simple_udelay_calibration() relies on x86_platform's calibration ops.
For KVM these ops are set late in setup_arch() and so
simple_udelay_calibration() ends up using native version.
Besides being possibly incorrect, this significantly increases kernel
boot time. For example, on my laptop executing start_kernel() by a guest
takes ~10 times more than when KVM's ops are used.
Since early_xdbc_setup_hardware() relies on calibration having been
performed move it too.
Signed-off-by: Boris Ostrovsky <boris.ostrovsky@oracle.com>
Cc: baolu.lu@linux.intel.com
Link: https://lkml.kernel.org/r/20170911185111.20636-1-boris.ostrovsky@oracle.com
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
A cold or warm boot through BIOS sets the APIC in default interrupt
delivery mode. A dump-capture kernel will not go through a BIOS reset and
leave the interrupt delivery mode in the state which was active on the
crashed kernel, but the dump kernel startup code assumes default delivery
mode which can result in interrupt delivery/handling to fail.
To solve this problem, it's required to set up the final interrupt delivery
mode as soon as possible. As IOAPIC setup needs the timer initialized for
verifying the timer interrupt delivery mode, the earliest point is right
after timer setup in late_time_init().
That results in the following init order:
1) Set up the legacy timer, if applicable on the platform
2) Set up APIC/IOAPIC which includes the verification of the legacy timer
interrupt delivery.
3) TSC calibration
4) Local APIC timer setup
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: yinghai@kernel.org
Cc: bhe@redhat.com
Link: https://lkml.kernel.org/r/1505293975-26005-12-git-send-email-douly.fnst@cn.fujitsu.com
timer_irq_works() is used to detects the timer IRQs. It calls mdelay(10) to
delay ten ticks and check whether the timer IRQ work or not.
mdelay() depends on the loops_per_jiffy which is set up in
calibrate_delay(), but the delay calibration depends on a working timer
interrupt, which causes a chicken and egg problem.
The correct solution is to set up the interrupt mode and making sure that
the timer interrupt is delivered correctly before invoking calibrate_delay().
That means that mdelay() cannot be used in timer_irq_works().
Provide helper functions to make a rough delay estimate which is good enough
to prove that the timer interrupt is working. Either use TSC or a simple
delay loop and assume that 4GHz is the maximum CPU frequency to base the
delay calculation on.
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: yinghai@kernel.org
Cc: bhe@redhat.com
Link: https://lkml.kernel.org/r/1505293975-26005-9-git-send-email-douly.fnst@cn.fujitsu.com
On a SMP-capable system, the kernel enables and sets up the APIC interrupt
delivery mode in native_smp_prepare_cpus(). The decision how to setup the
APIC is intermingled with the decision of setting up SMP or not.
Split the initialization of the APIC interrupt mode independent from other
decisions and have a separate apic_intr_mode_init() function for it.
The invocation time stays the same for now.
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: yinghai@kernel.org
Cc: bhe@redhat.com
Link: https://lkml.kernel.org/r/1505293975-26005-6-git-send-email-douly.fnst@cn.fujitsu.com
There are three places which initialize the interrupt delivery modes:
1) init_bsp_APIC() which is called early might setup the through-local-APIC
virtual wire mode on non SMP systems.
2) In an SMP-capable system, native_smp_prepare_cpus() tries to switch to
symmetric I/O model.
3) In UP system with UP_LATE_INIT=y, the local APIC and I/O APIC are set up
in smp_init().
There is no technical reason to make these initializations at random places
and run the kernel with the potentially wrong mode through the early boot
stage, but it has a problematic side effect: The late switch to symmetric
I/O mode causes dump-capture kernel to hang when the kernel command line
option 'notsc' is active.
Provide a new function to unify that three positions. Preparatory patch to
initialize an interrupt mode directly.
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: yinghai@kernel.org
Cc: bhe@redhat.com
Link: https://lkml.kernel.org/r/1505293975-26005-3-git-send-email-douly.fnst@cn.fujitsu.com
There are quite some switches which are used to determine the final
interrupt delivery mode, as shown below:
1) Kconfig: CONFIG_X86_64; CONFIG_X86_LOCAL_APIC; CONFIG_x86_IO_APIC
2) Command line options: disable_apic; skip_ioapic_setup
3) CPU Capability: boot_cpu_has(X86_FEATURE_APIC)
4) MP table: smp_found_config
5) ACPI: acpi_lapic; acpi_ioapic; nr_ioapic
These switches are disordered and scattered and there are also some
dependencies between them. These make the code difficult to maintain and
read.
