PIC, if not there is no PIC and we can safely set the legacy_pic to null_legacy_pic. Since the read from an unconnected I/O port returns 0xff, we will use ~(1 << PIC_CASCADE_IR) (0xfb: mask all lines except the cascade line) to probe for the existence of the PIC. In version V1 of the patch, I had cleaned up the code based on comments from Peter. In version V2 of the patch, I have addressed additional comments from Peter. In version V3 of the patch, I have addressed Jan's comments (JBeulich@suse.com). In version V4 of the patch, I have addressed additional comments from Peter. Signed-off-by: K. Y. Srinivasan <kys@microsoft.com> Link: http://lkml.kernel.org/r/1397501029-29286-1-git-send-email-kys@microsoft.com Cc: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2014-04-14Merge tag 'v3.15-rc1' into perf/urgentIngo Molnar
Pick up the latest fixes. Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-14x86/irq: Clean up VECTOR_UNDEFINED and VECTOR_RETRIGGERED definitionPrarit Bhargava
During another patch review, David Rientjes noted that VECTOR_UNDEFINED and VECTOR_RETRIGGERED should be defined with ()s so that they are not erroneously used in an arithmetic operation. Suggested-by: David Rientjes <rientjes@google.com> Signed-off-by: Prarit Bhargava <prarit@redhat.com> Cc: Seiji Aguchi <seiji.aguchi@hds.com> Cc: Yang Zhang <yang.z.zhang@Intel.com> Link: http://lkml.kernel.org/r/1396440827-18352-1-git-send-email-prarit@redhat.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-14x86/build: Supress "Nothing to be done for ..." messagesMasahiro Yamada
When we build an already built kernel again, arch/x86/syscalls/Makefile and arch/x86/tools/Makefile emits "Nothing to be done for ..." messages. Here is the command log: $ make defconfig [ snip ] $ make [ snip ] $ make make[1]: Nothing to be done for `all'. <----- make[1]: Nothing to be done for `relocs'. <----- CHK include/config/kernel.release CHK include/generated/uapi/linux/version.h Besides not emitting those, "all" and "relocs" should be added to PHONY as well. Signed-off-by: Masahiro Yamada <yamada.m@jp.panasonic.com> Acked-by: Peter Foley <pefoley2@pefoley.com> Acked-by: Michal Marek <mmarek@suse.cz> Link: http://lkml.kernel.org/r/1397093742-11144-1-git-send-email-yamada.m@jp.panasonic.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-14x86/gpu: Fix sign extension issue in Intel graphics stolen memory quirksVille Syrjälä
Have the KB(),MB(),GB() macros produce unsigned longs to avoid unintended sign extension issues with the gen2 memory size detection. What happens is first the uint8_t returned by read_pci_config_byte() gets promoted to an int which gets multiplied by another int from the MB() macro, and finally the result gets sign extended to size_t. Although this shouldn't be a problem in practice as all affected gen2 platforms are 32bit AFAIK, so size_t will be 32 bits. Reported-by: Bjorn Helgaas <bhelgaas@google.com> Suggested-by: H. Peter Anvin <hpa@zytor.com> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com> Cc: linux-kernel@vger.kernel.org Link: http://lkml.kernel.org/r/1397382303-17525-1-git-send-email-ville.syrjala@linux.intel.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-14x86/platform: Fix "make O=dir kvmconfig"Antonio Borneo
Running: make O=dir x86_64_defconfig make O=dir kvmconfig the second command dirties the source tree with file ".config", symlink "source" and objects in folder "scripts". Fixed by using properly prefixed paths in the arch Makefile. Signed-off-by: Antonio Borneo <borneo.antonio@gmail.com> Acked-by: Borislav Petkov <bp@suse.de> Cc: Pekka Enberg <penberg@kernel.org> Link: http://lkml.kernel.org/r/1397377568-8375-1-git-send-email-borneo.antonio@gmail.com Signed-off-by: Ingo Molnar <mingo@kernel.org>
2014-04-12Merge tag 'llvmlinux-for-v3.15' of ↵Linus Torvalds
git://git.linuxfoundation.org/llvmlinux/kernel Pull llvm patches from Behan Webster: "These are some initial updates to support compiling the kernel with clang. These patches have been through the proper reviews to the best of my ability, and have been soaking in linux-next for a few weeks. These patches by themselves still do not completely allow clang to be used with the kernel code, but lay the foundation for other patches which are still under review. Several other of the LLVMLinux patches have been already added via maintainer trees" * tag 'llvmlinux-for-v3.15' of git://git.linuxfoundation.org/llvmlinux/kernel: x86: LLVMLinux: Fix "incomplete type const struct x86cpu_device_id" x86 kbuild: LLVMLinux: More cc-options added for clang x86, acpi: LLVMLinux: Remove nested functions from Thinkpad ACPI LLVMLinux: Add support for clang to compiler.h and new compiler-clang.h LLVMLinux: Remove warning about returning an uninitialized variable kbuild: LLVMLinux: Fix LINUX_COMPILER definition script for compilation with clang Documentation: LLVMLinux: Update Documentation/dontdiff kbuild: LLVMLinux: Adapt warnings for compilation with clang kbuild: LLVMLinux: Add Kbuild support for building kernel with Clang
2014-04-12Merge git://git.infradead.org/users/eparis/auditLinus Torvalds
Pull audit updates from Eric Paris. * git://git.infradead.org/users/eparis/audit: (28 commits) AUDIT: make audit_is_compat depend on CONFIG_AUDIT_COMPAT_GENERIC audit: renumber AUDIT_FEATURE_CHANGE into the 1300 range audit: do not cast audit_rule_data pointers pointlesly AUDIT: Allow login in non-init namespaces audit: define audit_is_compat in kernel internal header kernel: Use RCU_INIT_POINTER(x, NULL) in audit.c sched: declare pid_alive as inline audit: use uapi/linux/audit.h for AUDIT_ARCH declarations syscall_get_arch: remove useless function arguments audit: remove stray newline from audit_log_execve_info() audit_panic() call audit: remove stray newlines from audit_log_lost messages audit: include subject in login records audit: remove superfluous new- prefix in AUDIT_LOGIN messages audit: allow user processes to log from another PID namespace audit: anchor all pid references in the initial pid namespace audit: convert PPIDs to the inital PID namespace. pid: get pid_t ppid of task in init_pid_ns audit: rename the misleading audit_get_context() to audit_take_context() audit: Add generic compat syscall support audit: Add CONFIG_HAVE_ARCH_AUDITSYSCALL ...
