Skip to contentSkip to conversation

5.15.226-rc1 review

PATCH In review Greg Kroah-Hartman 5 participants 2,219 messages Nov 18 – Nov 19 Page 12 of 16

Subsystems (7)
lore
Share & tools
lore
author maintainer
#1462 lore raw

PATCH 1459/2214 Kbuild: use -Wdeclaration-after-statement

In reply to Greg Kroah-Hartman

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

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

From: Mark Rutland <[email protected]>

[ Upstream commit 4d94f910e79a349b00a4f8aab6f3ae87129d8c5a ]

The kernel is moving from using -std=gnu89 to -std=gnu11, permitting the use of additional C11 features such as for-loop initial declarations.

One contentious aspect of C99 is that it permits mixed declarations and code, and for now at least, it seems preferable to enforce that declarations must come first.

These warnings were already enabled in the kernel itself, but not for KBUILD_USERCFLAGS or the compat VDSO on arch/arm64, which uses a separate set of CFLAGS.

This patch fixes an existing violation in modpost.c, which is not reported because of the missing flag in KBUILD_USERCFLAGS:

| scripts/mod/modpost.c: In function ‘match’:
| scripts/mod/modpost.c:837:3: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement]
|   837 |   const char *endp = p + strlen(p) - 1;
|       |   ^~~~~

Stable-dep-of: e6b49f8016ee ("dmaengine: dw-axi-dmac: Fix LLI dump out-of-bounds access")

---
 Makefile                          | 3 ++-
 arch/arm64/kernel/vdso32/Makefile | 1 +
 scripts/mod/modpost.c             | 4 +++-
 3 files changed, 6 insertions(+), 2 deletions(-)
Makefile +2 -1
index 215bf5d254d37..5a3fb7729baaf 100644
@@ -440,7 +440,8 @@ endif
 HOSTPKG_CONFIG	= pkg-config
 
 export KBUILD_USERCFLAGS := -Wall -Wmissing-prototypes -Wstrict-prototypes \
-			      -O2 -fomit-frame-pointer -std=gnu89
+			      -O2 -fomit-frame-pointer -std=gnu89 \
+			      -Wdeclaration-after-statement
 export KBUILD_USERLDFLAGS :=
 
 KBUILD_HOSTCFLAGS   := $(KBUILD_USERCFLAGS) $(HOST_LFS_CFLAGS) $(HOSTCFLAGS)
arch/arm64/kernel/vdso32/Makefile +1 -0
index 83e9399e38368..50cb1ec092ae5 100644
@@ -76,6 +76,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \
                -fno-strict-aliasing -fno-common \
                -Werror-implicit-function-declaration \
                -Wno-format-security \
+               -Wdeclaration-after-statement \
                -std=gnu89
 VDSO_CFLAGS  += -O2
 # Some useful compiler-dependent flags from top-level Makefile
scripts/mod/modpost.c +3 -1
index c846928c2a74e..6aa50e7e05d82 100644
@@ -837,8 +837,10 @@ static int match(const char *sym, const char * const pat[])
 {
 	const char *p;
 	while (*pat) {
+		const char *endp;
+
 		p = *pat++;
-		const char *endp = p + strlen(p) - 1;
+		endp = p + strlen(p) - 1;
 
 		/* "*foo*" */
 		if (*p == '*' && *endp == '*') {
Signature · 1 line
2.53.0
author maintainer
#1463 lore raw

PATCH 1460/2214 Kbuild: move to -std=gnu11

In reply to Greg Kroah-Hartman

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

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

From: Arnd Bergmann <[email protected]>

[ Upstream commit e8c07082a810fbb9db303a2b66b66b8d7e588b53 ]

During a patch discussion, Linus brought up the option of changing the C standard version from gnu89 to gnu99, which allows using variable declaration inside of a for() loop. While the C99, C11 and later standards introduce many other features, most of these are already available in gnu89 as GNU extensions as well.

An earlier attempt to do this when gcc-5 started defaulting to -std=gnu11 failed because at the time that caused warnings about designated initializers with older compilers. Now that gcc-5.1 is the minimum compiler version used for building kernels, that is no longer a concern. Similarly, the behavior of 'inline' functions changes between gnu89 using gnu_inline behavior and gnu11 using standard c99+ behavior, but this was taken care of by defining 'inline' to include __attribute__((gnu_inline)) in order to allow building with clang a while ago.

Nathan Chancellor reported a new -Wdeclaration-after-statement warning that appears in a system header on arm, this still needs a workaround.

The differences between gnu99, gnu11, gnu1x and gnu17 are fairly minimal and mainly impact warnings at the -Wpedantic level that the kernel never enables. Between these, gnu11 is the newest version that is supported by all supported compiler versions, though it is only the default on gcc-5, while all other supported versions of gcc or clang default to gnu1x/gnu17.

Stable-dep-of: e6b49f8016ee ("dmaengine: dw-axi-dmac: Fix LLI dump out-of-bounds access")

---
 Documentation/process/programming-language.rst              | 6 +++---
 .../translations/it_IT/process/programming-language.rst     | 4 ++--
 .../translations/zh_CN/process/programming-language.rst     | 3 +--
 .../translations/zh_TW/process/programming-language.rst     | 3 +--
 Makefile                                                    | 4 ++--
 arch/arm64/kernel/vdso32/Makefile                           | 2 +-
 6 files changed, 10 insertions(+), 12 deletions(-)
Documentation/process/programming-language.rst +3 -3
index ec474a70a02fa..5fc9160ca1fa5 100644
@@ -5,9 +5,9 @@ Programming Language
 
 The kernel is written in the C programming language [c-language]_.
 More precisely, the kernel is typically compiled with ``gcc`` [gcc]_
-under ``-std=gnu89`` [gcc-c-dialect-options]_: the GNU dialect of ISO C90
-(including some C99 features). ``clang`` [clang]_ is also supported, see
-docs on :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.
+under ``-std=gnu11`` [gcc-c-dialect-options]_: the GNU dialect of ISO C11.
+``clang`` [clang]_ is also supported, see docs on
+:ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.
 
 This dialect contains many extensions to the language [gnu-extensions]_,
 and many of them are used within the kernel as a matter of course.
Documentation/translations/it_IT/process/programming-language.rst +2 -2
index 41db2598ce119..c1a9b481a6f99 100644
@@ -10,8 +10,8 @@ Linguaggio di programmazione
 
 Il kernel è scritto nel linguaggio di programmazione C [it-c-language]_.
 Più precisamente, il kernel viene compilato con ``gcc`` [it-gcc]_ usando
-l'opzione ``-std=gnu89`` [it-gcc-c-dialect-options]_: il dialetto GNU
-dello standard ISO C90 (con l'aggiunta di alcune funzionalità da C99).
+l'opzione ``-std=gnu11`` [it-gcc-c-dialect-options]_: il dialetto GNU
+dello standard ISO C11.
 Linux supporta anche ``clang`` [it-clang]_, leggete la documentazione
 :ref:`Building Linux with Clang/LLVM <kbuild_llvm>`.
 
Documentation/translations/zh_CN/process/programming-language.rst +1 -2
index 2a47a1d2ec20f..fabdc338dbfbc 100644
@@ -9,8 +9,7 @@
 ============
 
 内核是用C语言 :ref:`c-language <cn_c-language>` 编写的。更准确地说,内核通常是用 :ref:`gcc <cn_gcc>`
-在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C90的 GNU 方言(
-包括一些C99特性)
+在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <cn_gcc-c-dialect-options>` 下编译的:ISO C11的 GNU 方言
 
 这种方言包含对语言 :ref:`gnu-extensions <cn_gnu-extensions>` 的许多扩展,当然,它们许多都在内核中使用。
 
Documentation/translations/zh_TW/process/programming-language.rst +1 -2
index 54e3699eadf85..144bdaf81a416 100644
@@ -12,8 +12,7 @@
 ============
 
 內核是用C語言 :ref:`c-language <tw_c-language>` 編寫的。更準確地說,內核通常是用 :ref:`gcc <tw_gcc>`
-在 ``-std=gnu89`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C90的 GNU 方言(
-包括一些C99特性)
+在 ``-std=gnu11`` :ref:`gcc-c-dialect-options <tw_gcc-c-dialect-options>` 下編譯的:ISO C11的 GNU 方言
 
 這種方言包含對語言 :ref:`gnu-extensions <tw_gnu-extensions>` 的許多擴展,當然,它們許多都在內核中使用。
 
Makefile +2 -2
index 5a3fb7729baaf..577eefed806f7 100644
@@ -524,7 +524,7 @@ KBUILD_CFLAGS   := -Wall -Wundef -Werror=strict-prototypes -Wno-trigraphs \
 		   -fno-strict-aliasing -fno-common -fshort-wchar -fno-PIE \
 		   -Werror=implicit-function-declaration -Werror=implicit-int \
 		   -Werror=return-type -Wno-format-security \
-		   -std=gnu89
+		   -std=gnu11
 KBUILD_CPPFLAGS := -D__KERNEL__
 KBUILD_AFLAGS_KERNEL :=
 KBUILD_CFLAGS_KERNEL :=
@@ -808,7 +808,7 @@ KBUILD_CFLAGS-$(CONFIG_WERROR) += -Werror
 
 ifdef CONFIG_CC_IS_CLANG
 KBUILD_CPPFLAGS += -Qunused-arguments
-# The kernel builds with '-std=gnu89' so use of GNU extensions is acceptable.
+# The kernel builds with '-std=gnu11' so use of GNU extensions is acceptable.
 KBUILD_CFLAGS += -Wno-gnu
 # CLANG uses a _MergedGlobals as optimization, but this breaks modpost, as the
 # source of a reference will be _MergedGlobals and not on of the whitelisted names.
arch/arm64/kernel/vdso32/Makefile +1 -1
index 50cb1ec092ae5..d7f5b140a5d2a 100644
@@ -77,7 +77,7 @@ VDSO_CFLAGS += -Wall -Wundef -Wstrict-prototypes -Wno-trigraphs \
                -Werror-implicit-function-declaration \
                -Wno-format-security \
                -Wdeclaration-after-statement \
-               -std=gnu89
+               -std=gnu11
 VDSO_CFLAGS  += -O2
 # Some useful compiler-dependent flags from top-level Makefile
 VDSO_CFLAGS += $(call cc32-option,-Wdeclaration-after-statement,)
Signature · 1 line
2.53.0
author maintainer
#1464 lore raw

PATCH 1461/2214 dmaengine: dw-axi-dmac: Fix LLI dump out-of-bounds access

In reply to Greg Kroah-Hartman

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

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

From: Jia Wang <[email protected]>

[ Upstream commit e6b49f8016ee8b642e8e495d1704342ee0b35d72 ]

axi_chan_list_dump_lli() uses the channel-wide descs_allocated count to walk the hw_desc[] array of a single transaction. If multiple transactions have allocated LLIs, the channel count can exceed the transaction-local nr_hw_descs and make the DMA error path read past the end of hw_desc[].

Use the descriptor-local nr_hw_descs count when dumping LLIs.

Fixes ef6fb2d6f1ab ("dmaengine: dw-axi-dmac: simplify descriptor management")
---
 drivers/dma/dw-axi-dmac/dw-axi-dmac-platform.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
drivers/dma/dw-axi-dmac/dw-axi-dmac-platform.c +1 -4
index 32faa698eab88..afae2f218bed1 100644
@@ -1001,10 +1001,7 @@ static void axi_chan_dump_lli(struct axi_dma_chan *chan,
 static void axi_chan_list_dump_lli(struct axi_dma_chan *chan,
 				   struct axi_dma_desc *desc_head)
 {
-	int count = atomic_read(&chan->descs_allocated);
-	int i;
-
-	for (i = 0; i < count; i++)
+	for (unsigned int i = 0; i < desc_head->nr_hw_descs; i++)
 		axi_chan_dump_lli(chan, &desc_head->hw_desc[i]);
 }
 
Signature · 1 line
2.53.0
author maintainer
#1465 lore raw

PATCH 1462/2214 dmaengine: dw-axi-dmac: Fix CH_CFG2 channel priority position

In reply to Greg Kroah-Hartman

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

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

From: Jia Wang <[email protected]>

[ Upstream commit 2a9013c20c46216b92ff09e2283b5244a91c59da ]

The CH_CFG2 channel priority field occupies bits 51:47 of the 64-bit channel configuration register, corresponding to bits 19:15 of CH_CFG2_H.

The driver currently shifts the priority value by 20, programming the wrong bits. Use bit 15 as the field position.

Fixes 824351668a41 ("dmaengine: dw-axi-dmac: support DMAX_NUM_CHANNELS > 8")
---
 drivers/dma/dw-axi-dmac/dw-axi-dmac.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/dma/dw-axi-dmac/dw-axi-dmac.h +1 -1
index a51e35df996ec..a84945d2a508c 100644
@@ -321,7 +321,7 @@ enum {
 #define CH_CFG2_H_TT_FC_POS		0
 #define CH_CFG2_H_HS_SEL_SRC_POS	3
 #define CH_CFG2_H_HS_SEL_DST_POS	4
-#define CH_CFG2_H_PRIORITY_POS		20
+#define CH_CFG2_H_PRIORITY_POS		15
 
 /**
  * DW AXI DMA channel interrupts
Signature · 1 line
2.53.0
author maintainer
#1466 lore raw

PATCH 1463/2214 PCI: endpoint: pci-epf-vntb: Serialize virtual PCI bus scan

In reply to Greg Kroah-Hartman

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

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

From: Koichiro Den <[email protected]>

[ Upstream commit 84c3520c9e8c14b09a5cf645cc4fa7c6d9d61127 ]

The virtual PCI bus is scanned without holding the PCI rescan/remove lock. This can race with other PCI topology changes.

Hold the lock across the bus scan and device addition.

---
 drivers/pci/endpoint/functions/pci-epf-vntb.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/pci/endpoint/functions/pci-epf-vntb.c +5 -0
index 30c003d8b0dd6..390bb9c9e8836 100644
@@ -977,14 +977,19 @@ static int vpci_scan_bus(void *sysdata)
 	struct pci_bus *vpci_bus;
 	struct epf_ntb *ndev = sysdata;
 
+	pci_lock_rescan_remove();
+
 	vpci_bus = pci_scan_bus(ndev->vbus_number, &vpci_ops, sysdata);
 	if (!vpci_bus) {
 		pr_err("create pci bus failed\n");
+		pci_unlock_rescan_remove();
 		return -EINVAL;
 	}
 
 	pci_bus_add_devices(vpci_bus);
 
+	pci_unlock_rescan_remove();
+
 	return 0;
 }
 
Signature · 1 line
2.53.0
author maintainer
#1467 lore raw

PATCH 1464/2214 regulator: ab8500: Fix AB8505 VANA voltage selectors

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit dfdfda2a95843706428a9d95ab546cb10983e684 ]

The AB8505 VANA voltage table assumes selector 0 represents 1.05 V and all eight selectors increase linearly. Selector 0 actually represents 1.2 V. Selectors 1 through 6 cover 1.05 V through 1.175 V, and selector 7 represents 1.225 V.

Correct the table so each selector reports and programs the documented voltage.

Reference: AB8505 User Manual, DM00046744 Rev 3, p. 257

Fixes 8a3b1b8703fe ("regulator: ab8500: Add voltage selection for AUDIO and ANA on AB8505")
Assisted-by LLM
---
 drivers/regulator/ab8500.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/regulator/ab8500.c +1 -1
index 23a401734a987..d603d7b9f6821 100644
@@ -265,13 +265,13 @@ static const unsigned int fixed_2050000_voltage[] = {
 };
 
 static const unsigned int ldo_vana_voltages[] = {
+	1200000,
 	1050000,
 	1075000,
 	1100000,
 	1125000,
 	1150000,
 	1175000,
-	1200000,
 	1225000,
 };
 
Signature · 1 line
2.53.0
author maintainer
#1468 lore raw

PATCH 1465/2214 regulator: ab8500: Remove stale expand_register kernel-doc entry

In reply to Greg Kroah-Hartman

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

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

From: Babanpreet Singh <[email protected]>

[ Upstream commit 0991f3b4624ed713d1d5eb56520c1c9b2deaab7a ]

Commit aeee55b76bfd ("regulator: ab8500: Remove unused embedded struct expand_register") deleted the expand_register member from struct ab8500_regulator_info and, in the same hunk, added an empty "@expand_register:" line to the kernel-doc block. That traded one W=1 warning for another:

drivers/regulator/ab8500.c:196 Excess struct member 'expand_register' description in 'ab8500_regulator_info'

Drop the leftover line; the remaining @member entries all match the struct.

No functional changes.

Stable-dep-of: ac72c091a981 ("regulator: ab8500: Add AB8505 VAUX3 3.05 V setting")

---
 drivers/regulator/ab8500.c | 1 -
 1 file changed, 1 deletion(-)
drivers/regulator/ab8500.c +0 -1
index d603d7b9f6821..066ce04cbca4b 100644
@@ -173,7 +173,6 @@ struct ab8500_shared_mode {
  * @voltage_bank: bank to control regulator voltage
  * @voltage_reg: register to control regulator voltage
  * @voltage_mask: mask to control regulator voltage
- * @expand_register: 
  */
 struct ab8500_regulator_info {
 	struct device		*dev;
Signature · 1 line
2.53.0
author maintainer
#1469 lore raw

PATCH 1466/2214 regulator: ab8500: Add AB8505 VAUX3 3.05 V setting

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit ac72c091a9810493ccd9209414d4f859276727ab ]

AB8505 has an additional VAUX3 voltage setting which is not encoded in the normal three-bit selector. ArmRegu2.Vaux3Sel3 overrides that selector and selects 3.05 V.

Add the missing voltage and use the override bit as an extended selector. Program the ordinary selector before clearing the override so VAUX3 does not briefly switch to a stale voltage.

Reference: AB8505 User Manual, DM00046744 Rev 3, pp. 229 and 254

---
 drivers/regulator/ab8500.c | 101 +++++++++++++++++++++++++++++++++++--
 1 file changed, 98 insertions(+), 3 deletions(-)
drivers/regulator/ab8500.c +98 -3
index 066ce04cbca4b..51a4ccb710b1a 100644
@@ -173,6 +173,7 @@ struct ab8500_shared_mode {
  * @voltage_bank: bank to control regulator voltage
  * @voltage_reg: register to control regulator voltage
  * @voltage_mask: mask to control regulator voltage
+ * @expand_register: additional register used to select an extra voltage
  */
 struct ab8500_regulator_info {
 	struct device		*dev;
@@ -193,6 +194,12 @@ struct ab8500_regulator_info {
 	u8 voltage_bank;
 	u8 voltage_reg;
 	u8 voltage_mask;
+	struct {
+		u8 voltage_limit;
+		u8 voltage_bank;
+		u8 voltage_reg;
+		u8 voltage_mask;
+	} expand_register;
 };
 
 /* voltage tables for the vauxn/vintcore supplies */
@@ -226,6 +233,18 @@ static const unsigned int ldo_vaux3_voltages[] = {
 	2910000,
 };
 
+static const unsigned int ldo_vaux3_ab8505_voltages[] = {
+	1200000,
+	1500000,
+	1800000,
+	2100000,
+	2500000,
+	2750000,
+	2790000,
+	2910000,
+	3050000,
+};
+
 static const unsigned int ldo_vaux56_voltages[] = {
 	1800000,
 	1050000,
@@ -593,6 +612,64 @@ static int ab8500_regulator_set_voltage_sel(struct regulator_dev *rdev,
 	return ret;
 }
 
+static int ab8500_regulator_get_voltage_sel_expand(struct regulator_dev *rdev)
+{
+	struct ab8500_regulator_info *info = rdev_get_drvdata(rdev);
+	u8 regval;
+	int ret;
+
+	if (!info)
+		return -EINVAL;
+
+	ret = abx500_get_register_interruptible(info->dev,
+			info->expand_register.voltage_bank,
+			info->expand_register.voltage_reg, &regval);
+	if (ret < 0) {
+		dev_err(rdev_get_dev(rdev),
+			"couldn't read voltage expand reg for regulator\n");
+		return ret;
+	}
+
+	if (regval & info->expand_register.voltage_mask)
+		return info->expand_register.voltage_limit;
+
+	return ab8500_regulator_get_voltage_sel(rdev);
+}
+
+static int ab8500_regulator_set_voltage_sel_expand(struct regulator_dev *rdev,
+						   unsigned int selector)
+{
+	struct ab8500_regulator_info *info = rdev_get_drvdata(rdev);
+	u8 regval;
+	int ret;
+
+	if (!info)
+		return -EINVAL;
+
+	if (selector > info->expand_register.voltage_limit)
+		return -EINVAL;
+
+	if (selector < info->expand_register.voltage_limit) {
+		ret = ab8500_regulator_set_voltage_sel(rdev, selector);
+		if (ret < 0)
+			return ret;
+
+		regval = 0;
+	} else {
+		regval = info->expand_register.voltage_mask;
+	}
+
+	ret = abx500_mask_and_set_register_interruptible(info->dev,
+			info->expand_register.voltage_bank,
+			info->expand_register.voltage_reg,
+			info->expand_register.voltage_mask, regval);
+	if (ret < 0)
+		dev_err(rdev_get_dev(rdev),
+			"couldn't set voltage expand reg for regulator\n");
+
+	return ret;
+}
+
 static const struct regulator_ops ab8500_regulator_volt_mode_ops = {
 	.enable			= ab8500_regulator_enable,
 	.disable		= ab8500_regulator_disable,
@@ -605,6 +682,18 @@ static const struct regulator_ops ab8500_regulator_volt_mode_ops = {
 	.list_voltage		= regulator_list_voltage_table,
 };
 
+static const struct regulator_ops ab8500_regulator_volt_mode_expand_ops = {
+	.enable			= ab8500_regulator_enable,
+	.disable		= ab8500_regulator_disable,
+	.is_enabled		= ab8500_regulator_is_enabled,
+	.get_optimum_mode	= ab8500_regulator_get_optimum_mode,
+	.set_mode		= ab8500_regulator_set_mode,
+	.get_mode		= ab8500_regulator_get_mode,
+	.get_voltage_sel	= ab8500_regulator_get_voltage_sel_expand,
+	.set_voltage_sel	= ab8500_regulator_set_voltage_sel_expand,
+	.list_voltage		= regulator_list_voltage_table,
+};
+
 static const struct regulator_ops ab8500_regulator_volt_ops = {
 	.enable		= ab8500_regulator_enable,
 	.disable	= ab8500_regulator_disable,
@@ -908,12 +997,12 @@ static struct ab8500_regulator_info
 	[AB8505_LDO_AUX3] = {
 		.desc = {
 			.name		= "LDO-AUX3",
-			.ops		= &ab8500_regulator_volt_mode_ops,
+			.ops		= &ab8500_regulator_volt_mode_expand_ops,
 			.type		= REGULATOR_VOLTAGE,
 			.id		= AB8505_LDO_AUX3,
 			.owner		= THIS_MODULE,
-			.n_voltages	= ARRAY_SIZE(ldo_vaux3_voltages),
-			.volt_table	= ldo_vaux3_voltages,
+			.n_voltages	= ARRAY_SIZE(ldo_vaux3_ab8505_voltages),
+			.volt_table	= ldo_vaux3_ab8505_voltages,
 		},
 		.load_lp_uA		= 5000,
 		.update_bank		= 0x04,
@@ -925,6 +1014,12 @@ static struct ab8500_regulator_info
 		.voltage_bank		= 0x04,
 		.voltage_reg		= 0x21,
 		.voltage_mask		= 0x07,
+		.expand_register = {
+			.voltage_limit	= 8,
+			.voltage_bank	= 0x04,
+			.voltage_reg	= 0x01,
+			.voltage_mask	= 0x10,
+		},
 	},
 	[AB8505_LDO_AUX4] = {
 		.desc = {
Signature · 1 line
2.53.0
author maintainer
#1470 lore raw

PATCH 1467/2214 regulator: ab8500: Handle AB8505 VINTCORE selector 7

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit 4ec4df7e54182edb6955caace37684e535e462dc ]

The AB8505 VINTCORE table exposes only selectors 0 through 6. The hardware also accepts selector 7 and maps it to 1.35 V, just like selector 6. Omitting it can make an OTP-programmed selector 7 appear invalid to the regulator core.

