On 08/14/2018 05:36 PM, Sebastian Gottschall wrote:

if SWAP is disabled in kernel config, the following compile error will raise up with this release

arch/x86/built-in.o: in function `max_swapfile_size': (.text+0x3bba1): undefined reference to `generic_max_swapfile_size'

of course this is simple to fix. the function max_swapfile_size must be excluded if CONFIG_SWAP is disabled

Fixed with upstream commit 792adb90fa72 ("x86/init: fix build with CONFIG_SWAP=n"). Also needed in other stable branches (at least in 4.4).

Guenter

Sebastian

Am 14.08.2018 um 19:16 schrieb Greg Kroah-Hartman:

...

This is the start of the stable review cycle for the 4.9.120 release. There are 107 patches in this series, all will be posted as a response to this one.  If anyone has any issues with these being applied, please let me know.

Responses should be made by Thu Aug 16 17:14:53 UTC 2018. Anything received after that time might be too late.

The whole patch series can be found in one patch at:     https://www.kernel.org/pub/linux/kernel/v4.x/stable-review/patch-4.9.120-rc1... or in the git tree and branch at:     git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-4.9.y and the diffstat can be found below.

thanks,

greg k-h

---

Pseudo-Shortlog of commits:

Greg Kroah-Hartman gregkh@linuxfoundation.org      Linux 4.9.120-rc1

Josh Poimboeuf jpoimboe@redhat.com      x86/microcode: Allow late microcode loading with SMT disabled

Ashok Raj ashok.raj@intel.com      x86/microcode: Do not upload microcode if CPUs are offline

David Woodhouse dwmw@amazon.co.uk      tools headers: Synchronise x86 cpufeatures.h for L1TF additions

Andi Kleen ak@linux.intel.com      x86/mm/kmmio: Make the tracer robust against L1TF

Andi Kleen ak@linux.intel.com      x86/mm/pat: Make set_memory_np() L1TF safe

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Make pmd/pud_mknotpresent() invert

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Invert all not present mappings

Thomas Gleixner tglx@linutronix.de      cpu/hotplug: Fix SMT supported evaluation

Paolo Bonzini pbonzini@redhat.com      KVM: VMX: Tell the nested hypervisor to skip L1D flush on vmentry

Paolo Bonzini pbonzini@redhat.com      x86/speculation: Use ARCH_CAPABILITIES to skip L1D flush on vmentry

Paolo Bonzini pbonzini@redhat.com      x86/speculation: Simplify sysfs report of VMX L1TF vulnerability

Paolo Bonzini pbonzini@redhat.com      KVM: VMX: support MSR_IA32_ARCH_CAPABILITIES as a feature MSR

Wanpeng Li wanpengli@tencent.com      KVM: X86: Allow userspace to define the microcode version

Wanpeng Li wanpengli@tencent.com      KVM: X86: Introduce kvm_get_msr_feature()

Tom Lendacky thomas.lendacky@amd.com      KVM: SVM: Add MSR-based feature support for serializing LFENCE

Tom Lendacky thomas.lendacky@amd.com      KVM: x86: Add a framework for supporting MSR-based features

Thomas Gleixner tglx@linutronix.de      Documentation/l1tf: Remove Yonah processors from not vulnerable list

Nicolai Stange nstange@suse.de      x86/KVM/VMX: Don't set l1tf_flush_l1d from vmx_handle_external_intr()

Nicolai Stange nstange@suse.de      x86/irq: Let interrupt handlers set kvm_cpu_l1tf_flush_l1d

Nicolai Stange nstange@suse.de      x86: Don't include linux/irq.h from asm/hardirq.h

Nicolai Stange nstange@suse.de      x86/KVM/VMX: Introduce per-host-cpu analogue of l1tf_flush_l1d

Nicolai Stange nstange@suse.de      x86/irq: Demote irq_cpustat_t::__softirq_pending to u16

Nicolai Stange nstange@suse.de      x86/KVM/VMX: Move the l1tf_flush_l1d test to vmx_l1d_flush()

