The quilt patch titled
Subject: mm/hugetlb: fix hugetlb_pmd_shared()
has been removed from the -mm tree. Its filename was
mm-hugetlb-fix-hugetlb_pmd_shared.patch
This patch was dropped because it was merged into the mm-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: "David Hildenbrand (Red Hat)" <david(a)kernel.org>
Subject: mm/hugetlb: fix hugetlb_pmd_shared()
Date: Fri, 5 Dec 2025 22:35:55 +0100
Patch series "mm/hugetlb: fixes for PMD table sharing (incl. using
mmu_gather)".
One functional fix, one performance regression fix, and two related
comment fixes.
I cleaned up my prototype I recently shared [1] for the performance fix,
deferring most of the cleanups I had in the prototype to a later point.
While doing that I identified the other things.
The goal of this patch set is to be backported to stable trees "fairly"
easily. At least patch #1 and #4.
Patch #1 fixes hugetlb_pmd_shared() not detecting any sharing
Patch #2 + #3 are simple comment fixes that patch #4 interacts with.
Patch #4 is a fix for the reported performance regression due to excessive
IPI broadcasts during fork()+exit().
The last patch is all about TLB flushes, IPIs and mmu_gather. Read:
complicated
I added as much comments + description that I possibly could, and I am
hoping for review from Jann.
There are plenty of cleanups in the future to be had + one reasonable
optimization on x86. But that's all out of scope for this series.
This patch (of 4):
We switched from (wrongly) using the page count to an independent shared
count. Now, shared page tables have a refcount of 1 (excluding
speculative references) and instead use ptdesc->pt_share_count to identify
sharing.
We didn't convert hugetlb_pmd_shared(), so right now, we would never
detect a shared PMD table as such, because sharing/unsharing no longer
touches the refcount of a PMD table.
Page migration, like mbind() or migrate_pages() would allow for migrating
folios mapped into such shared PMD tables, even though the folios are not
exclusive. In smaps we would account them as "private" although they are
"shared", and we would be wrongly setting the PM_MMAP_EXCLUSIVE in the
pagemap interface.
Fix it by properly using ptdesc_pmd_is_shared() in hugetlb_pmd_shared().
Link: https://lkml.kernel.org/r/20251205213558.2980480-1-david@kernel.org
Link: https://lkml.kernel.org/r/20251205213558.2980480-2-david@kernel.org
Link: https://lore.kernel.org/all/8cab934d-4a56-44aa-b641-bfd7e23bd673@kernel.org/ [1]
Fixes: 59d9094df3d7 ("mm: hugetlb: independent PMD page table shared count")
Signed-off-by: David Hildenbrand (Red Hat) <david(a)kernel.org>
Tested-by: Laurence Oberman <loberman(a)redhat.com>
Reviewed-by: Rik van Riel <riel(a)surriel.com>
Reviewed-by: Lance Yang <lance.yang(a)linux.dev>
Cc: Liu Shixin <liushixin2(a)huawei.com>
Cc: "Aneesh Kumar K.V" <aneesh.kumar(a)kernel.org>
Cc: Arnd Bergmann <arnd(a)arndb.de>
Cc: Jann Horn <jannh(a)google.com>
Cc: Liam Howlett <liam.howlett(a)oracle.com>
Cc: Lorenzo Stoakes <lorenzo.stoakes(a)oracle.com>
Cc: Muchun Song <muchun.song(a)linux.dev>
Cc: Nadav Amit <nadav.amit(a)gmail.com>
Cc: Nicholas Piggin <npiggin(a)gmail.com>
Cc: Oscar Salvador <osalvador(a)suse.de>
Cc: Peter Zijlstra <peterz(a)infradead.org>
Cc: Prakash Sangappa <prakash.sangappa(a)oracle.com>
Cc: Vlastimil Babka <vbabka(a)suse.cz>
Cc: Will Deacon <will(a)kernel.org>
Cc: Uschakow, Stanislav" <suschako(a)amazon.de>
Cc: <stable(a)vger.kernel.org>
Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org>
---
include/linux/hugetlb.h | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/include/linux/hugetlb.h~mm-hugetlb-fix-hugetlb_pmd_shared
+++ a/include/linux/hugetlb.h
@@ -1326,7 +1326,7 @@ static inline __init void hugetlb_cma_re
#ifdef CONFIG_HUGETLB_PMD_PAGE_TABLE_SHARING
static inline bool hugetlb_pmd_shared(pte_t *pte)
{
- return page_count(virt_to_page(pte)) > 1;
+ return ptdesc_pmd_is_shared(virt_to_ptdesc(pte));
}
#else
static inline bool hugetlb_pmd_shared(pte_t *pte)
_
Patches currently in -mm which might be from david(a)kernel.org are
The quilt patch titled
Subject: powerpc/pseries/cmm: adjust BALLOON_MIGRATE when migrating pages
has been removed from the -mm tree. Its filename was
powerpc-pseries-cmm-adjust-balloon_migrate-when-migrating-pages.patch
This patch was dropped because it was merged into the mm-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: David Hildenbrand <david(a)redhat.com>
Subject: powerpc/pseries/cmm: adjust BALLOON_MIGRATE when migrating pages
Date: Tue, 21 Oct 2025 12:06:06 +0200
Let's properly adjust BALLOON_MIGRATE like the other drivers.