Construct a selector to unify them into a single function, then, Use this
selector to get an interrupt delivery mode directly.
Signed-off-by: Dou Liyang <douly.fnst@cn.fujitsu.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: yinghai@kernel.org
Cc: bhe@redhat.com
Link: https://lkml.kernel.org/r/1505293975-26005-2-git-send-email-douly.fnst@cn.fujitsu.com
On x86, userspace can use the ptrace() or rt_sigreturn() system calls to
set a task's extended state (xstate) or "FPU" registers. ptrace() can
set them for another task using the PTRACE_SETREGSET request with
NT_X86_XSTATE, while rt_sigreturn() can set them for the current task.
In either case, registers can be set to any value, but the kernel
assumes that the XSAVE area itself remains valid in the sense that the
CPU can restore it.
However, in the case where the kernel is using the uncompacted xstate
format (which it does whenever the XSAVES instruction is unavailable),
it was possible for userspace to set the xcomp_bv field in the
xstate_header to an arbitrary value. However, all bits in that field
are reserved in the uncompacted case, so when switching to a task with
nonzero xcomp_bv, the XRSTOR instruction failed with a #GP fault. This
caused the WARN_ON_FPU(err) in copy_kernel_to_xregs() to be hit. In
addition, since the error is otherwise ignored, the FPU registers from
the task previously executing on the CPU were leaked.
Fix the bug by checking that the user-supplied value of xcomp_bv is 0 in
the uncompacted case, and returning an error otherwise.
The reason for validating xcomp_bv rather than simply overwriting it
with 0 is that we want userspace to see an error if it (incorrectly)
provides an XSAVE area in compacted format rather than in uncompacted
format.
Note that as before, in case of error we clear the task's FPU state.
This is perhaps non-ideal, especially for PTRACE_SETREGSET; it might be
better to return an error before changing anything. But it seems the
"clear on error" behavior is fine for now, and it's a little tricky to
do otherwise because it would mean we couldn't simply copy the full
userspace state into kernel memory in one __copy_from_user().
This bug was found by syzkaller, which hit the above-mentioned
WARN_ON_FPU():
WARNING: CPU: 1 PID: 0 at ./arch/x86/include/asm/fpu/internal.h:373 __switch_to+0x5b5/0x5d0
CPU: 1 PID: 0 Comm: swapper/1 Not tainted 4.13.0 #453
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011
task: ffff9ba2bc8e42c0 task.stack: ffffa78cc036c000
RIP: 0010:__switch_to+0x5b5/0x5d0
RSP: 0000:ffffa78cc08bbb88 EFLAGS: 00010082
RAX: 00000000fffffffe RBX: ffff9ba2b8bf2180 RCX: 00000000c0000100
RDX: 00000000ffffffff RSI: 000000005cb10700 RDI: ffff9ba2b8bf36c0
RBP: ffffa78cc08bbbd0 R08: 00000000929fdf46 R09: 0000000000000001
R10: 0000000000000000 R11: 0000000000000000 R12: ffff9ba2bc8e42c0
R13: 0000000000000000 R14: ffff9ba2b8bf3680 R15: ffff9ba2bf5d7b40
FS: 00007f7e5cb10700(0000) GS:ffff9ba2bf400000(0000) knlGS:0000000000000000
CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 00000000004005cc CR3: 0000000079fd5000 CR4: 00000000001406e0
Call Trace:
Code: 84 00 00 00 00 00 e9 11 fd ff ff 0f ff 66 0f 1f 84 00 00 00 00 00 e9 e7 fa ff ff 0f ff 66 0f 1f 84 00 00 00 00 00 e9 c2 fa ff ff <0f> ff 66 0f 1f 84 00 00 00 00 00 e9 d4 fc ff ff 66 66 2e 0f 1f
Here is a C reproducer. The expected behavior is that the program spin
forever with no output. However, on a buggy kernel running on a
processor with the "xsave" feature but without the "xsaves" feature
(e.g. Sandy Bridge through Broadwell for Intel), within a second or two
the program reports that the xmm registers were corrupted, i.e. were not
restored correctly. With CONFIG_X86_DEBUG_FPU=y it also hits the above
kernel warning.
