mmotm 2018-10-30-16-08 uploaded

mmotm 2018-10-30-16-08 uploaded

[akpm]

The mm-of-the-moment snapshot 2018-10-30-16-08 has been uploaded to

   http://www.ozlabs.org/~akpm/mmotm/

mmotm-readme.txt says

README for mm-of-the-moment:

http://www.ozlabs.org/~akpm/mmotm/

This is a snapshot of my -mm patch queue.  Uploaded at random hopefully
more than once a week.

You will need quilt to apply these patches to the latest Linus release (4.x
or 4.x-rcY).  The series file is in broken-out.tar.gz and is duplicated in
http://ozlabs.org/~akpm/mmotm/series

The file broken-out.tar.gz contains two datestamp files: .DATE and
.DATE-yyyy-mm-dd-hh-mm-ss.  Both contain the string yyyy-mm-dd-hh-mm-ss,
followed by the base kernel version against which this patch series is to
be applied.

This tree is partially included in linux-next.  To see which patches are
included in linux-next, consult the `series' file.  Only the patches
within the #NEXT_PATCHES_START/#NEXT_PATCHES_END markers are included in
linux-next.

A git tree which contains the memory management portion of this tree is
maintained at git://git.kernel.org/pub/scm/linux/kernel/git/mhocko/mm.git
by Michal Hocko.  It contains the patches which are between the
"#NEXT_PATCHES_START mm" and "#NEXT_PATCHES_END" markers, from the series
file, http://www.ozlabs.org/~akpm/mmotm/series.


A full copy of the full kernel tree with the linux-next and mmotm patches
already applied is available through git within an hour of the mmotm
release.  Individual mmotm releases are tagged.  The master branch always
points to the latest release, so it's constantly rebasing.

http://git.cmpxchg.org/cgit.cgi/linux-mmotm.git/

To develop on top of mmotm git:

  $ git remote add mmotm git://git.kernel.org/pub/scm/linux/kernel/git/mhocko/mm.git
  $ git remote update mmotm
  $ git checkout -b topic mmotm/master
  <make changes, commit>
  $ git send-email mmotm/master.. [...]

To rebase a branch with older patches to a new mmotm release:

  $ git remote update mmotm
  $ git rebase --onto mmotm/master <topic base> topic




The directory http://www.ozlabs.org/~akpm/mmots/ (mm-of-the-second)
contains daily snapshots of the -mm tree.  It is updated more frequently
than mmotm, and is untested.

A git copy of this tree is available at

	http://git.cmpxchg.org/cgit.cgi/linux-mmots.git/

and use of this tree is similar to
http://git.cmpxchg.org/cgit.cgi/linux-mmotm.git/, described above.


This mmotm tree contains the following patches against 4.19:
(patches marked "*" will be included in linux-next)

