From: Kan Liang <kan.liang(a)linux.intel.com>
Currently, the Sapphire Rapids and Granite Rapids share the same PMU
name, sapphire_rapids. Because from the kernel’s perspective, GNR is
similar to SPR. The only key difference is that they support different
extra MSRs. The code path and the PMU name are shared.
However, from end users' perspective, they are quite different. Besides
the extra MSRs, GNR has a newer PEBS format, supports Retire Latency,
supports new CPUID enumeration architecture, doesn't required the
load-latency AUX event, has additional TMA Level 1 Architectural Events,
etc. The differences can be enumerated by CPUID or the PERF_CAPABILITIES
MSR. They weren't reflected in the model-specific kernel setup.
But it is worth to have a distinct PMU name for GNR.
Fixes: a6742cb90b56 ("perf/x86/intel: Fix the FRONTEND encoding on GNR and MTL")
Suggested-by: Ahmad Yasin <ahmad.yasin(a)intel.com>
Signed-off-by: Kan Liang <kan.liang(a)linux.intel.com>
Cc: stable(a)vger.kernel.org
---
arch/x86/events/intel/core.c | 9 ++++++---
1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/arch/x86/events/intel/core.c b/arch/x86/events/intel/core.c
index f3315f13f920..da38a16b2cbc 100644
--- a/arch/x86/events/intel/core.c
+++ b/arch/x86/events/intel/core.c
@@ -6768,12 +6768,17 @@ __init int intel_pmu_init(void)
case INTEL_FAM6_EMERALDRAPIDS_X:
x86_pmu.flags |= PMU_FL_MEM_LOADS_AUX;
x86_pmu.extra_regs = intel_glc_extra_regs;
+ pr_cont("Sapphire Rapids events, ");
+ name = "sapphire_rapids";
fallthrough;
case INTEL_FAM6_GRANITERAPIDS_X:
case INTEL_FAM6_GRANITERAPIDS_D:
intel_pmu_init_glc(NULL);
- if (!x86_pmu.extra_regs)
+ if (!x86_pmu.extra_regs) {
x86_pmu.extra_regs = intel_rwc_extra_regs;
+ pr_cont("Granite Rapids events, ");
+ name = "granite_rapids";
+ }
x86_pmu.pebs_ept = 1;
x86_pmu.hw_config = hsw_hw_config;
x86_pmu.get_event_constraints = glc_get_event_constraints;
@@ -6784,8 +6789,6 @@ __init int intel_pmu_init(void)
td_attr = glc_td_events_attrs;
tsx_attr = glc_tsx_events_attrs;
intel_pmu_pebs_data_source_skl(true);
- pr_cont("Sapphire Rapids events, ");
- name = "sapphire_rapids";
break;
case INTEL_FAM6_ALDERLAKE:
--
2.35.1
Qualcomm Bluetooth controllers may not have been provisioned with a
valid device address and instead end up using the default address
00:00:00:00:5a:ad.
This was previously believed to be due to lack of persistent storage for
the address but it may also be due to integrators opting to not use the
on-chip OTP memory and instead store the address elsewhere (e.g. in
storage managed by secure world firmware).
According to Qualcomm, at least WCN6750, WCN6855 and WCN7850 have
on-chip OTP storage for the address.
As the device type alone cannot be used to determine when the address is
valid, instead read back the address during setup() and only set the
HCI_QUIRK_USE_BDADDR_PROPERTY flag when needed.
This specifically makes sure that controllers that have been provisioned
with an address do not start as unconfigured.
