[PATCH 4.13 00/11] 4.13.9-stable review

[PATCH 4.13 00/11] 4.13.9-stable review

[Greg Kroah-Hartman]

This is the start of the stable review cycle for the 4.13.9 release.
There are 11 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 Sat Oct 21 13:16:37 UTC 2017.
Anything received after that time might be too late.

The whole patch series can be found in one patch at:
	kernel.org/pub/linux/kernel/v4.x/stable-review/patch-4.13.9-rc1.gz
or in the git tree and branch at:
  git://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable-rc.git linux-4.13.y
and the diffstat can be found below.

thanks,

greg k-h

-------------
Pseudo-Shortlog of commits:

Greg Kroah-Hartman <gregkh@linuxfoundation.org>
    Linux 4.13.9-rc1

Stephen Hemminger <stephen@networkplumber.org>
    vmbus: more host signalling avoidance

Stephen Hemminger <stephen@networkplumber.org>
    vmbus: eliminate duplicate cached index

Stephen Hemminger <stephen@networkplumber.org>
    vmbus: refactor hv_signal_on_read

Stephen Hemminger <stephen@networkplumber.org>
    vmbus: simplify hv_ringbuffer_read

K. Y. Srinivasan <kys@microsoft.com>
    Drivers: hv: vmbus: Fix bugs in rescind handling

K. Y. Srinivasan <kys@microsoft.com>
    Drivers: hv: vmbus: Fix rescind handling issues

Alex Manoussakis <amanou@gnu.org>
    HID: hid-elecom: extend to fix descriptor for HUGE trackball

Will Deacon <will.deacon@arm.com>
    mm: page_vma_mapped: ensure pmd is loaded with READ_ONCE outside of lock

Mark Rutland <mark.rutland@arm.com>
    perf pmu: Unbreak perf record for arm/arm64 with events with explicit PMU

Paolo Bonzini <pbonzini@redhat.com>
    x86/apic: Silence "FW_BUG TSC_DEADLINE disabled due to Errata" on hypervisors

Hans de Goede <hdegoede@redhat.com>
    x86/apic: Silence "FW_BUG TSC_DEADLINE disabled due to Errata" on CPUs without the feature


-------------

Diffstat:

 Makefile                       |   4 +-
 arch/x86/kernel/apic/apic.c    |   7 +-
 drivers/hid/Kconfig            |   1 +
 drivers/hid/hid-core.c         |   2 +
 drivers/hid/hid-elecom.c       |  13 +++-
 drivers/hid/hid-ids.h          |   2 +
 drivers/hv/channel.c           |  20 ++++-
 drivers/hv/channel_mgmt.c      |  54 ++++++++++----
 drivers/hv/ring_buffer.c       | 162 ++++++++++++++---------------------------
 drivers/hv/vmbus_drv.c         |   6 +-
 include/linux/hyperv.h         |  67 +----------------
 mm/page_vma_mapped.c           |  25 +++----
 tools/perf/util/parse-events.c |   9 ++-
 tools/perf/util/pmu.c          |  56 ++++++++++----
 tools/perf/util/pmu.h          |   1 +
 15 files changed, 195 insertions(+), 234 deletions(-)

[PATCH 4.13 01/11] x86/apic: Silence "FW_BUG TSC_DEADLINE disabled due to Errata" on CPUs without the feature

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Hans de Goede <hdegoede@redhat.com>

commit 594a30fb12424717a41c62323d2a8bf167dbccad upstream.

When booting 4.13 on a VirtualBox VM on a Skylake host the following
error shows up in the logs:

 [    0.000000] [Firmware Bug]: TSC_DEADLINE disabled due to Errata;
                please update microcode to version: 0xb2 (or later)

This is caused by apic_check_deadline_errata() only checking CPU model
and not the X86_FEATURE_TSC_DEADLINE_TIMER flag (which VirtualBox does
NOT export to the guest), combined with VirtualBox not exporting the
micro-code version to the guest.

This commit adds a check for X86_FEATURE_TSC_DEADLINE_TIMER to
apic_check_deadline_errata(), silencing this error on VirtualBox VMs.

Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Acked-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Frank Mehnert <frank.mehnert@oracle.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Michael Thayer <michael.thayer@oracle.com>
Cc: Michal Necasek <michal.necasek@oracle.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Fixes: bd9240a18e ("x86/apic: Add TSC_DEADLINE quirk due to errata")
Link: http://lkml.kernel.org/r/20170830105811.27539-1-hdegoede@redhat.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 arch/x86/kernel/apic/apic.c |    6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -599,9 +599,13 @@ static const struct x86_cpu_id deadline_
 
 static void apic_check_deadline_errata(void)
 {
-	const struct x86_cpu_id *m = x86_match_cpu(deadline_match);
+	const struct x86_cpu_id *m;
 	u32 rev;
 
+	if (!boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+		return;
+
+	m = x86_match_cpu(deadline_match);
 	if (!m)
 		return;
 

[PATCH 4.13 02/11] x86/apic: Silence "FW_BUG TSC_DEADLINE disabled due to Errata" on hypervisors

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Paolo Bonzini <pbonzini@redhat.com>

commit cc6afe2240298049585e86b1ade85efc8a7f225d upstream.

Commit 594a30fb1242 ("x86/apic: Silence "FW_BUG TSC_DEADLINE disabled
due to Errata" on CPUs without the feature", 2017-08-30) was also about
silencing the warning on VirtualBox; however, KVM does expose the TSC
deadline timer, and it's virtualized so that it is immune from CPU errata.

Therefore, booting 4.13 with "-cpu Haswell" shows this in the logs:

     [    0.000000] [Firmware Bug]: TSC_DEADLINE disabled due to Errata;
                    please update microcode to version: 0xb2 (or later)

Even if you had a hypervisor that does _not_ virtualize the TSC deadline
and rather exposes the hardware one, it should be the hypervisors task
to update microcode and possibly hide the flag from CPUID.  So just
hide the message when running on _any_ hypervisor, not just those that
do not support the TSC deadline timer.

The older check still makes sense, so keep it.

Fixes: bd9240a18e ("x86/apic: Add TSC_DEADLINE quirk due to errata")
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Hans de Goede <hdegoede@redhat.com>
Cc: kvm@vger.kernel.org
Link: https://lkml.kernel.org/r/1507630377-54471-1-git-send-email-pbonzini@redhat.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 arch/x86/kernel/apic/apic.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

--- a/arch/x86/kernel/apic/apic.c
+++ b/arch/x86/kernel/apic/apic.c
@@ -602,7 +602,8 @@ static void apic_check_deadline_errata(v
 	const struct x86_cpu_id *m;
 	u32 rev;
 
-	if (!boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER))
+	if (!boot_cpu_has(X86_FEATURE_TSC_DEADLINE_TIMER) ||
+	    boot_cpu_has(X86_FEATURE_HYPERVISOR))
 		return;
 
 	m = x86_match_cpu(deadline_match);

[PATCH 4.13 03/11] perf pmu: Unbreak perf record for arm/arm64 with events with explicit PMU

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Mark Rutland <mark.rutland@arm.com>

commit 66ec11919a0f96e936bb731fdbc2851316077d26 upstream.