  origin.patch
* mm-hmm-fix-utf8.patch
* mm-rmap-map_pte-was-not-handling-private-zone_device-page-properly-v3.patch
* mm-hmm-fix-race-between-hmm_mirror_unregister-and-mmu_notifier-callback.patch
* mm-hmm-properly-handle-migration-pmd-v3.patch
* mm-hmm-use-a-structure-for-update-callback-parameters-v2.patch
* mm-hmm-invalidate-device-page-table-at-start-of-invalidation.patch
* mm-gup_benchmark-prevent-integer-overflow-in-ioctl.patch
* fs-proc-vmcorec-convert-to-use-vmf_error.patch
* include-linux-compilerh-add-version-detection-to-asm_volatile_goto.patch
* add-oleksij-rempel-to-mailmap.patch
* treewide-remove-current_text_addr.patch
* error-injection-remove-meaningless-null-pointer-check-before-debugfs_remove_recursive.patch
* lib-bitmapc-remove-wrong-documentation.patch
* linux-bitmaph-handle-constant-zero-size-bitmaps-correctly.patch
* linux-bitmaph-remove-redundant-uses-of-small_const_nbits.patch
* linux-bitmaph-fix-type-of-nbits-in-bitmap_shift_right.patch
* linux-bitmaph-relax-comment-on-compile-time-constant-nbits.patch
* lib-bitmapc-fix-remaining-space-computation-in-bitmap_print_to_pagebuf.patch
* lib-bitmapc-simplify-bitmap_print_to_pagebuf.patch
* lib-parserc-switch-match_strdup-over-to-use-kmemdup_nul.patch
* lib-parserc-switch-match_u64int-over-to-use-match_strdup.patch
* lib-parserc-switch-match_number-over-to-use-match_strdup.patch
* zlib-remove-fall-through-warnings.patch
* lib-sg_pool-remove-unnecessary-null-check-when-free-the-object.patch
* lib-rbtreec-fix-typo-in-comment-of-rb_insert_augmented.patch
* kstrtox-delete-unnecessary-casts.patch
* compat-mark-expected-switch-fall-throughs.patch
* checkpatch-remove-gcc_binary_constant-warning.patch
* init-do_mountsc-add-root=partlabel=name-support.patch
* hfsplus-prevent-btree-data-loss-on-root-split.patch
* hfsplus-fix-bug-on-bnode-parent-update.patch
* hfs-prevent-btree-data-loss-on-root-split.patch
* hfs-fix-bug-on-bnode-parent-update.patch
* hfsplus-prevent-btree-data-loss-on-enospc.patch
* hfs-prevent-btree-data-loss-on-enospc.patch
* hfsplus-fix-return-value-of-hfsplus_get_block.patch
* hfs-fix-return-value-of-hfs_get_block.patch
* hfsplus-update-timestamps-on-truncate.patch
* hfs-update-timestamp-on-truncate.patch
* hfs-fix-array-out-of-bounds-read-of-array-extent.patch
* reiserfs-propagate-errors-from-fill_with_dentries-properly.patch
* reiserfs-remove-workaround-code-for-gcc-3x.patch
* fat-expand-a-slightly-out-of-date-comment.patch
* fat-create-a-function-to-calculate-the-timezone-offest.patch
* fat-add-functions-to-update-and-truncate-timestamps-appropriately.patch
* fat-change-timestamp-updates-to-use-fat_truncate_time.patch
* fat-truncate-inode-timestamp-updates-in-setattr.patch
* kernel-fix-a-comment-error.patch
* signal-mark-expected-switch-fall-throughs.patch
* kernel-panic-do-not-append-newline-to-the-stack-protector-panic-string.patch
* kernel-panic-filter-out-a-potential-trailing-newline.patch
* ipc-ipcmni-limit-check-for-msgmni-and-shmmni.patch
* ipc-ipcmni-limit-check-for-semmni.patch
* lib-lz4-update-lz4-decompressor-module.patch
* kbuild-fix-kernel-boundsc-w=1-warning.patch
* percpu-cleanup-per_cpu_def_attributes-macro.patch
* mm-remove-config_no_bootmem.patch
* mm-remove-config_have_memblock.patch
* mm-remove-bootmem-allocator-implementation.patch
* mm-nobootmem-remove-dead-code.patch
* memblock-rename-memblock_alloc_nid_try_nid-to-memblock_phys_alloc.patch
* memblock-remove-_virt-from-apis-returning-virtual-address.patch
* memblock-replace-alloc_bootmem_align-with-memblock_alloc.patch
* memblock-replace-alloc_bootmem_low-with-memblock_alloc_low.patch
* memblock-replace-__alloc_bootmem_node_nopanic-with-memblock_alloc_try_nid_nopanic.patch
* memblock-replace-alloc_bootmem_pages_nopanic-with-memblock_alloc_nopanic.patch
* memblock-replace-alloc_bootmem_low-with-memblock_alloc_low-2.patch
* memblock-replace-__alloc_bootmem_nopanic-with-memblock_alloc_from_nopanic.patch
* memblock-add-align-parameter-to-memblock_alloc_node.patch
* memblock-replace-alloc_bootmem_pages_node-with-memblock_alloc_node.patch
* memblock-replace-__alloc_bootmem_node-with-appropriate-memblock_-api.patch
* memblock-replace-alloc_bootmem_node-with-memblock_alloc_node.patch
* memblock-replace-alloc_bootmem_low_pages-with-memblock_alloc_low.patch
* memblock-replace-alloc_bootmem_pages-with-memblock_alloc.patch
* memblock-replace-__alloc_bootmem-with-memblock_alloc_from.patch
* memblock-replace-alloc_bootmem-with-memblock_alloc.patch
* mm-nobootmem-remove-bootmem-allocation-apis.patch
* memblock-replace-free_bootmem_node-with-memblock_free.patch
* memblock-replace-free_bootmem_late-with-memblock_free_late.patch
* memblock-rename-free_all_bootmem-to-memblock_free_all.patch
* memblock-rename-__free_pages_bootmem-to-memblock_free_pages.patch
* mm-remove-nobootmem.patch
* memblock-replace-bootmem_alloc_-with-memblock-variants.patch
* mm-remove-include-linux-bootmemh.patch
* docs-boot-time-mm-remove-bootmem-documentation.patch
* memblock-stop-using-implicit-alignement-to-smp_cache_bytes.patch
* memblock-warn-if-zero-alignment-was-requested.patch
* android-binder-replace-vm_insert_page-with-vmf_insert_page.patch
* mm-memory_hotplug-make-remove_memory-take-the-device_hotplug_lock.patch
* mm-memory_hotplug-make-add_memory-take-the-device_hotplug_lock.patch
* mm-memory_hotplug-fix-online-offline_pages-called-wo-mem_hotplug_lock.patch
* powerpc-powernv-hold-device_hotplug_lock-when-calling-device_online.patch
* powerpc-powernv-hold-device_hotplug_lock-when-calling-memtrace_offline_pages.patch
* memory-hotplugtxt-add-some-details-about-locking-internals.patch
* mm-fix-warning-in-insert_pfn.patch
* mm-fix-__get_user_pages_fast-comment.patch
* mm-handle-no-memcg-case-in-memcg_kmem_charge-properly.patch
* kernel-srcu-fix-ctags.patch
* mm-dont-reclaim-inodes-with-many-attached-pages.patch
* mm-thp-always-specify-disabled-vmas-as-nh-in-smaps.patch
* mm-thp-relax-__gfp_thisnode-for-madv_hugepage-mappings.patch
* arm-arch-arm-include-asm-pageh-needs-personalityh.patch
* ocfs2-get-rid-of-ocfs2_is_o2cb_active-function.patch
* ocfs2-without-quota-support-try-to-avoid-calling-quota-recovery.patch
* ocfs2-dont-use-iocb-when-eiocbqueued-returns.patch
* ocfs2-fix-a-misuse-a-of-brelse-after-failing-ocfs2_check_dir_entry.patch
* ocfs2-dont-put-and-assigning-null-to-bh-allocated-outside.patch
* ocfs2-dlmglue-clean-up-timestamp-handling.patch
* fix-dead-lock-caused-by-ocfs2_defrag_extent.patch
* ocfs2-fix-dead-lock-caused-by-ocfs2_defrag_extent.patch
* fix-clusters-leak-in-ocfs2_defrag_extent.patch
* fix-clusters-leak-in-ocfs2_defrag_extent-fix.patch
* block-restore-proc-partitions-to-not-display-non-partitionable-removable-devices.patch
* vfs-allow-dedupe-of-user-owned-read-only-files.patch
* vfs-dedupe-should-return-eperm-if-permission-is-not-granted.patch
  mm.patch
* mm-thp-consolidate-thp-gfp-handling-into-alloc_hugepage_direct_gfpmask.patch
* memory_hotplug-free-pages-as-higher-order.patch
* memory_hotplug-free-pages-as-higher-order-fix.patch
* mm-page_alloc-remove-software-prefetching-in-__free_pages_core.patch
* z3fold-fix-wrong-handling-of-headless-pages.patch
* mm-swap-fix-race-between-swapoff-and-some-swap-operations.patch
* mm-swap-fix-race-between-swapoff-and-some-swap-operations-v6.patch
* mm-fix-race-between-swapoff-and-mincore.patch
* list_lru-prefetch-neighboring-list-entries-before-acquiring-lock.patch
* list_lru-prefetch-neighboring-list-entries-before-acquiring-lock-fix.patch
* mm-add-strictlimit-knob-v2.patch
* mm-dont-expose-page-to-fast-gup-before-its-ready.patch
* mm-page_owner-align-with-pageblock_nr_pages.patch
* mm-page_owner-align-with-pageblock_nr-pages.patch
* info-task-hung-in-generic_file_write_iter.patch
* kernel-kexec_file-remove-some-duplicated-include-file.patch
* kernel-sysctlc-remove-duplicated-include.patch
* bfs-add-sanity-check-at-bfs_fill_super.patch
* ipc-allow-boot-time-extension-of-ipcmni-from-32k-to-8m.patch
* ipc-allow-boot-time-extension-of-ipcmni-from-32k-to-8m-checkpatch-fixes.patch
* ipc-conserve-sequence-numbers-in-extended-ipcmni-mode.patch
  linux-next.patch
  linux-next-rejects.patch
* vfs-replace-current_kernel_time64-with-ktime-equivalent.patch
* fix-read-buffer-overflow-in-delta-ipc.patch
  make-sure-nobodys-leaking-resources.patch
  releasing-resources-with-children.patch
  mutex-subsystem-synchro-test-module.patch
  kernel-forkc-export-kernel_thread-to-modules.patch
  slab-leaks3-default-y.patch
  workaround-for-a-pci-restoring-bug.patch