Give AB8505 its own eight-entry selector table while leaving the AB8500 table unchanged.

Reference: AB8505 User Manual, DM00046744 Rev 3, p. 223

---
 drivers/regulator/ab8500.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)
drivers/regulator/ab8500.c +13 -2
index 51a4ccb710b1a..662434cccd217 100644
@@ -266,6 +266,17 @@ static const unsigned int ldo_vintcore_voltages[] = {
 	1350000,
 };
 
+static const unsigned int ldo_vintcore_ab8505_voltages[] = {
+	1200000,
+	1225000,
+	1250000,
+	1275000,
+	1300000,
+	1325000,
+	1350000,
+	1350000,
+};
+
 static const unsigned int fixed_1200000_voltage[] = {
 	1200000,
 };
@@ -1095,8 +1106,8 @@ static struct ab8500_regulator_info
 			.type		= REGULATOR_VOLTAGE,
 			.id		= AB8505_LDO_INTCORE,
 			.owner		= THIS_MODULE,
-			.n_voltages	= ARRAY_SIZE(ldo_vintcore_voltages),
-			.volt_table	= ldo_vintcore_voltages,
+			.n_voltages	= ARRAY_SIZE(ldo_vintcore_ab8505_voltages),
+			.volt_table	= ldo_vintcore_ab8505_voltages,
 		},
 		.load_lp_uA		= 5000,
 		.update_bank		= 0x03,
Signature · 1 line
2.53.0
author maintainer
#1471 lore raw

PATCH 1468/2214 regulator: ab8500: Treat cut 1.0 VAUX3 as fixed

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit b9fc8a8d31b914b09329240f532ff0e9c00cd727 ]

The early-cut workaround currently gives AB8500 cut 1.0 the 16 programmable VAUX3 settings introduced with cut 1.1. On cut 1.0 the selector is not programmable and VAUX3 is fixed at 1.2 V.

Register VAUX3 as a fixed-voltage regulator on cut 1.0 and retain the 16-value workaround only for cut 1.1.

Reference: AB8500 User Manual, UM0836 Rev 3, p. 240

---
 drivers/regulator/ab8500.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)
drivers/regulator/ab8500.c +8 -3
index 662434cccd217..110c660e899d7 100644
@@ -1783,9 +1783,14 @@ static int ab8500_regulator_register(struct platform_device *pdev,
 	config.driver_data = info;
 	config.of_node = np;
 
-	/* fix for hardware before ab8500v2.0 */
-	if (is_ab8500_1p1_or_earlier(ab8500)) {
-		if (info->desc.id == AB8500_LDO_AUX3) {
+	/* Handle the different VAUX3 implementations in early AB8500 cuts. */
+	if (info->desc.id == AB8500_LDO_AUX3) {
+		if (is_ab8500_1p0_or_earlier(ab8500)) {
+			info->desc.ops = &ab8500_regulator_mode_ops;
+			info->desc.n_voltages = 1;
+			info->desc.volt_table = fixed_1200000_voltage;
+			info->voltage_mask = 0;
+		} else if (is_ab8500_1p1_or_earlier(ab8500)) {
 			info->desc.n_voltages =
 				ARRAY_SIZE(ldo_vauxn_voltages);
 			info->desc.volt_table = ldo_vauxn_voltages;
Signature · 1 line
2.53.0
author maintainer
#1472 lore raw

PATCH 1469/2214 regulator: ab8500: Test dedicated enable bits only

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit a5ee78846015512b4d3d32f8aa463024a23234ce ]

Some regulator control registers have independent enable and low-power bits. is_enabled() currently tests their combined update mask, so an off regulator with its low-power bit set is incorrectly reported as enabled.

Add an optional enable mask and use it for VINTCORE, AB8500 TVOUT, AB8505 ADC, and AB8505 VAUX5/6. Regulators whose two-bit field encodes the complete operating state continue to test the full update mask.

References: AB8500 User Manual, UM0836 Rev 3, p. 214; AB8505 User Manual, DM00046744 Rev 3, pp. 171-172 and 223

Fixes 65e03ed2d0cd ("regulators: Fixed errors in ab8500 register mapping")
Assisted-by LLM
---
 drivers/regulator/ab8500.c | 15 +++++++++++++--
 1 file changed, 13 insertions(+), 2 deletions(-)
drivers/regulator/ab8500.c +13 -2
index 110c660e899d7..c00bd8bf1e058 100644
@@ -162,6 +162,7 @@ struct ab8500_shared_mode {
  * @update_bank: bank to control on/off
  * @update_reg: register to control on/off
  * @update_mask: mask to enable/disable and set mode of regulator
+ * @enable_mask: optional mask for an enable bit separate from the mode bit
  * @update_val: bits holding the regulator current mode
  * @update_val_idle: bits to enable the regulator in idle (low power) mode
  * @update_val_normal: bits to enable the regulator in normal (high power) mode
@@ -183,6 +184,7 @@ struct ab8500_regulator_info {
 	u8 update_bank;
 	u8 update_reg;
 	u8 update_mask;
+	u8 enable_mask;
 	u8 update_val;
 	u8 update_val_idle;
 	u8 update_val_normal;
@@ -377,6 +379,7 @@ static int ab8500_regulator_is_enabled(struct regulator_dev *rdev)
 {
 	int ret;
 	struct ab8500_regulator_info *info = rdev_get_drvdata(rdev);
+	u8 enable_mask;
 	u8 regval;
 
 	if (info == NULL) {
@@ -392,13 +395,15 @@ static int ab8500_regulator_is_enabled(struct regulator_dev *rdev)
 		return ret;
 	}
 
+	enable_mask = info->enable_mask ? info->enable_mask : info->update_mask;
+
 	dev_vdbg(rdev_get_dev(rdev),
 		"%s-is_enabled (bank, reg, mask, value): 0x%x, 0x%x, 0x%x,"
 		" 0x%x\n",
 		info->desc.name, info->update_bank, info->update_reg,
-		info->update_mask, regval);
+		enable_mask, regval);
 
-	if (regval & info->update_mask)
+	if (regval & enable_mask)
 		return 1;
 	else
 		return 0;
@@ -833,6 +838,7 @@ static struct ab8500_regulator_info
 		.update_bank		= 0x03,
 		.update_reg		= 0x80,
 		.update_mask		= 0x44,
+		.enable_mask		= 0x04,
 		.update_val		= 0x44,
 		.update_val_idle	= 0x44,
 		.update_val_normal	= 0x04,
@@ -861,6 +867,7 @@ static struct ab8500_regulator_info
 		.update_bank		= 0x03,
 		.update_reg		= 0x80,
 		.update_mask		= 0x82,
+		.enable_mask		= 0x02,
 		.update_val		= 0x02,
 		.update_val_idle	= 0x82,
 		.update_val_normal	= 0x02,
@@ -1070,6 +1077,7 @@ static struct ab8500_regulator_info
 		.update_bank		= 0x01,
 		.update_reg		= 0x55,
 		.update_mask		= 0x18,
+		.enable_mask		= 0x10,
 		.update_val		= 0x10,
 		.update_val_idle	= 0x18,
 		.update_val_normal	= 0x10,
@@ -1092,6 +1100,7 @@ static struct ab8500_regulator_info
 		.update_bank		= 0x01,
 		.update_reg		= 0x56,
 		.update_mask		= 0x18,
+		.enable_mask		= 0x10,
 		.update_val		= 0x10,
 		.update_val_idle	= 0x18,
 		.update_val_normal	= 0x10,
@@ -1113,6 +1122,7 @@ static struct ab8500_regulator_info
 		.update_bank		= 0x03,
 		.update_reg		= 0x80,
 		.update_mask		= 0x44,
+		.enable_mask		= 0x04,
 		.update_val		= 0x04,
 		.update_val_idle	= 0x44,
 		.update_val_normal	= 0x04,
@@ -1141,6 +1151,7 @@ static struct ab8500_regulator_info
 		.update_bank		= 0x03,
 		.update_reg		= 0x80,
 		.update_mask		= 0x82,
+		.enable_mask		= 0x02,
 		.update_val		= 0x02,
 		.update_val_idle	= 0x82,
 		.update_val_normal	= 0x02,
Signature · 1 line
2.53.0
author maintainer
#1473 lore raw

PATCH 1470/2214 regulator: ab8500: Propagate mode enable read errors

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit 71ac8878573058e2b85210760ccaeaac90796c35 ]

For regulators whose enable and mode share a state field, set_mode() first reads that field so changing the requested mode does not enable a disabled rail. A register read error is currently treated as true and the driver proceeds to write the new mode, potentially enabling a rail whose state is unknown.

Return the read error without changing the register or cached mode.

References: AB8500 User Manual, UM0836 Rev 3, p. 227; AB8505 User Manual, DM00046744 Rev 3, p. 237

Fixes 438e695b87e0 ("regulator: ab8500: Get rid of is_enabled from struct ab8500_regulator_info")
Assisted-by LLM
---
 drivers/regulator/ab8500.c | 14 ++++++++++++--
 1 file changed, 12 insertions(+), 2 deletions(-)
drivers/regulator/ab8500.c +12 -2
index c00bd8bf1e058..3a7576e3f0706 100644
@@ -433,7 +433,7 @@ static unsigned int ab8500_regulator_get_optimum_mode(
 static int ab8500_regulator_set_mode(struct regulator_dev *rdev,
 				     unsigned int mode)
 {
-	int ret = 0;
+	int enabled, ret = 0;
 	u8 bank, reg, mask, val;
 	bool lp_mode_req = false;
 	struct ab8500_regulator_info *info = rdev_get_drvdata(rdev);
@@ -490,7 +490,17 @@ static int ab8500_regulator_set_mode(struct regulator_dev *rdev,
 		goto out_unlock;
 	}
 
-	if (info->mode_mask || ab8500_regulator_is_enabled(rdev)) {
+	if (info->mode_mask) {
+		enabled = 1;
+	} else {
+		enabled = ab8500_regulator_is_enabled(rdev);
+		if (enabled < 0) {
+			ret = enabled;
+			goto out_unlock;
+		}
+	}
+
+	if (enabled) {
 		ret = abx500_mask_and_set_register_interruptible(info->dev,
 			bank, reg, mask, val);
 		if (ret < 0) {
Signature · 1 line
2.53.0
author maintainer
#1474 lore raw

PATCH 1471/2214 sit: fix UAF in ipip6_tunnel_del_prl()

In reply to Greg Kroah-Hartman

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit 5145bb6a2c2683d541828f86d5321e0565f2e30c ]

When flushing all potential router list (PRL) entries, ipip6_tunnel_del_prl() called call_rcu(&x->rcu_head, prl_list_destroy_rcu) before clearing t->prl.

A concurrent reader in isatap_chksrc() could enter an RCU read-side critical section after call_rcu() but before t->prl is set to NULL, allowing prl_list_destroy_rcu() to free nodes while the reader traverses them.

Clear t->prl using RCU_INIT_POINTER() before invoking call_rcu(). Also use rcu_assign_pointer() when unlinking individual entries.

Fixes ef9a9d1183b3 ("ipv6 sit: RCU conversion phase I")
---
 net/ipv6/sit.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/ipv6/sit.c +2 -2
index d700789fb9d8e..52001a858c1f8 100644
@@ -432,7 +432,7 @@ ipip6_tunnel_del_prl(struct ip_tunnel *t, struct ip_tunnel_prl *a)
 		     (x = rtnl_dereference(*p)) != NULL;
 		     p = &x->next) {
 			if (x->addr == a->addr) {
-				*p = x->next;
+				rcu_assign_pointer(*p, rtnl_dereference(x->next));
 				kfree_rcu(x, rcu_head);
 				t->prl_count--;
 				goto out;
@@ -443,8 +443,8 @@ ipip6_tunnel_del_prl(struct ip_tunnel *t, struct ip_tunnel_prl *a)
 		x = rtnl_dereference(t->prl);
 		if (x) {
 			t->prl_count = 0;
+			RCU_INIT_POINTER(t->prl, NULL);
 			call_rcu(&x->rcu_head, prl_list_destroy_rcu);
-			t->prl = NULL;
 		}
 	}
 out:
Signature · 1 line
2.53.0
author maintainer
#1475 lore raw

PATCH 1472/2214 ip_tunnel: use WRITE_ONCE in ip_tunnel_encap_setup

In reply to Greg Kroah-Hartman

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

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

From: Eric Dumazet <[email protected]>

[ Upstream commit 88b84cae6b94087d3a9679845fe28ad19c7bef8b ]

Update ip_tunnel_encap_setup() to use WRITE_ONCE() when writing to encap fields (type, sport, dport, flags) and hlen fields. This ensures that concurrent lockless readers (like fill_info) do not see torn writes.

Also remove the unsafe memset() on t->encap which could cause concurrent readers to transiently see zeroed fields. Removing it also fixes a bug where t->encap was left cleared even if ip_encap_hlen() failed, resulting in partial configuration.

Fixes 56328486539d ("net: Changes to ip_tunnel to support foo-over-udp encapsulation")
---
 net/ipv4/ip_tunnel.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)
net/ipv4/ip_tunnel.c +6 -8
index cbc3f8bdf0cb4..62fdfdbec7845 100644
@@ -476,19 +476,17 @@ int ip_tunnel_encap_setup(struct ip_tunnel *t,
 {
 	int hlen;
 
-	memset(&t->encap, 0, sizeof(t->encap));
-
 	hlen = ip_encap_hlen(ipencap);
 	if (hlen < 0)
 		return hlen;
 
-	t->encap.type = ipencap->type;
-	t->encap.sport = ipencap->sport;
-	t->encap.dport = ipencap->dport;
-	t->encap.flags = ipencap->flags;
+	WRITE_ONCE(t->encap.type, ipencap->type);
+	WRITE_ONCE(t->encap.sport, ipencap->sport);
+	WRITE_ONCE(t->encap.dport, ipencap->dport);
+	WRITE_ONCE(t->encap.flags, ipencap->flags);
 
-	t->encap_hlen = hlen;
-	t->hlen = t->encap_hlen + t->tun_hlen;
+	WRITE_ONCE(t->encap_hlen, hlen);
+	WRITE_ONCE(t->hlen, hlen + t->tun_hlen);
 
 	return 0;
 }
Signature · 1 line
2.53.0
author maintainer
#1476 lore raw

PATCH 1473/2214 crypto: amlogic - Fix dma_map_sg error check

In reply to Greg Kroah-Hartman

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

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

From: Jack Wang <[email protected]>

[ Upstream commit 45fa321e7de604146603e24ee5bf3c0b766efe46 ]

dma_map_sg return 0 on error.

Stable-dep-of: 6c2373dd818b ("crypto: amlogic - Fix DMA memory leak in cipher error path")

---
 drivers/crypto/amlogic/amlogic-gxl-cipher.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/crypto/amlogic/amlogic-gxl-cipher.c +3 -3
index e79514fce731f..af017a087ebf0 100644
@@ -177,7 +177,7 @@ static int meson_cipher(struct skcipher_request *areq)
 	if (areq->src == areq->dst) {
 		nr_sgs = dma_map_sg(mc->dev, areq->src, sg_nents(areq->src),
 				    DMA_BIDIRECTIONAL);
-		if (nr_sgs < 0) {
+		if (!nr_sgs) {
 			dev_err(mc->dev, "Invalid SG count %d\n", nr_sgs);
 			err = -EINVAL;
 			goto theend;
@@ -186,14 +186,14 @@ static int meson_cipher(struct skcipher_request *areq)
 	} else {
 		nr_sgs = dma_map_sg(mc->dev, areq->src, sg_nents(areq->src),
 				    DMA_TO_DEVICE);
-		if (nr_sgs < 0 || nr_sgs > MAXDESC - 3) {
+		if (!nr_sgs || nr_sgs > MAXDESC - 3) {
 			dev_err(mc->dev, "Invalid SG count %d\n", nr_sgs);
 			err = -EINVAL;
 			goto theend;
 		}
 		nr_sgd = dma_map_sg(mc->dev, areq->dst, sg_nents(areq->dst),
 				    DMA_FROM_DEVICE);
-		if (nr_sgd < 0 || nr_sgd > MAXDESC - 3) {
+		if (!nr_sgd || nr_sgd > MAXDESC - 3) {
 			dev_err(mc->dev, "Invalid SG count %d\n", nr_sgd);
 			err = -EINVAL;
 			goto theend;
Signature · 1 line
2.53.0
author maintainer
#1477 lore raw

PATCH 1474/2214 crypto: amlogic - Fix DMA memory leak in cipher error path

In reply to Greg Kroah-Hartman

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

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

From: Mohamad Raizudeen <[email protected]>

[ Upstream commit 6c2373dd818b777d0a81eb02acebc03940552dcd ]

A DMA memory leak occurs in meson_cipher() on the mapping error paths. The driver jumps to the end of the function without unmapping the previously mapped source scatterlist and key/IV buffer when the destination apping fails.

Additionally, a memory leak occurs when a scatterlist mapping succeeds but the returned count exceeds the driver's MAXDESC limit. In this case, the driver rejects the mapping without unmapping it. The BIDIRECTIONAL mapping branch also lacks the required 'MAXDESC -3' upper bound check.

Fix this by introducing proper error labels, error_src and error_keyiv and unmap resources immediately inside the calidation checks to ensure all successfully mapped resources are cleaned up before returning the error.

Fixes 48fe583fe541 ("crypto: amlogic - Add crypto accelerator for amlogic GXL")
---
 drivers/crypto/amlogic/amlogic-gxl-cipher.c | 25 ++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)
drivers/crypto/amlogic/amlogic-gxl-cipher.c +20 -5
index af017a087ebf0..c9eb7c5586322 100644
@@ -177,10 +177,13 @@ static int meson_cipher(struct skcipher_request *areq)
 	if (areq->src == areq->dst) {
 		nr_sgs = dma_map_sg(mc->dev, areq->src, sg_nents(areq->src),
 				    DMA_BIDIRECTIONAL);
-		if (!nr_sgs) {
-			dev_err(mc->dev, "Invalid SG count %d\n", nr_sgs);
+		if (!nr_sgs || nr_sgs > MAXDESC - 3) {
+			dev_err(mc->dev, "Invalid BIDIR SG count %d\n", nr_sgs);
 			err = -EINVAL;
-			goto theend;
+
+			if (nr_sgs)
+				dma_unmap_sg(mc->dev, areq->src, sg_nents(areq->src), DMA_BIDIRECTIONAL);
+			goto error_keyiv;
 		}
 		nr_sgd = nr_sgs;
 	} else {
@@ -189,14 +192,20 @@ static int meson_cipher(struct skcipher_request *areq)
 		if (!nr_sgs || nr_sgs > MAXDESC - 3) {
 			dev_err(mc->dev, "Invalid SG count %d\n", nr_sgs);
 			err = -EINVAL;
-			goto theend;
+
+			if (nr_sgs)
+				dma_unmap_sg(mc->dev, areq->src, sg_nents(areq->src), DMA_TO_DEVICE);
+			goto error_keyiv;
 		}
 		nr_sgd = dma_map_sg(mc->dev, areq->dst, sg_nents(areq->dst),
 				    DMA_FROM_DEVICE);
 		if (!nr_sgd || nr_sgd > MAXDESC - 3) {
 			dev_err(mc->dev, "Invalid SG count %d\n", nr_sgd);
 			err = -EINVAL;
-			goto theend;
+
+			if (nr_sgd)
+				dma_unmap_sg(mc->dev, areq->dst, sg_nents(areq->dst), DMA_FROM_DEVICE);
+			goto error_src;
 		}
 	}
 
@@ -251,6 +260,12 @@ static int meson_cipher(struct skcipher_request *areq)
 						 ivsize, 0);
 		}
 	}
+	goto theend;
+
+error_src:
+	dma_unmap_sg(mc->dev, areq->src, sg_nents(areq->src), DMA_TO_DEVICE);
+error_keyiv:
+	dma_unmap_single(mc->dev, phykeyiv, keyivlen, DMA_TO_DEVICE);
 theend:
 	kfree_sensitive(bkeyiv);
 	kfree_sensitive(backup_iv);
Signature · 1 line
2.53.0
author maintainer
#1478 lore raw

PATCH 1475/2214 crypto: qat - avoid OOB read when stripping RSA leading zeros

In reply to Greg Kroah-Hartman

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

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

From: Ahsan Atta <[email protected]>

[ Upstream commit ac508238898058f0b09b18a7891fb31279617781 ]

The logical AND operator uses short-circuit evaluation: the first operand is always evaluated, while the second is evaluated only if the first is true. Because the first operand dereferences a pointer, it may cause an out-of-bounds access before the bounds check in the second operand is evaluated.

Resolve this by checking the vlen prior to the pointer being dereferenced.

Fixes a990532023b9 ("crypto: qat - Add support for RSA algorithm")
---
 drivers/crypto/qat/qat_common/qat_asym_algs.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/crypto/qat/qat_common/qat_asym_algs.c +4 -4
index 4437c3a5c6dd7..fc0613f37b7e1 100644
@@ -934,7 +934,7 @@ static int qat_rsa_set_n(struct qat_rsa_ctx *ctx, const char *value,
 	const char *ptr = value;
 	int ret;
 
-	while (!*ptr && vlen) {
+	while (vlen && !*ptr) {
 		ptr++;
 		vlen--;
 	}
@@ -965,7 +965,7 @@ static int qat_rsa_set_e(struct qat_rsa_ctx *ctx, const char *value,
 	struct device *dev = &GET_DEV(inst->accel_dev);
 	const char *ptr = value;
 
-	while (!*ptr && vlen) {
+	while (vlen && !*ptr) {
 		ptr++;
 		vlen--;
 	}
@@ -991,7 +991,7 @@ static int qat_rsa_set_d(struct qat_rsa_ctx *ctx, const char *value,
 	const char *ptr = value;
 	int ret;
 
-	while (!*ptr && vlen) {
+	while (vlen && !*ptr) {
 		ptr++;
 		vlen--;
 	}
@@ -1014,7 +1014,7 @@ static int qat_rsa_set_d(struct qat_rsa_ctx *ctx, const char *value,
 
 static void qat_rsa_drop_leading_zeros(const char **ptr, unsigned int *len)
 {
-	while (!**ptr && *len) {
+	while (*len && !**ptr) {
 		(*ptr)++;
 		(*len)--;
 	}
Signature · 1 line
2.53.0
author maintainer
#1479 lore raw

PATCH 1476/2214 memory: exynos5422-dmc: handle clock enable failures

In reply to Greg Kroah-Hartman

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

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

From: Jiawen Liu <[email protected]>

[ Upstream commit cce46902ea9fcd585cd200fffa3166f63bc571f1 ]

Check the return values of clk_prepare_enable() for fout_bpll and mout_bpll.

If enabling fout_bpll fails, propagate the error. If enabling mout_bpll fails, disable fout_bpll before returning. This keeps the clock enable count balanced on probe error paths.