Nicolai Stange nstange@suse.de      x86/KVM/VMX: Replace 'vmx_l1d_flush_always' with 'vmx_l1d_flush_cond'

Nicolai Stange nstange@suse.de      x86/KVM/VMX: Don't set l1tf_flush_l1d to true from vmx_l1d_flush()

Josh Poimboeuf jpoimboe@redhat.com      cpu/hotplug: detect SMT disabled by BIOS

Tony Luck tony.luck@intel.com      Documentation/l1tf: Fix typos

Nicolai Stange nstange@suse.de      x86/KVM/VMX: Initialize the vmx_l1d_flush_pages' content

Thomas Gleixner tglx@linutronix.de      Documentation: Add section about CPU vulnerabilities

Jiri Kosina jkosina@suse.cz      x86/bugs, kvm: Introduce boot-time control of L1TF mitigations

Thomas Gleixner tglx@linutronix.de      cpu/hotplug: Set CPU_SMT_NOT_SUPPORTED early

Jiri Kosina jkosina@suse.cz      cpu/hotplug: Expose SMT control init function

Thomas Gleixner tglx@linutronix.de      x86/kvm: Allow runtime control of L1D flush

Thomas Gleixner tglx@linutronix.de      x86/kvm: Serialize L1D flush parameter setter

Thomas Gleixner tglx@linutronix.de      x86/kvm: Add static key for flush always

Thomas Gleixner tglx@linutronix.de      x86/kvm: Move l1tf setup function

Thomas Gleixner tglx@linutronix.de      x86/l1tf: Handle EPT disabled state proper

Thomas Gleixner tglx@linutronix.de      x86/kvm: Drop L1TF MSR list approach

Thomas Gleixner tglx@linutronix.de      x86/litf: Introduce vmx status variable

Thomas Gleixner tglx@linutronix.de      cpu/hotplug: Online siblings when SMT control is turned on

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/KVM/VMX: Use MSR save list for IA32_FLUSH_CMD if required

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/KVM/VMX: Extend add_atomic_switch_msr() to allow VMENTER only MSRs

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/KVM/VMX: Separate the VMX AUTOLOAD guest/host number accounting

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/KVM/VMX: Add find_msr() helper function

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/KVM/VMX: Split the VMX MSR LOAD structures to have an host/guest numbers

Jim Mattson jmattson@google.com      kvm: nVMX: Update MSR load counts on a VMCS switch

Paolo Bonzini pbonzini@redhat.com      x86/KVM/VMX: Add L1D flush logic

Paolo Bonzini pbonzini@redhat.com      x86/KVM/VMX: Add L1D MSR based flush

Paolo Bonzini pbonzini@redhat.com      x86/KVM/VMX: Add L1D flush algorithm

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/KVM/VMX: Add module argument for L1TF mitigation

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/KVM: Warn user if KVM is loaded SMT and L1TF CPU bug being present

Thomas Gleixner tglx@linutronix.de      cpu/hotplug: Boot HT siblings at least once

Thomas Gleixner tglx@linutronix.de      Revert "x86/apic: Ignore secondary threads if nosmt=force"

Michal Hocko mhocko@suse.cz      x86/speculation/l1tf: Fix up pte->pfn conversion for PAE

Vlastimil Babka vbabka@suse.cz      x86/speculation/l1tf: Protect PAE swap entries against L1TF

Borislav Petkov bp@suse.de      x86/CPU/AMD: Move TOPOEXT reenablement before reading smp_num_siblings

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/cpufeatures: Add detection of L1D cache flush support.