Re: mmotm 2018-10-30-16-08 uploaded (arch/x86/kernel/vsmp_64.c)

[Randy Dunlap]

On 10/30/18 4:09 PM, akpm@linux-foundation.org wrote:
> The mm-of-the-moment snapshot 2018-10-30-16-08 has been uploaded to
> 
>    http://www.ozlabs.org/~akpm/mmotm/
> 
> mmotm-readme.txt says
> 
> README for mm-of-the-moment:
> 
> http://www.ozlabs.org/~akpm/mmotm/
> 
> This is a snapshot of my -mm patch queue.  Uploaded at random hopefully
> more than once a week.


Build error on x86_64 from origin.patch (i.e., not mmotm)
when CONFIG_PCI is not enabled:
Oh:  CONFIG_X86_VSMP is also not enabled, but
arch/x86/kernel/Makefile always tries to build vsmp_64.o.


ld: arch/x86/kernel/vsmp_64.o: in function `vsmp_cap_cpus':
vsmp_64.c:(.init.text+0x1e): undefined reference to `read_pci_config'


-- 
~Randy

[PATCH] x86/build: Build VSMP support only if selected

[Thomas Gleixner]

VSMP support is built even if CONFIG_X86_VSMP is not set. This leads to a build
breakage when CONFIG_PCI is disabled as well.

Build VSMP code only when selected.

Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
diff --git a/arch/x86/include/asm/setup.h b/arch/x86/include/asm/setup.h
index ae13bc974416..b6b911c4c7f3 100644
--- a/arch/x86/include/asm/setup.h
+++ b/arch/x86/include/asm/setup.h
@@ -33,7 +33,7 @@
 extern u64 relocated_ramdisk;
 
 /* Interrupt control for vSMPowered x86_64 systems */
-#ifdef CONFIG_X86_64
+#if defined(CONFIG_X86_64) && defined(CONFIG_X86_VSMP)
 void vsmp_init(void);
 #else
 static inline void vsmp_init(void) { }
diff --git a/arch/x86/kernel/Makefile b/arch/x86/kernel/Makefile
index 8824d01c0c35..647ce52b17d5 100644
--- a/arch/x86/kernel/Makefile
+++ b/arch/x86/kernel/Makefile
@@ -148,5 +148,5 @@ ifeq ($(CONFIG_X86_64),y)
 	obj-$(CONFIG_CALGARY_IOMMU)	+= pci-calgary_64.o tce_64.o
 
 	obj-$(CONFIG_MMCONF_FAM10H)	+= mmconf-fam10h_64.o
-	obj-y				+= vsmp_64.o
+	obj-$(CONFIG_X86_VSMP)		+= vsmp_64.o
 endif

RE: [PATCH] x86/build: Build VSMP support only if selected

[Shai Fultheim (Shai@ScaleMP.com)]

On 01/11/18 11:37, Thomas Gleixner wrote:

> VSMP support is built even if CONFIG_X86_VSMP is not set. This leads to a build
> breakage when CONFIG_PCI is disabled as well.
> 
> Build VSMP code only when selected.

This patch disables detect_vsmp_box() on systems without CONFIG_X86_VSMP, due to
the recent 6da63eb241a05b0e676d68975e793c0521387141.  This is significant
regression that will affect significant number of deployments.

We will reply shortly with an updated patch that fix the dependency on pv_irq_ops,
and revert to CONFIG_PARAVIRT, with proper protection for CONFIG_PCI.

Re: [PATCH] x86/build: Build VSMP support only if selected

[Eial Czerwacki]

Greetings,

On 11/01/2018 12:39 PM, Shai Fultheim (Shai@ScaleMP.com) wrote:
> On 01/11/18 11:37, Thomas Gleixner wrote:
> 
>> VSMP support is built even if CONFIG_X86_VSMP is not set. This leads to a build
>> breakage when CONFIG_PCI is disabled as well.
>>
>> Build VSMP code only when selected.
> 
> This patch disables detect_vsmp_box() on systems without CONFIG_X86_VSMP, due to
> the recent 6da63eb241a05b0e676d68975e793c0521387141.  This is significant
> regression that will affect significant number of deployments.
> 
> We will reply shortly with an updated patch that fix the dependency on pv_irq_ops,
> and revert to CONFIG_PARAVIRT, with proper protection for CONFIG_PCI.
> 

here is the proper patch which fixes the issue on hand:
>From ebff534f8cfa55d7c3ab798c44abe879f3fbe2b8 Mon Sep 17 00:00:00 2001
From: Eial Czerwacki <eial@scalemp.com>
Date: Thu, 1 Nov 2018 15:08:32 +0200
Subject: [PATCH] x86/build: Build VSMP support only if CONFIG_PCI is
selected

vsmp dependency of pv_irq_ops removed some years ago, so now let's clean
it up from vsmp_64.c.

In short, "cap & ctl & (1 << 4)" was always returning 0, as such we can
remove all the PARAVIRT/PARAVIRT_XXL code handling that.

However, the rest of the code depends on CONFIG_PCI, so fix it accordingly.

Signed-off-by: Eial Czerwacki <eial@scalemp.com>
Acked-by: Shai Fultheim <shai@scalemp.com>
---
 arch/x86/Kconfig          |  1 -
 arch/x86/kernel/vsmp_64.c | 80
+++--------------------------------------------
 2 files changed, 5 insertions(+), 76 deletions(-)

diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
index c51c989..4b187ca 100644
--- a/arch/x86/Kconfig
+++ b/arch/x86/Kconfig
@@ -524,7 +524,6 @@ config X86_VSMP
        bool "ScaleMP vSMP"
        select HYPERVISOR_GUEST
        select PARAVIRT
-       select PARAVIRT_XXL
        depends on X86_64 && PCI
        depends on X86_EXTENDED_PLATFORM
        depends on SMP
diff --git a/arch/x86/kernel/vsmp_64.c b/arch/x86/kernel/vsmp_64.c
index 1eae5af..c6d2b76 100644
--- a/arch/x86/kernel/vsmp_64.c
+++ b/arch/x86/kernel/vsmp_64.c
@@ -26,64 +26,7 @@