---
 drivers/memory/samsung/exynos5422-dmc.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)
drivers/memory/samsung/exynos5422-dmc.c +9 -2
index c491cd549644f..85c05c9c75332 100644
@@ -1304,8 +1304,15 @@ static int exynos5_dmc_init_clks(struct exynos5_dmc *dmc)
 	if (ret)
 		return ret;
 
-	clk_prepare_enable(dmc->fout_bpll);
-	clk_prepare_enable(dmc->mout_bpll);
+	ret = clk_prepare_enable(dmc->fout_bpll);
+	if (ret)
+		return ret;
+
+	ret = clk_prepare_enable(dmc->mout_bpll);
+	if (ret) {
+		clk_disable_unprepare(dmc->fout_bpll);
+		return ret;
+	}
 
 	/*
 	 * Some bootloaders do not set clock routes correctly.
Signature · 1 line
2.53.0
author maintainer
#1480 lore raw

PATCH 1477/2214 gfs2: Fix journaled truncate range end offset

In reply to Greg Kroah-Hartman

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

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

From: Jan Kara <[email protected]>

[ Upstream commit a52652efd8f985397e87669f505af83f06e335cf ]

The third argument to truncate_pagecache_range() is the inclusive end offset, but gfs2_journaled_truncate_range() passes the length of each chunk. A hole punch at an offset larger than its length therefore creates an inverted range.

Commit 86066914edff ("gfs2: Don't support fallocate on jdata files") disables fallocate for jdata files. Otherwise, the following bug could occur:

The MM truncation path then underflows the unsigned partial-folio length and passes an oversized range to folio_zero_range(), triggering a BUG. Pass the last byte of the chunk so page-cache invalidation covers the intended range.

Fixes 4e56a6411fbc ("gfs2: Implement fallocate(FALLOC_FL_PUNCH_HOLE)")
Assisted-by LLM
---
 fs/gfs2/bmap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/gfs2/bmap.c +1 -1
index 1e95d777f6737..9a68e9b7c2ce5 100644
@@ -2387,7 +2387,7 @@ static int gfs2_journaled_truncate_range(struct inode *inode, loff_t offset,
 		if (offs && chunk > PAGE_SIZE)
 			chunk = offs + ((chunk - offs) & PAGE_MASK);
 
-		truncate_pagecache_range(inode, offset, chunk);
+		truncate_pagecache_range(inode, offset, offset + chunk - 1);
 		offset += chunk;
 		length -= chunk;
 
Signature · 1 line
2.53.0
author maintainer
#1481 lore raw

PATCH 1478/2214 ACPI: glue: Fix up and adjust acpi_unbind_one()

In reply to Greg Kroah-Hartman

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

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

From: Rafael J. Wysocki <[email protected]>

[ Upstream commit 693546852656075e0247ac637b05625ead134eff ]

Since none of the acpi_unbind_one() callers check its return value and it always returns 0 anyway, make it void.

Also notice that unlocking physical_node_lock for the given ACPI device should be carried out before dropping the reference to it in case that reference is the last one (highly unlikely), so rearrange the code to make that happen.

Fixes 3e3327837c18 ("ACPI: Use list_for_each_entry() in acpi_unbind_one()")
---
 drivers/acpi/glue.c     | 10 ++++++----
 include/acpi/acpi_bus.h |  2 +-
 2 files changed, 7 insertions(+), 5 deletions(-)
drivers/acpi/glue.c +6 -4
index 7a33a6d985f89..f696f0da49318 100644
@@ -252,13 +252,13 @@ int acpi_bind_one(struct device *dev, struct acpi_device *acpi_dev)
 }
 EXPORT_SYMBOL_GPL(acpi_bind_one);
 
-int acpi_unbind_one(struct device *dev)
+void acpi_unbind_one(struct device *dev)
 {
 	struct acpi_device *acpi_dev = ACPI_COMPANION(dev);
 	struct acpi_device_physical_node *entry;
 
 	if (!acpi_dev)
-		return 0;
+		return;
 
 	mutex_lock(&acpi_dev->physical_node_lock);
 
@@ -273,15 +273,17 @@ int acpi_unbind_one(struct device *dev)
 			sysfs_remove_link(&acpi_dev->dev.kobj, physnode_name);
 			sysfs_remove_link(&dev->kobj, "firmware_node");
 			ACPI_COMPANION_SET(dev, NULL);
+
+			mutex_unlock(&acpi_dev->physical_node_lock);
+
 			/* Drop references taken by acpi_bind_one(). */
 			put_device(dev);
 			acpi_dev_put(acpi_dev);
 			kfree(entry);
-			break;
+			return;
 		}
 
 	mutex_unlock(&acpi_dev->physical_node_lock);
-	return 0;
 }
 EXPORT_SYMBOL_GPL(acpi_unbind_one);
 
include/acpi/acpi_bus.h +1 -1
index 7a85ae6b7b005..7ce4e01b2406e 100644
@@ -576,7 +576,7 @@ struct acpi_bus_type {
 int register_acpi_bus_type(struct acpi_bus_type *);
 int unregister_acpi_bus_type(struct acpi_bus_type *);
 int acpi_bind_one(struct device *dev, struct acpi_device *adev);
-int acpi_unbind_one(struct device *dev);
+void acpi_unbind_one(struct device *dev);
 
 struct acpi_pci_root {
 	struct acpi_device * device;
Signature · 1 line
2.53.0
author maintainer
#1482 lore raw

PATCH 1479/2214 firewire: cdev: Fix FW_CDEV_IOC_CREATE_ISO_CONTEXT references in kernel-doc

In reply to Greg Kroah-Hartman

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

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

From: Karl Mehltretter <[email protected]>

[ Upstream commit 0a6cf2b7adcbd362039e9d282f73aff3f24022ca ]

The fw_cdev_event_iso_interrupt and fw_cdev_event_iso_interrupt_mc comments refer to FW_CDEV_CREATE_ISO_CONTEXT, but the ioctl is named FW_CDEV_IOC_CREATE_ISO_CONTEXT.

Correct both references. The misspellings were introduced separately with the userspace interface documentation and multichannel reception support.

Fixes 7ada60e36764 ("firewire: Document userspace ioctl interface.")
Fixes 872e330e3880 ("firewire: add isochronous multichannel reception")
Assisted-by LLM
---
 include/uapi/linux/firewire-cdev.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
include/uapi/linux/firewire-cdev.h +2 -2
index 5effa9832802c..75fe7db52b6b4 100644
@@ -211,7 +211,7 @@ struct fw_cdev_event_request2 {
 /**
  * struct fw_cdev_event_iso_interrupt - Sent when an iso packet was completed
  * @closure:	See &fw_cdev_event_common;
- *		set by %FW_CDEV_CREATE_ISO_CONTEXT ioctl
+ *		set by %FW_CDEV_IOC_CREATE_ISO_CONTEXT ioctl
  * @type:	See &fw_cdev_event_common; always %FW_CDEV_EVENT_ISO_INTERRUPT
  * @cycle:	Cycle counter of the last completed packet
  * @header_length: Total length of following headers, in bytes
@@ -271,7 +271,7 @@ struct fw_cdev_event_iso_interrupt {
 /**
  * struct fw_cdev_event_iso_interrupt_mc - An iso buffer chunk was completed
  * @closure:	See &fw_cdev_event_common;
- *		set by %FW_CDEV_CREATE_ISO_CONTEXT ioctl
+ *		set by %FW_CDEV_IOC_CREATE_ISO_CONTEXT ioctl
  * @type:	%FW_CDEV_EVENT_ISO_INTERRUPT_MULTICHANNEL
  * @completed:	Offset into the receive buffer; data before this offset is valid
  *
Signature · 1 line
2.53.0
author maintainer
#1483 lore raw

PATCH 1480/2214 firewire: core: remove unused link field in fw_packet structure

In reply to Greg Kroah-Hartman

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

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

From: Takashi Sakamoto <[email protected]>

[ Upstream commit 00d4d31e4fdfe51312f0184f8d162b8d4de172dc ]

The link field in fw_packet structure has never been used since it was added.

Fixes 3038e353cfaf ("firewire: Add core firewire stack.")
---
 include/linux/firewire.h | 1 -
 1 file changed, 1 deletion(-)
include/linux/firewire.h +0 -1
index aec8f30ab200d..6395605e91fd0 100644
@@ -309,7 +309,6 @@ struct fw_packet {
 	 */
 	fw_packet_callback_t callback;
 	int ack;
-	struct list_head link;
 	void *driver_data;
 };
 
Signature · 1 line
2.53.0
author maintainer
#1484 lore raw

PATCH 1481/2214 cpuidle: menu: Remove incorrect unlikely() annotation

In reply to Greg Kroah-Hartman

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

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

From: Breno Leitao <[email protected]>

[ Upstream commit fcbd7897b871e157ee5c595e950c8466d86c0cd5 ]

The unlikely() annotation on the early-return condition in menu_select() is incorrect on systems with only one idle state (e.g., ARM64 servers with a single ACPI LPI state). Branch profiling shows 100% misprediction on such systems since drv->state_count <= 1 is always true.

On platforms where only state0 is available, this path is the common case, not an unlikely edge case. Remove the misleading annotation to let the branch predictor learn the actual behavior.

Stable-dep-of: d594187cafd5 ("cpuidle: menu: Do not return a disabled idle state")

---
 drivers/cpuidle/governors/menu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/cpuidle/governors/menu.c +1 -1
index 956b42729c376..3188c55612ce7 100644
@@ -314,7 +314,7 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 	nr_iowaiters = nr_iowait_cpu(dev->cpu);
 	data->bucket = which_bucket(data->next_timer_ns, nr_iowaiters);
 
-	if (unlikely(drv->state_count <= 1 || latency_req == 0) ||
+	if (drv->state_count <= 1 || latency_req == 0 ||
 	    ((data->next_timer_ns < drv->states[1].target_residency_ns ||
 	      latency_req < drv->states[1].exit_latency_ns) &&
 	     !dev->states_usage[0].disable)) {
Signature · 1 line
2.53.0
author maintainer
#1485 lore raw

PATCH 1482/2214 cpuidle: menu: Remove single state handling

In reply to Greg Kroah-Hartman

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

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

From: Christian Loehle <[email protected]>

[ Upstream commit 93983a9f3beea791c21d77c2425488ad327d4fda ]

cpuidle systems where the governor has no choice because there's only a single idle state are now handled by cpuidle core and bypass the governor, so remove the related handling.

[ rjw: Rebase on top of the cpuidle changes merged recently ]

Stable-dep-of: d594187cafd5 ("cpuidle: menu: Do not return a disabled idle state")

---
 drivers/cpuidle/governors/menu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/cpuidle/governors/menu.c +1 -1
index 3188c55612ce7..c764e9f8d1bc2 100644
@@ -314,7 +314,7 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 	nr_iowaiters = nr_iowait_cpu(dev->cpu);
 	data->bucket = which_bucket(data->next_timer_ns, nr_iowaiters);
 
-	if (drv->state_count <= 1 || latency_req == 0 ||
+	if (latency_req == 0 ||
 	    ((data->next_timer_ns < drv->states[1].target_residency_ns ||
 	      latency_req < drv->states[1].exit_latency_ns) &&
 	     !dev->states_usage[0].disable)) {
Signature · 1 line
2.53.0
author maintainer
#1486 lore raw

PATCH 1483/2214 cpuidle: menu: Do not return a disabled idle state

In reply to Greg Kroah-Hartman

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

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

From: Xueqin Luo <[email protected]>

[ Upstream commit d594187cafd5ad65afc5f00242f8fa20e1d8c0f1 ]

When idle state 0 is disabled and the PM QoS latency constraint is zero (latency_req == 0), the first branch of the || in the early-return condition of menu_select() is taken immediately, so state 0 is returned regardless of its disable flag.

Address this by making the !disable check a prerequisite for the entire early-return condition, so that when state 0 is disabled the normal state selection loop is used instead, which already skips disabled states.

Fixes 53812cdc9100 ("cpuidle: menu: Move the latency_req == 0 special case check")
---
 drivers/cpuidle/governors/menu.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
drivers/cpuidle/governors/menu.c +4 -4
index c764e9f8d1bc2..e20b27e4f1398 100644
@@ -314,10 +314,10 @@ static int menu_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 	nr_iowaiters = nr_iowait_cpu(dev->cpu);
 	data->bucket = which_bucket(data->next_timer_ns, nr_iowaiters);
 
-	if (latency_req == 0 ||
-	    ((data->next_timer_ns < drv->states[1].target_residency_ns ||
-	      latency_req < drv->states[1].exit_latency_ns) &&
-	     !dev->states_usage[0].disable)) {
+	if (!dev->states_usage[0].disable &&
+	    (latency_req == 0 ||
+	     data->next_timer_ns < drv->states[1].target_residency_ns ||
+	     latency_req < drv->states[1].exit_latency_ns)) {
 		/*
 		 * In this case state[0] will be used no matter what, so return
 		 * it right away and keep the tick running if state[0] is a
Signature · 1 line
2.53.0
author maintainer
#1487 lore raw

PATCH 1484/2214 ARM: dts: stm32: lxa-mc1: change stdout-path baud rate from 9600 to 115200

In reply to Greg Kroah-Hartman

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

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

From: Ahmad Fatoum <[email protected]>

[ Upstream commit fc3c0ffefc169719316eddb8c36e2427e11ce713 ]

The default baud rate when none is specified is up to the DT consumer.

In the case of the Linux STM32 serial driver, it defaults to 9600 baud, which differs from the 115200 baud that this board's barebox bootloader configured.

This went unnoticed, because barebox automatically fixes up a console= command-line option that looks like this on the LXA boards:

console=ttySTM0,115200n8

This had precedence over the 9600 fallback baud rate.

But when EFI booting a kernel via GRUB, we run into this issue, because the barebox-provided command-line is disregarded by GRUB.

Fix this by explicitly setting the baud rate to the correct 115200.

---
 arch/arm/boot/dts/stm32mp157c-lxa-mc1.dts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/stm32mp157c-lxa-mc1.dts +1 -1
index 1e9bf7eea0f1f..2a53bc455f455 100644
@@ -33,7 +33,7 @@ backlight: backlight {
 	};
 
 	chosen {
-		stdout-path = &uart4;
+		stdout-path = "serial0:115200n8";
 	};
 
 	led-controller-0 {
Signature · 1 line
2.53.0
author maintainer
#1488 lore raw

PATCH 1485/2214 cpuidle: teo: Do not return a disabled idle state

In reply to Greg Kroah-Hartman

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

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

From: Xueqin Luo <[email protected]>

[ Upstream commit 60d5fca628f94f8f23398ad62d80aa284dae6571 ]

If idle state 0 is disabled and none of the enabled idle states satisfies the PM QoS latency constraint, constraint_idx in teo_select() stays at its initial value of 0, because the loop that updates it starts at state 1, and so the candidate state index is capped at state 0, which is disabled, and that state is returned to the caller.

The cpuidle core does not validate the returned state index against the per-state disable flags, so the CPU ends up entering the disabled state 0.

To address this, ensure that constraint_idx is at least equal to the index of the first enabled idle state (idx0), so that the candidate state is never capped to a disabled one.

Fixes c410a9a142f1 ("cpuidle: teo: Change the main idle state selection logic")
---
 drivers/cpuidle/governors/teo.c | 11 +++++++++++
 1 file changed, 11 insertions(+)
drivers/cpuidle/governors/teo.c +11 -0
index d9262db79cae5..8de499ca89ea7 100644
@@ -445,6 +445,17 @@ static int teo_select(struct cpuidle_driver *drv, struct cpuidle_device *dev,
 		}
 	}
 
+	/*
+	 * If the latency constraint does not allow any of the enabled idle
+	 * states to be used, the candidate state index will be capped to 0
+	 * by the check below, but state 0 may be disabled.  To prevent the
+	 * selection of a disabled state in that case, ensure that
+	 * constraint_idx is at least equal to the index of the first
+	 * enabled idle state.
+	 */
+	if (constraint_idx < idx0)
+		constraint_idx = idx0;
+
 	/*
 	 * If there is a latency constraint, it may be necessary to select an
 	 * idle state shallower than the current candidate one.
Signature · 1 line
2.53.0
author maintainer
#1489 lore raw

PATCH 1486/2214 ARM: dts: stm32: Drop bias-disable from STM32MP15xx DHCOM SoM sdmmc1_b4_pins_a

In reply to Greg Kroah-Hartman

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

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

From: Marek Vasut <[email protected]>

[ Upstream commit 3572825773791f4258b553391f4e2187df321731 ]

Drop bias-disable property from STM32MP15xx DHCOM SoM sdmmc1_b4_pins_a to fix the following DT checker warning. The bias-pull-up property is the only bias-related property that has to be present in those pinctrl node subnodes.

"
arch/arm/boot/dts/st/stm32mp153c-dhcom-drc02.dtb: pinctrl@50002000 (st,stm32mp157-pinctrl): sdmmc1-b4-0:pins1: {'pinmux': [10253, 10509, 10765, 11021, 12813], 'slew-rate': 1, 'drive-push-pull': True, 'bias-disable': True, 'bias-pull-up': True} is valid under each of {'required': ['bias-pull-up']}, {'required': ['bias-disable']} "

Fixes 7cd8567d988a ("ARM: dts: stm32: Enable internal pull-ups for SDMMC1 on DHCOM SoM")
---
 arch/arm/boot/dts/stm32mp15xx-dhcom-som.dtsi | 2 ++
 1 file changed, 2 insertions(+)
arch/arm/boot/dts/stm32mp15xx-dhcom-som.dtsi +2 -0
index 8c41f819f7769..39503921b7b6b 100644
@@ -481,9 +481,11 @@ &sdmmc1_b4_pins_a {
 	 * - mandatory on SoMs without SD voltage translator
 	 */
 	pins1 {
+		/delete-property/ bias-disable;
 		bias-pull-up;
 	};
 	pins2 {
+		/delete-property/ bias-disable;
 		bias-pull-up;
 	};
 };
Signature · 1 line
2.53.0
author maintainer
#1490 lore raw

PATCH 1487/2214 ARM: dts: imx7-tqma7: add SPI-NOR flash

In reply to Greg Kroah-Hartman

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

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

From: Matthias Schiffer <[email protected]>

[ Upstream commit 61b2f7b15839f5d6b4838a44b6faaff9957e98e7 ]

The SPI-NOR flash on the SoM was missing from the device tree.

The TQMa7 as a designated QSPI_RESET# pin, however depending on the hardware configuration the pin may be unconnected, or be used for a different purpose. With this in mind, we mux the pin as a pullup and define an input hog for it, but keep it a separate pin group, so that it is easy for dependent Device Trees to modify the configuration.

Stable-dep-of: f322f8f68c2c ("ARM: dts: imx7-tqma7: correctly specify PMIC voltage constraints")

---
 arch/arm/boot/dts/imx7-tqma7.dtsi | 43 +++++++++++++++++++++++++++++++
 1 file changed, 43 insertions(+)
arch/arm/boot/dts/imx7-tqma7.dtsi +43 -0
index 8773344b54aad..22d0eaa4cef79 100644
@@ -19,6 +19,16 @@ &cpu0 {
 	cpu-supply = <&sw1a_reg>;
 };
 
+&gpio2 {
+	/* Configured as pullup by QSPI pin group */
+	qspi-reset-hog {
+		gpio-hog;
+		gpios = <4 GPIO_ACTIVE_LOW>;
+		input;
+		line-name = "qspi-reset";
+	};
+};
+
 &i2c1 {
 	pinctrl-names = "default";
 	pinctrl-0 = <&pinctrl_i2c1>;
@@ -160,6 +170,25 @@ MX7D_PAD_SD2_RESET_B__GPIO5_IO11	0x4000005C
 		>;
 	};
 
+	pinctrl_qspi: qspigrp {
+		fsl,pins = <
+			MX7D_PAD_EPDC_DATA00__QSPI_A_DATA0	0x5A
+			MX7D_PAD_EPDC_DATA01__QSPI_A_DATA1	0x5A
+			MX7D_PAD_EPDC_DATA02__QSPI_A_DATA2	0x5A
+			MX7D_PAD_EPDC_DATA03__QSPI_A_DATA3	0x5A
+			MX7D_PAD_EPDC_DATA05__QSPI_A_SCLK	0x11
+			MX7D_PAD_EPDC_DATA06__QSPI_A_SS0_B	0x54
+			MX7D_PAD_EPDC_DATA07__QSPI_A_SS1_B	0x54
+		>;
+	};
+
+	pinctrl_qspi_reset: qspi_resetgrp {
+		fsl,pins = <
+			/* #QSPI_RESET */
+			MX7D_PAD_EPDC_DATA04__GPIO2_IO4		0x52
+		>;
+	};
+
 	pinctrl_usdhc3: usdhc3grp {
 		fsl,pins = <
 			MX7D_PAD_SD3_CMD__SD3_CMD		0x59
@@ -217,6 +246,20 @@ MX7D_PAD_LPSR_GPIO1_IO00__WDOG1_WDOG_B	0x30
 	};
 };
 
+&qspi {
+	pinctrl-names = "default";
+	pinctrl-0 = <&pinctrl_qspi &pinctrl_qspi_reset>;
+	status = "okay";
+
+	flash0: flash@0 {
+		compatible = "jedec,spi-nor";
+		reg = <0>;
+		spi-max-frequency = <29000000>;
+		spi-rx-bus-width = <4>;
+		spi-tx-bus-width = <4>;
+	};
+};
+
 &sdma {
 	status = "okay";
 };
Signature · 1 line
2.53.0
author maintainer
#1491 lore raw

PATCH 1488/2214 ARM: dts: imx7-tqma7: correctly specify PMIC voltage constraints

In reply to Greg Kroah-Hartman

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

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

From: Nora Schiffer <[email protected]>

[ Upstream commit f322f8f68c2c3291633ead663effa3b25c18f7b1 ]

regulator-(min|max)-microvolt refer to the supported voltages according to board design, not the possible range of the regulators themselves. Update the configuration for the TQMa7x's PMIC and also add a comment regarding the PMIC's VCC_SD switching feature.