Reported-by: Janaki Ramaiah Thota <quic_janathot(a)quicinc.com>
Link: https://lore.kernel.org/r/124a7d54-5a18-4be7-9a76-a12017f6cce5@quicinc.com/
Fixes: 5971752de44c ("Bluetooth: hci_qca: Set HCI_QUIRK_USE_BDADDR_PROPERTY for wcn3990")
Fixes: e668eb1e1578 ("Bluetooth: hci_core: Don't stop BT if the BD address missing in dts")
Fixes: 6945795bc81a ("Bluetooth: fix use-bdaddr-property quirk")
Cc: stable(a)vger.kernel.org # 6.5
Cc: Matthias Kaehlcke <mka(a)chromium.org>
Signed-off-by: Johan Hovold <johan+linaro(a)kernel.org>
---
drivers/bluetooth/btqca.c | 38 +++++++++++++++++++++++++++++++++++++
drivers/bluetooth/hci_qca.c | 2 --
2 files changed, 38 insertions(+), 2 deletions(-)
Matthias and Doug,
As Chromium is the only known user of the 'local-bd-address' property,
could you please confirm that your controllers use the 00:00:00:00:5a:ad
address by default so that the quirk continues to be set as intended?
Johan
diff --git a/drivers/bluetooth/btqca.c b/drivers/bluetooth/btqca.c
index 19cfc342fc7b..216826c31ee3 100644
--- a/drivers/bluetooth/btqca.c
+++ b/drivers/bluetooth/btqca.c
@@ -15,6 +15,8 @@
#define VERSION "0.1"
+#define QCA_BDADDR_DEFAULT (&(bdaddr_t) {{ 0xad, 0x5a, 0x00, 0x00, 0x00, 0x00 }})
+
int qca_read_soc_version(struct hci_dev *hdev, struct qca_btsoc_version *ver,
enum qca_btsoc_type soc_type)
{
@@ -612,6 +614,38 @@ int qca_set_bdaddr_rome(struct hci_dev *hdev, const bdaddr_t *bdaddr)
}
EXPORT_SYMBOL_GPL(qca_set_bdaddr_rome);
+static int qca_check_bdaddr(struct hci_dev *hdev)
+{
+ struct hci_rp_read_bd_addr *bda;
+ struct sk_buff *skb;
+ int err;
+
+ if (bacmp(&hdev->public_addr, BDADDR_ANY))
+ return 0;
+
+ skb = __hci_cmd_sync(hdev, HCI_OP_READ_BD_ADDR, 0, NULL,
+ HCI_INIT_TIMEOUT);
+ if (IS_ERR(skb)) {
+ err = PTR_ERR(skb);
+ bt_dev_err(hdev, "Failed to read device address (%d)", err);
+ return err;
+ }
+
+ if (skb->len != sizeof(*bda)) {
+ bt_dev_err(hdev, "Device address length mismatch");
+ kfree_skb(skb);
+ return -EIO;
+ }
+
+ bda = (struct hci_rp_read_bd_addr *)skb->data;
+ if (!bacmp(&bda->bdaddr, QCA_BDADDR_DEFAULT))
+ set_bit(HCI_QUIRK_USE_BDADDR_PROPERTY, &hdev->quirks);
+
+ kfree_skb(skb);
+
+ return 0;
+}
+
static void qca_generate_hsp_nvm_name(char *fwname, size_t max_size,
struct qca_btsoc_version ver, u8 rom_ver, u16 bid)
{
@@ -818,6 +852,10 @@ int qca_uart_setup(struct hci_dev *hdev, uint8_t baudrate,
break;
}
+ err = qca_check_bdaddr(hdev);
+ if (err)
+ return err;
+
bt_dev_info(hdev, "QCA setup on UART is completed");
return 0;
diff --git a/drivers/bluetooth/hci_qca.c b/drivers/bluetooth/hci_qca.c
index ecbc52eaf101..92fa20f5ac7d 100644
--- a/drivers/bluetooth/hci_qca.c
+++ b/drivers/bluetooth/hci_qca.c
@@ -1905,8 +1905,6 @@ static int qca_setup(struct hci_uart *hu)
case QCA_WCN6750:
case QCA_WCN6855:
case QCA_WCN7850:
- set_bit(HCI_QUIRK_USE_BDADDR_PROPERTY, &hdev->quirks);
-
qcadev = serdev_device_get_drvdata(hu->serdev);
if (qcadev->bdaddr_property_broken)
set_bit(HCI_QUIRK_BDADDR_PROPERTY_BROKEN, &hdev->quirks);
--
2.43.2
The Elan eKTH5015M touch controller found on the Lenovo ThinkPad X13s
shares the VCC33 supply with other peripherals that may remain powered
during suspend (e.g. when enabled as wakeup sources).