 #define TOPOLOGY_REGISTER_OFFSET 0x10

-#if defined CONFIG_PCI && defined CONFIG_PARAVIRT_XXL
-/*
- * Interrupt control on vSMPowered systems:
- * ~AC is a shadow of IF.  If IF is 'on' AC should be 'off'
- * and vice versa.
- */
-
-asmlinkage __visible unsigned long vsmp_save_fl(void)
-{
-       unsigned long flags = native_save_fl();
-
-       if (!(flags & X86_EFLAGS_IF) || (flags & X86_EFLAGS_AC))
-               flags &= ~X86_EFLAGS_IF;
-       return flags;
-}
-PV_CALLEE_SAVE_REGS_THUNK(vsmp_save_fl);
-
-__visible void vsmp_restore_fl(unsigned long flags)
-{
-       if (flags & X86_EFLAGS_IF)
-               flags &= ~X86_EFLAGS_AC;
-       else
-               flags |= X86_EFLAGS_AC;
-       native_restore_fl(flags);
-}
-PV_CALLEE_SAVE_REGS_THUNK(vsmp_restore_fl);
-
-asmlinkage __visible void vsmp_irq_disable(void)
-{
-       unsigned long flags = native_save_fl();
-
-       native_restore_fl((flags & ~X86_EFLAGS_IF) | X86_EFLAGS_AC);
-}
-PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_disable);
-
-asmlinkage __visible void vsmp_irq_enable(void)
-{
-       unsigned long flags = native_save_fl();
-
-       native_restore_fl((flags | X86_EFLAGS_IF) & (~X86_EFLAGS_AC));
-}
-PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_enable);
-
-static unsigned __init vsmp_patch(u8 type, void *ibuf,
-                                 unsigned long addr, unsigned len)
-{
-       switch (type) {
-       case PARAVIRT_PATCH(irq.irq_enable):
-       case PARAVIRT_PATCH(irq.irq_disable):
-       case PARAVIRT_PATCH(irq.save_fl):
-       case PARAVIRT_PATCH(irq.restore_fl):
-               return paravirt_patch_default(type, ibuf, addr, len);
-       default:
-               return native_patch(type, ibuf, addr, len);
-       }
-
-}
-
+#if defined CONFIG_PCI
 static void __init set_vsmp_pv_ops(void)
 {
        void __iomem *address;
@@ -109,28 +52,12 @@ static void __init set_vsmp_pv_ops(void)
        }
 #endif

-       if (cap & ctl & (1 << 4)) {
-               /* Setup irq ops and turn on vSMP  IRQ fastpath handling */
-               pv_ops.irq.irq_disable = PV_CALLEE_SAVE(vsmp_irq_disable);
-               pv_ops.irq.irq_enable = PV_CALLEE_SAVE(vsmp_irq_enable);
-               pv_ops.irq.save_fl = PV_CALLEE_SAVE(vsmp_save_fl);
-               pv_ops.irq.restore_fl = PV_CALLEE_SAVE(vsmp_restore_fl);
-               pv_ops.init.patch = vsmp_patch;
-               ctl &= ~(1 << 4);
-       }
        writel(ctl, address + 4);
        ctl = readl(address + 4);
        pr_info("vSMP CTL: control set to:0x%08x\n", ctl);

        early_iounmap(address, 8);
 }
-#else
-static void __init set_vsmp_pv_ops(void)
-{
-}
-#endif
-
-#ifdef CONFIG_PCI
 static int is_vsmp = -1;

 static void __init detect_vsmp_box(void)
@@ -164,11 +91,14 @@ static int is_vsmp_box(void)
 {
        return 0;
 }
+static void __init set_vsmp_pv_ops(void)
+{
+}
 #endif

 static void __init vsmp_cap_cpus(void)
 {
-#if !defined(CONFIG_X86_VSMP) && defined(CONFIG_SMP)
+#if !defined(CONFIG_X86_VSMP) && defined(CONFIG_SMP) && defined(CONFIG_PCI)
        void __iomem *address;
        unsigned int cfg, topology, node_shift, maxcpus;

-- 
2.7.4

Re: [PATCH] x86/build: Build VSMP support only if selected

[Thomas Gleixner]

On Thu, 1 Nov 2018, Eial Czerwacki wrote:

> Greetings,
> 
> On 11/01/2018 12:39 PM, Shai Fultheim (Shai@ScaleMP.com) wrote:
> > On 01/11/18 11:37, Thomas Gleixner wrote:
> > 
> >> VSMP support is built even if CONFIG_X86_VSMP is not set. This leads to a build
> >> breakage when CONFIG_PCI is disabled as well.
> >>
> >> Build VSMP code only when selected.
> > 
> > This patch disables detect_vsmp_box() on systems without CONFIG_X86_VSMP, due to
> > the recent 6da63eb241a05b0e676d68975e793c0521387141.  This is significant
> > regression that will affect significant number of deployments.
> > 
> > We will reply shortly with an updated patch that fix the dependency on pv_irq_ops,
> > and revert to CONFIG_PARAVIRT, with proper protection for CONFIG_PCI.
> > 
> 
> here is the proper patch which fixes the issue on hand:
> >From ebff534f8cfa55d7c3ab798c44abe879f3fbe2b8 Mon Sep 17 00:00:00 2001
> From: Eial Czerwacki <eial@scalemp.com>
> Date: Thu, 1 Nov 2018 15:08:32 +0200
> Subject: [PATCH] x86/build: Build VSMP support only if CONFIG_PCI is
> selected
> 
> vsmp dependency of pv_irq_ops removed some years ago, so now let's clean
> it up from vsmp_64.c.
> 
> In short, "cap & ctl & (1 << 4)" was always returning 0, as such we can
> remove all the PARAVIRT/PARAVIRT_XXL code handling that.
> 
> However, the rest of the code depends on CONFIG_PCI, so fix it accordingly.
> 
> Signed-off-by: Eial Czerwacki <eial@scalemp.com>
> Acked-by: Shai Fultheim <shai@scalemp.com>

Nice cleanup!