Fixes a80a1af6ec8a ("ARM: dts: tq imx7 common board support")
---
 arch/arm/boot/dts/imx7-tqma7.dtsi | 40 ++++++++++++++++++-------------
 1 file changed, 23 insertions(+), 17 deletions(-)
arch/arm/boot/dts/imx7-tqma7.dtsi +23 -17
index 22d0eaa4cef79..9fe4819fd34c9 100644
@@ -43,8 +43,8 @@ pfuze3000: pmic@8 {
 
 		regulators {
 			sw1a_reg: sw1a {
-				regulator-min-microvolt = <700000>;
-				regulator-max-microvolt = <3300000>;
+				regulator-min-microvolt = <1000000>;
+				regulator-max-microvolt = <1225000>;
 				regulator-boot-on;
 				regulator-always-on;
 				regulator-ramp-delay = <6250>;
@@ -52,34 +52,33 @@ sw1a_reg: sw1a {
 
 			/* use sw1c_reg to align with pfuze100/pfuze200 */
 			sw1c_reg: sw1b {
-				regulator-min-microvolt = <700000>;
-				regulator-max-microvolt = <1475000>;
+				regulator-min-microvolt = <1000000>;
+				regulator-max-microvolt = <1000000>;
 				regulator-boot-on;
 				regulator-always-on;
 				regulator-ramp-delay = <6250>;
 			};
 
 			sw2_reg: sw2 {
-				regulator-min-microvolt = <1500000>;
-				regulator-max-microvolt = <1850000>;
+				regulator-min-microvolt = <1800000>;
+				regulator-max-microvolt = <1800000>;
 				regulator-boot-on;
 				regulator-always-on;
 			};
 
 			sw3a_reg: sw3 {
-				regulator-min-microvolt = <900000>;
-				regulator-max-microvolt = <1650000>;
+				regulator-min-microvolt = <1350000>;
+				regulator-max-microvolt = <1350000>;
 				regulator-boot-on;
 				regulator-always-on;
 			};
 
 			swbst_reg: swbst {
-				regulator-min-microvolt = <5000000>;
-				regulator-max-microvolt = <5150000>;
+				/* Not connected and left unconfigured */
 			};
 
 			snvs_reg: vsnvs {
-				regulator-min-microvolt = <1000000>;
+				regulator-min-microvolt = <3000000>;
 				regulator-max-microvolt = <3000000>;
 				regulator-boot-on;
 				regulator-always-on;
@@ -92,18 +91,25 @@ vref_reg: vrefddr {
 
 			vgen1_reg: vldo1 {
 				regulator-min-microvolt = <1800000>;
-				regulator-max-microvolt = <3300000>;
+				regulator-max-microvolt = <1800000>;
 				regulator-always-on;
 			};
 
 			vgen2_reg: vldo2 {
-				regulator-min-microvolt = <800000>;
-				regulator-max-microvolt = <1550000>;
+				regulator-min-microvolt = <1500000>;
+				regulator-max-microvolt = <1500000>;
 				regulator-always-on;
 			};
 
 			vgen3_reg: vccsd {
-				regulator-min-microvolt = <2850000>;
+				/*
+				 * Specified values refer to 3.3V operation; voltage for 1.8V
+				 * operation is derived from this.
+				 *
+				 * The base board should usually mux GPIO1_IO08 to SD1_VSELECT for
+				 * automatic switching of SD card IO voltage.
+				 */
+				regulator-min-microvolt = <3300000>;
 				regulator-max-microvolt = <3300000>;
 				regulator-always-on;
 			};
@@ -115,13 +121,13 @@ vgen4_reg: v33 {
 			};
 
 			vgen5_reg: vldo3 {
-				regulator-min-microvolt = <1800000>;
+				regulator-min-microvolt = <3300000>;
 				regulator-max-microvolt = <3300000>;
 				regulator-always-on;
 			};
 
 			vgen6_reg: vldo4 {
-				regulator-min-microvolt = <1800000>;
+				regulator-min-microvolt = <3300000>;
 				regulator-max-microvolt = <3300000>;
 				regulator-always-on;
 			};
Signature · 1 line
2.53.0
author maintainer
#1492 lore raw

PATCH 1489/2214 x86/svm: Avoid sign extension of SVM_EVTINJ_VALID in 64-bit expressions

In reply to Greg Kroah-Hartman

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

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

From: Chaithanya Lagisetty <[email protected]>

[ Upstream commit c5991239b148d40d4981901968b3a6a838ff517b ]

SVM_EVTINJ_VALID is defined as (1 << 31), i.e. as an int with the sign bit set. Whenever it is used in a 64-bit expression the int is sign extended, so bits 63:32 end up set as well. Two 64-bit users are affected.

verify_exception_info() uses the macro as a mask against a 64-bit value when it sanity checks hypervisor supplied exception information:

	u64 info = ghcb->save.sw_exit_info_2;
	...
	if ((info & SVM_EVTINJ_VALID) &&

The mask is 0xffffffff80000000, so the check also passes when any bit of the error code in 63:32 is set while the valid bit itself is clear.

svm_vmgexit_inject_exception() builds the value that KVM hands to an SEV-ES guest in GHCB SW_EXITINFO2:

	u64 data = SVM_EVTINJ_VALID | SVM_EVTINJ_TYPE_EXEPT | vector;

For vector == X86_TRAP_GP this yields 0xffffffff8000030d instead of 0x000000008000030d. The field uses the EVENTINJ format, which holds the error code in bits 63:32, so KVM asks the guest to inject #GP with an error code of 0xffffffff. Linux guests only consume the error code when SVM_EVTINJ_VALID_ERR is set, which it is not here, so they are not affected in practice, but the value KVM writes is not the one it intends.

Use BIT() instead, as is already done for most single bit definitions in this header. Every other user applies the macro to a u32, where the resulting bits are unchanged.

Opportunistically give SVM_EVTINJ_VALID_ERR the same treatment, there's zero reason to open code BIT().

Fixes 597cfe48212a ("x86/boot/compressed/64: Setup a GHCB-based VC Exception handler")
Fixes c3392d0ab714 ("KVM: SVM: Provide helpers to set the error code")
[sean: use BIT() for SVM_EVTINJ_VALID_ERR as well]
---
 arch/x86/include/asm/svm.h | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/x86/include/asm/svm.h +2 -2
index b00dbc5fac2b2..8c0dbd5a8e727 100644
@@ -372,8 +372,8 @@ struct vmcb {
 #define SVM_EVTINJ_TYPE_EXEPT (3 << SVM_EVTINJ_TYPE_SHIFT)
 #define SVM_EVTINJ_TYPE_SOFT (4 << SVM_EVTINJ_TYPE_SHIFT)
 
-#define SVM_EVTINJ_VALID (1 << 31)
-#define SVM_EVTINJ_VALID_ERR (1 << 11)
+#define SVM_EVTINJ_VALID	BIT(31)
+#define SVM_EVTINJ_VALID_ERR	BIT(11)
 
 #define SVM_EXITINTINFO_VEC_MASK SVM_EVTINJ_VEC_MASK
 #define SVM_EXITINTINFO_TYPE_MASK SVM_EVTINJ_TYPE_MASK
Signature · 1 line
2.53.0
author maintainer
#1493 lore raw

PATCH 1490/2214 spi: Fix the return value in spi_new_ancillary_device() kernel-doc

In reply to Greg Kroah-Hartman

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

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

From: Karl Mehltretter <[email protected]>

[ Upstream commit 3742867ce96539100ed5f4c754f04273d8aa00d7 ]

spi_new_ancillary_device() returns the new struct spi_device, or an ERR_PTR() on failure, but its kernel-doc says "0 on success; negative errno on failure", which was never true. Describe the pointer.

Fixes 0c79378c0199 ("spi: add ancillary device support")
Assisted-by LLM
---
 drivers/spi/spi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/spi/spi.c +1 -1
index 962598f2b109f..490dfaad46e1e 100644
@@ -2230,7 +2230,7 @@ static void of_register_spi_devices(struct spi_controller *ctlr) { }
  *
  * This may only be called from main SPI device's probe routine.
  *
- * Return: 0 on success; negative errno on failure
+ * Return: the new device on success; an ERR_PTR() on failure
  */
 struct spi_device *spi_new_ancillary_device(struct spi_device *spi,
 					     u8 chip_select)
Signature · 1 line
2.53.0
author maintainer
#1494 lore raw

PATCH 1491/2214 ata: libata-scsi: Fix ata_scsi_hotplug() reference in kernel-doc

In reply to Greg Kroah-Hartman

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

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

From: Karl Mehltretter <[email protected]>

[ Upstream commit 5915ab523e33057224a0240e27b88c6c2798b919 ]

The ata_scsi_remove_dev() comment misspells its caller as ata_eh_scsi_hotplug(). The caller is ata_scsi_hotplug(). Correct the reference, which predates libata's move from drivers/scsi to drivers/ata.

Fixes 580b2102327a ("[PATCH] libata-hp: implement SCSI part of hotplug")
Assisted-by LLM
---
 drivers/ata/libata-scsi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/ata/libata-scsi.c +1 -1
index 81ed39f68c148..231e1d2faa2fe 100644
@@ -4413,7 +4413,7 @@ int ata_scsi_offline_dev(struct ata_device *dev)
  *	ata_scsi_remove_dev - remove attached SCSI device
  *	@dev: ATA device to remove attached SCSI device for
  *
- *	This function is called from ata_eh_scsi_hotplug() and
+ *	This function is called from ata_scsi_hotplug() and
  *	responsible for removing the SCSI device attached to @dev.
  *
  *	LOCKING:
Signature · 1 line
2.53.0
author maintainer
#1495 lore raw

PATCH 1492/2214 gfs2: Fix the return value in gfs2_get_link() kernel-doc

In reply to Greg Kroah-Hartman

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

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

From: Karl Mehltretter <[email protected]>

[ Upstream commit 700a0a6892f7c0c541d5d23837cfab98cc744d17 ]

gfs2_get_link() returns the link target or an error pointer, but its kernel-doc still describes an integer status. The mismatch dates back to when gfs2_readlink() was replaced by generic_readlink() and persisted across the removal of nd_set_link().

Describe the returned pointer and error value.

Fixes c177c2ac8c5a ("Switch gfs2 to nd_set_link()")
Assisted-by LLM
---
 fs/gfs2/inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/gfs2/inode.c +1 -1
index 061372f84a64c..883592c7287be 100644
@@ -1783,7 +1783,7 @@ static int gfs2_rename2(struct user_namespace *mnt_userns, struct inode *odir,
  *
  * This can handle symlinks of any size.
  *
- * Returns: 0 on success or error code
+ * Returns: the link target on success, an ERR_PTR() on failure
  */
 
 static const char *gfs2_get_link(struct dentry *dentry,
Signature · 1 line
2.53.0
author maintainer
#1496 lore raw

PATCH 1493/2214 ALSA: control: Fix virtual master creation reference in kernel-doc

In reply to Greg Kroah-Hartman

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

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

From: Karl Mehltretter <[email protected]>

[ Upstream commit a347cc458d97ecd8adb5462b91559682a70aced6 ]

The snd_ctl_add_follower() comment refers to snd_ctl_create_virtual_master(), but virtual master controls are created with snd_ctl_make_virtual_master(). Correct the helper name.

Fixes 79c7cdd5441f ("ALSA: Add kernel-doc comments to vmaster stuff")
Assisted-by LLM
---
 include/sound/control.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/sound/control.h +1 -1
index a1fc7e0a47d95..b35ee47f4a6ba 100644
@@ -218,7 +218,7 @@ int _snd_ctl_add_follower(struct snd_kcontrol *master,
  * @follower: follower element to add
  *
  * Add a virtual follower control to the given master element created via
- * snd_ctl_create_virtual_master() beforehand.
+ * snd_ctl_make_virtual_master() beforehand.
  *
  * All followers must be the same type (returning the same information
  * via info callback).  The function doesn't check it, so it's your
Signature · 1 line
2.53.0
author maintainer
#1497 lore raw

PATCH 1494/2214 ACPI: video: Fix backlight unregistration ordering

In reply to Greg Kroah-Hartman

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

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

From: Rafael J. Wysocki <[email protected]>

[ Upstream commit 46577d36611953d1b99e86493289cfc824116ad2 ]

In acpi_video_dev_unregister_backlight(), the sysfs interface of the cooling class device may access the brightness object under the backlight device's ACPI companion, so that object cannot be freed before unregistering the cooling class device.

Adjust the code to take that into account.

Fixes 67b662e189f4 ("ACPI / video: seperate backlight control and event interface")
[ rjw: Fixed the Fixes: tag ]
---
 drivers/acpi/acpi_video.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)
drivers/acpi/acpi_video.c +6 -6
index d9c68fc74d611..473de787e8e82 100644
@@ -1916,6 +1916,12 @@ static int acpi_video_bus_register_backlight(struct acpi_video_bus *video)
 
 static void acpi_video_dev_unregister_backlight(struct acpi_video_device *device)
 {
+	if (device->cooling_dev) {
+		sysfs_remove_link(&device->dev->dev.kobj, "thermal_cooling");
+		sysfs_remove_link(&device->cooling_dev->device.kobj, "device");
+		thermal_cooling_device_unregister(device->cooling_dev);
+		device->cooling_dev = NULL;
+	}
 	if (device->backlight) {
 		backlight_device_unregister(device->backlight);
 		device->backlight = NULL;
@@ -1925,12 +1931,6 @@ static void acpi_video_dev_unregister_backlight(struct acpi_video_device *device
 		kfree(device->brightness);
 		device->brightness = NULL;
 	}
-	if (device->cooling_dev) {
-		sysfs_remove_link(&device->dev->dev.kobj, "thermal_cooling");
-		sysfs_remove_link(&device->cooling_dev->device.kobj, "device");
-		thermal_cooling_device_unregister(device->cooling_dev);
-		device->cooling_dev = NULL;
-	}
 }
 
 static int acpi_video_bus_unregister_backlight(struct acpi_video_bus *video)
Signature · 1 line
2.53.0
author maintainer
#1498 lore raw

PATCH 1495/2214 perf header: Fix potential memory corruption in cpu_cache_level__read

In reply to Greg Kroah-Hartman

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

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

From: Ian Rogers <[email protected]>

[ Upstream commit d964ffba01f5d0935dd6bc850029958a62c74491 ]

strim() may advance the pointer assigned to cache->size which causes later frees to crash.

Fix by performing the strim and then memmove-ing the potentially shifted string back over the original string.

The bug was introduced by the transition from rtrim() to strim(), as rtrim() wouldn't move on the left.

Fixes 13c230ab6e56c6ae ("perf tools: Ditch rtrim(), use strim() from tools/lib")
Assisted-by Antigravity:gemini-3.1-pro
---
 tools/perf/util/header.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)
tools/perf/util/header.c +15 -3
index 3e18acb719324..2875552767a9f 100644
@@ -1126,7 +1126,11 @@ static int cpu_cache_level__read(struct cpu_cache_level *cache, u32 cpu, u16 lev
 		return -1;
 
 	cache->type[len] = 0;
-	cache->type = strim(cache->type);
+	{
+		char *trimmed = strim(cache->type);
+
+		memmove(cache->type, trimmed, strlen(trimmed) + 1);
+	}
 
 	scnprintf(file, PATH_MAX, "%s/size", path);
 	if (sysfs__read_str(file, &cache->size, &len)) {
@@ -1135,7 +1139,11 @@ static int cpu_cache_level__read(struct cpu_cache_level *cache, u32 cpu, u16 lev
 	}
 
 	cache->size[len] = 0;
-	cache->size = strim(cache->size);
+	{
+		char *trimmed = strim(cache->size);
+
+		memmove(cache->size, trimmed, strlen(trimmed) + 1);
+	}
 
 	scnprintf(file, PATH_MAX, "%s/shared_cpu_list", path);
 	if (sysfs__read_str(file, &cache->map, &len)) {
@@ -1145,7 +1153,11 @@ static int cpu_cache_level__read(struct cpu_cache_level *cache, u32 cpu, u16 lev
 	}
 
 	cache->map[len] = 0;
-	cache->map = strim(cache->map);
+	{
+		char *trimmed = strim(cache->map);
+
+		memmove(cache->map, trimmed, strlen(trimmed) + 1);
+	}
 	return 0;
 }
 
Signature · 1 line
2.53.0
author maintainer
#1499 lore raw

PATCH 1496/2214 ntfs3: Remove calls to set/clear the error flag

In reply to Greg Kroah-Hartman

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

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

From: Matthew Wilcox (Oracle) <[email protected]>

[ Upstream commit 326a6fd9600c52016c3c18c1aeafca2fab254146 ]

Nobody checks the error flag on ntfs3 folios, so stop setting and clearing it.

Stable-dep-of: a32e893a077a ("fs/ntfs3: zero frame pages on read error")

---
 fs/ntfs3/frecord.c | 4 ----
 1 file changed, 4 deletions(-)
fs/ntfs3/frecord.c +0 -4
index 2957ba4ee0387..b3dbe21f48409 100644
@@ -2122,9 +2122,6 @@ int ni_readpage_cmpr(struct ntfs_inode *ni, struct page *page)
 	err = ni_read_frame(ni, frame_vbo, pages, pages_per_frame);
 
 out1:
-	if (err)
-		SetPageError(page);
-
 	for (i = 0; i < pages_per_frame; i++) {
 		pg = pages[i];
 		if (i == idx || !pg)
@@ -2708,7 +2705,6 @@ int ni_read_frame(struct ntfs_inode *ni, u64 frame_vbo, struct page **pages,
 	for (i = 0; i < pages_per_frame; i++) {
 		pg = pages[i];
 		kunmap(pg);
-		ClearPageError(pg);
 		SetPageUptodate(pg);
 	}
 
Signature · 1 line
2.53.0
author maintainer
#1500 lore raw

PATCH 1497/2214 ntfs: Do not kmap page cache pages for compression

In reply to Greg Kroah-Hartman

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

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

From: Matthew Wilcox (Oracle) <[email protected]>

[ Upstream commit 953b79a7a124c3dae86544cea581a5bc7655aa13 ]

These pages are accessed through vmap; they are not accessed by calling page_address(), so they do not need to be kmapped.

Stable-dep-of: a32e893a077a ("fs/ntfs3: zero frame pages on read error")

---
 fs/ntfs3/frecord.c | 11 -----------
 1 file changed, 11 deletions(-)
fs/ntfs3/frecord.c +0 -11
index b3dbe21f48409..07e6ebe736da3 100644
@@ -2446,9 +2446,6 @@ int ni_read_frame(struct ntfs_inode *ni, u64 frame_vbo, struct page **pages,
 	 * To simplify decompress algorithm do vmap for source
 	 * and target pages.
 	 */
-	for (i = 0; i < pages_per_frame; i++)
-		kmap(pages[i]);
-
 	frame_size = pages_per_frame << PAGE_SHIFT;
 	frame_mem = vmap(pages, pages_per_frame, VM_MAP, PAGE_KERNEL);
 	if (!frame_mem) {
@@ -2704,7 +2701,6 @@ int ni_read_frame(struct ntfs_inode *ni, u64 frame_vbo, struct page **pages,
 out:
 	for (i = 0; i < pages_per_frame; i++) {
 		pg = pages[i];
-		kunmap(pg);
 		SetPageUptodate(pg);
 	}
 
@@ -2791,9 +2787,6 @@ int ni_write_frame(struct ntfs_inode *ni, struct page **pages,
 		goto out1;
 	}
 
-	for (i = 0; i < pages_per_frame; i++)
-		kmap(pages[i]);
-
 	/* Map in-memory frame for read-only. */
 	frame_mem = vmap(pages, pages_per_frame, VM_MAP, PAGE_KERNEL_RO);
 	if (!frame_mem) {
@@ -2859,11 +2852,7 @@ int ni_write_frame(struct ntfs_inode *ni, struct page **pages,
 
 out3:
 	vunmap(frame_mem);
-
 out2:
-	for (i = 0; i < pages_per_frame; i++)
-		kunmap(pages[i]);
-
 	vunmap(frame_ondisk);
 out1:
 	for (i = 0; i < pages_per_frame; i++) {
Signature · 1 line
2.53.0
author maintainer
#1501 lore raw

PATCH 1498/2214 fs/ntfs3: zero frame pages on read error

In reply to Greg Kroah-Hartman

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

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

From: Weiming Shi <[email protected]>

[ Upstream commit a32e893a077a4b0d3b04c86cb7f69e354a187995 ]

ni_read_frame() marks every frame page uptodate at the 'out:' label regardless of the return value. The pages come from ntfs_lock_new_page() and are not zeroed, and several error paths reach 'out:' before anything is written to them (e.g. a failed decompress_lznt()/decompress_lzx_xpress() on a corrupted chunk, or an allocation failure).

A page marked uptodate is served directly from the page cache, so a later read() of the file returns the uninitialized page contents to userspace. On a crafted compressed image this leaks kernel memory, including pointers.

Zero each page on the error path before marking it uptodate. The success path is unchanged.

Fixes 4342306f0f0d ("fs/ntfs3: Add file operations and implementation")
Assisted-by Claude:claude-opus-4-8
---
 fs/ntfs3/frecord.c | 2 ++
 1 file changed, 2 insertions(+)
fs/ntfs3/frecord.c +2 -0
index 07e6ebe736da3..c358e3d632977 100644
@@ -2701,6 +2701,8 @@ int ni_read_frame(struct ntfs_inode *ni, u64 frame_vbo, struct page **pages,
 out:
 	for (i = 0; i < pages_per_frame; i++) {
 		pg = pages[i];
+		if (err)
+			clear_highpage(pg);
 		SetPageUptodate(pg);
 	}
 
Signature · 1 line
2.53.0
author maintainer
#1502 lore raw

PATCH 1499/2214 microblaze: uaccess: Zero out destination on failed get_user()

In reply to Greg Kroah-Hartman

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

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

From: Thomas Weißschuh <[email protected]>

[ Upstream commit 54c35131e628e1a18f02365495a24a73a1fb515e ]

On failure get_user() is supposed to zero out the destination variable. This is documented in the kdoc of the microblaze get_user() implementation and validated in lib/tests/usercopy_kunit.c.

Currently that zeroing is missing.

Add it.

---
 arch/microblaze/include/asm/uaccess.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
arch/microblaze/include/asm/uaccess.h +2 -1
index 3fe96979d2c62..d09e6430785af 100644
@@ -125,7 +125,8 @@ extern long __user_bad(void);
 #define get_user(x, ptr) ({				\
 	const typeof(*(ptr)) __user *__gu_ptr = (ptr);	\
 	access_ok(__gu_ptr, sizeof(*__gu_ptr)) ?	\
-		__get_user(x, __gu_ptr) : -EFAULT;	\
+		__get_user(x, __gu_ptr) :		\
+		((x) = 0, -EFAULT);			\
 })
 
 #define __get_user(x, ptr)						\
Signature · 1 line
2.53.0
author maintainer
#1503 lore raw

PATCH 1500/2214 microblaze: Pass -m{big,little}-endian through KBUILD_CPPFLAGS

In reply to Greg Kroah-Hartman

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

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

From: Thomas Weißschuh <[email protected]>

[ Upstream commit b9eababe8c0b9f71ce7f35f903879a7fcb6daebe ]

These parameters are also necessary for preprocessed files which do not contain code. One example is arch/microblaze/kernel/vmlinux.lds.S which inspects __MICROBLAZEEL__ for the definition of 'jiffies_64'. As currently the endianness is not passed from the build system for this file, the preprocessor will instead use the default endianness of the toolchain. When a default big-endian toolchain is used to build a little-endian kernel, that kernel will hang during delay calibration as 'jiffies' never changes.

Make sure all preprocessor invocations know about the correct endianness by passing it through KBUILD_CPPFLAGS. This will also be inherited by CFLAGS and AFLAGS.

Fixes 02b08045a030 ("microblaze: Add support for little-endian Microblaze")
---
 arch/microblaze/Makefile | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
arch/microblaze/Makefile +2 -4
index 9adc6b6434dfe..a46e259611b71 100644
@@ -34,12 +34,10 @@ CPUFLAGS-$(CONFIG_XILINX_MICROBLAZE0_USE_BARREL) += -mxl-barrel-shift
 CPUFLAGS-$(CONFIG_XILINX_MICROBLAZE0_USE_PCMP_INSTR) += -mxl-pattern-compare
 
 ifdef CONFIG_CPU_BIG_ENDIAN
-KBUILD_CFLAGS += -mbig-endian
-KBUILD_AFLAGS += -mbig-endian
+KBUILD_CPPFLAGS += -mbig-endian
 KBUILD_LDFLAGS += -EB
 else
-KBUILD_CFLAGS += -mlittle-endian
-KBUILD_AFLAGS += -mlittle-endian
+KBUILD_CPPFLAGS += -mlittle-endian
 KBUILD_LDFLAGS += -EL
 endif
 
Signature · 1 line
2.53.0
author maintainer
#1504 lore raw

PATCH 1501/2214 wifi: mac80211: Fix the return value in mesh_path_add() kernel-doc

In reply to Greg Kroah-Hartman

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

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

From: Karl Mehltretter <[email protected]>

[ Upstream commit 5c7b043cb1ef4bca9aa2e793dab862c0989d59c3 ]

mesh_path_add() has returned the new (or the already existing) struct mesh_path, or an ERR_PTR(), since commit ae76eef027f7 ("mac80211: return new mpath from mesh_path_add()"), but its kernel-doc still says "Returns: 0 on success". Describe the pointer.