#define _GNU_SOURCE
#include <stdbool.h>
#include <inttypes.h>
#include <linux/elf.h>
#include <stdio.h>
#include <sys/ptrace.h>
#include <sys/uio.h>
#include <sys/wait.h>
#include <unistd.h>
int main(void)
{
int pid = fork();
uint64_t xstate[512];
struct iovec iov = { .iov_base = xstate, .iov_len = sizeof(xstate) };
if (pid == 0) {
bool tracee = true;
for (int i = 0; i < sysconf(_SC_NPROCESSORS_ONLN) && tracee; i++)
tracee = (fork() != 0);
uint32_t xmm0[4] = { [0 ... 3] = tracee ? 0x00000000 : 0xDEADBEEF };
asm volatile(" movdqu %0, %%xmm0\n"
" mov %0, %%rbx\n"
"1: movdqu %%xmm0, %0\n"
" mov %0, %%rax\n"
" cmp %%rax, %%rbx\n"
" je 1b\n"
: "+m" (xmm0) : : "rax", "rbx", "xmm0");
printf("BUG: xmm registers corrupted! tracee=%d, xmm0=%08X%08X%08X%08X\n",
tracee, xmm0[0], xmm0[1], xmm0[2], xmm0[3]);
} else {
usleep(100000);
ptrace(PTRACE_ATTACH, pid, 0, 0);
wait(NULL);
ptrace(PTRACE_GETREGSET, pid, NT_X86_XSTATE, &iov);
xstate[65] = -1;
ptrace(PTRACE_SETREGSET, pid, NT_X86_XSTATE, &iov);
ptrace(PTRACE_CONT, pid, 0, 0);
wait(NULL);
}
return 1;
}
Note: the program only tests for the bug using the ptrace() system call.
The bug can also be reproduced using the rt_sigreturn() system call, but
only when called from a 32-bit program, since for 64-bit programs the
kernel restores the FPU state from the signal frame by doing XRSTOR
directly from userspace memory (with proper error checking).
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Eric Biggers <ebiggers@google.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Rik van Riel <riel@redhat.com>
Acked-by: Dave Hansen <dave.hansen@linux.intel.com>
Cc: <stable@vger.kernel.org> [v3.17+]
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Eric Biggers <ebiggers3@gmail.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Kevin Hao <haokexin@gmail.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Michael Halcrow <mhalcrow@google.com>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Wanpeng Li <wanpeng.li@hotmail.com>
Cc: Yu-cheng Yu <yu-cheng.yu@intel.com>
Cc: kernel-hardening@lists.openwall.com
Fixes: 0b29643a58 ("x86/xsaves: Change compacted format xsave area header")
Link: http://lkml.kernel.org/r/20170922174156.16780-2-ebiggers3@gmail.com
Link: http://lkml.kernel.org/r/20170923130016.21448-25-mingo@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Right now there's a confusing mixture of 'offset' and 'size' parameters:
- __copy_xstate_to_*() input parameter 'end_pos' not not really an offset,
but the full size of the copy to be performed.
- input parameter 'count' to copy_xstate_to_*() shadows that of
__copy_xstate_to_*()'s 'count' parameter name - but the roles
are different: the first one is the total number of bytes to
be copied, while the second one is a partial copy size.
To unconfuse all this, use a consistent set of parameter names:
- 'size' is the partial copy size within a single xstate component
- 'size_total' is the total copy requested
- 'offset_start' is the requested starting offset.
- 'offset' is the offset within an xstate component.
No change in functionality.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Eric Biggers <ebiggers3@gmail.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Yu-cheng Yu <yu-cheng.yu@intel.com>
Link: http://lkml.kernel.org/r/20170923130016.21448-9-mingo@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The 'copyin/copyout' nomenclature needlessly departs from what the modern FPU code
uses, which is:
copy_fpregs_to_fpstate()
copy_fpstate_to_sigframe()
copy_fregs_to_user()
copy_fxregs_to_kernel()
copy_fxregs_to_user()
copy_kernel_to_fpregs()
copy_kernel_to_fregs()
copy_kernel_to_fxregs()
copy_kernel_to_xregs()
copy_user_to_fregs()
copy_user_to_fxregs()
copy_user_to_xregs()
copy_xregs_to_kernel()
copy_xregs_to_user()
I.e. according to this pattern, the following rename should be done:
copyin_to_xsaves() -> copy_user_to_xstate()
copyout_from_xsaves() -> copy_xstate_to_user()
or, if we want to be pedantic, denote that that the user-space format is ptrace:
copyin_to_xsaves() -> copy_user_ptrace_to_xstate()
copyout_from_xsaves() -> copy_xstate_to_user_ptrace()
But I'd suggest the shorter, non-pedantic name.
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Andy Lutomirski <luto@kernel.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Eric Biggers <ebiggers3@gmail.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Rik van Riel <riel@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Yu-cheng Yu <yu-cheng.yu@intel.com>
Link: http://lkml.kernel.org/r/20170923130016.21448-2-mingo@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