The reset line is also wired so that it can be left deasserted when the
supply is off.
This is important as it avoids holding the controller in reset for
extended periods of time when it remains powered, which can lead to
increased power consumption, and also avoids leaking current through the
X13s reset circuitry during suspend (and after driver unbind).
Use the new 'no-reset-on-power-off' devicetree property to determine
when reset needs to be asserted on power down.
Notably this also avoids wasting power on machine variants without a
touchscreen for which the driver would otherwise exit probe with reset
asserted.
Fixes: bd3cba00dcc6 ("HID: i2c-hid: elan: Add support for Elan eKTH6915 i2c-hid touchscreens")
Cc: stable(a)vger.kernel.org # 6.0
Cc: Douglas Anderson <dianders(a)chromium.org>
Signed-off-by: Johan Hovold <johan+linaro(a)kernel.org>
---
drivers/hid/i2c-hid/i2c-hid-of-elan.c | 37 ++++++++++++++++++++-------
1 file changed, 28 insertions(+), 9 deletions(-)
diff --git a/drivers/hid/i2c-hid/i2c-hid-of-elan.c b/drivers/hid/i2c-hid/i2c-hid-of-elan.c
index 5b91fb106cfc..8a905027d5e9 100644
--- a/drivers/hid/i2c-hid/i2c-hid-of-elan.c
+++ b/drivers/hid/i2c-hid/i2c-hid-of-elan.c
@@ -31,6 +31,7 @@ struct i2c_hid_of_elan {
struct regulator *vcc33;
struct regulator *vccio;
struct gpio_desc *reset_gpio;
+ bool no_reset_on_power_off;
const struct elan_i2c_hid_chip_data *chip_data;
};
@@ -40,17 +41,17 @@ static int elan_i2c_hid_power_up(struct i2chid_ops *ops)
container_of(ops, struct i2c_hid_of_elan, ops);
int ret;
+ gpiod_set_value_cansleep(ihid_elan->reset_gpio, 1);
+
if (ihid_elan->vcc33) {
ret = regulator_enable(ihid_elan->vcc33);
if (ret)
- return ret;
+ goto err_deassert_reset;
}
ret = regulator_enable(ihid_elan->vccio);
- if (ret) {
- regulator_disable(ihid_elan->vcc33);
- return ret;
- }
+ if (ret)
+ goto err_disable_vcc33;
if (ihid_elan->chip_data->post_power_delay_ms)
msleep(ihid_elan->chip_data->post_power_delay_ms);
@@ -60,6 +61,15 @@ static int elan_i2c_hid_power_up(struct i2chid_ops *ops)
msleep(ihid_elan->chip_data->post_gpio_reset_on_delay_ms);
return 0;
+
+err_disable_vcc33:
+ if (ihid_elan->vcc33)
+ regulator_disable(ihid_elan->vcc33);
+err_deassert_reset:
+ if (ihid_elan->no_reset_on_power_off)
+ gpiod_set_value_cansleep(ihid_elan->reset_gpio, 0);
+
+ return ret;
}
static void elan_i2c_hid_power_down(struct i2chid_ops *ops)
@@ -67,7 +77,14 @@ static void elan_i2c_hid_power_down(struct i2chid_ops *ops)
struct i2c_hid_of_elan *ihid_elan =
container_of(ops, struct i2c_hid_of_elan, ops);
- gpiod_set_value_cansleep(ihid_elan->reset_gpio, 1);
+ /*
+ * Do not assert reset when the hardware allows for it to remain
+ * deasserted regardless of the state of the (shared) power supply to
+ * avoid wasting power when the supply is left on.
