When creating a new monitoring group, the RMID allocated for it may have
been used by a group which was previously removed. In this case, the
hardware counters will have non-zero values which should be deducted
from what is reported in the new group's counts.
resctrl_arch_reset_rmid() initializes the prev_msr value for counters to
0, causing the initial count to be charged to the new group. Resurrect
__rmid_read() and use it to initialize prev_msr correctly.
Unlike before, __rmid_read() checks for error bits in the MSR read so
that callers don't need to.
Fixes: 1d81d15db39c ("x86/resctrl: Move mbm_overflow_count() into resctrl_arch_rmid_read()")
Signed-off-by: Peter Newman <peternewman(a)google.com>
Reviewed-by: Reinette Chatre <reinette.chatre(a)intel.com>
Cc: stable(a)vger.kernel.org
---
v3:
- add changelog
- CC stable
v2:
- move error bit processing into __rmid_read()
v1: https://lore.kernel.org/lkml/20221207112924.3602960-1-peternewman@google.co…
v2: https://lore.kernel.org/lkml/20221214160856.2164207-1-peternewman@google.co…
---
arch/x86/kernel/cpu/resctrl/monitor.c | 49 ++++++++++++++++++---------
1 file changed, 33 insertions(+), 16 deletions(-)
diff --git a/arch/x86/kernel/cpu/resctrl/monitor.c b/arch/x86/kernel/cpu/resctrl/monitor.c
index efe0c30d3a12..77538abeb72a 100644
--- a/arch/x86/kernel/cpu/resctrl/monitor.c
+++ b/arch/x86/kernel/cpu/resctrl/monitor.c
@@ -146,6 +146,30 @@ static inline struct rmid_entry *__rmid_entry(u32 rmid)
return entry;
}
+static int __rmid_read(u32 rmid, enum resctrl_event_id eventid, u64 *val)
+{
+ u64 msr_val;
+
+ /*
+ * As per the SDM, when IA32_QM_EVTSEL.EvtID (bits 7:0) is configured
+ * with a valid event code for supported resource type and the bits
+ * IA32_QM_EVTSEL.RMID (bits 41:32) are configured with valid RMID,
+ * IA32_QM_CTR.data (bits 61:0) reports the monitored data.
+ * IA32_QM_CTR.Error (bit 63) and IA32_QM_CTR.Unavailable (bit 62)
+ * are error bits.
+ */
+ wrmsr(MSR_IA32_QM_EVTSEL, eventid, rmid);
+ rdmsrl(MSR_IA32_QM_CTR, msr_val);
+
+ if (msr_val & RMID_VAL_ERROR)
+ return -EIO;
+ if (msr_val & RMID_VAL_UNAVAIL)
+ return -EINVAL;
+
+ *val = msr_val;
+ return 0;
+}
+
static struct arch_mbm_state *get_arch_mbm_state(struct rdt_hw_domain *hw_dom,
u32 rmid,
enum resctrl_event_id eventid)
@@ -172,8 +196,12 @@ void resctrl_arch_reset_rmid(struct rdt_resource *r, struct rdt_domain *d,
struct arch_mbm_state *am;
am = get_arch_mbm_state(hw_dom, rmid, eventid);
- if (am)
+ if (am) {
memset(am, 0, sizeof(*am));
+
+ /* Record any initial, non-zero count value. */
+ __rmid_read(rmid, eventid, &am->prev_msr);
+ }
}
static u64 mbm_overflow_count(u64 prev_msr, u64 cur_msr, unsigned int width)
@@ -191,25 +219,14 @@ int resctrl_arch_rmid_read(struct rdt_resource *r, struct rdt_domain *d,
struct rdt_hw_domain *hw_dom = resctrl_to_arch_dom(d);
struct arch_mbm_state *am;
u64 msr_val, chunks;
+ int ret;
if (!cpumask_test_cpu(smp_processor_id(), &d->cpu_mask))
return -EINVAL;
- /*
- * As per the SDM, when IA32_QM_EVTSEL.EvtID (bits 7:0) is configured
- * with a valid event code for supported resource type and the bits
- * IA32_QM_EVTSEL.RMID (bits 41:32) are configured with valid RMID,
- * IA32_QM_CTR.data (bits 61:0) reports the monitored data.
- * IA32_QM_CTR.Error (bit 63) and IA32_QM_CTR.Unavailable (bit 62)
- * are error bits.
- */
- wrmsr(MSR_IA32_QM_EVTSEL, eventid, rmid);
- rdmsrl(MSR_IA32_QM_CTR, msr_val);
-
- if (msr_val & RMID_VAL_ERROR)
- return -EIO;
- if (msr_val & RMID_VAL_UNAVAIL)
- return -EINVAL;
+ ret = __rmid_read(rmid, eventid, &msr_val);
+ if (ret)
+ return ret;
am = get_arch_mbm_state(hw_dom, rmid, eventid);
if (am) {
base-commit: 830b3c68c1fb1e9176028d02ef86f3cf76aa2476
--
2.39.0.314.g84b9a713c41-goog
The preallocation of the maple tree nodes may leak if the error path to
"error_just_free" is taken. Fix this by moving the freeing of the maple
tree nodes to a shared location for all error paths.