2014-04-11i386: Wire up the renameat2() syscallMiklos Szeredi
The renameat2() system call was only wired up for x86-64. Signed-off-by: Miklos Szeredi <mszeredi@suse.cz> Link: http://lkml.kernel.org/r/1397211951-20549-2-git-send-email-miklos@szeredi.hu Cc: H. Peter Anvin <hpa@zytor.com> Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2014-04-11Merge tag 'pm+acpi-3.15-rc1-3' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm Pull more ACPI and power management fixes and updates from Rafael Wysocki: "This is PM and ACPI material that has emerged over the last two weeks and one fix for a CPU hotplug regression introduced by the recent CPU hotplug notifiers registration series. Included are intel_idle and turbostat updates from Len Brown (these have been in linux-next for quite some time), a new cpufreq driver for powernv (that might spend some more time in linux-next, but BenH was asking me so nicely to push it for 3.15 that I couldn't resist), some cpufreq fixes and cleanups (including fixes for some silly breakage in a couple of cpufreq drivers introduced during the 3.14 cycle), assorted ACPI cleanups, wakeup framework documentation fixes, a new sysfs attribute for cpuidle and a new command line argument for power domains diagnostics. Specifics: - Fix for a recently introduced CPU hotplug regression in ARM KVM from Ming Lei. - Fixes for breakage in the at32ap, loongson2_cpufreq, and unicore32 cpufreq drivers introduced during the 3.14 cycle (-stable material) from Chen Gang and Viresh Kumar. - New powernv cpufreq driver from Vaidyanathan Srinivasan, with bits from Gautham R Shenoy and Srivatsa S Bhat. - Exynos cpufreq driver fix preventing it from being included into multiplatform builds that aren't supported by it from Sachin Kamat. - cpufreq cleanups related to the usage of the driver_data field in struct cpufreq_frequency_table from Viresh Kumar. - cpufreq ppc driver cleanup from Sachin Kamat. - Intel BayTrail support for intel_idle and ACPI idle from Len Brown. - Intel CPU model 54 (Atom N2000 series) support for intel_idle from Jan Kiszka. - intel_idle fix for Intel Ivy Town residency targets from Len Brown. - turbostat updates (Intel Broadwell support and output cleanups) from Len Brown. - New cpuidle sysfs attribute for exporting C-states' target residency information to user space from Daniel Lezcano. - New kernel command line argument to prevent power domains enabled by the bootloader from being turned off even if they are not in use (for diagnostics purposes) from Tushar Behera. - Fixes for wakeup sysfs attributes documentation from Geert Uytterhoeven. - New ACPI video blacklist entry for ThinkPad Helix from Stephen Chandler Paul. - Assorted ACPI cleanups and a Kconfig help update from Jonghwan Choi, Zhihui Zhang, Hanjun Guo" * tag 'pm+acpi-3.15-rc1-3' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm: (28 commits) ACPI: Update the ACPI spec information in Kconfig arm, kvm: fix double lock on cpu_add_remove_lock cpuidle: sysfs: Export target residency information cpufreq: ppc: Remove duplicate inclusion of fsl_soc.h cpufreq: create another field .flags in cpufreq_frequency_table cpufreq: use kzalloc() to allocate memory for cpufreq_frequency_table cpufreq: don't print value of .driver_data from core cpufreq: ia64: don't set .driver_data to index cpufreq: powernv: Select CPUFreq related Kconfig options for powernv cpufreq: powernv: Use cpufreq_frequency_table.driver_data to store pstate ids cpufreq: powernv: cpufreq driver for powernv platform cpufreq: at32ap: don't declare local variable as static cpufreq: loongson2_cpufreq: don't declare local variable as static cpufreq: unicore32: fix typo issue for 'clk' cpufreq: exynos: Disable on multiplatform build PM / wakeup: Correct presence vs. emptiness of wakeup_* attributes PM / domains: Add pd_ignore_unused to keep power domains enabled ACPI / dock: Drop dock_device_ids[] table ACPI / video: Favor native backlight interface for ThinkPad Helix ACPI / thermal: Fix wrong variable usage in debug statement ...
2014-04-11Merge branch 'x86-platform-for-linus' of ↵Linus Torvalds
git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip Pullx86 core platform updates from Peter Anvin: "This is the x86/platform branch with the objectionable IOSF patches removed. What is left is proper memory handling for Intel GPUs, and a change to the Calgary IOMMU code which will be required to make kexec work sanely on those platforms after some upcoming kexec changes" * 'x86-platform-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip: x86, calgary: Use 8M TCE table size by default x86/gpu: Print the Intel graphics stolen memory range x86/gpu: Add Intel graphics stolen memory quirk for gen2 platforms x86/gpu: Add vfunc for Intel graphics stolen memory base address