Currently, perf record is broken on arm/arm64 systems when the PMU is
specified explicitly as part of the event, e.g.

$ ./perf record -e armv8_cortex_a53/cpu_cycles/u true

In such cases, perf record fails to open events unless
perf_event_paranoid is set to -1, even if the PMU in question supports
mode exclusion. Further, even when perf_event_paranoid is toggled, no
samples are recorded.

This is an unintended side effect of commit:

  e3ba76deef23064f ("perf tools: Force uncore events to system wide monitoring)

... which assumes that if a PMU has an associated cpu_map, it is an
uncore PMU, and forces events for such PMUs to be system-wide.

This is not true for arm/arm64 systems, which can have heterogeneous
CPUs. To account for this, multiple CPU PMUs are exposed, each with a
"cpus" field under sysfs, which the perf tool parses into a cpu_map. ARM
PMUs do not have a "cpumask" file, and only have a "cpus" file. For the
gory details as to why, see commit:

 7e3fcffe95544010 ("perf pmu: Support alternative sysfs cpumask")

Given all of this, we can instead identify uncore PMUs by explicitly
checking for a "cpumask" file, and restore arm/arm64 PMU support back to
a working state. This patch does so, adding a new perf_pmu::is_uncore
field, and splitting the existing cpumask parsing so that it can be
reused.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Tested-by Will Deacon <will.deacon@arm.com>
Acked-by: Jiri Olsa <jolsa@kernel.org>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: David Ahern <dsahern@gmail.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Fixes: e3ba76deef23064f ("perf tools: Force uncore events to system wide monitoring)
Link: http://lkml.kernel.org/r/1507315102-5942-1-git-send-email-mark.rutland@arm.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 tools/perf/util/parse-events.c |    9 +++---
 tools/perf/util/pmu.c          |   56 ++++++++++++++++++++++++++++++-----------
 tools/perf/util/pmu.h          |    1 
 3 files changed, 47 insertions(+), 19 deletions(-)

--- a/tools/perf/util/parse-events.c
+++ b/tools/perf/util/parse-events.c
@@ -309,10 +309,11 @@ static char *get_config_name(struct list
 static struct perf_evsel *
 __add_event(struct list_head *list, int *idx,
 	    struct perf_event_attr *attr,
-	    char *name, struct cpu_map *cpus,
+	    char *name, struct perf_pmu *pmu,
 	    struct list_head *config_terms)
 {
 	struct perf_evsel *evsel;
+	struct cpu_map *cpus = pmu ? pmu->cpus : NULL;
 
 	event_attr_init(attr);
 
@@ -323,7 +324,7 @@ __add_event(struct list_head *list, int
 	(*idx)++;
 	evsel->cpus        = cpu_map__get(cpus);
 	evsel->own_cpus    = cpu_map__get(cpus);
-	evsel->system_wide = !!cpus;
+	evsel->system_wide = pmu ? pmu->is_uncore : false;
 
 	if (name)
 		evsel->name = strdup(name);
@@ -1232,7 +1233,7 @@ int parse_events_add_pmu(struct parse_ev
 
 	if (!head_config) {
 		attr.type = pmu->type;
-		evsel = __add_event(list, &data->idx, &attr, NULL, pmu->cpus, NULL);
+		evsel = __add_event(list, &data->idx, &attr, NULL, pmu, NULL);
 		return evsel ? 0 : -ENOMEM;
 	}
 
@@ -1253,7 +1254,7 @@ int parse_events_add_pmu(struct parse_ev
 		return -EINVAL;
 
 	evsel = __add_event(list, &data->idx, &attr,
-			    get_config_name(head_config), pmu->cpus,
+			    get_config_name(head_config), pmu,
 			    &config_terms);
 	if (evsel) {
 		evsel->unit = info.unit;
--- a/tools/perf/util/pmu.c
+++ b/tools/perf/util/pmu.c
@@ -470,17 +470,36 @@ static void pmu_read_sysfs(void)
 	closedir(dir);
 }
 
+static struct cpu_map *__pmu_cpumask(const char *path)
+{
+	FILE *file;
+	struct cpu_map *cpus;
+
+	file = fopen(path, "r");
+	if (!file)
+		return NULL;
+
+	cpus = cpu_map__read(file);
+	fclose(file);
+	return cpus;
+}
+
+/*
+ * Uncore PMUs have a "cpumask" file under sysfs. CPU PMUs (e.g. on arm/arm64)
+ * may have a "cpus" file.
+ */
+#define CPUS_TEMPLATE_UNCORE	"%s/bus/event_source/devices/%s/cpumask"
+#define CPUS_TEMPLATE_CPU	"%s/bus/event_source/devices/%s/cpus"
+
 static struct cpu_map *pmu_cpumask(const char *name)
 {
-	struct stat st;
 	char path[PATH_MAX];
-	FILE *file;
 	struct cpu_map *cpus;
 	const char *sysfs = sysfs__mountpoint();
 	const char *templates[] = {
-		 "%s/bus/event_source/devices/%s/cpumask",
-		 "%s/bus/event_source/devices/%s/cpus",
-		 NULL
+		CPUS_TEMPLATE_UNCORE,
+		CPUS_TEMPLATE_CPU,
+		NULL
 	};
 	const char **template;
 
@@ -489,20 +508,25 @@ static struct cpu_map *pmu_cpumask(const
 
 	for (template = templates; *template; template++) {
 		snprintf(path, PATH_MAX, *template, sysfs, name);
-		if (stat(path, &st) == 0)
-			break;
+		cpus = __pmu_cpumask(path);
+		if (cpus)
+			return cpus;
 	}
 
-	if (!*template)
-		return NULL;
+	return NULL;
+}
 
-	file = fopen(path, "r");
-	if (!file)
-		return NULL;
+static bool pmu_is_uncore(const char *name)
+{
+	char path[PATH_MAX];
+	struct cpu_map *cpus;
+	const char *sysfs = sysfs__mountpoint();
 
-	cpus = cpu_map__read(file);
-	fclose(file);
-	return cpus;
+	snprintf(path, PATH_MAX, CPUS_TEMPLATE_UNCORE, sysfs, name);
+	cpus = __pmu_cpumask(path);
+	cpu_map__put(cpus);
+
+	return !!cpus;
 }
 
 /*
@@ -617,6 +641,8 @@ static struct perf_pmu *pmu_lookup(const
 
 	pmu->cpus = pmu_cpumask(name);
 
+	pmu->is_uncore = pmu_is_uncore(name);
+
 	INIT_LIST_HEAD(&pmu->format);
 	INIT_LIST_HEAD(&pmu->aliases);
 	list_splice(&format, &pmu->format);
--- a/tools/perf/util/pmu.h
+++ b/tools/perf/util/pmu.h
@@ -22,6 +22,7 @@ struct perf_pmu {
 	char *name;
 	__u32 type;
 	bool selectable;
+	bool is_uncore;
 	struct perf_event_attr *default_config;
 	struct cpu_map *cpus;
 	struct list_head format;  /* HEAD struct perf_pmu_format -> list */

[PATCH 4.13 04/11] mm: page_vma_mapped: ensure pmd is loaded with READ_ONCE outside of lock

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Will Deacon <will.deacon@arm.com>

commit a7b100953aa33a5bbdc3e5e7f2241b9c0704606e upstream.