Cc: stable(a)vger.kernel.org
Fixes: 8220543df148 ("nommu: remove uses of VMA linked list")
Signed-off-by: Liam R. Howlett <Liam.Howlett(a)oracle.com>
---
Changes since v1:
- Added 'Cc: stable(a)vger.kernel.org' to commit message
mm/nommu.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/mm/nommu.c b/mm/nommu.c
index 214c70e1d059..c8252f01d5db 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -1240,6 +1240,7 @@ unsigned long do_mmap(struct file *file,
error_just_free:
up_write(&nommu_region_sem);
error:
+ mas_destroy(&mas);
if (region->vm_file)
fput(region->vm_file);
kmem_cache_free(vm_region_jar, region);
@@ -1250,7 +1251,6 @@ unsigned long do_mmap(struct file *file,
sharing_violation:
up_write(&nommu_region_sem);
- mas_destroy(&mas);
pr_warn("Attempt to share mismatched mappings\n");
ret = -EINVAL;
goto error;
--
2.35.1
When removing a VMA from the tree fails due to no memory, do not free
the VMA since a reference still exists.
Cc: stable(a)vger.kernel.org
Fixes: 8220543df148 ("nommu: remove uses of VMA linked list")
Signed-off-by: Liam R. Howlett <Liam.Howlett(a)oracle.com>
---
Changes since v1:
- Added 'Cc: stable(a)vger.kernel.org' to commit message
mm/nommu.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/mm/nommu.c b/mm/nommu.c
index c8252f01d5db..844af5be7640 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -1509,7 +1509,8 @@ int do_munmap(struct mm_struct *mm, unsigned long start, size_t len, struct list
erase_whole_vma:
if (delete_vma_from_mm(vma))
ret = -ENOMEM;
- delete_vma(mm, vma);
+ else
+ delete_vma(mm, vma);
return ret;
}
--
2.35.1
During the maple tree conversion of nommu, an error in counting the VMAs
was introduced by counting the existing VMA again. The counting used to
be decremented by one and incremented by two, but now it only increments
by two. Fix the counting error by moving the increment outside the
setup_vma_to_mm() function to the callers.
Cc: stable(a)vger.kernel.org
Fixes: 8220543df148 ("nommu: remove uses of VMA linked list")
Signed-off-by: Liam R. Howlett <Liam.Howlett(a)oracle.com>
---
Changes since v1:
- Added 'Cc: stable(a)vger.kernel.org' to commit message
mm/nommu.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/mm/nommu.c b/mm/nommu.c
index 844af5be7640..5b83938ecb67 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -559,7 +559,6 @@ void vma_mas_remove(struct vm_area_struct *vma, struct ma_state *mas)
static void setup_vma_to_mm(struct vm_area_struct *vma, struct mm_struct *mm)
{
- mm->map_count++;
vma->vm_mm = mm;
/* add the VMA to the mapping */
@@ -587,6 +586,7 @@ static void mas_add_vma_to_mm(struct ma_state *mas, struct mm_struct *mm,
BUG_ON(!vma->vm_region);
setup_vma_to_mm(vma, mm);
+ mm->map_count++;
/* add the VMA to the tree */
vma_mas_store(vma, mas);
@@ -1347,6 +1347,7 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
if (vma->vm_file)
return -ENOMEM;
+ mm = vma->vm_mm;
if (mm->map_count >= sysctl_max_map_count)
return -ENOMEM;
@@ -1398,6 +1399,7 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
mas_set_range(&mas, vma->vm_start, vma->vm_end - 1);
mas_store(&mas, vma);
vma_mas_store(new, &mas);
+ mm->map_count++;
return 0;
err_mas_preallocate:
--
2.35.1
During the maple tree conversion of nommu, an error in counting the VMAs
was introduced by counting the existing VMA again. The counting used to
be decremented by one and incremented by two, but now it only increments
by two. Fix the counting error by moving the increment outside the
setup_vma_to_mm() function to the callers.