Vlastimil Babka vbabka@suse.cz      x86/speculation/l1tf: Extend 64bit swap file size limit

Thomas Gleixner tglx@linutronix.de      x86/apic: Ignore secondary threads if nosmt=force

Thomas Gleixner tglx@linutronix.de      x86/cpu/AMD: Evaluate smp_num_siblings early

Borislav Petkov bp@suse.de      x86/CPU/AMD: Do not check CPUID max ext level before parsing SMP info

Thomas Gleixner tglx@linutronix.de      x86/cpu/intel: Evaluate smp_num_siblings early

Thomas Gleixner tglx@linutronix.de      x86/cpu/topology: Provide detect_extended_topology_early()

Thomas Gleixner tglx@linutronix.de      x86/cpu/common: Provide detect_ht_early()

Thomas Gleixner tglx@linutronix.de      x86/cpu/AMD: Remove the pointless detect_ht() call

Thomas Gleixner tglx@linutronix.de      x86/cpu: Remove the pointless CPU printout

Thomas Gleixner tglx@linutronix.de      cpu/hotplug: Provide knobs to control SMT

Thomas Gleixner tglx@linutronix.de      cpu/hotplug: Split do_cpu_down()

Thomas Gleixner tglx@linutronix.de      cpu/hotplug: Make bringup/teardown of smp threads symmetric

Thomas Gleixner tglx@linutronix.de      x86/topology: Provide topology_smt_supported()

Thomas Gleixner tglx@linutronix.de      x86/smp: Provide topology_is_primary_thread()

Konrad Rzeszutek Wilk konrad.wilk@oracle.com      x86/bugs: Move the l1tf function and define pr_fmt properly

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Limit swap file size to MAX_PA/2

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Disallow non privileged high MMIO PROT_NONE mappings

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Add sysfs reporting for l1tf

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Make sure the first page is always reserved

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Protect PROT_NONE PTEs against speculation

Linus Torvalds torvalds@linux-foundation.org      x86/speculation/l1tf: Protect swap entries against L1TF

Linus Torvalds torvalds@linux-foundation.org      x86/speculation/l1tf: Change order of offset/type in swap entry

Naoya Horiguchi n-horiguchi@ah.jp.nec.com      mm: x86: move _PAGE_SWP_SOFT_DIRTY from bit 7 to bit 1

Andi Kleen ak@linux.intel.com      x86/speculation/l1tf: Increase 32bit PAE __PHYSICAL_PAGE_SHIFT

Nick Desaulniers ndesaulniers@google.com      x86/irqflags: Provide a declaration for native_save_fl

Masami Hiramatsu mhiramat@kernel.org      kprobes/x86: Fix %p uses in error messages

Jiri Kosina jkosina@suse.cz      x86/speculation: Protect against userspace-userspace spectreRSB

Peter Zijlstra peterz@infradead.org      x86/paravirt: Fix spectre-v2 mitigations for paravirt guests

Oleksij Rempel o.rempel@pengutronix.de      ARM: dts: imx6sx: fix irq for pcie bridge

Michael Mera dev@michaelmera.com      IB/ocrdma: fix out of bounds access to local buffer

Fabio Estevam fabio.estevam@nxp.com      mtd: nand: qcom: Add a NULL check for devm_kasprintf()

Jack Morgenstein jackm@dev.mellanox.co.il      IB/mlx4: Mark user MR as writable if actual virtual memory is writable

Jack Morgenstein jackm@dev.mellanox.co.il      IB/core: Make testing MR flags for writability a static inline function

Eric W. Biederman ebiederm@xmission.com      proc: Fix proc_sys_prune_dcache to hold a sb reference

Eric W. Biederman ebiederm@xmission.com      proc/sysctl: Don't grab i_lock under sysctl_lock.

Konstantin Khlebnikov khlebnikov@yandex-team.ru      proc/sysctl: prune stale dentries during unregistering

Al Viro viro@zeniv.linux.org.uk      fix __legitimize_mnt()/mntput() race

Al Viro viro@zeniv.linux.org.uk      fix mntput/mntput race

Al Viro viro@zeniv.linux.org.uk      make sure that __dentry_kill() always invalidates d_seq, unhashed or not

Al Viro viro@zeniv.linux.org.uk      root dentries need RCU-delayed freeing

Linus Torvalds torvalds@linux-foundation.org      init: rename and re-order boot_cpu_state_init()

Bart Van Assche bart.vanassche@wdc.com      scsi: sr: Avoid that opening a CD-ROM hangs with runtime power management enabled