Acked-by: Thomas Gleixner <tglx@linutronix.de>

Re: [PATCH] x86/build: Build VSMP support only if selected

[Juergen Gross]

On 01/11/2018 14:10, Eial Czerwacki wrote:
> Greetings,
> 
> On 11/01/2018 12:39 PM, Shai Fultheim (Shai@ScaleMP.com) wrote:
>> On 01/11/18 11:37, Thomas Gleixner wrote:
>>
>>> VSMP support is built even if CONFIG_X86_VSMP is not set. This leads to a build
>>> breakage when CONFIG_PCI is disabled as well.
>>>
>>> Build VSMP code only when selected.
>>
>> This patch disables detect_vsmp_box() on systems without CONFIG_X86_VSMP, due to
>> the recent 6da63eb241a05b0e676d68975e793c0521387141.  This is significant
>> regression that will affect significant number of deployments.
>>
>> We will reply shortly with an updated patch that fix the dependency on pv_irq_ops,
>> and revert to CONFIG_PARAVIRT, with proper protection for CONFIG_PCI.
>>
> 
> here is the proper patch which fixes the issue on hand:
> From ebff534f8cfa55d7c3ab798c44abe879f3fbe2b8 Mon Sep 17 00:00:00 2001
> From: Eial Czerwacki <eial@scalemp.com>
> Date: Thu, 1 Nov 2018 15:08:32 +0200
> Subject: [PATCH] x86/build: Build VSMP support only if CONFIG_PCI is
> selected
> 
> vsmp dependency of pv_irq_ops removed some years ago, so now let's clean
> it up from vsmp_64.c.
> 
> In short, "cap & ctl & (1 << 4)" was always returning 0, as such we can
> remove all the PARAVIRT/PARAVIRT_XXL code handling that.
> 
> However, the rest of the code depends on CONFIG_PCI, so fix it accordingly.
> 
> Signed-off-by: Eial Czerwacki <eial@scalemp.com>
> Acked-by: Shai Fultheim <shai@scalemp.com>
> ---
>  arch/x86/Kconfig          |  1 -
>  arch/x86/kernel/vsmp_64.c | 80
> +++--------------------------------------------
>  2 files changed, 5 insertions(+), 76 deletions(-)
> 
> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
> index c51c989..4b187ca 100644
> --- a/arch/x86/Kconfig
> +++ b/arch/x86/Kconfig
> @@ -524,7 +524,6 @@ config X86_VSMP
>         bool "ScaleMP vSMP"
>         select HYPERVISOR_GUEST
>         select PARAVIRT

Do you really still need PARAVIRT and HYPERVISOR_GUEST?
Maybe you want IRQ_REMAP instead?

> -       select PARAVIRT_XXL
>         depends on X86_64 && PCI
>         depends on X86_EXTENDED_PLATFORM
>         depends on SMP
> diff --git a/arch/x86/kernel/vsmp_64.c b/arch/x86/kernel/vsmp_64.c
> index 1eae5af..c6d2b76 100644
> --- a/arch/x86/kernel/vsmp_64.c
> +++ b/arch/x86/kernel/vsmp_64.c
> @@ -26,64 +26,7 @@
> 
>  #define TOPOLOGY_REGISTER_OFFSET 0x10
> 
> -#if defined CONFIG_PCI && defined CONFIG_PARAVIRT_XXL
> -/*
> - * Interrupt control on vSMPowered systems:
> - * ~AC is a shadow of IF.  If IF is 'on' AC should be 'off'
> - * and vice versa.
> - */
> -
> -asmlinkage __visible unsigned long vsmp_save_fl(void)
> -{
> -       unsigned long flags = native_save_fl();
> -
> -       if (!(flags & X86_EFLAGS_IF) || (flags & X86_EFLAGS_AC))
> -               flags &= ~X86_EFLAGS_IF;
> -       return flags;
> -}
> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_save_fl);
> -
> -__visible void vsmp_restore_fl(unsigned long flags)
> -{
> -       if (flags & X86_EFLAGS_IF)
> -               flags &= ~X86_EFLAGS_AC;
> -       else
> -               flags |= X86_EFLAGS_AC;
> -       native_restore_fl(flags);
> -}
> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_restore_fl);
> -
> -asmlinkage __visible void vsmp_irq_disable(void)
> -{
> -       unsigned long flags = native_save_fl();
> -
> -       native_restore_fl((flags & ~X86_EFLAGS_IF) | X86_EFLAGS_AC);
> -}
> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_disable);
> -
> -asmlinkage __visible void vsmp_irq_enable(void)
> -{
> -       unsigned long flags = native_save_fl();
> -
> -       native_restore_fl((flags | X86_EFLAGS_IF) & (~X86_EFLAGS_AC));
> -}
> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_enable);
> -
> -static unsigned __init vsmp_patch(u8 type, void *ibuf,
> -                                 unsigned long addr, unsigned len)
> -{
> -       switch (type) {
> -       case PARAVIRT_PATCH(irq.irq_enable):
> -       case PARAVIRT_PATCH(irq.irq_disable):
> -       case PARAVIRT_PATCH(irq.save_fl):
> -       case PARAVIRT_PATCH(irq.restore_fl):
> -               return paravirt_patch_default(type, ibuf, addr, len);
> -       default:
> -               return native_patch(type, ibuf, addr, len);
> -       }
> -
> -}
> -
> +#if defined CONFIG_PCI
>  static void __init set_vsmp_pv_ops(void)