Fixes ae76eef027f7 ("mac80211: return new mpath from mesh_path_add()")
Assisted-by LLM
---
 net/mac80211/mesh_pathtbl.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
net/mac80211/mesh_pathtbl.c +1 -1
index f2970f33f96c5..0a4641f06bb1f 100644
@@ -386,7 +386,7 @@ struct mesh_path *mesh_path_new(struct ieee80211_sub_if_data *sdata,
  * @dst: destination address of the path (ETH_ALEN length)
  * @sdata: local subif
  *
- * Returns: 0 on success
+ * Returns: the new or the existing path on success, an ERR_PTR() on failure
  *
  * State: the initial state of the new path is set to 0
  */
Signature · 1 line
2.53.0
author maintainer
#1505 lore raw

PATCH 1502/2214 firmware: imx: dsp: fix mailbox channel leak on kasprintf failure

In reply to Greg Kroah-Hartman

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

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

From: Linkai Gong <[email protected]>

[ Upstream commit e3631d0242842940f9e902afaea98202e56277e8 ]

If kasprintf() fails while setting up mailbox channels, already requested channels from earlier iterations were never freed. Route the failure through the existing cleanup path.

Fixes ffbf23d50353 ("firmware: imx: Add DSP IPC protocol interface")
---
 drivers/firmware/imx/imx-dsp.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/firmware/imx/imx-dsp.c +4 -2
index 0f656e4191d5c..e0419f908f6bd 100644
@@ -100,8 +100,10 @@ static int imx_dsp_setup_channels(struct imx_dsp_ipc *dsp_ipc)
 		else
 			chan_name = kasprintf(GFP_KERNEL, "rxdb%d", i - 2);
 
-		if (!chan_name)
-			return -ENOMEM;
+		if (!chan_name) {
+			ret = -ENOMEM;
+			goto out;
+		}
 
 		dsp_chan = &dsp_ipc->chans[i];
 		dsp_chan->name = chan_name;
Signature · 1 line
2.53.0
author maintainer
#1506 lore raw

PATCH 1503/2214 phy: cadence: Sierra: Do not modify register when getting parent clock

In reply to Greg Kroah-Hartman

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

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

From: Gregory CLEMENT <[email protected]>

[ Upstream commit 955e68eb3067842067b7b5395a3f8c278428802a ]

The get_parent() callback for the PLL_CMNLC1 clock was incorrectly writing to the register while determining the parent clock index. This unintended register access forces the PHY back into training mode. If the PHY is already configured, this unexpected change prevents it from exiting training mode.

Remove the register write operation to ensure the PHY remains stable during the get_parent() callback.

Fixes da08aab940092 ("phy: cadence: Sierra: Fix to get correct parent for mux clocks")
---
 drivers/phy/cadence/phy-cadence-sierra.c | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)
drivers/phy/cadence/phy-cadence-sierra.c +2 -9
index 3e2d096d54fd7..4b062976267d6 100644
@@ -439,23 +439,16 @@ static const struct phy_ops ops = {
 static u8 cdns_sierra_pll_mux_get_parent(struct clk_hw *hw)
 {
 	struct cdns_sierra_pll_mux *mux = to_cdns_sierra_pll_mux(hw);
-	struct regmap_field *plllc1en_field = mux->plllc1en_field;
-	struct regmap_field *termen_field = mux->termen_field;
 	struct regmap_field *field = mux->pfdclk_sel_preg;
 	unsigned int val;
 	int index;
 
 	regmap_field_read(field, &val);
 
-	if (strstr(clk_hw_get_name(hw), clk_names[CDNS_SIERRA_PLL_CMNLC1])) {
+	if (strstr(clk_hw_get_name(hw), clk_names[CDNS_SIERRA_PLL_CMNLC1]))
 		index = clk_mux_val_to_index(hw, cdns_sierra_pll_mux_table[CMN_PLLLC1], 0, val);
-		if (index == 1) {
-			regmap_field_write(plllc1en_field, 1);
-			regmap_field_write(termen_field, 1);
-		}
-	} else {
+	else
 		index = clk_mux_val_to_index(hw, cdns_sierra_pll_mux_table[CMN_PLLLC], 0, val);
-	}
 
 	return index;
 }
Signature · 1 line
2.53.0
author maintainer
#1507 lore raw

PATCH 1504/2214 memory: emif: Convert to platform remove callback returning void

In reply to Greg Kroah-Hartman

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

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

From: Uwe Kleine-König <[email protected]>

[ Upstream commit c8a53461990cb697ca494d6671fab9e196d20ce4 ]

The .remove() callback for a platform driver returns an int which makes many driver authors wrongly assume it's possible to do error handling by returning an error code. However the value returned is ignored (apart from emitting a warning) and this typically results in resource leaks.

To improve here there is a quest to make the remove callback return void. In the first step of this quest all drivers are converted to .remove_new(), which already returns void. Eventually after all drivers are converted, .remove_new() will be renamed to .remove().

Trivially convert this driver from always returning zero in the remove callback to the void returning variant.

Stable-dep-of: 314ef67d6732 ("memory: emif: fix device_node reference leak on np_ddr")

---
 drivers/memory/emif.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)
drivers/memory/emif.c +2 -4
index ecc78d6f89ed2..ca800aa3c59df 100644
@@ -1170,13 +1170,11 @@ static int __init_or_module emif_probe(struct platform_device *pdev)
 	return -ENODEV;
 }
 
-static int __exit emif_remove(struct platform_device *pdev)
+static void __exit emif_remove(struct platform_device *pdev)
 {
 	struct emif_data *emif = platform_get_drvdata(pdev);
 
 	emif_debugfs_exit(emif);
-
-	return 0;
 }
 
 static void emif_shutdown(struct platform_device *pdev)
@@ -1196,7 +1194,7 @@ MODULE_DEVICE_TABLE(of, emif_of_match);
 #endif
 
 static struct platform_driver emif_driver = {
-	.remove		= __exit_p(emif_remove),
+	.remove_new	= __exit_p(emif_remove),
 	.shutdown	= emif_shutdown,
 	.driver = {
 		.name = "emif",
Signature · 1 line
2.53.0
author maintainer
#1508 lore raw

PATCH 1505/2214 memory: emif: fix device_node reference leak on np_ddr

In reply to Greg Kroah-Hartman

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

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

From: Miles Krause <[email protected]>

[ Upstream commit 314ef67d673284d4c070455bd62d332f6d0f8634 ]

of_get_memory_device_details() takes a reference on the DDR device node via of_parse_phandle() and stores it in emif->np_ddr for later use, but drops that reference without calling of_node_put() on two error paths: the devm_kzalloc() allocation-failure check and the is_dev_data_valid() check. Both return NULL immediately, leaking the node reference each time either failure triggers.

The same reference is also never released on the success path: it is kept for the life of the bound device (compared against a later instance's np_ddr in the duplicate-DDR-geometry check) but emif_remove() never calls of_node_put() on it, leaking it every time an EMIF device unbinds. of_node_put() on a NULL pointer is a no-op, so this is also safe for the non-DT get_device_details() path, where np_ddr is left NULL by devm_kzalloc().

Fixes e6b42eb6a66c ("memory: emif: add device tree support to emif driver")
---
 drivers/memory/emif.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/memory/emif.c +3 -0
index ca800aa3c59df..8a7440b735290 100644
@@ -937,6 +937,7 @@ static struct emif_data * __init_or_module of_get_memory_device_details(
 	if (!emif || !pd || !dev_info) {
 		dev_err(dev, "%s: Out of memory!!\n",
 			__func__);
+		of_node_put(np_ddr);
 		goto error;
 	}
 
@@ -961,6 +962,7 @@ static struct emif_data * __init_or_module of_get_memory_device_details(
 			pd->device_info->io_width, pd->phy_type, pd->ip_rev,
 			emif->dev)) {
 		dev_err(dev, "%s: invalid device data!!\n", __func__);
+		of_node_put(np_ddr);
 		goto error;
 	}
 	/*
@@ -1175,6 +1177,7 @@ static void __exit emif_remove(struct platform_device *pdev)
 	struct emif_data *emif = platform_get_drvdata(pdev);
 
 	emif_debugfs_exit(emif);
+	of_node_put(emif->np_ddr);
 }
 
 static void emif_shutdown(struct platform_device *pdev)
Signature · 1 line
2.53.0
author maintainer
#1509 lore raw

PATCH 1506/2214 RDMA/uverbs: Make CQ handle mandatory for WQ creation

In reply to Greg Kroah-Hartman

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

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

From: Li RongQing <[email protected]>

[ Upstream commit f7d28791433d373ecdab7b00f60ddb504e858e0a ]

UVERBS_ATTR_CREATE_WQ_CQ_HANDLE is declared UA_OPTIONAL in the ioctl method definition, so the mandatory attribute bitmap does not enforce its presence. When userspace omits it, uverbs_attr_get_obj() returns ERR_PTR(-ENOENT) and the handler stores that error pointer into wq_init_attr.cq without validation.

The bogus cq pointer is then passed to the driver's create_wq callback. In the mlx5 case, create_rq() calls to_mcq(init_attr->cq) which applies container_of to the ERR_PTR value, producing a near-NULL pointer. The subsequent access in get_rq_ts_format() triggers a kernel NULL pointer dereference:

  BUG: kernel NULL pointer dereference, address: 0000000000000296
  RIP: 0010:create_rq+0x32/0x550 [mlx5_ib]
  Call Trace:
   mlx5_ib_create_wq+0x14a/0x210 [mlx5_ib]
   ib_uverbs_handler_UVERBS_METHOD_WQ_CREATE+0x1f0/0x320 [ib_uverbs]
   ib_uverbs_run_method+0x296/0x320 [ib_uverbs]
   ib_uverbs_cmd_verbs+0x1a0/0x260 [ib_uverbs]
   ib_uverbs_ioctl+0xa8/0x120 [ib_uverbs]

A WQ without a CQ was never valid; the legacy write path always required one via uobj_get_obj_read() in ib_uverbs_ex_create_wq(). Declare the attribute UA_MANDATORY so the uverbs framework rejects the ioctl early when the CQ handle is missing, before the handler ever runs.

Fixes ef3bc084a8ed ("IB/uverbs: Introduce create/destroy WQ commands over ioctl")
---
 drivers/infiniband/core/uverbs_std_types_wq.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/core/uverbs_std_types_wq.c +1 -1
index 7ded8339346f3..c23f455231bf0 100644
@@ -140,7 +140,7 @@ DECLARE_UVERBS_NAMED_METHOD(
 	UVERBS_ATTR_IDR(UVERBS_ATTR_CREATE_WQ_CQ_HANDLE,
 			UVERBS_OBJECT_CQ,
 			UVERBS_ACCESS_READ,
-			UA_OPTIONAL),
+			UA_MANDATORY),
 	UVERBS_ATTR_FD(UVERBS_ATTR_CREATE_WQ_EVENT_FD,
 		       UVERBS_OBJECT_ASYNC_EVENT,
 		       UVERBS_ACCESS_READ,
Signature · 1 line
2.53.0
author maintainer
#1510 lore raw

PATCH 1507/2214 remoteproc: core: Set trace buffer name for diagnostics

In reply to Greg Kroah-Hartman

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

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

From: Yonghao Zhang <[email protected]>

[ Upstream commit e23eab619b0d2dd796618406ba25da62f4d539a2 ]

rproc_trace_read() (in remoteproc_debugfs.c) prints the name of a trace buffer whose device address could not be resolved, but nothing ever assigns trace_mem.name, so that the diagnostics always prints an empty string:

Trace not available

Copy the generated "trace%d" name into the memory entry, so the diagnostics identifies the buffer failed to translate.

Fixes a987e6b91a5a ("remoteproc: fix trace buffer va initialization")
Assisted-by Claude-Code:glm-5.3
---
 drivers/remoteproc/remoteproc_core.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/remoteproc/remoteproc_core.c +3 -0
index 0e12d45b551f5..998eae2f87430 100644
@@ -690,6 +690,9 @@ static int rproc_handle_trace(struct rproc *rproc, void *ptr,
 		return -EINVAL;
 	}
 
+	/* keep the name for the diagnostic in rproc_trace_read() */
+	strscpy(trace->trace_mem.name, name, sizeof(trace->trace_mem.name));
+
 	list_add_tail(&trace->node, &rproc->traces);
 
 	rproc->num_traces++;
Signature · 1 line
2.53.0
author maintainer
#1511 lore raw

PATCH 1508/2214 dmaengine: mv_xor: kill tasklet on channel add error path

In reply to Greg Kroah-Hartman

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

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

From: Rosen Penev <[email protected]>

[ Upstream commit bcb4d469b19d53dc21e4a1da12d4438e7489e150 ]

If request_irq() succeeds but a later step (self-test or device registration) fails, the error path at err_free_irq frees the IRQ but does not kill the tasklet. A previously scheduled tasklet can still run after the channel is freed, accessing freed memory.

Add tasklet_kill() after free_irq() on the error path.

Fixes ff7b04796d98 ("dmaengine: DMA engine driver for Marvell XOR engine")
Assisted-by opencode:big-pickle
---
 drivers/dma/mv_xor.c | 1 +
 1 file changed, 1 insertion(+)
drivers/dma/mv_xor.c +1 -0
index 8b215cbca1186..7c1fdb2c91ce1 100644
@@ -1162,6 +1162,7 @@ mv_xor_channel_add(struct mv_xor_device *xordev,
 
 err_free_irq:
 	free_irq(mv_chan->irq, mv_chan);
+	tasklet_kill(&mv_chan->irq_tasklet);
 err_free_dma:
 	dma_free_wc(&pdev->dev, MV_XOR_POOL_SIZE,
 			  mv_chan->dma_desc_pool_virt, mv_chan->dma_desc_pool);
Signature · 1 line
2.53.0
author maintainer
#1512 lore raw

PATCH 1509/2214 loadpin: Fix stale KEXEC_VERIFY_SIG reference in documentation

In reply to Greg Kroah-Hartman

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

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

From: Karl Mehltretter <[email protected]>

[ Upstream commit eb0b0d222a832b6ec851e56d4922a775ac762955 ]

Commit 99d5cadfde2b ("kexec_file: split KEXEC_VERIFY_SIG into KEXEC_SIG and KEXEC_SIG_FORCE") removed CONFIG_KEXEC_VERIFY_SIG but left the LoadPin documentation pointing at it. Refer to CONFIG_KEXEC_SIG instead.

While at it, clean up the grammar of the sentence.

Fixes 99d5cadfde2b ("kexec_file: split KEXEC_VERIFY_SIG into KEXEC_SIG and KEXEC_SIG_FORCE")
Assisted-by LLM

Message-ID: <[email protected]>

---
 Documentation/admin-guide/LSM/LoadPin.rst | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)
Documentation/admin-guide/LSM/LoadPin.rst +5 -5
index dd3ca68b5df18..a90fd9dac9606 100644
@@ -23,9 +23,9 @@ sysctl allows for easy testing on systems with a mutable filesystem.)
 It's also possible to exclude specific file types from LoadPin using kernel
 command line option "``loadpin.exclude``". By default, all files are
 included, but they can be excluded using kernel command line option such
-as "``loadpin.exclude=kernel-module,kexec-image``". This allows to use
+as "``loadpin.exclude=kernel-module,kexec-image``". This allows using
 different mechanisms such as ``CONFIG_MODULE_SIG`` and
-``CONFIG_KEXEC_VERIFY_SIG`` to verify kernel module and kernel image while
-still use LoadPin to protect the integrity of other files kernel loads. The
-full list of valid file types can be found in ``kernel_read_file_str``
-defined in ``include/linux/kernel_read_file.h``.
+``CONFIG_KEXEC_SIG`` to verify kernel modules and kernel images while
+still using LoadPin to protect the integrity of other files the kernel
+loads. The full list of valid file types can be found in
+``kernel_read_file_str`` defined in ``include/linux/kernel_read_file.h``.
Signature · 1 line
2.53.0
author maintainer
#1513 lore raw

PATCH 1510/2214 ntfs3: fix out-of-bounds read in ntfs_read_mft()

In reply to Greg Kroah-Hartman

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

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

From: Weiming Shi <[email protected]>

[ Upstream commit 7a374d9be8267834dc534e55f01e5e01765cab4c ]

ntfs_read_mft() walks the attributes of an MFT record and, for each FILE_NAME attribute, compares the resident on-disk name with the name being looked up. The only size check before the compare is rsize >= SIZEOF_ATTRIBUTE_FILENAME, which covers the fixed header up to and including name_len, but not the variable-length name that follows it.

name_len comes straight from disk and is not validated against the resident data_size. A crafted image can set name_len large while keeping data_size at the header size, so ntfs_cmp_names_cpu() reads name_len le16 characters past the end of the MFT record buffer:

  BUG: KASAN: slab-out-of-bounds in ntfs_cmp_names_cpu (fs/ntfs3/upcase.c:84)
  Read of size 2 at addr ffff888031879c00 by task a.out
   ntfs_cmp_names_cpu (fs/ntfs3/upcase.c:84)
   ntfs_iget5 (fs/ntfs3/inode.c:209)
   dir_search_u (fs/ntfs3/dir.c:265)
   ntfs_lookup (fs/ntfs3/namei.c:85)
   __lookup_slow (fs/namei.c:1915)
   lookup_slow (fs/namei.c:1932)
   path_lookupat (fs/namei.c:2278)
   vfs_statx (fs/stat.c)
   vfs_fstatat (fs/stat.c)
   __do_sys_newfstatat (fs/stat.c)

Allocated by task:
mi_init (fs/ntfs3/record.c)
ntfs_iget5 (fs/ntfs3/inode.c)
dir_search_u (fs/ntfs3/dir.c)
ntfs_lookup (fs/ntfs3/namei.c)

Check that the name fits in the resident data before using it.

Fixes 82cae269cfa9 ("fs/ntfs3: Add initialization of super block")
Assisted-by Claude:claude-opus-4-8
---
 fs/ntfs3/inode.c | 5 +++++
 1 file changed, 5 insertions(+)
fs/ntfs3/inode.c +5 -0
index bccfd60db7de7..eb89fb3a7f4f1 100644
@@ -198,6 +198,11 @@ static struct inode *ntfs_read_mft(struct inode *inode,
 
 		names += 1;
 		fname = Add2Ptr(attr, roff);
+
+		/* Make sure the full name fits in the resident data. */
+		if (rsize < fname_full_size(fname))
+			goto out;
+
 		if (fname->type == FILE_NAME_DOS)
 			goto next_attr;
 
Signature · 1 line
2.53.0
author maintainer
#1514 lore raw

PATCH 1511/2214 mm/gup_test: reject wrapped user ranges

In reply to Greg Kroah-Hartman

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

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

From: Samuel Moelius <[email protected]>

[ Upstream commit e8ae6fd67021fafa9205330b3b248c9fc7216e0b ]

gup_test accepts an address and size from the debugfs ioctl and repeatedly compares against addr + size. If that addition wraps, the loop can be skipped and the ioctl returns success with size rewritten to zero.

Compute the end address once with overflow checking and use that checked end for the loop bounds.

Stable-dep-of: 44817a536aaa ("mm/gup_test: safely calculate GUP batch size")

---
 mm/gup_test.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)
mm/gup_test.c +8 -4
index d974dec19e1ca..15074dc772468 100644
@@ -102,11 +102,15 @@ static int __gup_test_ioctl(unsigned int cmd,
 	unsigned long i, nr_pages, addr, next;
 	long nr;
 	struct page **pages;
+	unsigned long end;
 	int ret = 0;
 	bool needs_mmap_lock =
 		cmd != GUP_FAST_BENCHMARK && cmd != PIN_FAST_BENCHMARK;
 
-	if (gup->size > ULONG_MAX)
+	if (gup->addr > ULONG_MAX || gup->size > ULONG_MAX)
+		return -EINVAL;
+	if (check_add_overflow((unsigned long)gup->addr,
+			       (unsigned long)gup->size, &end))
 		return -EINVAL;
 
 	nr_pages = gup->size / PAGE_SIZE;
@@ -122,13 +126,13 @@ static int __gup_test_ioctl(unsigned int cmd,
 	i = 0;
 	nr = gup->nr_pages_per_call;
 	start_time = ktime_get();
-	for (addr = gup->addr; addr < gup->addr + gup->size; addr = next) {
+	for (addr = gup->addr; addr < end; addr = next) {
 		if (nr != gup->nr_pages_per_call)
 			break;
 
 		next = addr + nr * PAGE_SIZE;
-		if (next > gup->addr + gup->size) {
-			next = gup->addr + gup->size;
+		if (next > end) {
+			next = end;
 			nr = (next - addr) / PAGE_SIZE;
 		}
 
Signature · 1 line
2.53.0
author maintainer
#1515 lore raw

PATCH 1512/2214 mm/gup_test: safely calculate GUP batch size

In reply to Greg Kroah-Hartman

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

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

From: Sarthak Sharma <[email protected]>

[ Upstream commit 44817a536aaa5bfac0d316242e93e489d0a69d5f ]

__gup_test_ioctl() calculates the end of a GUP batch using:

next = addr + nr * PAGE_SIZE;

If nr is too large, it can cause next to overflow and wrap around. If it wraps, the next > end check is bypassed and a large value of nr is passed to the GUP call, even though the pages array was allocated according to gup->size. This can lead to out of bounds writes.

Also, when fewer than PAGE_SIZE bytes remain, the calculated batch contains zero pages. The code still calls GUP functions with pages + i, which can point past the allocated array.

Calculate nr by taking the minimum of the number of pages per call and the pages remaining in the address range. Reject zero sized and non page aligned gup->size values and zero nr_pages_per_call value.

Fixes 64c349f4ae78 ("mm: add infrastructure for get_user_pages_fast() benchmarking")
---
 mm/gup_test.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
mm/gup_test.c +4 -5
index 15074dc772468..081440cee3245 100644
@@ -107,7 +107,8 @@ static int __gup_test_ioctl(unsigned int cmd,
 	bool needs_mmap_lock =
 		cmd != GUP_FAST_BENCHMARK && cmd != PIN_FAST_BENCHMARK;
 
-	if (gup->addr > ULONG_MAX || gup->size > ULONG_MAX)
+	if (gup->addr > ULONG_MAX || gup->size > ULONG_MAX || !gup->size ||
+	    !gup->nr_pages_per_call || !PAGE_ALIGNED(gup->size))
 		return -EINVAL;
 	if (check_add_overflow((unsigned long)gup->addr,
 			       (unsigned long)gup->size, &end))
@@ -130,11 +131,9 @@ static int __gup_test_ioctl(unsigned int cmd,
 		if (nr != gup->nr_pages_per_call)
 			break;
 
+		nr = min_t(unsigned long, nr, (end - addr) / PAGE_SIZE);
+
 		next = addr + nr * PAGE_SIZE;
-		if (next > end) {
-			next = end;
-			nr = (next - addr) / PAGE_SIZE;
-		}
 
 		switch (cmd) {
 		case GUP_FAST_BENCHMARK:
Signature · 1 line
2.53.0
author maintainer
#1516 lore raw

PATCH 1513/2214 mm/gup_test: report actual pinned bytes

In reply to Greg Kroah-Hartman

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

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

From: Sarthak Sharma <[email protected]>

[ Upstream commit 65a79b9b693d8c7bd7e5e4865c02cce922a7cc50 ]

__gup_test_ioctl() advances addr to the end of the current batch before checking if GUP pinned the entire requested batch. If GUP pins more than 0 pages but less than the requested batch size, addr still advances by the requested batch size.