Hans de Goede hdegoede@redhat.com      ACPI / LPSS: Add missing prv_offset setting for byt/cht PWM devices

Juergen Gross jgross@suse.com      xen/netfront: don't cache skb_shinfo()

Linus Torvalds torvalds@linux-foundation.org      Mark HI and TASKLET softirq synchronous

Andrey Konovalov andreyknvl@google.com      kasan: add no_sanitize attribute for clang builds

John David Anglin dave.anglin@bell.net      parisc: Define mb() and add memory barriers to assembler unlock sequences

Helge Deller deller@gmx.de      parisc: Enable CONFIG_MLONGCALLS by default

Tadeusz Struk tadeusz.struk@intel.com      tpm: fix race condition in tpm_common_write()

Theodore Ts'o tytso@mit.edu      ext4: fix check to prevent initializing reserved inodes

---

Diffstat:

Documentation/ABI/testing/sysfs-devices-system-cpu |  24 +   Documentation/index.rst                            |   1 +   Documentation/kernel-parameters.txt                |  78 +++   Documentation/l1tf.rst                             | 610 +++++++++++++++++++++   Documentation/virtual/kvm/api.txt                  |  40 +-   Makefile                                           |   4 +-   arch/Kconfig                                       |   3 +   arch/arm/boot/dts/imx6sx.dtsi                      |   2 +-   arch/parisc/Kconfig                                |   2 +-   arch/parisc/include/asm/barrier.h                  |  32 ++   arch/parisc/kernel/entry.S                         |   2 +   arch/parisc/kernel/pacache.S                       |   1 +   arch/parisc/kernel/syscall.S                       |   4 +   arch/x86/Kconfig                                   |   1 +   arch/x86/include/asm/apic.h                        |  10 +   arch/x86/include/asm/cpufeatures.h                 |   4 +-   arch/x86/include/asm/dmi.h                         |   2 +-   arch/x86/include/asm/hardirq.h                     |  26 +-   arch/x86/include/asm/irqflags.h                    |   2 +   arch/x86/include/asm/kvm_host.h                    |   9 +   arch/x86/include/asm/msr-index.h                   |   7 +   arch/x86/include/asm/page_32_types.h               |   9 +-   arch/x86/include/asm/pgtable-2level.h              |  17 +   arch/x86/include/asm/pgtable-3level.h              |  37 +-   arch/x86/include/asm/pgtable-invert.h              |  32 ++   arch/x86/include/asm/pgtable.h                     |  85 ++-   arch/x86/include/asm/pgtable_64.h                  |  48 +-   arch/x86/include/asm/pgtable_types.h               |  10 +-   arch/x86/include/asm/processor.h                   |  17 +   arch/x86/include/asm/topology.h                    |   6 +-   arch/x86/include/asm/vmx.h                         |  11 +   arch/x86/kernel/apic/apic.c                        |  17 +   arch/x86/kernel/apic/htirq.c                       |   2 +   arch/x86/kernel/apic/io_apic.c                     |   1 +   arch/x86/kernel/apic/msi.c                         |   1 +   arch/x86/kernel/apic/vector.c                      |   1 +   arch/x86/kernel/cpu/amd.c                          |  53 +-   arch/x86/kernel/cpu/bugs.c                         | 171 ++++--   arch/x86/kernel/cpu/common.c                       |  56 +-   arch/x86/kernel/cpu/cpu.h                          |   2 +   arch/x86/kernel/cpu/intel.c                        |   7 +   arch/x86/kernel/cpu/microcode/core.c               |  26 +   arch/x86/kernel/cpu/topology.c                     |  41 +-   arch/x86/kernel/fpu/core.c                         |   1 +   arch/x86/kernel/ftrace.c                           |   1 +   arch/x86/kernel/hpet.c                             |   1 +   arch/x86/kernel/i8259.c                            |   1 +   arch/x86/kernel/irq.c                              |   1 +   arch/x86/kernel/irq_32.c                           |   1 +   arch/x86/kernel/irq_64.c                           |   1 +   arch/x86/kernel/irqinit.c                          |   1 +   arch/x86/kernel/kprobes/core.c                     |   5 +-   arch/x86/kernel/kprobes/opt.c                      |   1 +   arch/x86/kernel/paravirt.c                         |  14 +-   arch/x86/kernel/setup.c                            |   6 +   arch/x86/kernel/smp.c                              |   1 +   arch/x86/kernel/smpboot.c                          |  18 +   arch/x86/kernel/time.c                             |   1 +   arch/x86/kvm/svm.c                                 |  46 +-   arch/x86/kvm/vmx.c                                 | 426 ++++++++++++--   arch/x86/kvm/x86.c                                 | 133 ++++-   arch/x86/mm/fault.c                                |   1 +   arch/x86/mm/init.c                                 |  23 +   arch/x86/mm/kaiser.c                               |   1 +   arch/x86/mm/kmmio.c                                |  25 +-   arch/x86/mm/mmap.c                                 |  21 +   arch/x86/mm/pageattr.c                             |   8 +-   arch/x86/platform/efi/efi_64.c                     |   1 +   arch/x86/platform/efi/quirks.c                     |   1 +   .../intel-mid/device_libs/platform_mrfld_wdt.c     |   1 +   arch/x86/platform/uv/tlb_uv.c                      |   1 +   arch/x86/xen/enlighten.c                           |   1 +   arch/x86/xen/setup.c                               |   1 +   drivers/acpi/acpi_lpss.c                           |   2 +   drivers/base/cpu.c                                 |   8 +   drivers/char/tpm/tpm-dev.c                         |  43 +-   drivers/infiniband/core/umem.c                     |  11 +-   drivers/infiniband/hw/mlx4/mr.c                    |  50 +-   drivers/infiniband/hw/ocrdma/ocrdma_stats.c        |   2 +-   drivers/mtd/nand/qcom_nandc.c                      |   3 +   drivers/net/xen-netfront.c                         |   8 +-   drivers/pci/host/pci-hyperv.c                      |   2 +   drivers/scsi/sr.c                                  |  29 +-   fs/dcache.c                                        |  13 +-   fs/ext4/ialloc.c                                   |   5 +-   fs/ext4/super.c                                    |   8 +-   fs/namespace.c                                     |  28 +-   fs/proc/inode.c                                    |   3 +-   fs/proc/internal.h                                 |   7 +-   fs/proc/proc_sysctl.c                              |  83 ++-   include/asm-generic/pgtable.h                      |  12 +   include/linux/compiler-clang.h                     |   3 +   include/linux/cpu.h                                |  23 +-   include/linux/swapfile.h                           |   2 +   include/linux/sysctl.h                             |   1 +   include/rdma/ib_verbs.h                            |  14 +   include/uapi/linux/kvm.h                           |   2 +   init/main.c                                        |   2 +-   kernel/cpu.c                                       | 282 +++++++++-   kernel/smp.c                                       |   2 +   kernel/softirq.c                                   |  12 +-   mm/memory.c                                        |  29 +-   mm/mprotect.c                                      |  49 ++   mm/swapfile.c                                      |  46 +-   tools/arch/x86/include/asm/cpufeatures.h           |   4 +-   105 files changed, 2680 insertions(+), 366 deletions(-)

On 08/14/2018 11:14 PM, Greg Kroah-Hartman wrote:

On Tue, Aug 14, 2018 at 07:16:23PM +0200, Greg Kroah-Hartman wrote:

...

This is the start of the stable review cycle for the 4.9.120 release. There are 107 patches in this series, all will be posted as a response to this one. If anyone has any issues with these being applied, please let me know.

Responses should be made by Thu Aug 16 17:14:53 UTC 2018. Anything received after that time might be too late.

The whole patch series can be found in one patch at: https://www.kernel.org/pub/linux/kernel/v4.x/stable-review/patch-4.9.120-rc1...

-rc2 is now out: https://www.kernel.org/pub/linux/kernel/v4.x/stable-review/patch-4.9.120-rc2...