Loading the pmd without holding the pmd_lock exposes us to races with
concurrent updaters of the page tables but, worse still, it also allows
the compiler to cache the pmd value in a register and reuse it later on,
even if we've performed a READ_ONCE in between and seen a more recent
value.

In the case of page_vma_mapped_walk, this leads to the following crash
when the pmd loaded for the initial pmd_trans_huge check is all zeroes
and a subsequent valid table entry is loaded by check_pmd.  We then
proceed into map_pte, but the compiler re-uses the zero entry inside
pte_offset_map, resulting in a junk pointer being installed in
pvmw->pte:

  PC is at check_pte+0x20/0x170
  LR is at page_vma_mapped_walk+0x2e0/0x540
  [...]
  Process doio (pid: 2463, stack limit = 0xffff00000f2e8000)
  Call trace:
    check_pte+0x20/0x170
    page_vma_mapped_walk+0x2e0/0x540
    page_mkclean_one+0xac/0x278
    rmap_walk_file+0xf0/0x238
    rmap_walk+0x64/0xa0
    page_mkclean+0x90/0xa8
    clear_page_dirty_for_io+0x84/0x2a8
    mpage_submit_page+0x34/0x98
    mpage_process_page_bufs+0x164/0x170
    mpage_prepare_extent_to_map+0x134/0x2b8
    ext4_writepages+0x484/0xe30
    do_writepages+0x44/0xe8
    __filemap_fdatawrite_range+0xbc/0x110
    file_write_and_wait_range+0x48/0xd8
    ext4_sync_file+0x80/0x4b8
    vfs_fsync_range+0x64/0xc0
    SyS_msync+0x194/0x1e8

This patch fixes the problem by ensuring that READ_ONCE is used before
the initial checks on the pmd, and this value is subsequently used when
checking whether or not the pmd is present.  pmd_check is removed and
the pmd_present check is inlined directly.

Link: http://lkml.kernel.org/r/1507222630-5839-1-git-send-email-will.deacon@arm.com
Fixes: f27176cfc363 ("mm: convert page_mkclean_one() to use page_vma_mapped_walk()")
Signed-off-by: Will Deacon <will.deacon@arm.com>
Tested-by: Yury Norov <ynorov@caviumnetworks.com>
Tested-by: Richard Ruigrok <rruigrok@codeaurora.org>
Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: "Paul E. McKenney" <paulmck@linux.vnet.ibm.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
[will: backport to 4.13.y]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 mm/page_vma_mapped.c |   25 ++++++++++---------------
 1 file changed, 10 insertions(+), 15 deletions(-)

--- a/mm/page_vma_mapped.c
+++ b/mm/page_vma_mapped.c
@@ -6,17 +6,6 @@
 
 #include "internal.h"
 
-static inline bool check_pmd(struct page_vma_mapped_walk *pvmw)
-{
-	pmd_t pmde;
-	/*
-	 * Make sure we don't re-load pmd between present and !trans_huge check.
-	 * We need a consistent view.
-	 */
-	pmde = READ_ONCE(*pvmw->pmd);
-	return pmd_present(pmde) && !pmd_trans_huge(pmde);
-}
-
 static inline bool not_found(struct page_vma_mapped_walk *pvmw)
 {
 	page_vma_mapped_walk_done(pvmw);
@@ -106,6 +95,7 @@ bool page_vma_mapped_walk(struct page_vm
 	pgd_t *pgd;
 	p4d_t *p4d;
 	pud_t *pud;
+	pmd_t pmde;
 
 	/* The only possible pmd mapping has been handled on last iteration */
 	if (pvmw->pmd && !pvmw->pte)
@@ -138,7 +128,13 @@ restart:
 	if (!pud_present(*pud))
 		return false;
 	pvmw->pmd = pmd_offset(pud, pvmw->address);
-	if (pmd_trans_huge(*pvmw->pmd)) {
+	/*
+	 * Make sure the pmd value isn't cached in a register by the
+	 * compiler and used as a stale value after we've observed a
+	 * subsequent update.
+	 */
+	pmde = READ_ONCE(*pvmw->pmd);
+	if (pmd_trans_huge(pmde)) {
 		pvmw->ptl = pmd_lock(mm, pvmw->pmd);
 		if (!pmd_present(*pvmw->pmd))
 			return not_found(pvmw);
@@ -153,9 +149,8 @@ restart:
 			spin_unlock(pvmw->ptl);
 			pvmw->ptl = NULL;
 		}
-	} else {
-		if (!check_pmd(pvmw))
-			return false;
+	} else if (!pmd_present(pmde)) {
+		return false;
 	}
 	if (!map_pte(pvmw))
 		goto next_pte;

[PATCH 4.13 05/11] HID: hid-elecom: extend to fix descriptor for HUGE trackball

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Alex Manoussakis <amanou@gnu.org>

commit a0933a456ff83a3b5ffa3a1903e0b8de4a56adf5 upstream.

In addition to DEFT, Elecom introduced a larger trackball called HUGE, in
both wired (M-HT1URBK) and wireless (M-HT1DRBK) versions. It has the same
buttons and behavior as the DEFT. This patch adds the two relevant USB IDs
to enable operation of the three Fn buttons on the top of the device.

Cc: Diego Elio Petteno <flameeyes@flameeyes.eu>
Signed-off-by: Alex Manoussakis <amanou@gnu.org>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/hid/Kconfig      |    1 +
 drivers/hid/hid-core.c   |    2 ++
 drivers/hid/hid-elecom.c |   13 +++++++++----
 drivers/hid/hid-ids.h    |    2 ++
 4 files changed, 14 insertions(+), 4 deletions(-)

--- a/drivers/hid/Kconfig
+++ b/drivers/hid/Kconfig
@@ -281,6 +281,7 @@ config HID_ELECOM
 	Support for ELECOM devices:
 	  - BM084 Bluetooth Mouse
 	  - DEFT Trackball (Wired and wireless)
+	  - HUGE Trackball (Wired and wireless)
 
 config HID_ELO
 	tristate "ELO USB 4000/4500 touchscreen"
--- a/drivers/hid/hid-core.c
+++ b/drivers/hid/hid-core.c
@@ -2031,6 +2031,8 @@ static const struct hid_device_id hid_ha
 	{ HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_BM084) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_DEFT_WIRED) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_DEFT_WIRELESS) },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_HUGE_WIRED) },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_HUGE_WIRELESS) },
 #endif
 #if IS_ENABLED(CONFIG_HID_ELO)
 	{ HID_USB_DEVICE(USB_VENDOR_ID_ELO, 0x0009) },
--- a/drivers/hid/hid-elecom.c
+++ b/drivers/hid/hid-elecom.c
@@ -3,6 +3,7 @@
  *  Copyright (c) 2010 Richard Nauber <Richard.Nauber@gmail.com>
  *  Copyright (c) 2016 Yuxuan Shui <yshuiv7@gmail.com>
  *  Copyright (c) 2017 Diego Elio Pettenò <flameeyes@flameeyes.eu>
+ *  Copyright (c) 2017 Alex Manoussakis <amanou@gnu.org>
  */
 