The next iteration detects the partial pinning and breaks out of the loop. Again gup->size is calculated using addr - gup->addr, so it also includes the unpinned pages of the requested batch.

Calculate gup->size using the actual number of pages pinned multiplied by PAGE_SIZE.

Fixes 64c349f4ae78 ("mm: add infrastructure for get_user_pages_fast() benchmarking")
---
 mm/gup_test.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
mm/gup_test.c +1 -1
index 081440cee3245..5c52e0d6f032a 100644
@@ -180,7 +180,7 @@ static int __gup_test_ioctl(unsigned int cmd,
 	nr_pages = i;
 
 	gup->get_delta_usec = ktime_us_delta(end_time, start_time);
-	gup->size = addr - gup->addr;
+	gup->size = nr_pages * PAGE_SIZE;
 
 	/*
 	 * Take an un-benchmark-timed moment to verify DMA pinned
Signature · 1 line
2.53.0
author maintainer
#1517 lore raw

PATCH 1514/2214 arm64: dts: mediatek: mt8173: Fix MFG_ASYNC power domain clock

In reply to Greg Kroah-Hartman

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

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

From: Chen-Yu Tsai <[email protected]>

[ Upstream commit c9b3c82c9f1f86cdb2449d47e4ca66a19159ee12 ]

The MFG_ASYNC domain, which is likely associated to the whole MFG block, currently specifies clk26m as its domain clock. This is bogus, since the clock is an external crystal with no controls. Also, the MFG block has a independent CLK_TOP_AXI_MFG_IN_SEL clock, which according to the block diagram, gates access to the hardware registers. Having this one as the domain clock makes much more sense. This also fixes access to the MFGTOP registers.

Change the MFG_ASYNC domain clock to CLK_TOP_AXI_MFG_IN_SEL.

Fixes 8b6562644df9 ("arm64: dts: mediatek: Add mt8173 power domain controller")
---
 arch/arm64/boot/dts/mediatek/mt8173.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm64/boot/dts/mediatek/mt8173.dtsi +1 -1
index 9d05f908ddef5..75a7e40698bbf 100644
@@ -506,7 +506,7 @@ power-domain@MT8173_POWER_DOMAIN_USB {
 				};
 				mfg_async: power-domain@MT8173_POWER_DOMAIN_MFG_ASYNC {
 					reg = <MT8173_POWER_DOMAIN_MFG_ASYNC>;
-					clocks = <&clk26m>;
+					clocks = <&topckgen CLK_TOP_AXI_MFG_IN_SEL>;
 					clock-names = "mfg";
 					#address-cells = <1>;
 					#size-cells = <0>;
Signature · 1 line
2.53.0
author maintainer
#1518 lore raw

PATCH 1515/2214 ACPI: tables: handle failed initrd override mappings

In reply to Greg Kroah-Hartman

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

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

From: Slavin Liu <[email protected]>

[ Upstream commit 047c834d939df75386cfc0603c7ddeaa23a349a4 ]

The runtime AML table-loading path can invoke initrd override after permanent mappings are enabled. Mapping bookkeeping can then fail.

Return AE_NO_MEMORY in that case to prevent a NULL pointer dereference when attempting to read the table length.

Detected by static analysis and reviewed with AI-assisted source auditing.

Fixes b2a35003dfbc ("ACPI: Implement physical address table override")
Assisted-by LLM
[ rjw: Added empty code line after conditional, changelog edits ]
---
 drivers/acpi/tables.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/acpi/tables.c +3 -0
index 71419eb16e09f..cef62b6aa8284 100644
@@ -651,6 +651,9 @@ acpi_table_initrd_override(struct acpi_table_header *existing_table,
 	while (table_offset + ACPI_HEADER_SIZE <= all_tables_size) {
 		table = acpi_os_map_memory(acpi_tables_addr + table_offset,
 					   ACPI_HEADER_SIZE);
+		if (!table)
+			return AE_NO_MEMORY;
+
 		if (table_offset + table->length > all_tables_size) {
 			acpi_os_unmap_memory(table, ACPI_HEADER_SIZE);
 			WARN_ON(1);
Signature · 1 line
2.53.0
author maintainer
#1519 lore raw

PATCH 1516/2214 scsi: pm8001: Fix SATA FIS offset in struct hw_event_resp

In reply to Greg Kroah-Hartman

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

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

From: Bart Van Assche <[email protected]>

[ Upstream commit cdc97921de02a0c1a8324afa03e8b3e30a72bbe6 ]

In struct hw_event_resp, sas_identify was defined using struct sas_identify_frame, which is 32 bytes long because it includes a 4-byte CRC field. However, PM8001 and PM80xx hardware does not send the CRC in the event response frame, providing only the 28-byte identify payload.

Because sas_identify had a size of 32 bytes, the following sata_fis member was located at byte offset 44 instead of the actual hardware offset 40. In hw_event_sata_phy_up(), this discrepancy was worked around with negative pointer arithmetic: ((u8 *)&pPayload->sata_fis - 4). This triggers a compiler error under Clang with -Wstringop-overread:

error: 'memcpy' reading 20 bytes from a region of size 0 [-Werror,-Wstringop-overread]

          memcpy(phy->frame_rcvd, ((u8 *)&pPayload->sata_fis - 4),

Move struct sas_identify_frame_local (28 bytes) into pm8001_defs.h, and update struct hw_event_resp in both pm8001_hwi.h and pm80xx_hwi.h to use struct sas_identify_frame_local followed by sata_fis and a 4-byte padding word. This correctly aligns sata_fis at byte offset 40 while maintaining the 64-byte total payload size.

Update hw_event_sata_phy_up() in pm8001_hwi.c and pm80xx_hwi.c to copy directly from &pPayload->sata_fis.

Compile-tested only. Both the patch and its description have been generated by Gemini.

Fixes dbf9bfe61571 ("[SCSI] pm8001: add SAS/SATA HBA driver")
Fixes f5860992db55 ("[SCSI] pm80xx: Added SPCv/ve specific hardware functionalities and relevant changes in common files")
---
 drivers/scsi/pm8001/pm8001_defs.h |  98 +++++++++++++++++++++++++++++
 drivers/scsi/pm8001/pm8001_hwi.c  |   2 +-
 drivers/scsi/pm8001/pm8001_hwi.h  |   4 +-
 drivers/scsi/pm8001/pm80xx_hwi.c  |   2 +-
 drivers/scsi/pm8001/pm80xx_hwi.h  | 100 +-----------------------------
 5 files changed, 106 insertions(+), 100 deletions(-)
drivers/scsi/pm8001/pm8001_defs.h +98 -0
index 501b574239e8b..ff7b282090792 100644
@@ -41,6 +41,104 @@
 #ifndef _PM8001_DEFS_H_
 #define _PM8001_DEFS_H_
 
+#include <scsi/sas.h>
+
+#ifdef __LITTLE_ENDIAN_BITFIELD
+struct sas_identify_frame_local {
+	/* Byte 0 */
+	u8  frame_type:4;
+	u8  dev_type:3;
+	u8  _un0:1;
+
+	/* Byte 1 */
+	u8  _un1;
+
+	/* Byte 2 */
+	union {
+		struct {
+			u8  _un20:1;
+			u8  smp_iport:1;
+			u8  stp_iport:1;
+			u8  ssp_iport:1;
+			u8  _un247:4;
+		};
+		u8 initiator_bits;
+	};
+
+	/* Byte 3 */
+	union {
+		struct {
+			u8  _un30:1;
+			u8 smp_tport:1;
+			u8 stp_tport:1;
+			u8 ssp_tport:1;
+			u8 _un347:4;
+		};
+		u8 target_bits;
+	};
+
+	/* Byte 4 - 11 */
+	u8 _un4_11[8];
+
+	/* Byte 12 - 19 */
+	u8 sas_addr[SAS_ADDR_SIZE];
+
+	/* Byte 20 */
+	u8 phy_id;
+
+	u8 _un21_27[7];
+
+} __packed;
+
+#elif defined(__BIG_ENDIAN_BITFIELD)
+struct sas_identify_frame_local {
+	/* Byte 0 */
+	u8  _un0:1;
+	u8  dev_type:3;
+	u8  frame_type:4;
+
+	/* Byte 1 */
+	u8  _un1;
+
+	/* Byte 2 */
+	union {
+		struct {
+			u8  _un247:4;
+			u8  ssp_iport:1;
+			u8  stp_iport:1;
+			u8  smp_iport:1;
+			u8  _un20:1;
+		};
+		u8 initiator_bits;
+	};
+
+	/* Byte 3 */
+	union {
+		struct {
+			u8 _un347:4;
+			u8 ssp_tport:1;
+			u8 stp_tport:1;
+			u8 smp_tport:1;
+			u8 _un30:1;
+		};
+		u8 target_bits;
+	};
+
+	/* Byte 4 - 11 */
+	u8 _un4_11[8];
+
+	/* Byte 12 - 19 */
+	u8 sas_addr[SAS_ADDR_SIZE];
+
+	/* Byte 20 */
+	u8 phy_id;
+
+	u8 _un21_27[7];
+} __packed;
+#else
+#error "Bitfield order not defined!"
+#endif
+
 enum chip_flavors {
 	chip_8001,
 	chip_8008,
drivers/scsi/pm8001/pm8001_hwi.c +1 -1
index 537b762a2b834..97441653e355f 100644
@@ -3433,7 +3433,7 @@ hw_event_sata_phy_up(struct pm8001_hba_info *pm8001_ha, void *piomb)
 	phy->sas_phy.oob_mode = SATA_OOB_MODE;
 	sas_notify_phy_event(&phy->sas_phy, PHYE_OOB_DONE, GFP_ATOMIC);
 	spin_lock_irqsave(&phy->sas_phy.frame_rcvd_lock, flags);
-	memcpy(phy->frame_rcvd, ((u8 *)&pPayload->sata_fis - 4),
+	memcpy(phy->frame_rcvd, &pPayload->sata_fis,
 		sizeof(struct dev_to_host_fis));
 	phy->frame_rcvd_size = sizeof(struct dev_to_host_fis);
 	phy->identify.target_port_protocols = SAS_PROTOCOL_SATA;
drivers/scsi/pm8001/pm8001_hwi.h +3 -1
index d1f3aa93325bf..288d35029e7d6 100644
@@ -42,6 +42,7 @@
 
 #include <linux/types.h>
 #include <scsi/libsas.h>
+#include "pm8001_defs.h"
 
 
 /* for Request Opcode of IOMB */
@@ -229,8 +230,9 @@ struct hw_event_resp {
 	__le32	lr_evt_status_phyid_portid;
 	__le32	evt_param;
 	__le32	npip_portstate;
-	struct sas_identify_frame	sas_identify;
+	struct sas_identify_frame_local	sas_identify;
 	struct dev_to_host_fis	sata_fis;
+	u32	_r_a;
 } __attribute__((packed, aligned(4)));
 
 
drivers/scsi/pm8001/pm80xx_hwi.c +1 -1
index e5b5a730116a9..d3eb6eef9d6bf 100644
@@ -3420,7 +3420,7 @@ hw_event_sata_phy_up(struct pm8001_hba_info *pm8001_ha, void *piomb)
 	phy->sas_phy.oob_mode = SATA_OOB_MODE;
 	sas_notify_phy_event(&phy->sas_phy, PHYE_OOB_DONE, GFP_ATOMIC);
 	spin_lock_irqsave(&phy->sas_phy.frame_rcvd_lock, flags);
-	memcpy(phy->frame_rcvd, ((u8 *)&pPayload->sata_fis - 4),
+	memcpy(phy->frame_rcvd, &pPayload->sata_fis,
 		sizeof(struct dev_to_host_fis));
 	phy->frame_rcvd_size = sizeof(struct dev_to_host_fis);
 	phy->identify.target_port_protocols = SAS_PROTOCOL_SATA;
drivers/scsi/pm8001/pm80xx_hwi.h +3 -97

119 lines not on this page · raw message

Signature · 1 line
2.53.0
author maintainer
#1520 lore raw

PATCH 1517/2214 staging: iio: adc: ad7816: Fix DMA safety issues in SPI transfers

In reply to Greg Kroah-Hartman

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

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

From: Abdelnasser Hussein <[email protected]>

[ Upstream commit 69fa76f0af3414cc189c3b0b807cb59e327ecc00 ]

The SPI operations in this driver are not DMA safe:
1. spi_read() uses a stack-allocated buffer.
2. spi_write() in ad7816_spi_read() uses a struct member that shares a cacheline with other variables.
3. spi_write() in ad7816_spi_write() passes a stack parameter by reference.

Fix these violations by replacing all spi_read() and spi_write() calls with spi_write_then_read(). This safely handles DMA by internally allocating a bounce buffer for the transfers, avoiding cacheline sharing issues without needing dedicated aligned buffers. This implicitly corrects the size argument in read to sizeof(buf).

Fixes 7924425db04a ("staging: iio: adc: new driver for AD7816 devices")
---
 drivers/staging/iio/adc/ad7816.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/staging/iio/adc/ad7816.c +3 -3
index 081b17f498638..a46ca0559b6a2 100644
@@ -69,7 +69,7 @@ static int ad7816_spi_read(struct ad7816_chip_info *chip, u16 *data)
 
 	gpiod_set_value(chip->rdwr_pin, 1);
 	gpiod_set_value(chip->rdwr_pin, 0);
-	ret = spi_write(spi_dev, &chip->channel_id, sizeof(chip->channel_id));
+	ret = spi_write_then_read(spi_dev, &chip->channel_id, sizeof(chip->channel_id), NULL, 0);
 	if (ret < 0) {
 		dev_err(&spi_dev->dev, "SPI channel setting error\n");
 		return ret;
@@ -91,7 +91,7 @@ static int ad7816_spi_read(struct ad7816_chip_info *chip, u16 *data)
 
 	gpiod_set_value(chip->rdwr_pin, 0);
 	gpiod_set_value(chip->rdwr_pin, 1);
-	ret = spi_read(spi_dev, &buf, sizeof(*data));
+	ret = spi_write_then_read(spi_dev, NULL, 0, &buf, sizeof(buf));
 	if (ret < 0) {
 		dev_err(&spi_dev->dev, "SPI data read error\n");
 		return ret;
@@ -109,7 +109,7 @@ static int ad7816_spi_write(struct ad7816_chip_info *chip, u8 data)
 
 	gpiod_set_value(chip->rdwr_pin, 1);
 	gpiod_set_value(chip->rdwr_pin, 0);
-	ret = spi_write(spi_dev, &data, sizeof(data));
+	ret = spi_write_then_read(spi_dev, &data, sizeof(data), NULL, 0);
 	if (ret < 0)
 		dev_err(&spi_dev->dev, "SPI oti data write error\n");
 
Signature · 1 line
2.53.0
author maintainer
#1521 lore raw

PATCH 1518/2214 md: reject v0.90 arrays with too many devices

In reply to Greg Kroah-Hartman

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

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

From: Jérémy Jean <[email protected]>

[ Upstream commit 1a33a379e67853d6d778e6e80b214bdf1208f6f8 ]

The v0.90 superblock has room for MD_SB_DISKS device descriptors, but super_90_load() only rejects nonpositive raid_disks values. A crafted superblock can therefore copy a larger value into mddev->raid_disks. When metadata is next updated, super_90_sync() walks sb->disks[] up to that count and reads and writes beyond the 4 KiB superblock buffer.

A checksum-valid RAID1 member advertising 64 devices can be assembled by mdadm --incremental --run. The first write to the array then produces out-of-bounds writes from super_90_sync() under KASAN.

Reject device counts that cannot be represented by the v0.90 format while loading the superblock.

Assisted-by Codex:gpt-5
Fixes 1da177e4c3f4 ("Linux-2.6.12-rc2")
---
 drivers/md/md.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/md/md.c +1 -1
index d86c53606ed1d..bb3d9c52202b0 100644
@@ -1229,7 +1229,7 @@ static int super_90_load(struct md_rdev *rdev, struct md_rdev *refdev, int minor
 		goto abort;
 	}
 
-	if (sb->raid_disks <= 0)
+	if (sb->raid_disks <= 0 || sb->raid_disks > MD_SB_DISKS)
 		goto abort;
 
 	if (md_csum_fold(calc_sb_csum(sb)) != md_csum_fold(sb->sb_csum)) {
Signature · 1 line
2.53.0
author maintainer
#1522 lore raw

PATCH 1519/2214 md/raid5: dont BUG() on an inconsistent reshape state from the superblock

In reply to Greg Kroah-Hartman

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

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

From: Yogesh Gaur <[email protected]>

[ Upstream commit 0669c7dbd1d375f279e607a24addc2f1dae2d4f3 ]

raid5_run() branches on mddev->reshape_position. When it is MaxSector -- no reshape in progress -- the else branch asserts that nothing else describes one:

	BUG_ON(mddev->level != mddev->new_level);
	BUG_ON(mddev->layout != mddev->new_layout);
	BUG_ON(mddev->chunk_sectors != mddev->new_chunk_sectors);
	BUG_ON(mddev->delta_disks != 0);

Nothing enforces that invariant. Both superblock validators copy the reshape fields straight off disk without cross-checking them against each other: super_1_validate() takes reshape_position, delta_disks, new_level, new_layout and new_chunk from the superblock whenever MD_FEATURE_RESHAPE_ACTIVE is set, and super_90_validate() does the same for minor version 91. A superblock that sets the reshape feature while leaving reshape_position at the MaxSector sentinel therefore reaches the else branch with a non-zero delta_disks, and assembling the array takes the machine down:

kernel BUG at drivers/md/raid5.c:8117!

  Oops: invalid opcode: 0000 [#1] SMP KASAN NOPTI
  RIP: 0010:raid5_run+0x11a7/0x1670 drivers/md/raid5.c:8117
  Call Trace:
   md_run+0xc2f/0x2510 drivers/md/md.c:6779
   do_md_run+0x36/0x660 drivers/md/md.c:6880
   array_state_store+0x9c5/0xcf0 drivers/md/md.c:4765
   md_attr_store+0x1c5/0x330 drivers/md/md.c:6158
   sysfs_kf_write+0xf2/0x150 fs/sysfs/file.c:145
   kernfs_fop_write_iter+0x3e0/0x5f0 fs/kernfs/file.c:345

This is reachable by anyone who can present an md superblock, so BUG() is the wrong response. Refuse to start the array instead, the way the reshape_position != MaxSector branch a few lines above already refuses a reshape it cannot resume. Nothing has been allocated at this point -- the journal-and-bitmap check just above returns -EINVAL the same way -- so there is nothing to unwind.

---
 drivers/md/raid5.c | 16 +++++++++++-----
 1 file changed, 11 insertions(+), 5 deletions(-)
drivers/md/raid5.c +11 -5
index 943bbe0526542..0017e530745e7 100644
@@ -7593,11 +7593,17 @@ static int raid5_run(struct mddev *mddev)
 		}
 		pr_debug("md/raid:%s: reshape will continue\n", mdname(mddev));
 		/* OK, we should be able to continue; */
-	} else {
-		BUG_ON(mddev->level != mddev->new_level);
-		BUG_ON(mddev->layout != mddev->new_layout);
-		BUG_ON(mddev->chunk_sectors != mddev->new_chunk_sectors);
-		BUG_ON(mddev->delta_disks != 0);
+	} else if (mddev->level != mddev->new_level ||
+		   mddev->layout != mddev->new_layout ||
+		   mddev->chunk_sectors != mddev->new_chunk_sectors ||
+		   mddev->delta_disks != 0) {
+		/* No reshape is in progress, but the array describes one.
+		 * The superblock validators do not cross-check these against
+		 * reshape_position, so this is reachable from disk.
+		 */
+		pr_warn("md/raid:%s: inconsistent reshape state - aborting.\n",
+			mdname(mddev));
+		return -EINVAL;
 	}
 
 	if (test_bit(MD_HAS_JOURNAL, &mddev->flags) &&
Signature · 1 line
2.53.0
author maintainer
#1523 lore raw

PATCH 1520/2214 md/raid5: setup_conf: Fix null-ptr-def in error handling path

In reply to Greg Kroah-Hartman

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

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

From: Zhihao Cheng <[email protected]>

[ Upstream commit eba5f518701b1c7f2d883504f89c41611bd2de46 ]

For raid5 setup_conf, if an error(eg. bioset_init fails) happens before 'conf->mddev = mddev', following error handling path could trigger a null-ptr-def problem:

free_conf
log_exit
raid5_has_ppl

    test_bit(MD_HAS_PPL, &conf->mddev->flags) // conf->mddev is NULL !

Fix it by initializing conf->mddev in advance.

Fixes 3418d036c81d ("raid5-ppl: Partial Parity Log write logging implementation")
---
 drivers/md/raid5.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/md/raid5.c +1 -1
index 0017e530745e7..d5e5e3b4691f8 100644
@@ -7228,6 +7228,7 @@ static struct r5conf *setup_conf(struct mddev *mddev)
 	if (conf == NULL)
 		goto abort;
 
+	conf->mddev = mddev;
 #if PAGE_SIZE != DEFAULT_STRIPE_SIZE
 	conf->stripe_size = DEFAULT_STRIPE_SIZE;
 	conf->stripe_shift = ilog2(DEFAULT_STRIPE_SIZE) - 9;
@@ -7300,7 +7301,6 @@ static struct r5conf *setup_conf(struct mddev *mddev)
 	ret = bioset_init(&conf->bio_split, BIO_POOL_SIZE, 0, 0);
 	if (ret)
 		goto abort;
-	conf->mddev = mddev;
 
 	if ((conf->stripe_hashtbl = kzalloc(PAGE_SIZE, GFP_KERNEL)) == NULL)
 		goto abort;
Signature · 1 line
2.53.0
author maintainer
#1524 lore raw

PATCH 1521/2214 clk: milbeaut: Return divider update timeouts

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 92d0998090d1716ee6b47174c10eb9415a12fca8 ]

m10v_clk_divider_set_rate() logs when the hardware does not clear its write-valid request, but still returns success to the clock framework. It also uses the sleeping poll helper while holding the divider spinlock.

Use the atomic poll helper under the lock and return its timeout so the rate change is not reported as successful before the hardware accepts it.

The issue was found by our static-analysis tool and manually reviewed.