+ */
+ if (!ihid_elan->no_reset_on_power_off)
+ gpiod_set_value_cansleep(ihid_elan->reset_gpio, 1);
+
if (ihid_elan->chip_data->post_gpio_reset_off_delay_ms)
msleep(ihid_elan->chip_data->post_gpio_reset_off_delay_ms);
@@ -87,12 +104,14 @@ static int i2c_hid_of_elan_probe(struct i2c_client *client)
ihid_elan->ops.power_up = elan_i2c_hid_power_up;
ihid_elan->ops.power_down = elan_i2c_hid_power_down;
- /* Start out with reset asserted */
- ihid_elan->reset_gpio =
- devm_gpiod_get_optional(&client->dev, "reset", GPIOD_OUT_HIGH);
+ ihid_elan->reset_gpio = devm_gpiod_get_optional(&client->dev, "reset",
+ GPIOD_ASIS);
if (IS_ERR(ihid_elan->reset_gpio))
return PTR_ERR(ihid_elan->reset_gpio);
+ ihid_elan->no_reset_on_power_off = of_property_read_bool(client->dev.of_node,
+ "no-reset-on-power-off");
+
ihid_elan->vccio = devm_regulator_get(&client->dev, "vccio");
if (IS_ERR(ihid_elan->vccio))
return PTR_ERR(ihid_elan->vccio);
--
2.43.2
It turned out that KMSAN instruments READ_ONCE_NOCHECK(), resulting in
false positive reports, because __no_sanitize_or_inline enforced inlining.
Properly declare __no_sanitize_or_inline under __SANITIZE_MEMORY__,
so that it does not __always_inline the annotated function.
Reported-by: syzbot+355c5bb8c1445c871ee8(a)syzkaller.appspotmail.com
Link: https://lkml.kernel.org/r/000000000000826ac1061675b0e3@google.com
Fixes: 5de0ce85f5a4 ("kmsan: mark noinstr as __no_sanitize_memory")
Cc: stable(a)vger.kernel.org
Signed-off-by: Alexander Potapenko <glider(a)google.com>
Reviewed-by: Marco Elver <elver(a)google.com>
---
include/linux/compiler_types.h | 11 +++++++++++
1 file changed, 11 insertions(+)
diff --git a/include/linux/compiler_types.h b/include/linux/compiler_types.h
index 0caf354cb94b5..a6a28952836cb 100644
--- a/include/linux/compiler_types.h
+++ b/include/linux/compiler_types.h
@@ -278,6 +278,17 @@ struct ftrace_likely_data {
# define __no_kcsan
#endif
+#ifdef __SANITIZE_MEMORY__
+/*
+ * Similarly to KASAN and KCSAN, KMSAN loses function attributes of inlined
+ * functions, therefore disabling KMSAN checks also requires disabling inlining.
+ *
+ * __no_sanitize_or_inline effectively prevents KMSAN from reporting errors
+ * within the function and marks all its outputs as initialized.
+ */
+# define __no_sanitize_or_inline __no_kmsan_checks notrace __maybe_unused
+#endif
+
#ifndef __no_sanitize_or_inline
#define __no_sanitize_or_inline __always_inline
#endif
--
2.44.0.769.g3c40516874-goog
Correctly set the length of the drm_event to the size of the structure
that's actually used.
The length of the drm_event was set to the parent structure instead of
to the drm_vmw_event_fence which is supposed to be read. drm_read
uses the length parameter to copy the event to the user space thus
resuling in oob reads.