Heads up: We'll need (at least) one more fix for v4.{9,14,17}. I'll send out backports after testing.

Guenter

On 2018-08-15 19:01 +0200, Sebastian Gottschall wrote:

nother issue seen on xscale and as it affects all non SMP configured kernels

kernel/cpu.c: In function 'boot_cpu_hotplug_init': kernel/cpu.c:2204:28: error: 'struct cpuhp_cpu_state' has no member named 'booted_once'   this_cpu_write(cpuhp_state.booted_once, true);

I got a different error in 4.9.120 with CONFIG_SMP unset:

,---- | CC drivers/rtc/rtc-cmos.o | In file included from drivers/rtc/rtc-cmos.c:45:0: | ./arch/x86/include/asm/i8259.h: In function 'inb_pic': | ./arch/x86/include/asm/i8259.h:32:24: error: implicit declaration of function 'inb' [-Werror=implicit-function-declaration] | unsigned char value = inb(port); | ^~~ | ./arch/x86/include/asm/i8259.h: In function 'outb_pic': | ./arch/x86/include/asm/i8259.h:45:2: error: implicit declaration of function 'outb' [-Werror=implicit-function-declaration] | outb(value, port); | ^~~~ | In file included from ./include/linux/mc146818rtc.h:14:0, | from drivers/rtc/rtc-cmos.c:49: | ./arch/x86/include/asm/io.h: At top level: | ./arch/x86/include/asm/io.h:277:20: warning: conflicting types for 'outb' | static inline void out##bwl(unsigned type value, int port) \ | ^ | ./arch/x86/include/asm/io.h:316:1: note: in expansion of macro 'BUILDIO' | BUILDIO(b, b, char) | ^~~~~~~ | ./arch/x86/include/asm/io.h:277:20: error: static declaration of 'outb' follows non-static declaration | static inline void out##bwl(unsigned type value, int port) \ | ^ | ./arch/x86/include/asm/io.h:316:1: note: in expansion of macro 'BUILDIO' | BUILDIO(b, b, char) | ^~~~~~~ | In file included from drivers/rtc/rtc-cmos.c:45:0: | ./arch/x86/include/asm/i8259.h:45:2: note: previous implicit declaration of 'outb' was here | outb(value, port); | ^~~~ | In file included from ./include/linux/mc146818rtc.h:14:0, | from drivers/rtc/rtc-cmos.c:49: | ./arch/x86/include/asm/io.h:283:29: error: conflicting types for 'inb' | static inline unsigned type in##bwl(int port) \ | ^ | ./arch/x86/include/asm/io.h:316:1: note: in expansion of macro 'BUILDIO' | BUILDIO(b, b, char) | ^~~~~~~ | In file included from drivers/rtc/rtc-cmos.c:45:0: | ./arch/x86/include/asm/i8259.h:32:24: note: previous implicit declaration of 'inb' was here | unsigned char value = inb(port); `----

Cheers, Sven

On Wed, Aug 15, 2018 at 11:22 AM Sebastian Gottschall s.gottschall@dd-wrt.com wrote:

btw. i found another compile error with x86 :-)

This should already be fixed by commit d0055f351e64 ("x86/smp: fix non-SMP broken build due to redefinition of apic_id_is_primary_thread").

Linus

On Wed, Aug 15, 2018 at 08:33:30PM +0200, Sebastian Gottschall wrote:

Am 15.08.2018 um 20:26 schrieb Linus Torvalds:

...

On Wed, Aug 15, 2018 at 11:22 AM Sebastian Gottschall s.gottschall@dd-wrt.com wrote:

...

btw. i found another compile error with x86 :-)

This should already be fixed by commit d0055f351e64 ("x86/smp: fix non-SMP broken build due to redefinition of apic_id_is_primary_thread").