 /*
@@ -32,9 +33,11 @@ static __u8 *elecom_report_fixup(struct
 		break;
 	case USB_DEVICE_ID_ELECOM_DEFT_WIRED:
 	case USB_DEVICE_ID_ELECOM_DEFT_WIRELESS:
-		/* The DEFT trackball has eight buttons, but its descriptor only
-		 * reports five, disabling the three Fn buttons on the top of
-		 * the mouse.
+	case USB_DEVICE_ID_ELECOM_HUGE_WIRED:
+	case USB_DEVICE_ID_ELECOM_HUGE_WIRELESS:
+		/* The DEFT/HUGE trackball has eight buttons, but its descriptor
+		 * only reports five, disabling the three Fn buttons on the top
+		 * of the mouse.
 		 *
 		 * Apply the following diff to the descriptor:
 		 *
@@ -62,7 +65,7 @@ static __u8 *elecom_report_fixup(struct
 		 * End Collection,                     End Collection,
 		 */
 		if (*rsize == 213 && rdesc[13] == 5 && rdesc[21] == 5) {
-			hid_info(hdev, "Fixing up Elecom DEFT Fn buttons\n");
+			hid_info(hdev, "Fixing up Elecom DEFT/HUGE Fn buttons\n");
 			rdesc[13] = 8; /* Button/Variable Report Count */
 			rdesc[21] = 8; /* Button/Variable Usage Maximum */
 			rdesc[29] = 0; /* Button/Constant Report Count */
@@ -76,6 +79,8 @@ static const struct hid_device_id elecom
 	{ HID_BLUETOOTH_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_BM084) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_DEFT_WIRED) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_DEFT_WIRELESS) },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_HUGE_WIRED) },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_ELECOM, USB_DEVICE_ID_ELECOM_HUGE_WIRELESS) },
 	{ }
 };
 MODULE_DEVICE_TABLE(hid, elecom_devices);
--- a/drivers/hid/hid-ids.h
+++ b/drivers/hid/hid-ids.h
@@ -366,6 +366,8 @@
 #define USB_DEVICE_ID_ELECOM_BM084	0x0061
 #define USB_DEVICE_ID_ELECOM_DEFT_WIRED	0x00fe
 #define USB_DEVICE_ID_ELECOM_DEFT_WIRELESS	0x00ff
+#define USB_DEVICE_ID_ELECOM_HUGE_WIRED	0x010c
+#define USB_DEVICE_ID_ELECOM_HUGE_WIRELESS	0x010d
 
 #define USB_VENDOR_ID_DREAM_CHEEKY	0x1d34
 #define USB_DEVICE_ID_DREAM_CHEEKY_WN	0x0004

[PATCH 4.13 06/11] Drivers: hv: vmbus: Fix rescind handling issues

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: K. Y. Srinivasan <kys@microsoft.com>

commit 6f3d791f300618caf82a2be0c27456edd76d5164 upstream.

This patch handles the following issues that were observed when we are
handling racing channel offer message and rescind message for the same
offer:

1. Since the host does not respond to messages on a rescinded channel,
in the current code, we could be indefinitely blocked on the vmbus_open() call.

2. When a rescinded channel is being closed, if there is a pending interrupt on the
channel, we could end up freeing the channel that the interrupt handler would run on.

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Reviewed-by: Dexuan Cui <decui@microsoft.com>
Tested-by: Dexuan Cui <decui@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/hv/channel.c      |   14 ++++++++++++++
 drivers/hv/channel_mgmt.c |   29 ++++++++++++++++++++++++++---
 drivers/hv/vmbus_drv.c    |    3 +++
 include/linux/hyperv.h    |    2 ++
 4 files changed, 45 insertions(+), 3 deletions(-)

--- a/drivers/hv/channel.c
+++ b/drivers/hv/channel.c
@@ -177,6 +177,11 @@ int vmbus_open(struct vmbus_channel *new
 		      &vmbus_connection.chn_msg_list);
 	spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags);
 
+	if (newchannel->rescind) {
+		err = -ENODEV;
+		goto error_free_gpadl;
+	}
+
 	ret = vmbus_post_msg(open_msg,
 			     sizeof(struct vmbus_channel_open_channel), true);
 
@@ -421,6 +426,11 @@ int vmbus_establish_gpadl(struct vmbus_c
 
 	spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags);
 
+	if (channel->rescind) {
+		ret = -ENODEV;
+		goto cleanup;
+	}
+
 	ret = vmbus_post_msg(gpadlmsg, msginfo->msgsize -
 			     sizeof(*msginfo), true);
 	if (ret != 0)
@@ -494,6 +504,10 @@ int vmbus_teardown_gpadl(struct vmbus_ch
 	list_add_tail(&info->msglistentry,
 		      &vmbus_connection.chn_msg_list);
 	spin_unlock_irqrestore(&vmbus_connection.channelmsg_lock, flags);
+
+	if (channel->rescind)
+		goto post_msg_err;
+
 	ret = vmbus_post_msg(msg, sizeof(struct vmbus_channel_gpadl_teardown),
 			     true);
 
--- a/drivers/hv/channel_mgmt.c
+++ b/drivers/hv/channel_mgmt.c
@@ -451,6 +451,12 @@ static void vmbus_process_offer(struct v
 	/* Make sure this is a new offer */
 	mutex_lock(&vmbus_connection.channel_mutex);
 
+	/*
+	 * Now that we have acquired the channel_mutex,
+	 * we can release the potentially racing rescind thread.
+	 */
+	atomic_dec(&vmbus_connection.offer_in_progress);
+
 	list_for_each_entry(channel, &vmbus_connection.chn_list, listentry) {
 		if (!uuid_le_cmp(channel->offermsg.offer.if_type,
 			newchannel->offermsg.offer.if_type) &&
@@ -481,7 +487,6 @@ static void vmbus_process_offer(struct v
 			channel->num_sc++;
 			spin_unlock_irqrestore(&channel->lock, flags);
 		} else {
-			atomic_dec(&vmbus_connection.offer_in_progress);
 			goto err_free_chan;
 		}
 	}
@@ -510,7 +515,6 @@ static void vmbus_process_offer(struct v
 	if (!fnew) {
 		if (channel->sc_creation_callback != NULL)
 			channel->sc_creation_callback(newchannel);
-		atomic_dec(&vmbus_connection.offer_in_progress);
 		return;
 	}
 