Signed-off-by: Zack Rusin <zack.rusin(a)broadcom.com>
Fixes: 8b7de6aa8468 ("vmwgfx: Rework fence event action")
Reported-by: zdi-disclosures(a)trendmicro.com # ZDI-CAN-23566
Cc: David Airlie <airlied(a)gmail.com>
CC: Daniel Vetter <daniel(a)ffwll.ch>
Cc: Zack Rusin <zack.rusin(a)broadcom.com>
Cc: Broadcom internal kernel review list <bcm-kernel-feedback-list(a)broadcom.com>
Cc: dri-devel(a)lists.freedesktop.org
Cc: linux-kernel(a)vger.kernel.org
Cc: <stable(a)vger.kernel.org> # v3.4+
---
drivers/gpu/drm/vmwgfx/vmwgfx_fence.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
index 2a0cda324703..5efc6a766f64 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -991,7 +991,7 @@ static int vmw_event_fence_action_create(struct drm_file *file_priv,
}
event->event.base.type = DRM_VMW_EVENT_FENCE_SIGNALED;
- event->event.base.length = sizeof(*event);
+ event->event.base.length = sizeof(event->event);
event->event.user_data = user_data;
ret = drm_event_reserve_init(dev, file_priv, &event->base, &event->event.base);
--
2.40.1
Legacy DU was broken by the referenced fixes commit because the placement
and the busy_placement no longer pointed to the same object. This was later
fixed indirectly by commit a78a8da51b36c7a0c0c16233f91d60aac03a5a49
("drm/ttm: replace busy placement with flags v6") in v6.9.
Fixes: 39985eea5a6d ("drm/vmwgfx: Abstract placement selection")
Signed-off-by: Ian Forbes <ian.forbes(a)broadcom.com>
Cc: <stable(a)vger.kernel.org> # v6.4+
---
drivers/gpu/drm/vmwgfx/vmwgfx_bo.c | 1 +
1 file changed, 1 insertion(+)
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
index 2bfac3aad7b7..98e73eb0ccf1 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_bo.c
@@ -204,6 +204,7 @@ int vmw_bo_pin_in_start_of_vram(struct vmw_private *dev_priv,
VMW_BO_DOMAIN_VRAM,
VMW_BO_DOMAIN_VRAM);
buf->places[0].lpfn = PFN_UP(bo->resource->size);
+ buf->busy_places[0].lpfn = PFN_UP(bo->resource->size);
ret = ttm_bo_validate(bo, &buf->placement, &ctx);
/* For some reason we didn't end up at the start of vram */
--
2.34.1
The patch below does not apply to the 6.6-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>.
To reproduce the conflict and resubmit, you may use the following commands:
git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.6.y
git checkout FETCH_HEAD
git cherry-pick -x 35e351780fa9d8240dd6f7e4f245f9ea37e96c19
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2024042319-freeing-degree-ca0d@gregkh' --subject-prefix 'PATCH 6.6.y' HEAD^..
Possible dependencies:
35e351780fa9 ("fork: defer linking file vma until vma is fully initialized")
d24062914837 ("fork: use __mt_dup() to duplicate maple tree in dup_mmap()")
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
From 35e351780fa9d8240dd6f7e4f245f9ea37e96c19 Mon Sep 17 00:00:00 2001
From: Miaohe Lin <linmiaohe(a)huawei.com>
Date: Wed, 10 Apr 2024 17:14:41 +0800
Subject: [PATCH] fork: defer linking file vma until vma is fully initialized
Thorvald reported a WARNING [1]. And the root cause is below race:
CPU 1 CPU 2
fork hugetlbfs_fallocate
dup_mmap hugetlbfs_punch_hole
i_mmap_lock_write(mapping);
vma_interval_tree_insert_after -- Child vma is visible through i_mmap tree.
i_mmap_unlock_write(mapping);
hugetlb_dup_vma_private -- Clear vma_lock outside i_mmap_rwsem!
i_mmap_lock_write(mapping);
hugetlb_vmdelete_list
vma_interval_tree_foreach
hugetlb_vma_trylock_write -- Vma_lock is cleared.
tmp->vm_ops->open -- Alloc new vma_lock outside i_mmap_rwsem!
hugetlb_vma_unlock_write -- Vma_lock is assigned!!!
i_mmap_unlock_write(mapping);
hugetlb_dup_vma_private() and hugetlb_vm_op_open() are called outside
i_mmap_rwsem lock while vma lock can be used in the same time. Fix this
by deferring linking file vma until vma is fully initialized. Those vmas
should be initialized first before they can be used.