Note that the INFLATE/DEFLATE events are triggered from the core when
enqueueing/dequeueing pages.
This was found by code inspection.
Link: https://lkml.kernel.org/r/20251021100606.148294-3-david@redhat.com
Fixes: fe030c9b85e6 ("powerpc/pseries/cmm: Implement balloon compaction")
Signed-off-by: David Hildenbrand <david(a)redhat.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list(a)gmail.com>
Cc: Christophe Leroy <christophe.leroy(a)csgroup.eu>
Cc: Madhavan Srinivasan <maddy(a)linux.ibm.com>
Cc: Michael Ellerman <mpe(a)ellerman.id.au>
Cc: Nicholas Piggin <npiggin(a)gmail.com>
Cc: <stable(a)vger.kernel.org>
Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org>
---
arch/powerpc/platforms/pseries/cmm.c | 1 +
1 file changed, 1 insertion(+)
--- a/arch/powerpc/platforms/pseries/cmm.c~powerpc-pseries-cmm-adjust-balloon_migrate-when-migrating-pages
+++ a/arch/powerpc/platforms/pseries/cmm.c
@@ -532,6 +532,7 @@ static int cmm_migratepage(struct balloo
spin_lock_irqsave(&b_dev_info->pages_lock, flags);
balloon_page_insert(b_dev_info, newpage);
+ __count_vm_event(BALLOON_MIGRATE);
b_dev_info->isolated_pages--;
spin_unlock_irqrestore(&b_dev_info->pages_lock, flags);
_
Patches currently in -mm which might be from david(a)redhat.com are
The quilt patch titled
Subject: powerpc/pseries/cmm: call balloon_devinfo_init() also without CONFIG_BALLOON_COMPACTION
has been removed from the -mm tree. Its filename was
powerpc-pseries-cmm-call-balloon_devinfo_init-also-without-config_balloon_compaction.patch
This patch was dropped because it was merged into the mm-stable branch
of git://git.kernel.org/pub/scm/linux/kernel/git/akpm/mm
------------------------------------------------------
From: David Hildenbrand <david(a)redhat.com>
Subject: powerpc/pseries/cmm: call balloon_devinfo_init() also without CONFIG_BALLOON_COMPACTION
Date: Tue, 21 Oct 2025 12:06:05 +0200
Patch series "powerpc/pseries/cmm: two smaller fixes".
Two smaller fixes identified while doing a bigger rework.
This patch (of 2):
We always have to initialize the balloon_dev_info, even when compaction is
not configured in: otherwise the containing list and the lock are left
uninitialized.
Likely not many such configs exist in practice, but let's CC stable to
be sure.
This was found by code inspection.
Link: https://lkml.kernel.org/r/20251021100606.148294-1-david@redhat.com
Link: https://lkml.kernel.org/r/20251021100606.148294-2-david@redhat.com
Fixes: fe030c9b85e6 ("powerpc/pseries/cmm: Implement balloon compaction")
Signed-off-by: David Hildenbrand <david(a)redhat.com>
Reviewed-by: Ritesh Harjani (IBM) <ritesh.list(a)gmail.com>
Cc: Christophe Leroy <christophe.leroy(a)csgroup.eu>
Cc: Madhavan Srinivasan <maddy(a)linux.ibm.com>
Cc: Michael Ellerman <mpe(a)ellerman.id.au>
Cc: Nicholas Piggin <npiggin(a)gmail.com>
Cc: <stable(a)vger.kernel.org>
Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org>
---
arch/powerpc/platforms/pseries/cmm.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
--- a/arch/powerpc/platforms/pseries/cmm.c~powerpc-pseries-cmm-call-balloon_devinfo_init-also-without-config_balloon_compaction
+++ a/arch/powerpc/platforms/pseries/cmm.c
@@ -550,7 +550,6 @@ static int cmm_migratepage(struct balloo
static void cmm_balloon_compaction_init(void)
{
- balloon_devinfo_init(&b_dev_info);
b_dev_info.migratepage = cmm_migratepage;
}
#else /* CONFIG_BALLOON_COMPACTION */
@@ -572,6 +571,7 @@ static int cmm_init(void)
if (!firmware_has_feature(FW_FEATURE_CMO) && !simulate)
return -EOPNOTSUPP;
+ balloon_devinfo_init(&b_dev_info);
cmm_balloon_compaction_init();
rc = register_oom_notifier(&cmm_oom_nb);
_
Patches currently in -mm which might be from david(a)redhat.com are
This reverts commit 3b5e5185881edf4ee5a1af575e3aedac4a38a764.