@@ -541,7 +545,7 @@ static void vmbus_process_offer(struct v
 		goto err_deq_chan;
 	}
 
-	atomic_dec(&vmbus_connection.offer_in_progress);
+	newchannel->probe_done = true;
 	return;
 
 err_deq_chan:
@@ -882,8 +886,27 @@ static void vmbus_onoffer_rescind(struct
 	channel->rescind = true;
 	spin_unlock_irqrestore(&channel->lock, flags);
 
+	/*
+	 * Now that we have posted the rescind state, perform
+	 * rescind related cleanup.
+	 */
 	vmbus_rescind_cleanup(channel);
 
+	/*
+	 * Now wait for offer handling to complete.
+	 */
+	while (READ_ONCE(channel->probe_done) == false) {
+		/*
+		 * We wait here until any channel offer is currently
+		 * being processed.
+		 */
+		msleep(1);
+	}
+
+	/*
+	 * At this point, the rescind handling can proceed safely.
+	 */
+
 	if (channel->device_obj) {
 		if (channel->chn_rescind_callback) {
 			channel->chn_rescind_callback(channel);
--- a/drivers/hv/vmbus_drv.c
+++ b/drivers/hv/vmbus_drv.c
@@ -940,6 +940,9 @@ static void vmbus_chan_sched(struct hv_p
 			if (channel->offermsg.child_relid != relid)
 				continue;
 
+			if (channel->rescind)
+				continue;
+
 			switch (channel->callback_mode) {
 			case HV_CALL_ISR:
 				vmbus_channel_isr(channel);
--- a/include/linux/hyperv.h
+++ b/include/linux/hyperv.h
@@ -895,6 +895,8 @@ struct vmbus_channel {
 	 */
 	enum hv_numa_policy affinity_policy;
 
+	bool probe_done;
+
 };
 
 static inline bool is_hvsock_channel(const struct vmbus_channel *c)

[PATCH 4.13 07/11] Drivers: hv: vmbus: Fix bugs in rescind handling

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: K. Y. Srinivasan <kys@microsoft.com>

commit 192b2d78722ffea188e5ec6ae5d55010dce05a4b upstream.

This patch addresses the following bugs in the current rescind handling code:

1. Fixes a race condition where we may be invoking hv_process_channel_removal()
on an already freed channel.

2. Prevents indefinite wait when rescinding sub-channels by correctly setting
the probe_complete state.

I would like to thank Dexuan for patiently reviewing earlier versions of this
patch and identifying many of the issues fixed here.

Greg, please apply this to 4.14-final.

Fixes: '54a66265d675 ("Drivers: hv: vmbus: Fix rescind handling")'

Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Reviewed-by: Dexuan Cui <decui@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/hv/channel.c      |    6 +++---
 drivers/hv/channel_mgmt.c |   37 ++++++++++++++++++-------------------
 drivers/hv/vmbus_drv.c    |    3 +--
 include/linux/hyperv.h    |    2 +-
 4 files changed, 23 insertions(+), 25 deletions(-)

--- a/drivers/hv/channel.c
+++ b/drivers/hv/channel.c
@@ -640,6 +640,7 @@ void vmbus_close(struct vmbus_channel *c
 		 */
 		return;
 	}
+	mutex_lock(&vmbus_connection.channel_mutex);
 	/*
 	 * Close all the sub-channels first and then close the
 	 * primary channel.
@@ -648,16 +649,15 @@ void vmbus_close(struct vmbus_channel *c
 		cur_channel = list_entry(cur, struct vmbus_channel, sc_list);
 		vmbus_close_internal(cur_channel);
 		if (cur_channel->rescind) {
-			mutex_lock(&vmbus_connection.channel_mutex);
-			hv_process_channel_removal(cur_channel,
+			hv_process_channel_removal(
 					   cur_channel->offermsg.child_relid);
-			mutex_unlock(&vmbus_connection.channel_mutex);
 		}
 	}
 	/*
 	 * Now close the primary.
 	 */
 	vmbus_close_internal(channel);
+	mutex_unlock(&vmbus_connection.channel_mutex);
 }
 EXPORT_SYMBOL_GPL(vmbus_close);
 
--- a/drivers/hv/channel_mgmt.c
+++ b/drivers/hv/channel_mgmt.c
@@ -159,7 +159,7 @@ static void vmbus_rescind_cleanup(struct
 
 
 	spin_lock_irqsave(&vmbus_connection.channelmsg_lock, flags);
-
+	channel->rescind = true;
 	list_for_each_entry(msginfo, &vmbus_connection.chn_msg_list,
 				msglistentry) {
 
@@ -381,14 +381,21 @@ static void vmbus_release_relid(u32 reli
 		       true);
 }
 
-void hv_process_channel_removal(struct vmbus_channel *channel, u32 relid)
+void hv_process_channel_removal(u32 relid)
 {
 	unsigned long flags;
-	struct vmbus_channel *primary_channel;
+	struct vmbus_channel *primary_channel, *channel;
 
-	BUG_ON(!channel->rescind);
 	BUG_ON(!mutex_is_locked(&vmbus_connection.channel_mutex));
 
+	/*
+	 * Make sure channel is valid as we may have raced.
+	 */
+	channel = relid2channel(relid);
+	if (!channel)
+		return;
+
+	BUG_ON(!channel->rescind);
 	if (channel->target_cpu != get_cpu()) {
 		put_cpu();
 		smp_call_function_single(channel->target_cpu,
@@ -515,6 +522,7 @@ static void vmbus_process_offer(struct v
 	if (!fnew) {
 		if (channel->sc_creation_callback != NULL)
 			channel->sc_creation_callback(newchannel);
+		newchannel->probe_done = true;
 		return;
 	}
 
@@ -843,7 +851,6 @@ static void vmbus_onoffer_rescind(struct
 {
 	struct vmbus_channel_rescind_offer *rescind;
 	struct vmbus_channel *channel;
-	unsigned long flags;
 	struct device *dev;
 
 	rescind = (struct vmbus_channel_rescind_offer *)hdr;
@@ -882,16 +889,6 @@ static void vmbus_onoffer_rescind(struct
 		return;
 	}
 