       Linus

Hello Linus

unfortunatly there are numerous other non smp compile issues. this patch covers just one out of 3 found so far

Do you have patches for the others? Non smp builds are pretty rare these days :)

thanks,

greg k-h

On Wed, Aug 15, 2018 at 08:27:00PM +0200, Sebastian Gottschall wrote:

if i fix the other error (can be reproduced with disable smp on standard i386 build)

another raises up again related to smp. to be serious. this patchset of x86 patches is absolutelly broken and put together without any care. not a simple compile test has been done

sorry for beeing a little bit upset. i'm sure i will find other bugs if i go deeper

It might possibly help if you would consider to publish your configuration(s).

As for "not a simple compile test has been done", please have a look at http://kerneltests.org/builders. I guess those compile (and boot) tests don't count as "simple". I assume the same applies to all the builds done by 0day and all the other test builders.

Guenter

Am 15.08.2018 um 21:42 schrieb Linus Torvalds:

On Wed, Aug 15, 2018 at 12:26 PM Guenter Roeck linux@roeck-us.net wrote:

...

Also seen in mainline. Meaning my non-SMP test builds are broken. Hmm.

Grr. I think it's due mainly due to commit 447ae3166702 ("x86: Don't include linux/irq.h from asm/hardirq.h"), which exposes a number of files that had some dodgy include file dependenencies, and just randomly happened to get it right because of that odd include that caused problems.

correct.

That commit itself is definitely the right thing to do, but yes, we clearly had a lot of cases of things getting core header files included purely by luck.

And because this was all done under embargo, we didn't get the kind of test robot coverage we usually get.

Part of it can also be due to subtle merge issues - even if the original branch got good coverage, later changes sometimes ended up adding things like that.

For example, my merge of the L1TF code found that in the meantime, arch/x86/kernel/kvmclock.c had added a call to kzalloc(), which used to work just fine, but with the header cleanup it turned out that kvmclock.c had never included <linux/slab.h>, so now it didn't build right.

And that was just the one I noticed because of my limited build tests.

And yes, every single developer has CONFIG_SMP in their config, but perhaps equally importantly, I suspect CONFIG_SMP ends up getting more header files included almost by mistake, so it can _continue_ to hide these kinds of incomplete header file includes that just happen to work.

...

Anyway, care to send a proper patch ? I am sure Linus will be more than happy to apply it.

I think "happy" is too strong a word for my state of mind with all this, but yes, I'll apply it, and you'll get the glory of fixing some configuration that clearly didn't get properly tested.

In the meantime, maybe I should just do a "make allmodconfig" and then disable SMP and see if that shows anything for me.

if it would be helpfull i can do this quick here using the latest vanilla tree

Sebastian

           Linus

Am 15.08.2018 um 22:42 schrieb Linus Torvalds:

It would definitely be helpful, since I'm already quite busy with the normal merge window work. Having the embargo for L1TF end just as the merge window opened wasn't all that convenient.

all good so far. i checked them all and i will see how my nightly build goes out which does compilation on alot of mips, arm, powerpc etc. platforms

Of course, I have my own "I will delay 4.18 by a week" to blame for the timing.

So if you can find the places that break, and add the right header, and send me a patch for all the ones you find, that would be lovely.

will do so. but for now it seems that this was really the only one

It's not necessarilky always <linux/io.h> or <linux/irq.h> that makes sense to add as an include. As mentioned, sometimes it's <linux/slab.h> or similar.

mmh why? i prefer to include the correct header which references the function or macro instead of adding a header which finds the right way by multiple iterations. its more transparent

Sebastian

On Wed, Aug 15, 2018 at 08:27:00PM +0200, Sebastian Gottschall wrote:

if i fix the other error (can be reproduced with disable smp on standard i386 build)

another raises up again related to smp. to be serious. this patchset of x86 patches is absolutelly broken and put together without any care. not a simple compile test has been done

That's just not true at all. Gift horse, mouth.

sorry for beeing a little bit upset. i'm sure i will find other bugs if i go deeper

Testing is always great to have. Perhaps just move to 4.14.y or Linus's tree if 4.9 does not work for you? THere's only so much testing we could do before the embargo, sorry. We are very lucky to even have these patches for 4.9 in the first place.

thanks,

greg k-h