The REO queue lookup table feature was enabled in 6.12.y due to an
upstream backport, but it causes severe RX performance degradation on
QCN9274 hw2.0 devices.
With this feature enabled, the vast majority of received packets are
dropped, reducing throughput drastically and making the device nearly
unusable.
Reverting this change restores full RX performance.
Tested-on: QCN9274 hw2.0 PCI WLAN.WBE.1.5-01651-QCAHKSWPL_SILICONZ-1
Fixes: 3b5e5185881e ("wifi: ath12k: Enable REO queue lookup table feature on QCN9274 hw2.0")
Signed-off-by: Oliver Sedlbauer <os(a)dev.tdt.de>
---
Note:
This commit reverts a backport that was not a fix. The backported change
breaks previously working behavior on QCN9274 hw2.0 devices and should
not have been applied to the 6.12.y stable kernel.
drivers/net/wireless/ath/ath12k/hw.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/net/wireless/ath/ath12k/hw.c b/drivers/net/wireless/ath/ath12k/hw.c
index 057ef2d282b2..e3eb22bb9e1c 100644
--- a/drivers/net/wireless/ath/ath12k/hw.c
+++ b/drivers/net/wireless/ath/ath12k/hw.c
@@ -1084,7 +1084,7 @@ static const struct ath12k_hw_params ath12k_hw_params[] = {
.download_calib = true,
.supports_suspend = false,
.tcl_ring_retry = true,
- .reoq_lut_support = true,
+ .reoq_lut_support = false,
.supports_shadow_regs = false,
.num_tcl_banks = 48,
--
2.39.5
The following commit has been merged into the perf/urgent branch of tip:
Commit-ID: 01439286514ce9d13b8123f8ec3717d7135ff1d6
Gitweb: https://git.kernel.org/tip/01439286514ce9d13b8123f8ec3717d7135ff1d6
Author: Sandipan Das <sandipan.das(a)amd.com>
AuthorDate: Tue, 09 Dec 2025 13:56:38 +05:30
Committer: Ingo Molnar <mingo(a)kernel.org>
CommitterDate: Tue, 09 Dec 2025 12:21:47 +01:00
perf/x86/amd/uncore: Fix the return value of amd_uncore_df_event_init() on error
If amd_uncore_event_init() fails, return an error irrespective of the
pmu_version. Setting hwc->config should be safe even if there is an
error so use this opportunity to simplify the code.
Closes: https://lore.kernel.org/all/aTaI0ci3vZ44lmBn@stanley.mountain/
Fixes: d6389d3ccc13 ("perf/x86/amd/uncore: Refactor uncore management")
Reported-by: Dan Carpenter <dan.carpenter(a)linaro.org>
Signed-off-by: Sandipan Das <sandipan.das(a)amd.com>
Signed-off-by: Ingo Molnar <mingo(a)kernel.org>
Cc: Peter Zijlstra <peterz(a)infradead.org>
Cc: stable(a)vger.kernel.org
Link: https://patch.msgid.link/076935e23a70335d33bd6e23308b75ae0ad35ba2.176526866…
---
arch/x86/events/amd/uncore.c | 5 +----
1 file changed, 1 insertion(+), 4 deletions(-)
diff --git a/arch/x86/events/amd/uncore.c b/arch/x86/events/amd/uncore.c
index e8b6af1..9293ce5 100644
--- a/arch/x86/events/amd/uncore.c
+++ b/arch/x86/events/amd/uncore.c
@@ -656,14 +656,11 @@ static int amd_uncore_df_event_init(struct perf_event *event)
struct hw_perf_event *hwc = &event->hw;
int ret = amd_uncore_event_init(event);
- if (ret || pmu_version < 2)
- return ret;
-
hwc->config = event->attr.config &
(pmu_version >= 2 ? AMD64_PERFMON_V2_RAW_EVENT_MASK_NB :
AMD64_RAW_EVENT_MASK_NB);
- return 0;
+ return ret;
}
static int amd_uncore_df_add(struct perf_event *event, int flags)
On hardware based on Toradex Verdin AM62 the recovery mechanism added by
commit ad5c6ecef27e ("drm: bridge: ti-sn65dsi83: Add error recovery
mechanism") has been reported [0] to make the display turn on and off and
and the kernel logging "Unexpected link status 0x01".