-	spin_lock_irqsave(&channel->lock, flags);
-	channel->rescind = true;
-	spin_unlock_irqrestore(&channel->lock, flags);
-
-	/*
-	 * Now that we have posted the rescind state, perform
-	 * rescind related cleanup.
-	 */
-	vmbus_rescind_cleanup(channel);
-
 	/*
 	 * Now wait for offer handling to complete.
 	 */
@@ -910,6 +907,7 @@ static void vmbus_onoffer_rescind(struct
 	if (channel->device_obj) {
 		if (channel->chn_rescind_callback) {
 			channel->chn_rescind_callback(channel);
+			vmbus_rescind_cleanup(channel);
 			return;
 		}
 		/*
@@ -918,6 +916,7 @@ static void vmbus_onoffer_rescind(struct
 		 */
 		dev = get_device(&channel->device_obj->device);
 		if (dev) {
+			vmbus_rescind_cleanup(channel);
 			vmbus_device_unregister(channel->device_obj);
 			put_device(dev);
 		}
@@ -930,16 +929,16 @@ static void vmbus_onoffer_rescind(struct
 		 * 1. Close all sub-channels first
 		 * 2. Then close the primary channel.
 		 */
+		mutex_lock(&vmbus_connection.channel_mutex);
+		vmbus_rescind_cleanup(channel);
 		if (channel->state == CHANNEL_OPEN_STATE) {
 			/*
 			 * The channel is currently not open;
 			 * it is safe for us to cleanup the channel.
 			 */
-			mutex_lock(&vmbus_connection.channel_mutex);
-			hv_process_channel_removal(channel,
-						channel->offermsg.child_relid);
-			mutex_unlock(&vmbus_connection.channel_mutex);
+			hv_process_channel_removal(rescind->child_relid);
 		}
+		mutex_unlock(&vmbus_connection.channel_mutex);
 	}
 }
 
--- a/drivers/hv/vmbus_drv.c
+++ b/drivers/hv/vmbus_drv.c
@@ -768,8 +768,7 @@ static void vmbus_device_release(struct
 	struct vmbus_channel *channel = hv_dev->channel;
 
 	mutex_lock(&vmbus_connection.channel_mutex);
-	hv_process_channel_removal(channel,
-				   channel->offermsg.child_relid);
+	hv_process_channel_removal(channel->offermsg.child_relid);
 	mutex_unlock(&vmbus_connection.channel_mutex);
 	kfree(hv_dev);
 
--- a/include/linux/hyperv.h
+++ b/include/linux/hyperv.h
@@ -1455,7 +1455,7 @@ extern bool vmbus_prep_negotiate_resp(st
 				const int *srv_version, int srv_vercnt,
 				int *nego_fw_version, int *nego_srv_version);
 
-void hv_process_channel_removal(struct vmbus_channel *channel, u32 relid);
+void hv_process_channel_removal(u32 relid);
 
 void vmbus_setevent(struct vmbus_channel *channel);
 /*

[PATCH 4.13 08/11] vmbus: simplify hv_ringbuffer_read

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stephen Hemminger <stephen@networkplumber.org>

commit 4226ff69a3dff78bead7d9a270423cd21f8d40b8 upstream.

With new iterator functions (and the double mapping) the ring buffer
read function can be greatly simplified.

Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/hv/ring_buffer.c |  116 ++++++-----------------------------------------
 1 file changed, 16 insertions(+), 100 deletions(-)

--- a/drivers/hv/ring_buffer.c
+++ b/drivers/hv/ring_buffer.c
@@ -94,30 +94,6 @@ hv_set_next_write_location(struct hv_rin
 	ring_info->ring_buffer->write_index = next_write_location;
 }
 
-/* Get the next read location for the specified ring buffer. */
-static inline u32
-hv_get_next_read_location(const struct hv_ring_buffer_info *ring_info)
-{
-	return ring_info->ring_buffer->read_index;
-}
-
-/*
- * Get the next read location + offset for the specified ring buffer.
- * This allows the caller to skip.
- */
-static inline u32
-hv_get_next_readlocation_withoffset(const struct hv_ring_buffer_info *ring_info,
-				    u32 offset)
-{
-	u32 next = ring_info->ring_buffer->read_index;
-
-	next += offset;
-	if (next >= ring_info->ring_datasize)
-		next -= ring_info->ring_datasize;
-
-	return next;
-}
-
 /* Set the next read location for the specified ring buffer. */
 static inline void
 hv_set_next_read_location(struct hv_ring_buffer_info *ring_info,
@@ -142,29 +118,6 @@ hv_get_ring_bufferindices(struct hv_ring
 }
 
 /*
- * Helper routine to copy to source from ring buffer.
- * Assume there is enough room. Handles wrap-around in src case only!!
- */
-static u32 hv_copyfrom_ringbuffer(
-	const struct hv_ring_buffer_info *ring_info,
-	void				*dest,
-	u32				destlen,
-	u32				start_read_offset)
-{
-	void *ring_buffer = hv_get_ring_buffer(ring_info);
-	u32 ring_buffer_size = hv_get_ring_buffersize(ring_info);
-
-	memcpy(dest, ring_buffer + start_read_offset, destlen);
-
-	start_read_offset += destlen;
-	if (start_read_offset >= ring_buffer_size)
-		start_read_offset -= ring_buffer_size;
-
-	return start_read_offset;
-}
-
-
-/*
  * Helper routine to copy from source to ring buffer.
  * Assume there is enough room. Handles wrap-around in dest case only!!
  */
@@ -334,33 +287,22 @@ int hv_ringbuffer_write(struct vmbus_cha
 	return 0;
 }
 
-static inline void
-init_cached_read_index(struct hv_ring_buffer_info *rbi)
-{
-	rbi->cached_read_index = rbi->ring_buffer->read_index;
-}
-
 int hv_ringbuffer_read(struct vmbus_channel *channel,
 		       void *buffer, u32 buflen, u32 *buffer_actual_len,
 		       u64 *requestid, bool raw)
 {
-	u32 bytes_avail_toread;
-	u32 next_read_location;
-	u64 prev_indices = 0;
-	struct vmpacket_descriptor desc;
-	u32 offset;
-	u32 packetlen;
-	struct hv_ring_buffer_info *inring_info = &channel->inbound;
+	struct vmpacket_descriptor *desc;
+	u32 packetlen, offset;
 
-	if (buflen <= 0)
+	if (unlikely(buflen == 0))
 		return -EINVAL;
 
 	*buffer_actual_len = 0;
 	*requestid = 0;
 
-	bytes_avail_toread = hv_get_bytes_to_read(inring_info);
 	/* Make sure there is something to read */
-	if (bytes_avail_toread < sizeof(desc)) {
+	desc = hv_pkt_iter_first(channel);
+	if (desc == NULL) {
 		/*
 		 * No error is set when there is even no header, drivers are
 		 * supposed to analyze buffer_actual_len.
@@ -368,48 +310,22 @@ int hv_ringbuffer_read(struct vmbus_chan
 		return 0;
 	}
 
-	init_cached_read_index(inring_info);
-
-	next_read_location = hv_get_next_read_location(inring_info);
-	next_read_location = hv_copyfrom_ringbuffer(inring_info, &desc,
-						    sizeof(desc),
-						    next_read_location);
-
-	offset = raw ? 0 : (desc.offset8 << 3);
-	packetlen = (desc.len8 << 3) - offset;
+	offset = raw ? 0 : (desc->offset8 << 3);
+	packetlen = (desc->len8 << 3) - offset;
 	*buffer_actual_len = packetlen;
-	*requestid = desc.trans_id;
-
-	if (bytes_avail_toread < packetlen + offset)
-		return -EAGAIN;
+	*requestid = desc->trans_id;
 