Wouldn't be a rename of the function be appropriate now?


Juergen

Re: [PATCH] x86/build: Build VSMP support only if selected

[Eial Czerwacki]

Greetings,

On 11/01/2018 03:45 PM, Juergen Gross wrote:
> On 01/11/2018 14:10, Eial Czerwacki wrote:
>> Greetings,
>>
>> On 11/01/2018 12:39 PM, Shai Fultheim (Shai@ScaleMP.com) wrote:
>>> On 01/11/18 11:37, Thomas Gleixner wrote:
>>>
>>>> VSMP support is built even if CONFIG_X86_VSMP is not set. This leads to a build
>>>> breakage when CONFIG_PCI is disabled as well.
>>>>
>>>> Build VSMP code only when selected.
>>>
>>> This patch disables detect_vsmp_box() on systems without CONFIG_X86_VSMP, due to
>>> the recent 6da63eb241a05b0e676d68975e793c0521387141.  This is significant
>>> regression that will affect significant number of deployments.
>>>
>>> We will reply shortly with an updated patch that fix the dependency on pv_irq_ops,
>>> and revert to CONFIG_PARAVIRT, with proper protection for CONFIG_PCI.
>>>
>>
>> here is the proper patch which fixes the issue on hand:
>> From ebff534f8cfa55d7c3ab798c44abe879f3fbe2b8 Mon Sep 17 00:00:00 2001
>> From: Eial Czerwacki <eial@scalemp.com>
>> Date: Thu, 1 Nov 2018 15:08:32 +0200
>> Subject: [PATCH] x86/build: Build VSMP support only if CONFIG_PCI is
>> selected
>>
>> vsmp dependency of pv_irq_ops removed some years ago, so now let's clean
>> it up from vsmp_64.c.
>>
>> In short, "cap & ctl & (1 << 4)" was always returning 0, as such we can
>> remove all the PARAVIRT/PARAVIRT_XXL code handling that.
>>
>> However, the rest of the code depends on CONFIG_PCI, so fix it accordingly.
>>
>> Signed-off-by: Eial Czerwacki <eial@scalemp.com>
>> Acked-by: Shai Fultheim <shai@scalemp.com>
>> ---
>>  arch/x86/Kconfig          |  1 -
>>  arch/x86/kernel/vsmp_64.c | 80
>> +++--------------------------------------------
>>  2 files changed, 5 insertions(+), 76 deletions(-)
>>
>> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
>> index c51c989..4b187ca 100644
>> --- a/arch/x86/Kconfig
>> +++ b/arch/x86/Kconfig
>> @@ -524,7 +524,6 @@ config X86_VSMP
>>         bool "ScaleMP vSMP"
>>         select HYPERVISOR_GUEST
>>         select PARAVIRT
> 
> Do you really still need PARAVIRT and HYPERVISOR_GUEST?
> Maybe you want IRQ_REMAP instead?
> 
Better performance is achieved with PARAVIRTed kernel.   Hence we keep
them both in.