According to the report, the error recovery mechanism is triggered by the
PLL_UNLOCK error going active. Analysis suggested the board is unable to
provide the correct DSI clock neede by the SN65DSI84, to which the TI
SN65DSI84 reacts by raising the PLL_UNLOCK, while the display still works
apparently without issues.
On other hardware, where all the clocks are within the components
specifications, the PLL_UNLOCK bit does not trigger while the display is in
normal use. It can trigger for e.g. electromagnetic interference, which is
a transient event and exactly the reason why the error recovery mechanism
has been implemented.
Idelly the PLL_UNLOCK bit could be ignored when working out of
specification, but this requires to detect in software whether it triggers
because the device is working out of specification but visually correctly
for the user or for good reasons (e.g. EMI, or even because working out of
specifications but compromising the visual output).
The ongoing analysis as of this writing [1][2] has not yet found a way for
the driver to discriminate among the two cases. So as a temporary measure
mask the PLL_UNLOCK error bit unconditionally.
[0] https://lore.kernel.org/r/bhkn6hley4xrol5o3ytn343h4unkwsr26p6s6ltcwexnrsjsd…
[1] https://lore.kernel.org/all/b71e941c-fc8a-4ac1-9407-0fe7df73b412@gmail.com/
[2] https://lore.kernel.org/all/20251125103900.31750-1-francesco@dolcini.it/
Closes: https://lore.kernel.org/r/bhkn6hley4xrol5o3ytn343h4unkwsr26p6s6ltcwexnrsjsd…
Cc: stable(a)vger.kernel.org # 6.15+
Co-developed-by: Hervé Codina <herve.codina(a)bootlin.com>
Signed-off-by: Hervé Codina <herve.codina(a)bootlin.com>
Signed-off-by: Luca Ceresoli <luca.ceresoli(a)bootlin.com>
---
Francesco, Emanuele, João: can you please apply this patch and report
whether the display on the affected boards gets back to working as before?
Cc: João Paulo Gonçalves <jpaulo.silvagoncalves(a)gmail.com>
Cc: Francesco Dolcini <francesco(a)dolcini.it>
Cc: Emanuele Ghidoli <ghidoliemanuele(a)gmail.com>
---
drivers/gpu/drm/bridge/ti-sn65dsi83.c | 11 +++++++++--
1 file changed, 9 insertions(+), 2 deletions(-)
diff --git a/drivers/gpu/drm/bridge/ti-sn65dsi83.c b/drivers/gpu/drm/bridge/ti-sn65dsi83.c
index 033c44326552..fffb47b62f43 100644
--- a/drivers/gpu/drm/bridge/ti-sn65dsi83.c
+++ b/drivers/gpu/drm/bridge/ti-sn65dsi83.c
@@ -429,7 +429,14 @@ static void sn65dsi83_handle_errors(struct sn65dsi83 *ctx)
*/
ret = regmap_read(ctx->regmap, REG_IRQ_STAT, &irq_stat);
- if (ret || irq_stat) {
+
+ /*
+ * Some hardware (Toradex Verdin AM62) is known to report the
+ * PLL_UNLOCK error interrupt while working without visible
+ * problems. In lack of a reliable way to discriminate such cases
+ * from user-visible PLL_UNLOCK cases, ignore that bit entirely.
+ */
+ if (ret || irq_stat & ~REG_IRQ_STAT_CHA_PLL_UNLOCK) {
/*
* IRQ acknowledged is not always possible (the bridge can be in
* a state where it doesn't answer anymore). To prevent an
@@ -654,7 +661,7 @@ static void sn65dsi83_atomic_enable(struct drm_bridge *bridge,
if (ctx->irq) {
/* Enable irq to detect errors */
regmap_write(ctx->regmap, REG_IRQ_GLOBAL, REG_IRQ_GLOBAL_IRQ_EN);
- regmap_write(ctx->regmap, REG_IRQ_EN, 0xff);
+ regmap_write(ctx->regmap, REG_IRQ_EN, 0xff & ~REG_IRQ_EN_CHA_PLL_UNLOCK_EN);
} else {
/* Use the polling task */
sn65dsi83_monitor_start(ctx);
---
base-commit: c884ee70b15a8d63184d7c1e02eba99676a6fcf7
change-id: 20251126-drm-ti-sn65dsi83-ignore-pll-unlock-4a28aa29eb5c
Best regards,
--
Luca Ceresoli <luca.ceresoli(a)bootlin.com>