-	if (packetlen > buflen)
+	if (unlikely(packetlen > buflen))
 		return -ENOBUFS;
 
-	next_read_location =
-		hv_get_next_readlocation_withoffset(inring_info, offset);
+	/* since ring is double mapped, only one copy is necessary */
+	memcpy(buffer, (const char *)desc + offset, packetlen);
 
-	next_read_location = hv_copyfrom_ringbuffer(inring_info,
-						buffer,
-						packetlen,
-						next_read_location);
-
-	next_read_location = hv_copyfrom_ringbuffer(inring_info,
-						&prev_indices,
-						sizeof(u64),
-						next_read_location);
+	/* Advance ring index to next packet descriptor */
+	__hv_pkt_iter_next(channel, desc);
 
-	/*
-	 * Make sure all reads are done before we update the read index since
-	 * the writer may start writing to the read area once the read index
-	 * is updated.
-	 */
-	virt_mb();
-
-	/* Update the read index */
-	hv_set_next_read_location(inring_info, next_read_location);
-
-	hv_signal_on_read(channel);
+	/* Notify host of update */
+	hv_pkt_iter_close(channel);
 
 	return 0;
 }
@@ -442,7 +358,7 @@ struct vmpacket_descriptor *hv_pkt_iter_
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
 
 	/* set state for later hv_signal_on_read() */
-	init_cached_read_index(rbi);
+	rbi->cached_read_index = rbi->ring_buffer->read_index;
 
 	if (hv_pkt_iter_avail(rbi) < sizeof(struct vmpacket_descriptor))
 		return NULL;

[PATCH 4.13 09/11] vmbus: refactor hv_signal_on_read

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stephen Hemminger <stephen@networkplumber.org>

commit 8dd45f2ab005a1f3301296059b23b03ec3dbf79b upstream.

The function hv_signal_on_read was defined in hyperv.h and
only used in one place in ring_buffer code. Clearer to just
move it inline there.

Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/hv/ring_buffer.c |   32 ++++++++++++++++++++++++++++--
 include/linux/hyperv.h   |   49 -----------------------------------------------
 2 files changed, 30 insertions(+), 51 deletions(-)

--- a/drivers/hv/ring_buffer.c
+++ b/drivers/hv/ring_buffer.c
@@ -29,6 +29,7 @@
 #include <linux/uio.h>
 #include <linux/vmalloc.h>
 #include <linux/slab.h>
+#include <linux/prefetch.h>
 
 #include "hyperv_vmbus.h"
 
@@ -357,7 +358,7 @@ struct vmpacket_descriptor *hv_pkt_iter_
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
 
-	/* set state for later hv_signal_on_read() */
+	/* set state for later hv_pkt_iter_close */
 	rbi->cached_read_index = rbi->ring_buffer->read_index;
 
 	if (hv_pkt_iter_avail(rbi) < sizeof(struct vmpacket_descriptor))
@@ -400,6 +401,8 @@ EXPORT_SYMBOL_GPL(__hv_pkt_iter_next);
 void hv_pkt_iter_close(struct vmbus_channel *channel)
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
+	u32 cur_write_sz, cached_write_sz;
+	u32 pending_sz;
 
 	/*
 	 * Make sure all reads are done before we update the read index since
@@ -409,6 +412,31 @@ void hv_pkt_iter_close(struct vmbus_chan
 	virt_rmb();
 	rbi->ring_buffer->read_index = rbi->priv_read_index;
 
-	hv_signal_on_read(channel);
+	/*
+	 * Issue a full memory barrier before making the signaling decision.
+	 * Here is the reason for having this barrier:
+	 * If the reading of the pend_sz (in this function)
+	 * were to be reordered and read before we commit the new read
+	 * index (in the calling function)  we could
+	 * have a problem. If the host were to set the pending_sz after we
+	 * have sampled pending_sz and go to sleep before we commit the
+	 * read index, we could miss sending the interrupt. Issue a full
+	 * memory barrier to address this.
+	 */
+	virt_mb();
+
+	pending_sz = READ_ONCE(rbi->ring_buffer->pending_send_sz);
+	/* If the other end is not blocked on write don't bother. */
+	if (pending_sz == 0)
+		return;
+
+	cur_write_sz = hv_get_bytes_to_write(rbi);
+
+	if (cur_write_sz < pending_sz)
+		return;
+
+	cached_write_sz = hv_get_cached_bytes_to_write(rbi);
+	if (cached_write_sz < pending_sz)
+		vmbus_setevent(channel);
 }
 EXPORT_SYMBOL_GPL(hv_pkt_iter_close);
--- a/include/linux/hyperv.h
+++ b/include/linux/hyperv.h
@@ -1476,55 +1476,6 @@ hv_get_ring_buffer(const struct hv_ring_
 }
 
 /*
- * To optimize the flow management on the send-side,
- * when the sender is blocked because of lack of
- * sufficient space in the ring buffer, potential the
- * consumer of the ring buffer can signal the producer.
- * This is controlled by the following parameters:
- *
- * 1. pending_send_sz: This is the size in bytes that the
- *    producer is trying to send.
- * 2. The feature bit feat_pending_send_sz set to indicate if
- *    the consumer of the ring will signal when the ring
- *    state transitions from being full to a state where
- *    there is room for the producer to send the pending packet.
- */
-
-static inline  void hv_signal_on_read(struct vmbus_channel *channel)
-{
-	u32 cur_write_sz, cached_write_sz;
-	u32 pending_sz;
-	struct hv_ring_buffer_info *rbi = &channel->inbound;
-
-	/*
-	 * Issue a full memory barrier before making the signaling decision.
-	 * Here is the reason for having this barrier:
-	 * If the reading of the pend_sz (in this function)
-	 * were to be reordered and read before we commit the new read
-	 * index (in the calling function)  we could
-	 * have a problem. If the host were to set the pending_sz after we
-	 * have sampled pending_sz and go to sleep before we commit the
-	 * read index, we could miss sending the interrupt. Issue a full
-	 * memory barrier to address this.
-	 */
-	virt_mb();
-
-	pending_sz = READ_ONCE(rbi->ring_buffer->pending_send_sz);
-	/* If the other end is not blocked on write don't bother. */
-	if (pending_sz == 0)
-		return;
-
-	cur_write_sz = hv_get_bytes_to_write(rbi);
-
-	if (cur_write_sz < pending_sz)
-		return;
-
-	cached_write_sz = hv_get_cached_bytes_to_write(rbi);
-	if (cached_write_sz < pending_sz)
-		vmbus_setevent(channel);
-}
-
-/*
  * Mask off host interrupt callback notifications
  */
 static inline void hv_begin_read(struct hv_ring_buffer_info *rbi)

[PATCH 4.13 10/11] vmbus: eliminate duplicate cached index

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stephen Hemminger <stephen@networkplumber.org>

commit 05d00bc94ac27d220d8a78e365d7fa3a26dcca17 upstream.

Don't need cached read index anymore now that packet iterator
is used. The iterator has the original read index until the
visible read_index is updated.

Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/hv/ring_buffer.c |   17 ++++-------------
 include/linux/hyperv.h   |   14 --------------
 2 files changed, 4 insertions(+), 27 deletions(-)

--- a/drivers/hv/ring_buffer.c
+++ b/drivers/hv/ring_buffer.c
@@ -358,9 +358,6 @@ struct vmpacket_descriptor *hv_pkt_iter_
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
 
-	/* set state for later hv_pkt_iter_close */
-	rbi->cached_read_index = rbi->ring_buffer->read_index;
-
 	if (hv_pkt_iter_avail(rbi) < sizeof(struct vmpacket_descriptor))
 		return NULL;
 
@@ -388,10 +385,7 @@ __hv_pkt_iter_next(struct vmbus_channel
 		rbi->priv_read_index -= dsize;
 
 	/* more data? */
-	if (hv_pkt_iter_avail(rbi) < sizeof(struct vmpacket_descriptor))
-		return NULL;
-	else
-		return hv_get_ring_buffer(rbi) + rbi->priv_read_index;
+	return hv_pkt_iter_first(channel);
 }
 EXPORT_SYMBOL_GPL(__hv_pkt_iter_next);
 
@@ -401,7 +395,7 @@ EXPORT_SYMBOL_GPL(__hv_pkt_iter_next);
 void hv_pkt_iter_close(struct vmbus_channel *channel)
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
-	u32 cur_write_sz, cached_write_sz;
+	u32 orig_write_sz = hv_get_bytes_to_write(rbi);
 	u32 pending_sz;
 
 	/*
@@ -430,13 +424,10 @@ void hv_pkt_iter_close(struct vmbus_chan
 	if (pending_sz == 0)
 		return;
 
-	cur_write_sz = hv_get_bytes_to_write(rbi);
-
-	if (cur_write_sz < pending_sz)
+	if (hv_get_bytes_to_write(rbi) < pending_sz)
 		return;
 
-	cached_write_sz = hv_get_cached_bytes_to_write(rbi);
-	if (cached_write_sz < pending_sz)
+	if (orig_write_sz < pending_sz)
 		vmbus_setevent(channel);
 }
 EXPORT_SYMBOL_GPL(hv_pkt_iter_close);
--- a/include/linux/hyperv.h
+++ b/include/linux/hyperv.h
@@ -127,7 +127,6 @@ struct hv_ring_buffer_info {
 	u32 ring_data_startoffset;
 	u32 priv_write_index;
 	u32 priv_read_index;
-	u32 cached_read_index;
 };
 
 /*
@@ -180,19 +179,6 @@ static inline u32 hv_get_bytes_to_write(
 	return write;
 }
 
-static inline u32 hv_get_cached_bytes_to_write(
-	const struct hv_ring_buffer_info *rbi)
-{
-	u32 read_loc, write_loc, dsize, write;
-
-	dsize = rbi->ring_datasize;
-	read_loc = rbi->cached_read_index;
-	write_loc = rbi->ring_buffer->write_index;
-
-	write = write_loc >= read_loc ? dsize - (write_loc - read_loc) :
-		read_loc - write_loc;
-	return write;
-}
 /*
  * VMBUS version is 32 bit entity broken up into
  * two 16 bit quantities: major_number. minor_number.

[PATCH 4.13 11/11] vmbus: more host signalling avoidance

[Greg Kroah-Hartman]

4.13-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Stephen Hemminger <stephen@networkplumber.org>

commit 03bad714a1619c0074eb44d6f217c505fe27030f upstream.

Don't signal host if it has disabled interrupts for that
ring buffer. Check the feature bit to see if host supports
pending send size flag.

Signed-off-by: Stephen Hemminger <sthemmin@microsoft.com>
Signed-off-by: K. Y. Srinivasan <kys@microsoft.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

---
 drivers/hv/ring_buffer.c |   27 +++++++++++++++++++--------
 1 file changed, 19 insertions(+), 8 deletions(-)

--- a/drivers/hv/ring_buffer.c
+++ b/drivers/hv/ring_buffer.c
@@ -396,7 +396,6 @@ void hv_pkt_iter_close(struct vmbus_chan
 {
 	struct hv_ring_buffer_info *rbi = &channel->inbound;
 	u32 orig_write_sz = hv_get_bytes_to_write(rbi);
-	u32 pending_sz;
 
 	/*
 	 * Make sure all reads are done before we update the read index since
@@ -419,15 +418,27 @@ void hv_pkt_iter_close(struct vmbus_chan
 	 */
 	virt_mb();
 
-	pending_sz = READ_ONCE(rbi->ring_buffer->pending_send_sz);
-	/* If the other end is not blocked on write don't bother. */
-	if (pending_sz == 0)
+	/* If host has disabled notifications then skip */
+	if (rbi->ring_buffer->interrupt_mask)
 		return;
 
-	if (hv_get_bytes_to_write(rbi) < pending_sz)
-		return;
+	if (rbi->ring_buffer->feature_bits.feat_pending_send_sz) {
+		u32 pending_sz = READ_ONCE(rbi->ring_buffer->pending_send_sz);
+
+		/*
+		 * If there was space before we began iteration,
+		 * then host was not blocked. Also handles case where
+		 * pending_sz is zero then host has nothing pending
+		 * and does not need to be signaled.
+		 */
+		if (orig_write_sz > pending_sz)
+			return;
+
+		/* If pending write will not fit, don't give false hope. */
+		if (hv_get_bytes_to_write(rbi) < pending_sz)
+			return;
+	}
 
-	if (orig_write_sz < pending_sz)
-		vmbus_setevent(channel);
+	vmbus_setevent(channel);
 }
 EXPORT_SYMBOL_GPL(hv_pkt_iter_close);

Re: [PATCH 4.13 00/11] 4.13.9-stable review

[Guenter Roeck]

On 10/19/2017 06:38 AM, Greg Kroah-Hartman wrote:
> This is the start of the stable review cycle for the 4.13.9 release.
> There are 11 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 Sat Oct 21 13:16:37 UTC 2017.
> Anything received after that time might be too late.
> 


Build results:
	total: 145 pass: 145 fail: 0
Qemu test results:
	total: 123 pass: 123 fail: 0

Detauls are available at http://kerneltests.org/builders.

Guenter

Re: [PATCH 4.13 00/11] 4.13.9-stable review

[Greg Kroah-Hartman]

On Fri, Oct 20, 2017 at 06:06:35AM -0700, Guenter Roeck wrote:
> On 10/19/2017 06:38 AM, Greg Kroah-Hartman wrote:
> > This is the start of the stable review cycle for the 4.13.9 release.
> > There are 11 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 Sat Oct 21 13:16:37 UTC 2017.
> > Anything received after that time might be too late.
> > 
> 
> 
> Build results:
> 	total: 145 pass: 145 fail: 0
> Qemu test results:
> 	total: 123 pass: 123 fail: 0
> 
> Detauls are available at http://kerneltests.org/builders.

Thanks for testing all of these and letting me know.

greg k-h