Link: https://lkml.kernel.org/r/20240410091441.3539905-1-linmiaohe@huawei.com
Fixes: 8d9bfb260814 ("hugetlb: add vma based lock for pmd sharing")
Signed-off-by: Miaohe Lin <linmiaohe(a)huawei.com>
Reported-by: Thorvald Natvig <thorvald(a)google.com>
Closes: https://lore.kernel.org/linux-mm/20240129161735.6gmjsswx62o4pbja@revolver/T/ [1]
Reviewed-by: Jane Chu <jane.chu(a)oracle.com>
Cc: Christian Brauner <brauner(a)kernel.org>
Cc: Heiko Carstens <hca(a)linux.ibm.com>
Cc: Kent Overstreet <kent.overstreet(a)linux.dev>
Cc: Liam R. Howlett <Liam.Howlett(a)oracle.com>
Cc: Mateusz Guzik <mjguzik(a)gmail.com>
Cc: Matthew Wilcox (Oracle) <willy(a)infradead.org>
Cc: Miaohe Lin <linmiaohe(a)huawei.com>
Cc: Muchun Song <muchun.song(a)linux.dev>
Cc: Oleg Nesterov <oleg(a)redhat.com>
Cc: Peng Zhang <zhangpeng.00(a)bytedance.com>
Cc: Tycho Andersen <tandersen(a)netflix.com>
Cc: <stable(a)vger.kernel.org>
Signed-off-by: Andrew Morton <akpm(a)linux-foundation.org>
diff --git a/kernel/fork.c b/kernel/fork.c
index 39a5046c2f0b..aebb3e6c96dc 100644
--- a/kernel/fork.c
+++ b/kernel/fork.c
@@ -714,6 +714,23 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
} else if (anon_vma_fork(tmp, mpnt))
goto fail_nomem_anon_vma_fork;
vm_flags_clear(tmp, VM_LOCKED_MASK);
+ /*
+ * Copy/update hugetlb private vma information.
+ */
+ if (is_vm_hugetlb_page(tmp))
+ hugetlb_dup_vma_private(tmp);
+
+ /*
+ * Link the vma into the MT. After using __mt_dup(), memory
+ * allocation is not necessary here, so it cannot fail.
+ */
+ vma_iter_bulk_store(&vmi, tmp);
+
+ mm->map_count++;
+
+ if (tmp->vm_ops && tmp->vm_ops->open)
+ tmp->vm_ops->open(tmp);
+
file = tmp->vm_file;
if (file) {
struct address_space *mapping = file->f_mapping;
@@ -730,25 +747,9 @@ static __latent_entropy int dup_mmap(struct mm_struct *mm,
i_mmap_unlock_write(mapping);
}
- /*
- * Copy/update hugetlb private vma information.
- */
- if (is_vm_hugetlb_page(tmp))
- hugetlb_dup_vma_private(tmp);
-
- /*
- * Link the vma into the MT. After using __mt_dup(), memory
- * allocation is not necessary here, so it cannot fail.
- */
- vma_iter_bulk_store(&vmi, tmp);
-
- mm->map_count++;
if (!(tmp->vm_flags & VM_WIPEONFORK))
retval = copy_page_range(tmp, mpnt);
- if (tmp->vm_ops && tmp->vm_ops->open)
- tmp->vm_ops->open(tmp);
-
if (retval) {
mpnt = vma_next(&vmi);
goto loop_out;
The patch below does not apply to the 6.8-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>.