Fixes 6a6ba5b55a72 ("clock: milbeaut: Add Milbeaut M10V clock controller")
Assisted-by gpt 5
---
 drivers/clk/clk-milbeaut.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)
drivers/clk/clk-milbeaut.c +6 -4
index 80b9d78493bcd..eacc37f5b0bbd 100644
@@ -411,8 +411,8 @@ static int m10v_clk_divider_set_rate(struct clk_hw *hw, unsigned long rate,
 				unsigned long parent_rate)
 {
 	struct m10v_clk_divider *divider = to_m10v_div(hw);
-	int value;
 	unsigned long flags = 0;
+	int value, ret = 0;
 	u32 val;
 	u32 write_en = BIT(divider->width - 1);
 
@@ -434,8 +434,10 @@ static int m10v_clk_divider_set_rate(struct clk_hw *hw, unsigned long rate,
 
 	if (divider->write_valid_reg) {
 		writel(M10V_DCHREQ, divider->write_valid_reg);
-		if (readl_poll_timeout(divider->write_valid_reg, val,
-			!val, M10V_UPOLL_RATE, M10V_UTIMEOUT))
+		ret = readl_poll_timeout_atomic(divider->write_valid_reg, val,
+						!val, M10V_UPOLL_RATE,
+						M10V_UTIMEOUT);
+		if (ret)
 			pr_err("%s:%s couldn't stabilize\n",
 				__func__, clk_hw_get_name(hw));
 	}
@@ -445,7 +447,7 @@ static int m10v_clk_divider_set_rate(struct clk_hw *hw, unsigned long rate,
 	else
 		__release(divider->lock);
 
-	return 0;
+	return ret;
 }
 
 static const struct clk_ops m10v_clk_divider_ops = {
Signature · 1 line
2.53.0
author maintainer
#1525 lore raw

PATCH 1522/2214 drm/amd/pm: bound SCLK FCW range entries to the table size

In reply to Greg Kroah-Hartman

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

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

From: Hari Mishal <[email protected]>

[ Upstream commit f35aa5c77f5fca20fc269d6221df19c7b96704a5 ]

atomctrl_get_smc_sclk_range_table() copies ucSclkEntryNum entries from the VBIOS SMU_Info table into pp_atom_ctrl_sclk_range_table. Both the source (asSclkFcwRangeEntry[]) and destination (table->entry[]) arrays are fixed at MAX_SCLK_RANGE (8) entries, but ucSclkEntryNum is an unchecked u8 read straight from the VBIOS, so it can be up to 255. The destination is stack-allocated in both callers (polaris10_get_sclk_range_table(), vegam_get_sclk_range_table()), so a count above 8 overflows a stack buffer, not just a heap read.

Reject counts above MAX_SCLK_RANGE rather than clamping to it. Only one version of this VBIOS struct (ATOM_SMU_INFO_V2_1) has ever existed, so a count above 8 corresponds to no legitimate table format and means either a corrupted or a tampered VBIOS.

Fixes a23eefa2f461 ("drm/amd/powerplay: enable dpm for baffin.")
Assisted-by gkh_clanker_t1000
---
 drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/amd/pm/powerplay/hwmgr/ppatomctrl.c +3 -0
index e981cc3d2f1a5..23a520f402b9d 100644
@@ -1453,6 +1453,9 @@ int atomctrl_get_smc_sclk_range_table(struct pp_hwmgr *hwmgr, struct pp_atom_ctr
 	if (!psmu_info)
 		return -EINVAL;
 
+	if (psmu_info->ucSclkEntryNum > MAX_SCLK_RANGE)
+		return -EINVAL;
+
 	for (i = 0; i < psmu_info->ucSclkEntryNum; i++) {
 		table->entry[i].ucVco_setting = psmu_info->asSclkFcwRangeEntry[i].ucVco_setting;
 		table->entry[i].ucPostdiv = psmu_info->asSclkFcwRangeEntry[i].ucPostdiv;
Signature · 1 line
2.53.0
author maintainer
#1526 lore raw

PATCH 1523/2214 ata: sata_dwc_460ex: Use devm_platform_*ioremap_resource() APIs

In reply to Greg Kroah-Hartman

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

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

From: Andy Shevchenko <[email protected]>

[ Upstream commit f713961de5057765bc663db72b7e540f27827750 ]

Use devm_platform_get_and_ioremap_resource() and devm_platform_ioremap_resource() APIs instead of their open coded analogues.

Stable-dep-of: e137035fcd66 ("ata: sata_dwc_460ex: Don't enable interrupt if libATA core registration fails")

---
 drivers/ata/sata_dwc_460ex.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)
drivers/ata/sata_dwc_460ex.c +2 -5
index 1692c29227a02..7b0b320e4e825 100644
@@ -241,7 +241,6 @@ static int sata_dwc_dma_init_old(struct platform_device *pdev,
 				 struct sata_dwc_device *hsdev)
 {
 	struct device_node *np = pdev->dev.of_node;
-	struct resource *res;
 
 	hsdev->dma = devm_kzalloc(&pdev->dev, sizeof(*hsdev->dma), GFP_KERNEL);
 	if (!hsdev->dma)
@@ -258,8 +257,7 @@ static int sata_dwc_dma_init_old(struct platform_device *pdev,
 	}
 
 	/* Get physical SATA DMA register base address */
-	res = platform_get_resource(pdev, IORESOURCE_MEM, 1);
-	hsdev->dma->regs = devm_ioremap_resource(&pdev->dev, res);
+	hsdev->dma->regs = devm_platform_ioremap_resource(pdev, 1);
 	if (IS_ERR(hsdev->dma->regs))
 		return PTR_ERR(hsdev->dma->regs);
 
@@ -1224,8 +1222,7 @@ static int sata_dwc_probe(struct platform_device *ofdev)
 	host->private_data = hsdev;
 
 	/* Ioremap SATA registers */
-	res = platform_get_resource(ofdev, IORESOURCE_MEM, 0);
-	base = devm_ioremap_resource(&ofdev->dev, res);
+	base = devm_platform_get_and_ioremap_resource(ofdev, 0, &res);
 	if (IS_ERR(base))
 		return PTR_ERR(base);
 	dev_dbg(&ofdev->dev, "ioremap done for SATA register address\n");
Signature · 1 line
2.53.0
author maintainer
#1527 lore raw

PATCH 1524/2214 ata: sata_dwc_460ex: Use temporary variable for struct device

In reply to Greg Kroah-Hartman

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

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

From: Andy Shevchenko <[email protected]>

[ Upstream commit f1550f27f8a92a4d29329aeeb28b743365abceae ]

Use temporary variable for struct device to make code neater.

Stable-dep-of: e137035fcd66 ("ata: sata_dwc_460ex: Don't enable interrupt if libATA core registration fails")

---
 drivers/ata/sata_dwc_460ex.c | 37 ++++++++++++++++++------------------
 1 file changed, 19 insertions(+), 18 deletions(-)
drivers/ata/sata_dwc_460ex.c +19 -18
index 7b0b320e4e825..41a37a3dddb9e 100644
@@ -219,9 +219,10 @@ static int sata_dwc_dma_get_channel_old(struct sata_dwc_device_port *hsdevp)
 {
 	struct sata_dwc_device *hsdev = hsdevp->hsdev;
 	struct dw_dma_slave *dws = &sata_dwc_dma_dws;
+	struct device *dev = hsdev->dev;
 	dma_cap_mask_t mask;
 
-	dws->dma_dev = hsdev->dev;
+	dws->dma_dev = dev;
 
 	dma_cap_zero(mask);
 	dma_cap_set(DMA_SLAVE, mask);
@@ -229,8 +230,7 @@ static int sata_dwc_dma_get_channel_old(struct sata_dwc_device_port *hsdevp)
 	/* Acquire DMA channel */
 	hsdevp->chan = dma_request_channel(mask, sata_dwc_dma_filter, hsdevp);
 	if (!hsdevp->chan) {
-		dev_err(hsdev->dev, "%s: dma channel unavailable\n",
-			 __func__);
+		dev_err(dev, "%s: dma channel unavailable\n", __func__);
 		return -EAGAIN;
 	}
 
@@ -240,19 +240,20 @@ static int sata_dwc_dma_get_channel_old(struct sata_dwc_device_port *hsdevp)
 static int sata_dwc_dma_init_old(struct platform_device *pdev,
 				 struct sata_dwc_device *hsdev)
 {
-	struct device_node *np = pdev->dev.of_node;
+	struct device *dev = &pdev->dev;
+	struct device_node *np = dev->of_node;
 
-	hsdev->dma = devm_kzalloc(&pdev->dev, sizeof(*hsdev->dma), GFP_KERNEL);
+	hsdev->dma = devm_kzalloc(dev, sizeof(*hsdev->dma), GFP_KERNEL);
 	if (!hsdev->dma)
 		return -ENOMEM;
 
-	hsdev->dma->dev = &pdev->dev;
+	hsdev->dma->dev = dev;
 	hsdev->dma->id = pdev->id;
 
 	/* Get SATA DMA interrupt number */
 	hsdev->dma->irq = irq_of_parse_and_map(np, 1);
 	if (hsdev->dma->irq == NO_IRQ) {
-		dev_err(&pdev->dev, "no SATA DMA irq\n");
+		dev_err(dev, "no SATA DMA irq\n");
 		return -ENODEV;
 	}
 
@@ -1201,6 +1202,8 @@ static const struct ata_port_info sata_dwc_port_info[] = {
 
 static int sata_dwc_probe(struct platform_device *ofdev)
 {
+	struct device *dev = &ofdev->dev;
+	struct device_node *np = dev->of_node;
 	struct sata_dwc_device *hsdev;
 	u32 idr, versionr;
 	char *ver = (char *)&versionr;
@@ -1210,12 +1213,11 @@ static int sata_dwc_probe(struct platform_device *ofdev)
 	struct ata_host *host;
 	struct ata_port_info pi = sata_dwc_port_info[0];
 	const struct ata_port_info *ppi[] = { &pi, NULL };
-	struct device_node *np = ofdev->dev.of_node;
 	struct resource *res;
 
 	/* Allocate DWC SATA device */
-	host = ata_host_alloc_pinfo(&ofdev->dev, ppi, SATA_DWC_MAX_PORTS);
-	hsdev = devm_kzalloc(&ofdev->dev, sizeof(*hsdev), GFP_KERNEL);
+	host = ata_host_alloc_pinfo(dev, ppi, SATA_DWC_MAX_PORTS);
+	hsdev = devm_kzalloc(dev, sizeof(*hsdev), GFP_KERNEL);
 	if (!host || !hsdev)
 		return -ENOMEM;
 
@@ -1225,7 +1227,7 @@ static int sata_dwc_probe(struct platform_device *ofdev)
 	base = devm_platform_get_and_ioremap_resource(ofdev, 0, &res);
 	if (IS_ERR(base))
 		return PTR_ERR(base);
-	dev_dbg(&ofdev->dev, "ioremap done for SATA register address\n");
+	dev_dbg(dev, "ioremap done for SATA register address\n");
 
 	/* Synopsys DWC SATA specific Registers */
 	hsdev->sata_dwc_regs = base + SATA_DWC_REG_OFFSET;
@@ -1239,16 +1241,15 @@ static int sata_dwc_probe(struct platform_device *ofdev)
 	/* Read the ID and Version Registers */
 	idr = sata_dwc_readl(&hsdev->sata_dwc_regs->idr);
 	versionr = sata_dwc_readl(&hsdev->sata_dwc_regs->versionr);
-	dev_notice(&ofdev->dev, "id %d, controller version %c.%c%c\n",
-		   idr, ver[0], ver[1], ver[2]);
+	dev_notice(dev, "id %d, controller version %c.%c%c\n", idr, ver[0], ver[1], ver[2]);
 
 	/* Save dev for later use in dev_xxx() routines */
-	hsdev->dev = &ofdev->dev;
+	hsdev->dev = dev;
 
 	/* Get SATA interrupt number */
 	irq = irq_of_parse_and_map(np, 0);
 	if (irq == NO_IRQ) {
-		dev_err(&ofdev->dev, "no SATA DMA irq\n");
+		dev_err(dev, "no SATA DMA irq\n");
 		return -ENODEV;
 	}
 
@@ -1260,7 +1261,7 @@ static int sata_dwc_probe(struct platform_device *ofdev)
 	}
 #endif
 
-	hsdev->phy = devm_phy_optional_get(hsdev->dev, "sata-phy");
+	hsdev->phy = devm_phy_optional_get(dev, "sata-phy");
 	if (IS_ERR(hsdev->phy))
 		return PTR_ERR(hsdev->phy);
 
@@ -1275,7 +1276,7 @@ static int sata_dwc_probe(struct platform_device *ofdev)
 	 */
 	err = ata_host_activate(host, irq, sata_dwc_isr, 0, &sata_dwc_sht);
 	if (err)
-		dev_err(&ofdev->dev, "failed to activate host");
+		dev_err(dev, "failed to activate host");
 
 	/* Enable SATA Interrupts */
 	sata_dwc_enable_interrupts(hsdev);
@@ -1301,7 +1302,7 @@ static int sata_dwc_remove(struct platform_device *ofdev)
 	sata_dwc_dma_exit_old(hsdev);
 #endif
 
-	dev_dbg(&ofdev->dev, "done\n");
+	dev_dbg(dev, "done\n");
 	return 0;
 }
 
Signature · 1 line
2.53.0
author maintainer
#1528 lore raw

PATCH 1525/2214 ata: sata_dwc_460ex: Dont enable interrupt if libATA core registration fails

In reply to Greg Kroah-Hartman

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

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

From: Nikolay Borisov <[email protected]>

[ Upstream commit e137035fcd66cb454c479bb989968561bbc60341 ]

Recent 4bbc16a353a9 commit changed the timing when interrupts are enabled. However it didn't account for the libATA registration step. So it's possible that interrupts are enabled when the device is not registered with libATA core.

Resolve the issue by failing the entire registration process in case of failure to register with libATA.

Fixes 4bbc16a353a9 ("ata: sata_dwc_460ex: enable SATA interrupts only after IRQ handler is registered")
---
 drivers/ata/sata_dwc_460ex.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/ata/sata_dwc_460ex.c +4 -1
index 41a37a3dddb9e..e830a4158d0ec 100644
@@ -1275,11 +1275,14 @@ static int sata_dwc_probe(struct platform_device *ofdev)
 	 * error_handler() to execute a dummy Softreset EH session
 	 */
 	err = ata_host_activate(host, irq, sata_dwc_isr, 0, &sata_dwc_sht);
-	if (err)
+	if (err) {
 		dev_err(dev, "failed to activate host");
+		goto error_out;
+	}
 
 	/* Enable SATA Interrupts */
 	sata_dwc_enable_interrupts(hsdev);
+
 	return 0;
 
 error_out:
Signature · 1 line
2.53.0
author maintainer
#1529 lore raw

PATCH 1526/2214 PCI: cadence: Preserve all error codes in cdns_plat_pcie_probe()

In reply to Greg Kroah-Hartman

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

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

From: Vladimir Oltean <[email protected]>

[ Upstream commit fe108800a865b268eec103e4940c248c85a86e9f ]

The blamed commit functionally changed the error path of cdns_pcie_host_probe(), now cdns_plat_pcie_probe().

When the old code path executed "goto err_get_sync", the PCIe controller probe function propagated the pm_runtime_get_sync() error code. The new code doesn't, and returns 0.

Similarly for the "goto err_init" previously triggered by cdns_pcie_host_init() errors, and now triggered by cdns_pcie_host_setup() and cdns_pcie_ep_setup() errors. These are not propagated and will result in probing success, which is incorrect.

Fixes bd22885aa188 ("PCI: cadence: Refactor driver to use as a core library")
---
 drivers/pci/controller/cadence/pcie-cadence-plat.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/pci/controller/cadence/pcie-cadence-plat.c +1 -1
index 181b36a5d159d..d34bc117f8c27 100644
@@ -122,7 +122,7 @@ static int cdns_plat_pcie_probe(struct platform_device *pdev)
 	while (phy_count--)
 		device_link_del(pcie->link[phy_count]);
 
-	return 0;
+	return ret;
 }
 
 static void cdns_plat_pcie_shutdown(struct platform_device *pdev)
Signature · 1 line
2.53.0
author maintainer
#1530 lore raw

PATCH 1527/2214 kconfig: remove unneeded symbol_empty variable

In reply to Greg Kroah-Hartman

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

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

From: Masahiro Yamada <[email protected]>

[ Upstream commit 4e244c10eab345a735c5052688e4a55bddce5bf7 ]

This is used only for initializing other variables.

Use the empty string "" directly.

Please note newval.tri is unused for S_INT/HEX/STRING.

Stable-dep-of: dbe1ae2304a3 ("kconfig: fix extra output from savedefconfig on out-of-range defaults")

---
 scripts/kconfig/symbol.c | 11 ++---------
 1 file changed, 2 insertions(+), 9 deletions(-)
scripts/kconfig/symbol.c +2 -9
index d1e9c06456ae6..b978f1d09fba1 100644
@@ -32,12 +32,6 @@ struct symbol symbol_no = {
 	.flags = SYMBOL_CONST|SYMBOL_VALID,
 };
 
-static struct symbol symbol_empty = {
-	.name = "",
-	.curr = { "", no },
-	.flags = SYMBOL_VALID,
-};
-
 struct symbol *modules_sym;
 static tristate modules_val;
 static int sym_warnings;
@@ -359,7 +353,7 @@ void sym_calc_value(struct symbol *sym)
 	case S_INT:
 	case S_HEX:
 	case S_STRING:
-		newval = symbol_empty.curr;
+		newval.val = "";
 		break;
 	case S_BOOLEAN:
 	case S_TRISTATE:
@@ -710,13 +704,12 @@ const char *sym_get_string_default(struct symbol *sym)
 {
 	struct property *prop;
 	struct symbol *ds;
-	const char *str;
+	const char *str = "";
 	tristate val;
 
 	sym_calc_visibility(sym);
 	sym_calc_value(modules_sym);
 	val = symbol_no.curr.tri;
-	str = symbol_empty.curr.val;
 
 	/* If symbol has a default value look it up */
 	prop = sym_get_default_prop(sym);
Signature · 1 line
2.53.0
author maintainer
#1531 lore raw

PATCH 1528/2214 kconfig: default to zero if int/hex symbol lacks default property

In reply to Greg Kroah-Hartman

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

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

From: Masahiro Yamada <[email protected]>

[ Upstream commit 6262afa10ef7cc8fdf39b81a36f9546b68810431 ]

When a default property is missing in an int or hex symbol, it defaults to an empty string, which is not a valid symbol value.

It results in an incorrect .config, and can also lead to an infinite loop in scripting.

Use "0" for int and "0x0" for hex as a default value.

Stable-dep-of: dbe1ae2304a3 ("kconfig: fix extra output from savedefconfig on out-of-range defaults")

---
 scripts/kconfig/symbol.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)
scripts/kconfig/symbol.c +12 -5
index b978f1d09fba1..4b658116fdef1 100644
@@ -351,7 +351,11 @@ void sym_calc_value(struct symbol *sym)
 
 	switch (sym->type) {
 	case S_INT:
+		newval.val = "0";
+		break;
 	case S_HEX:
+		newval.val = "0x0";
+		break;
 	case S_STRING:
 		newval.val = "";
 		break;
@@ -759,14 +763,17 @@ const char *sym_get_string_default(struct symbol *sym)
 		case yes: return "y";
 		}
 	case S_INT:
+		if (!str[0])
+			str = "0";
+		break;
 	case S_HEX:
-		return str;
-	case S_STRING:
-		return str;
-	case S_UNKNOWN:
+		if (!str[0])
+			str = "0x0";
+		break;
+	default:
 		break;
 	}
-	return "";
+	return str;
 }
 
 const char *sym_get_string_value(struct symbol *sym)
Signature · 1 line
2.53.0
author maintainer
#1532 lore raw

PATCH 1529/2214 kconfig: fix extra output from savedefconfig on out-of-range defaults

In reply to Greg Kroah-Hartman

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

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

From: Julian Braha <[email protected]>

[ Upstream commit dbe1ae2304a3b02c3bf6ae82895ff232a29f3fee ]

The Kconfig interpreter currently allows defaults that are outside of the range bounds.

In these cases, the 'sym_validate_range' function will adjust the default value to the nearest range bound. For example, see this example:

config A
int
range 1 2
default 16

Here, since the default value of 16 is greater than the bounds, the effective default value gets adjusted down to the upper bound, 2.

However, 'savedefconfig' writes non-default values, and without being aware of the automatic adjustment to the range bound, it would write: A=2

This limitation is also documented in a comment: "The following fails to handle the situation where a default value is further limited by the valid range."

To resolve this, let's factor out the default-range adjustment logic from the existing 'sym_validate_range' function into its own 'sym_get_near_range_bound' function for 'savedefconfig' to use too, so that it compares against the effective value.

Adds tests, accordingly.

---
 scripts/kconfig/symbol.c                      | 39 ++++++++----
 .../kconfig/tests/savedefconfig_range/Kconfig | 60 +++++++++++++++++++
 .../tests/savedefconfig_range/__init__.py     |  8 +++
 .../kconfig/tests/savedefconfig_range/config  |  7 +++
 .../savedefconfig_range/expected_defconfig    |  0
 5 files changed, 102 insertions(+), 12 deletions(-)
 create mode 100644 scripts/kconfig/tests/savedefconfig_range/Kconfig
 create mode 100644 scripts/kconfig/tests/savedefconfig_range/__init__.py
 create mode 100644 scripts/kconfig/tests/savedefconfig_range/config
 create mode 100644 scripts/kconfig/tests/savedefconfig_range/expected_defconfig
scripts/kconfig/symbol.c +27 -12
index 4b658116fdef1..e939a92b92369 100644
@@ -117,7 +117,12 @@ static long long sym_get_range_val(struct symbol *sym, int base)
 	return strtoll(sym->curr.val, NULL, base);
 }
 
-static void sym_validate_range(struct symbol *sym)
+/*
+ * Return the nearest range bound for an out-of-range default value.
+ * Return NULL if the value is valid or the symbol has no active range.
+ */
+static struct symbol *sym_get_near_range_bound(struct symbol *sym,
+					  const char *value)
 {
 	struct property *prop;
 	struct symbol *range_sym;
@@ -132,21 +137,31 @@ static void sym_validate_range(struct symbol *sym)
 		base = 16;
 		break;
 	default:
-		return;
+		return NULL;
 	}
 	prop = sym_get_range_prop(sym);
 	if (!prop)
-		return;
-	val = strtoll(sym->curr.val, NULL, base);
+		return NULL;
+	val = strtoll(value, NULL, base);
 	range_sym = prop->expr->left.sym;
 	val2 = sym_get_range_val(range_sym, base);
 	if (val >= val2) {
 		range_sym = prop->expr->right.sym;
 		val2 = sym_get_range_val(range_sym, base);
 		if (val <= val2)
-			return;
+			return NULL;
 	}
-	sym->curr.val = range_sym->curr.val;
+
+	return range_sym;
+}
+
+static void sym_validate_range(struct symbol *sym)
+{
+	struct symbol *range_sym;
+
+	range_sym = sym_get_near_range_bound(sym, sym->curr.val);
+	if (range_sym)
+		sym->curr.val = range_sym->curr.val;
 }
 
 static void sym_set_changed(struct symbol *sym)
@@ -707,7 +722,7 @@ bool sym_set_string_value(struct symbol *sym, const char *newval)
 const char *sym_get_string_default(struct symbol *sym)
 {
 	struct property *prop;
-	struct symbol *ds;
+	struct symbol *ds, *range_sym;
 	const char *str = "";
 	tristate val;
 
@@ -725,11 +740,6 @@ const char *sym_get_string_default(struct symbol *sym)
 			val = EXPR_AND(expr_calc_value(prop->expr), prop->visible.tri);
 			break;
 		default:
-			/*
-			 * The following fails to handle the situation
-			 * where a default value is further limited by
-			 * the valid range.
-			 */
 			ds = prop_get_symbol(prop);
 			if (ds != NULL) {
 				sym_calc_value(ds);
@@ -773,6 +783,11 @@ const char *sym_get_string_default(struct symbol *sym)
 	default:
 		break;
 	}
+
+	range_sym = sym_get_near_range_bound(sym, str);
+	if (range_sym)
+		str = range_sym->curr.val;
+
 	return str;
 }
 
scripts/kconfig/tests/savedefconfig_range/Kconfig new file +60 -0

60 lines not on this page · raw message

scripts/kconfig/tests/savedefconfig_range/__init__.py new file +8 -0
new file mode 100644
index 0000000000000..961454be732c2
@@ -0,0 +1,8 @@
+# SPDX-License-Identifier: GPL-2.0
+"""Test savedefconfig with numerical defaults outside of ranges."""
+
+
+def test(conf):
+    assert conf._run_conf('--savedefconfig=defconfig', dot_config='config',
+                          out_file='defconfig') == 0
+    assert conf.config_matches('expected_defconfig')
scripts/kconfig/tests/savedefconfig_range/config new file +7 -0
new file mode 100644
index 0000000000000..d939cfe5fd349
@@ -0,0 +1,7 @@
+CONFIG_INT_DEFAULT_ABOVE_RANGE=1
+CONFIG_INT_DEFAULT_BELOW_RANGE=4
+CONFIG_INT_DYNAMIC_RANGE=3
+CONFIG_HEX_DEFAULT_ABOVE_RANGE=0x20
+CONFIG_INT_IMPLICIT_DEFAULT_ZERO=1
+CONFIG_INT_CONDITIONAL_RANGE=2
+CONFIG_INT_CONDITIONAL_DEFAULT=2
scripts/kconfig/tests/savedefconfig_range/expected_defconfig new file +0 -0
new file mode 100644
index 0000000000000..e69de29bb2d1d
Signature · 1 line
2.53.0
author maintainer
#1533 lore raw

PATCH 1530/2214 crypto: atmel-sha - reject initialization without a device

In reply to Greg Kroah-Hartman

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

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

From: Slavin Liu <[email protected]>

[ Upstream commit a9a2152d292fd0bc73814eb4a127278ee7dd3f6a ]

A transform may be initialized for the first time after the last device has been removed from the driver list. Return -ENODEV when the lookup is empty, before the debug message reads dd->dev.