Cc: stable(a)vger.kernel.org
Fixes: 8220543df148 ("nommu: remove uses of VMA linked list")
Signed-off-by: Liam R. Howlett <Liam.Howlett(a)oracle.com>
---
Changes since v1:
- Added 'Cc: stable(a)vger.kernel.org' to commit message
mm/nommu.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/mm/nommu.c b/mm/nommu.c
index 844af5be7640..5b83938ecb67 100644
--- a/mm/nommu.c
+++ b/mm/nommu.c
@@ -559,7 +559,6 @@ void vma_mas_remove(struct vm_area_struct *vma, struct ma_state *mas)
static void setup_vma_to_mm(struct vm_area_struct *vma, struct mm_struct *mm)
{
- mm->map_count++;
vma->vm_mm = mm;
/* add the VMA to the mapping */
@@ -587,6 +586,7 @@ static void mas_add_vma_to_mm(struct ma_state *mas, struct mm_struct *mm,
BUG_ON(!vma->vm_region);
setup_vma_to_mm(vma, mm);
+ mm->map_count++;
/* add the VMA to the tree */
vma_mas_store(vma, mas);
@@ -1347,6 +1347,7 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
if (vma->vm_file)
return -ENOMEM;
+ mm = vma->vm_mm;
if (mm->map_count >= sysctl_max_map_count)
return -ENOMEM;
@@ -1398,6 +1399,7 @@ int split_vma(struct mm_struct *mm, struct vm_area_struct *vma,
mas_set_range(&mas, vma->vm_start, vma->vm_end - 1);
mas_store(&mas, vma);
vma_mas_store(new, &mas);
+ mm->map_count++;
return 0;
err_mas_preallocate:
--
2.35.1
The patch below does not apply to the 6.0-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
Possible dependencies:
b389286d0234 ("drm/mgag200: Fix PLL setup for G200_SE_A rev >=4")
877507bb954e ("drm/mgag200: Provide per-device callbacks for PIXPLLC")
8aeeb3144fe2 ("drm/mgag200: Provide per-device callbacks for BMC synchronization")
f639f74a7895 ("drm/mgag200: Add per-device callbacks")
1baf9127c482 ("drm/mgag200: Replace simple-KMS with regular atomic helpers")
4f4dc37e374c ("drm/mgag200: Reorganize before dropping simple-KMS helpers")
ed2ef21f1089 ("drm/mgag200: Store primary plane's color format in CRTC state")
2d70b9a1482e ("drm/mgag200: Acquire I/O-register lock in atomic_commit_tail function")
1ee181fe958a ("drm/mgag200: Move DAC-register setup into model-specific code")
44373151ab42 ("drm/mgag200: Split mgag200_modeset_init()")
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
From b389286d0234e1edbaf62ed8bc0892a568c33662 Mon Sep 17 00:00:00 2001
From: Jocelyn Falempe <jfalempe(a)redhat.com>
Date: Thu, 13 Oct 2022 15:28:10 +0200
Subject: [PATCH] drm/mgag200: Fix PLL setup for G200_SE_A rev >=4
For G200_SE_A, PLL M setting is wrong, which leads to blank screen,
or "signal out of range" on VGA display.
previous code had "m |= 0x80" which was changed to
m |= ((pixpllcn & BIT(8)) >> 1);
Tested on G200_SE_A rev 42
This line of code was moved to another file with
commit 877507bb954e ("drm/mgag200: Provide per-device callbacks for
PIXPLLC") but can be easily backported before this commit.
v2: * put BIT(7) First to respect MSB-to-LSB (Thomas)
* Add a comment to explain that this bit must be set (Thomas)
Fixes: 2dd040946ecf ("drm/mgag200: Store values (not bits) in struct mgag200_pll_values")
Cc: stable(a)vger.kernel.org
Signed-off-by: Jocelyn Falempe <jfalempe(a)redhat.com>
Reviewed-by: Thomas Zimmermann <tzimmermann(a)suse.de>
Link: https://patchwork.freedesktop.org/patch/msgid/20221013132810.521945-1-jfale…
diff --git a/drivers/gpu/drm/mgag200/mgag200_g200se.c b/drivers/gpu/drm/mgag200/mgag200_g200se.c
index be389ed91cbd..bd6e573c9a1a 100644
--- a/drivers/gpu/drm/mgag200/mgag200_g200se.c
+++ b/drivers/gpu/drm/mgag200/mgag200_g200se.c
@@ -284,7 +284,8 @@ static void mgag200_g200se_04_pixpllc_atomic_update(struct drm_crtc *crtc,
pixpllcp = pixpllc->p - 1;
pixpllcs = pixpllc->s;
- xpixpllcm = pixpllcm | ((pixpllcn & BIT(8)) >> 1);
+ // For G200SE A, BIT(7) should be set unconditionally.
+ xpixpllcm = BIT(7) | pixpllcm;
xpixpllcn = pixpllcn;
xpixpllcp = (pixpllcs << 3) | pixpllcp;
[Public]
Hi,
The following commit fixes some s2idle wakeup problems with certain firmware on AMD Rembrandt systems with Qualcomm WLAN cards.
3f9b09ccf7d5 ("wifi: ath11k: Send PME message during wakeup from D3cold")
Can you please bring this to 6.1.y?
Thanks,