To reproduce the conflict and resubmit, you may use the following commands:
git fetch https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/ linux-6.8.y
git checkout FETCH_HEAD
git cherry-pick -x ca7c52ac7ad384bcf299d89482c45fec7cd00da9
# <resolve conflicts, build, test, etc.>
git commit -s
git send-email --to '<stable(a)vger.kernel.org>' --in-reply-to '2024042358-esteemed-fastball-c2d8@gregkh' --subject-prefix 'PATCH 6.8.y' HEAD^..
Possible dependencies:
ca7c52ac7ad3 ("drm/xe/vm: prevent UAF with asid based lookup")
0eb2a18a8fad ("drm/xe: Implement VM snapshot support for BO's and userptr")
be7d51c5b468 ("drm/xe: Add batch buffer addresses to devcoredump")
4376cee62092 ("drm/xe: Print more device information in devcoredump")
98fefec8c381 ("drm/xe: Change devcoredump functions parameters to xe_sched_job")
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
From ca7c52ac7ad384bcf299d89482c45fec7cd00da9 Mon Sep 17 00:00:00 2001
From: Matthew Auld <matthew.auld(a)intel.com>
Date: Fri, 12 Apr 2024 12:31:45 +0100
Subject: [PATCH] drm/xe/vm: prevent UAF with asid based lookup
The asid is only erased from the xarray when the vm refcount reaches
zero, however this leads to potential UAF since the xe_vm_get() only
works on a vm with refcount != 0. Since the asid is allocated in the vm
create ioctl, rather erase it when closing the vm, prior to dropping the
potential last ref. This should also work when user closes driver fd
without explicit vm destroy.
Fixes: dd08ebf6c352 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Closes: https://gitlab.freedesktop.org/drm/xe/kernel/-/issues/1594
Signed-off-by: Matthew Auld <matthew.auld(a)intel.com>
Cc: Matthew Brost <matthew.brost(a)intel.com>
Cc: <stable(a)vger.kernel.org> # v6.8+
Reviewed-by: Matthew Brost <matthew.brost(a)intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240412113144.259426-4-matth…
(cherry picked from commit 83967c57320d0d01ae512f10e79213f81e4bf594)
Signed-off-by: Lucas De Marchi <lucas.demarchi(a)intel.com>
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 62d1ef8867a8..3d4c8f342e21 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -1577,6 +1577,16 @@ void xe_vm_close_and_put(struct xe_vm *vm)
xe->usm.num_vm_in_fault_mode--;
else if (!(vm->flags & XE_VM_FLAG_MIGRATION))
xe->usm.num_vm_in_non_fault_mode--;
+
+ if (vm->usm.asid) {
+ void *lookup;
+
+ xe_assert(xe, xe->info.has_asid);
+ xe_assert(xe, !(vm->flags & XE_VM_FLAG_MIGRATION));
+
+ lookup = xa_erase(&xe->usm.asid_to_vm, vm->usm.asid);
+ xe_assert(xe, lookup == vm);
+ }
mutex_unlock(&xe->usm.lock);
for_each_tile(tile, xe, id)
@@ -1592,24 +1602,15 @@ static void vm_destroy_work_func(struct work_struct *w)
struct xe_device *xe = vm->xe;
struct xe_tile *tile;
u8 id;
- void *lookup;
/* xe_vm_close_and_put was not called? */
xe_assert(xe, !vm->size);
mutex_destroy(&vm->snap_mutex);
- if (!(vm->flags & XE_VM_FLAG_MIGRATION)) {
+ if (!(vm->flags & XE_VM_FLAG_MIGRATION))
xe_device_mem_access_put(xe);
- if (xe->info.has_asid && vm->usm.asid) {
- mutex_lock(&xe->usm.lock);
- lookup = xa_erase(&xe->usm.asid_to_vm, vm->usm.asid);
- xe_assert(xe, lookup == vm);
- mutex_unlock(&xe->usm.lock);
- }
- }
-
for_each_tile(tile, xe, id)
XE_WARN_ON(vm->pt_root[id]);