>> -       select PARAVIRT_XXL
>>         depends on X86_64 && PCI
>>         depends on X86_EXTENDED_PLATFORM
>>         depends on SMP
>> diff --git a/arch/x86/kernel/vsmp_64.c b/arch/x86/kernel/vsmp_64.c
>> index 1eae5af..c6d2b76 100644
>> --- a/arch/x86/kernel/vsmp_64.c
>> +++ b/arch/x86/kernel/vsmp_64.c
>> @@ -26,64 +26,7 @@
>>
>>  #define TOPOLOGY_REGISTER_OFFSET 0x10
>>
>> -#if defined CONFIG_PCI && defined CONFIG_PARAVIRT_XXL
>> -/*
>> - * Interrupt control on vSMPowered systems:
>> - * ~AC is a shadow of IF.  If IF is 'on' AC should be 'off'
>> - * and vice versa.
>> - */
>> -
>> -asmlinkage __visible unsigned long vsmp_save_fl(void)
>> -{
>> -       unsigned long flags = native_save_fl();
>> -
>> -       if (!(flags & X86_EFLAGS_IF) || (flags & X86_EFLAGS_AC))
>> -               flags &= ~X86_EFLAGS_IF;
>> -       return flags;
>> -}
>> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_save_fl);
>> -
>> -__visible void vsmp_restore_fl(unsigned long flags)
>> -{
>> -       if (flags & X86_EFLAGS_IF)
>> -               flags &= ~X86_EFLAGS_AC;
>> -       else
>> -               flags |= X86_EFLAGS_AC;
>> -       native_restore_fl(flags);
>> -}
>> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_restore_fl);
>> -
>> -asmlinkage __visible void vsmp_irq_disable(void)
>> -{
>> -       unsigned long flags = native_save_fl();
>> -
>> -       native_restore_fl((flags & ~X86_EFLAGS_IF) | X86_EFLAGS_AC);
>> -}
>> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_disable);
>> -
>> -asmlinkage __visible void vsmp_irq_enable(void)
>> -{
>> -       unsigned long flags = native_save_fl();
>> -
>> -       native_restore_fl((flags | X86_EFLAGS_IF) & (~X86_EFLAGS_AC));
>> -}
>> -PV_CALLEE_SAVE_REGS_THUNK(vsmp_irq_enable);
>> -
>> -static unsigned __init vsmp_patch(u8 type, void *ibuf,
>> -                                 unsigned long addr, unsigned len)
>> -{
>> -       switch (type) {
>> -       case PARAVIRT_PATCH(irq.irq_enable):
>> -       case PARAVIRT_PATCH(irq.irq_disable):
>> -       case PARAVIRT_PATCH(irq.save_fl):
>> -       case PARAVIRT_PATCH(irq.restore_fl):
>> -               return paravirt_patch_default(type, ibuf, addr, len);
>> -       default:
>> -               return native_patch(type, ibuf, addr, len);
>> -       }
>> -
>> -}
>> -
>> +#if defined CONFIG_PCI
>>  static void __init set_vsmp_pv_ops(void)
> 
> Wouldn't be a rename of the function be appropriate now?
you are correct, I'll fix and resend the patch.

> 
> 
> Juergen
> 

Re: [PATCH] x86/build: Build VSMP support only if selected

[Juergen Gross]

On 01/11/2018 16:09, Eial Czerwacki wrote:
> Greetings,
> 
> On 11/01/2018 03:45 PM, Juergen Gross wrote:
>> On 01/11/2018 14:10, Eial Czerwacki wrote:
>>> Greetings,
>>>
>>> On 11/01/2018 12:39 PM, Shai Fultheim (Shai@ScaleMP.com) wrote:
>>>> On 01/11/18 11:37, Thomas Gleixner wrote:
>>>>
>>>>> VSMP support is built even if CONFIG_X86_VSMP is not set. This leads to a build
>>>>> breakage when CONFIG_PCI is disabled as well.
>>>>>
>>>>> Build VSMP code only when selected.
>>>>
>>>> This patch disables detect_vsmp_box() on systems without CONFIG_X86_VSMP, due to
>>>> the recent 6da63eb241a05b0e676d68975e793c0521387141.  This is significant
>>>> regression that will affect significant number of deployments.
>>>>
>>>> We will reply shortly with an updated patch that fix the dependency on pv_irq_ops,
>>>> and revert to CONFIG_PARAVIRT, with proper protection for CONFIG_PCI.
>>>>
>>>
>>> here is the proper patch which fixes the issue on hand:
>>> From ebff534f8cfa55d7c3ab798c44abe879f3fbe2b8 Mon Sep 17 00:00:00 2001
>>> From: Eial Czerwacki <eial@scalemp.com>
>>> Date: Thu, 1 Nov 2018 15:08:32 +0200
>>> Subject: [PATCH] x86/build: Build VSMP support only if CONFIG_PCI is
>>> selected
>>>
>>> vsmp dependency of pv_irq_ops removed some years ago, so now let's clean
>>> it up from vsmp_64.c.
>>>
>>> In short, "cap & ctl & (1 << 4)" was always returning 0, as such we can
>>> remove all the PARAVIRT/PARAVIRT_XXL code handling that.
>>>
>>> However, the rest of the code depends on CONFIG_PCI, so fix it accordingly.
>>>
>>> Signed-off-by: Eial Czerwacki <eial@scalemp.com>
>>> Acked-by: Shai Fultheim <shai@scalemp.com>
>>> ---
>>>  arch/x86/Kconfig          |  1 -
>>>  arch/x86/kernel/vsmp_64.c | 80
>>> +++--------------------------------------------
>>>  2 files changed, 5 insertions(+), 76 deletions(-)
>>>
>>> diff --git a/arch/x86/Kconfig b/arch/x86/Kconfig
>>> index c51c989..4b187ca 100644
>>> --- a/arch/x86/Kconfig
>>> +++ b/arch/x86/Kconfig
>>> @@ -524,7 +524,6 @@ config X86_VSMP
>>>         bool "ScaleMP vSMP"
>>>         select HYPERVISOR_GUEST
>>>         select PARAVIRT
>>
>> Do you really still need PARAVIRT and HYPERVISOR_GUEST?
>> Maybe you want IRQ_REMAP instead?
>>
> Better performance is achieved with PARAVIRTed kernel.   Hence we keep
> them both in.

Do you have an explanation for that? Normally PARAVIRT is expected
to have a small negative impact on performance.


Juergen