Detected by static analysis and reviewed with AI-assisted source auditing.

Fixes ebc82efa1cd6 ("crypto: atmel - add Atmel SHA1/SHA256 driver")
Assisted-by LLM
---
 drivers/crypto/atmel-sha.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/crypto/atmel-sha.c +3 -0
index 1b13f601fd959..6eea72ed92669 100644
@@ -430,6 +430,9 @@ static int atmel_sha_init(struct ahash_request *req)
 	struct atmel_sha_reqctx *ctx = ahash_request_ctx(req);
 	struct atmel_sha_dev *dd = atmel_sha_find_dev(tctx);
 
+	if (!dd)
+		return -ENODEV;
+
 	ctx->dd = dd;
 
 	ctx->flags = 0;
Signature · 1 line
2.53.0
author maintainer
#1534 lore raw

PATCH 1531/2214 RDMA/uverbs: Drop restrack ref on ib_init_ucontext() failure in GET_CONTEXT

In reply to Greg Kroah-Hartman

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

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

From: Li RongQing <[email protected]>

[ Upstream commit 9dcacdc41083cccff74292851bf9fde999e0731a ]

The UVERBS_METHOD_GET_CONTEXT handler allocates the ucontext via ib_alloc_ucontext(), which calls rdma_restrack_new() (initialising the kref to 1) and rdma_restrack_set_name(NULL), the latter attaching the current task and taking a task_struct reference. When ib_init_ucontext() subsequently fails, the error path only does:

	kfree(attrs->context);
	attrs->context = NULL;

without first calling rdma_restrack_put() on &attrs->context->res. The kref therefore never reaches zero, restrack_release() is never invoked, and put_task_struct() is never called, so the task_struct reference taken during set_name is leaked permanently.

A local user with access to an RDMA device can repeat this path (e.g. by hitting an RDMA cgroup limit or supplying invalid ucaps) and leak one task_struct reference per attempt, eventually preventing those processes from being reaped or exhausting memory.

The legacy ib_uverbs_get_context() path already handles this correctly at its err_ucontext label, where rdma_restrack_put() precedes kfree(). Mirror that ordering here.

Fixes a1123418ba10 ("RDMA/uverbs: Add ioctl command to get a device context")
---
 drivers/infiniband/core/uverbs_std_types_device.c | 1 +
 1 file changed, 1 insertion(+)
drivers/infiniband/core/uverbs_std_types_device.c +1 -0
index 049684880ae03..05cc541c82090 100644
@@ -230,6 +230,7 @@ static int UVERBS_HANDLER(UVERBS_METHOD_GET_CONTEXT)(
 		return ret;
 	ret = ib_init_ucontext(attrs);
 	if (ret) {
+		rdma_restrack_put(&attrs->context->res);
 		kfree(attrs->context);
 		attrs->context = NULL;
 		return ret;
Signature · 1 line
2.53.0
author maintainer
#1535 lore raw

PATCH 1532/2214 ACPI: CPPC: Validate the _CPC package header

In reply to Greg Kroah-Hartman

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

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

From: Christian Loehle <[email protected]>

[ Upstream commit fb345025f161d17fe7c8a6048c7e80e467fa8ddf ]

The _CPC NumEntries field includes every package element, including NumEntries and Revision. acpi_cppc_processor_probe() nevertheless reads those first two elements before checking that they exist and trusts NumEntries when walking the remaining elements.

Reject packages with fewer than two elements and NumEntries values which exceed the package count before accessing Revision or iterating over register descriptors. Although the specification defines NumEntries as the number of package elements, tolerate additional trailing elements because ignoring them is safe and avoids rejecting padded firmware.

Revision is specified as a BYTE, but the parser assigns its 64-bit AML Integer to an unsigned int before validating it. Reject values above U8_MAX before conversion so, for example, 0x100000004 cannot truncate to revision 4.

---
 drivers/acpi/cppc_acpi.c | 22 ++++++++++++++++++----
 1 file changed, 18 insertions(+), 4 deletions(-)
drivers/acpi/cppc_acpi.c +18 -4
index c51183c2fd6f3..f9e3232067bc1 100644
@@ -693,6 +693,11 @@ int acpi_cppc_processor_probe(struct acpi_processor *pr)
 	}
 
 	out_obj = (union acpi_object *) output.pointer;
+	if (out_obj->package.count < 2) {
+		pr_debug("Unexpected _CPC package count (%u) for CPU:%d\n",
+			 out_obj->package.count, pr->id);
+		goto out_buf_free;
+	}
 
 	cpc_ptr = kzalloc(sizeof(struct cpc_desc), GFP_KERNEL);
 	if (!cpc_ptr) {
@@ -703,12 +708,15 @@ int acpi_cppc_processor_probe(struct acpi_processor *pr)
 	/* First entry is NumEntries. */
 	cpc_obj = &out_obj->package.elements[0];
 	if (cpc_obj->type == ACPI_TYPE_INTEGER)	{
-		num_ent = cpc_obj->integer.value;
-		if (num_ent <= 1) {
-			pr_debug("Unexpected _CPC NumEntries value (%d) for CPU:%d\n",
-				 num_ent, pr->id);
+		if (cpc_obj->integer.value < 2 ||
+		    cpc_obj->integer.value > out_obj->package.count) {
+			pr_debug("Invalid _CPC NumEntries (%llu) for package count (%u) on CPU:%d\n",
+				 cpc_obj->integer.value, out_obj->package.count,
+				 pr->id);
 			goto out_free;
 		}
+
+		num_ent = cpc_obj->integer.value;
 	} else {
 		pr_debug("Unexpected entry type(%d) for NumEntries\n",
 				cpc_obj->type);
@@ -718,6 +726,12 @@ int acpi_cppc_processor_probe(struct acpi_processor *pr)
 	/* Second entry should be revision. */
 	cpc_obj = &out_obj->package.elements[1];
 	if (cpc_obj->type == ACPI_TYPE_INTEGER)	{
+		if (cpc_obj->integer.value > U8_MAX) {
+			pr_debug("Invalid _CPC Revision (%llu) for CPU:%d\n",
+				 cpc_obj->integer.value, pr->id);
+			ret = -EINVAL;
+			goto out_free;
+		}
 		cpc_rev = cpc_obj->integer.value;
 	} else {
 		pr_debug("Unexpected entry type(%d) for Revision\n",
Signature · 1 line
2.53.0
author maintainer
#1536 lore raw

PATCH 1533/2214 spi: mxic: fix ps_clk leak in runtime resume

In reply to Greg Kroah-Hartman

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

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

From: Felix Gu <[email protected]>

[ Upstream commit 648e3c436032b675b339662ea9c0edd9221add76 ]

mxic_spi_runtime_resume() enables ps_clk before calling mxic_spi_clk_enable(). If the latter fails, ps_clk was not disabled.

Unwind ps_clk with clk_disable_unprepare() before returning the error.

Fixes b942d80b0a39 ("spi: Add MXIC controller driver")
---
 drivers/spi/spi-mxic.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/spi/spi-mxic.c +7 -1
index 03fce4493aa79..99015d0ae1ebb 100644
@@ -529,7 +529,13 @@ static int __maybe_unused mxic_spi_runtime_resume(struct device *dev)
 		return ret;
 	}
 
-	return mxic_spi_clk_enable(mxic);
+	ret = mxic_spi_clk_enable(mxic);
+	if (ret) {
+		clk_disable_unprepare(mxic->ps_clk);
+		return ret;
+	}
+
+	return 0;
 }
 
 static const struct dev_pm_ops mxic_spi_dev_pm_ops = {
Signature · 1 line
2.53.0
author maintainer
#1537 lore raw

PATCH 1534/2214 dt-bindings: vendor-prefixes: add ESS Technology

In reply to Greg Kroah-Hartman

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

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

From: Karl Asseily <[email protected]>

[ Upstream commit beb56d1d3de522885b30bb97fee83825041d65c2 ]

Add the vendor prefix for ESS Technology, Inc., maker of the SABRE family of audio DACs. The prefix follows the company's domain, esstech.com.

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index 38fc30b90f6cd..2ca68ee8602cf 100644
@@ -379,6 +379,8 @@ patternProperties:
     description: Seiko Epson Corp.
   "^esp,.*":
     description: Espressif Systems Co. Ltd.
+  "^esstech,.*":
+    description: ESS Technology, Inc.
   "^est,.*":
     description: ESTeem Wireless Modems
   "^ettus,.*":
Signature · 1 line
2.53.0
author maintainer
#1538 lore raw

PATCH 1535/2214 clk: raspberrypi: Handle cpufreq device registration failure

In reply to Greg Kroah-Hartman

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

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

From: Triet Hoang <[email protected]>

[ Upstream commit e98e3d712023859af7cd78bb7fdb87ea3a9ffb83 ]

Check the return value of platform_device_register_data() when registering the raspberrypi-cpufreq device and propagate the error if registration fails.

This prevents the clock driver from reporting a successful probe when the cpufreq device could not be registered.

Fixes e2bb18347c8e5 ("clk: raspberrypi: register platform device for raspberrypi-cpufreq")
---
 drivers/clk/bcm/clk-raspberrypi.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/clk/bcm/clk-raspberrypi.c +2 -0
index 280fd7a5ac75d..7d4476372e324 100644
@@ -351,6 +351,8 @@ static int raspberrypi_clk_probe(struct platform_device *pdev)
 
 	rpi->cpufreq = platform_device_register_data(dev, "raspberrypi-cpufreq",
 						     -1, NULL, 0);
+	if (IS_ERR(rpi->cpufreq))
+		return PTR_ERR(rpi->cpufreq);
 
 	return 0;
 }
Signature · 1 line
2.53.0
author maintainer
#1539 lore raw

PATCH 1536/2214 bpf: Factor out helpers for ctx access checking

In reply to Greg Kroah-Hartman

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

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

From: Hou Tao <[email protected]>

[ Upstream commit 35346ab64132d0f5919b06932d708c0d10360553 ]

Factor out two helpers to check the read access of ctx for raw tp and BTF function. bpf_tracing_ctx_access() is used to check the read access to argument is valid, and bpf_tracing_btf_ctx_access() checks whether the btf type of argument is valid besides the checking of argument read. bpf_tracing_btf_ctx_access() will be used by the following patch.

Stable-dep-of: c42da1d3ba9d ("bpf, lsm: Reject writes into the BPF LSM program context")

---
 include/linux/bpf.h      | 23 +++++++++++++++++++++++
 kernel/trace/bpf_trace.c | 16 ++--------------
 net/ipv4/bpf_tcp_ca.c    |  9 +--------
 3 files changed, 26 insertions(+), 22 deletions(-)
include/linux/bpf.h +23 -0
index e9081b1ec4367..5f26d4cd2272e 100644
@@ -1808,6 +1808,29 @@ bool bpf_prog_test_check_kfunc_call(u32 kfunc_id);
 bool btf_ctx_access(int off, int size, enum bpf_access_type type,
 		    const struct bpf_prog *prog,
 		    struct bpf_insn_access_aux *info);
+
+static inline bool bpf_tracing_ctx_access(int off, int size,
+					  enum bpf_access_type type)
+{
+	if (off < 0 || off >= sizeof(__u64) * MAX_BPF_FUNC_ARGS)
+		return false;
+	if (type != BPF_READ)
+		return false;
+	if (off % size != 0)
+		return false;
+	return true;
+}
+
+static inline bool bpf_tracing_btf_ctx_access(int off, int size,
+					      enum bpf_access_type type,
+					      const struct bpf_prog *prog,
+					      struct bpf_insn_access_aux *info)
+{
+	if (!bpf_tracing_ctx_access(off, size, type))
+		return false;
+	return btf_ctx_access(off, size, type, prog, info);
+}
+
 int btf_struct_access(struct bpf_verifier_log *log, const struct btf *btf,
 		      const struct btf_type *t, int off, int size,
 		      enum bpf_access_type atype,
kernel/trace/bpf_trace.c +2 -14
index e08928f4a862f..e4e6f91aa3c94 100644
@@ -1575,13 +1575,7 @@ static bool raw_tp_prog_is_valid_access(int off, int size,
 					const struct bpf_prog *prog,
 					struct bpf_insn_access_aux *info)
 {
-	if (off < 0 || off >= sizeof(__u64) * MAX_BPF_FUNC_ARGS)
-		return false;
-	if (type != BPF_READ)
-		return false;
-	if (off % size != 0)
-		return false;
-	return true;
+	return bpf_tracing_ctx_access(off, size, type);
 }
 
 static bool tracing_prog_is_valid_access(int off, int size,
@@ -1589,13 +1583,7 @@ static bool tracing_prog_is_valid_access(int off, int size,
 					 const struct bpf_prog *prog,
 					 struct bpf_insn_access_aux *info)
 {
-	if (off < 0 || off >= sizeof(__u64) * MAX_BPF_FUNC_ARGS)
-		return false;
-	if (type != BPF_READ)
-		return false;
-	if (off % size != 0)
-		return false;
-	return btf_ctx_access(off, size, type, prog, info);
+	return bpf_tracing_btf_ctx_access(off, size, type, prog, info);
 }
 
 int __weak bpf_prog_test_run_tracing(struct bpf_prog *prog,
net/ipv4/bpf_tcp_ca.c +1 -8
index d3a2dbd13ea6b..0675f87a68a51 100644
@@ -81,14 +81,7 @@ static bool bpf_tcp_ca_is_valid_access(int off, int size,
 				       const struct bpf_prog *prog,
 				       struct bpf_insn_access_aux *info)
 {
-	if (off < 0 || off >= sizeof(__u64) * MAX_BPF_FUNC_ARGS)
-		return false;
-	if (type != BPF_READ)
-		return false;
-	if (off % size != 0)
-		return false;
-
-	if (!btf_ctx_access(off, size, type, prog, info))
+	if (!bpf_tracing_btf_ctx_access(off, size, type, prog, info))
 		return false;
 
 	if (info->reg_type == PTR_TO_BTF_ID && info->btf_id == sock_id)
Signature · 1 line
2.53.0
author maintainer
#1540 lore raw

PATCH 1537/2214 bpf, lsm: Reject writes into the BPF LSM program context

In reply to Greg Kroah-Hartman

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

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

From: Daniel Borkmann <[email protected]>

[ Upstream commit c42da1d3ba9d5067cf3f0676b98bcb3f5f9e3ce1 ]

The verifier types a load from a tracing-style context by its offset
alone: btf_ctx_access() looks the offset up in the attach hook's prototype, and check_mem_access() stamps the resulting type onto the destination register. Nothing tracks what the slot actually holds. That is sound only as long as the context is read-only, and for BPF LSM it is not. lsm_verifier_ops wires ->is_valid_access() straight to btf_ctx_access(), which never looks at enum bpf_access_type, so a store into the context passes verification. Every other user of the helper either goes through bpf_tracing_btf_ctx_access() or open-codes the same type != BPF_READ test first; BPF LSM is the only one that does neither.

A program can therefore store a scalar of its choosing into a context slot, read the slot back, and get a register the verifier believes is a trusted pointer of the hook's argument type. Handing that to a kfunc or helper which dereferences a trusted argument turns it into an arbitrary kernel access. Route BPF LSM through bpf_tracing_btf_ctx_access() like everyone else.

---
 kernel/bpf/bpf_lsm.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
kernel/bpf/bpf_lsm.c +2 -2
index 02913d8610414..07be0ecfd020d 100644
@@ -223,6 +223,6 @@ const struct bpf_prog_ops lsm_prog_ops = {
 };
 
 const struct bpf_verifier_ops lsm_verifier_ops = {
-	.get_func_proto = bpf_lsm_func_proto,
-	.is_valid_access = btf_ctx_access,
+	.get_func_proto		= bpf_lsm_func_proto,
+	.is_valid_access	= bpf_tracing_btf_ctx_access,
 };
Signature · 1 line
2.53.0
author maintainer
#1541 lore raw

PATCH 1538/2214 fs/ntfs3: check INDEX_ROOT resident size before use in ntfs_read_mft()

In reply to Greg Kroah-Hartman

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

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

From: Xiang Mei (Microsoft) <[email protected]>

[ Upstream commit d0ac69ce3af4dbae01d4f064b7482f0c5440dbde ]

When ntfs_read_mft() parses an ATTR_ROOT attribute it computes

	root = Add2Ptr(attr, roff);

using the on-disk resident data offset (roff = attr->res.data_off) and then dereferences root->type and root->rule, and later passes root on to indx_init(). Unlike the ATTR_STD and ATTR_NAME cases, which validate that the resident data is large enough (asize >= sizeof(struct X) + roff and rsize >= sizeof(struct X)) before accessing it, the ATTR_ROOT case has no such check.

mi_enum_attr() only guarantees data_off + data_size <= attribute size, so a crafted image can set data_off such that root lands at (or past) the end of the MFT record's slab allocation while still passing that check (e.g. data_size == 0). The subsequent read of root->type is then a slab-out-of-bounds read, reported by KASAN when mounting the image:

  BUG: KASAN: slab-out-of-bounds in ntfs_iget5 (fs/ntfs3/inode.c:283)
  Read of size 4 at addr ffff88800c06a400 by task exploit/142
  Call Trace:
   ...  [KASAN report internals]
   ntfs_iget5 (fs/ntfs3/inode.c:283 fs/ntfs3/inode.c:542)
   ntfs_fill_super (fs/ntfs3/super.c:1690)
   get_tree_bdev_flags (fs/super.c:1634)
   vfs_get_tree (fs/super.c:1694)
   fc_mount (fs/namespace.c:1198)
   path_mount (fs/namespace.c:3765 ...)
   __x64_sys_mount (fs/namespace.c:4174 ...)
   do_syscall_64 (arch/x86/entry/syscall_64.c:63)
   entry_SYSCALL_64_after_hwframe (arch/x86/entry/entry_64.S:121)
  The buggy address is located 0 bytes to the right of
   allocated 1024-byte region [ffff88800c06a000, ffff88800c06a400)
   which belongs to the cache kmalloc-1k of size 1024

Validate the resident data size in the ATTR_ROOT case the same way the neighbouring cases do, so that a truncated or mis-offset INDEX_ROOT is rejected before root is dereferenced.

Fixes 82cae269cfa9 ("fs/ntfs3: Add initialization of super block")
Reported-by [email protected]
---
 fs/ntfs3/inode.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
fs/ntfs3/inode.c +3 -1
index eb89fb3a7f4f1..02de0ae1da8bf 100644
@@ -273,7 +273,9 @@ static struct inode *ntfs_read_mft(struct inode *inode,
 		break;
 
 	case ATTR_ROOT:
-		if (attr->non_res)
+		if (attr->non_res ||
+		    asize < sizeof(struct INDEX_ROOT) + roff ||
+		    rsize < sizeof(struct INDEX_ROOT))
 			goto out;
 
 		root = Add2Ptr(attr, roff);
Signature · 1 line
2.53.0
author maintainer
#1542 lore raw

PATCH 1539/2214 fs/ntfs3: fix valid_size not being updated in attr_allocate_frame()

In reply to Greg Kroah-Hartman

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

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

From: Zhan Xusheng <[email protected]>

[ Upstream commit a1972501ffb15313f7ae598989bb2cd32829e625 ]

At the end of attr_allocate_frame() the on-disk valid_size is meant to be updated to the value passed in via @new_valid:

	valid_size = le64_to_cpu(attr_b->nres.valid_size);
	if (new_valid != valid_size) {
		attr_b->nres.valid_size = cpu_to_le64(valid_size);
		mi_b->dirty = true;
	}

but it writes back the old valid_size that was just read instead of new_valid, so the field is never actually changed even though the record is marked dirty -- the intended update is a no-op.

ni_update_parent() spells the same read-compare-write out correctly a few hundred lines away, storing the newly derived value:

	valid_le = cpu_to_le64(new_valid);
	if (valid_le != attr->nres.valid_size) {
		attr->nres.valid_size = valid_le;

In practice that path, reached from ni_write_inode(), re-derives nres.valid_size from ni->i_valid on inode write-back, which normally masks this, so the visible impact is limited. The assignment here is still wrong though; store @new_valid as intended.

Fixes be71b5cba2e6 ("fs/ntfs3: Add attrib operations")
---
 fs/ntfs3/attrib.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/ntfs3/attrib.c +1 -1
index 872586e9598a7..5a210bea00641 100644
@@ -1686,7 +1686,7 @@ int attr_allocate_frame(struct ntfs_inode *ni, CLST frame, size_t compr_size,
 
 		valid_size = le64_to_cpu(attr_b->nres.valid_size);
 		if (new_valid != valid_size) {
-			attr_b->nres.valid_size = cpu_to_le64(valid_size);
+			attr_b->nres.valid_size = cpu_to_le64(new_valid);
 			mi_b->dirty = true;
 		}
 	}
Signature · 1 line
2.53.0
author maintainer
#1543 lore raw

PATCH 1540/2214 fs/ntfs3: validate EA entry layout against ef->size in ntfs_read_ea

In reply to Greg Kroah-Hartman

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

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

From: Konstantin Komarov <[email protected]>

[ Upstream commit 01e56820ef52d14a5560fb7c0c6f2f75ef6f9393 ]

In ntfs_read_ea(), the ef->size branch only checks that the declared entry size fits the remaining buffer; unlike the zero-size branch it never verifies that the entry's own name and value fit inside ef->size. A crafted image can thus present an in-bounds ef->size with an arbitrary elength, which find_ea()/ntfs_get_ea() later use as the memcpy length, reading past the allocation and leaking adjacent heap memory:

  BUG: KASAN: slab-out-of-bounds in ntfs_get_ea (fs/ntfs3/xattr.c:302)
  Read of size 65535 at addr ffff888010770b0f by task exploit/142
   ...
   __asan_memcpy (mm/kasan/shadow.c:105)
   ntfs_get_ea (fs/ntfs3/xattr.c:302)
   ntfs_getxattr (fs/ntfs3/xattr.c:848)
   __vfs_getxattr (fs/xattr.c:446)
   vfs_getxattr (fs/xattr.c:479)
   do_getxattr (fs/xattr.c:808)
   filename_getxattr (fs/xattr.c:837)
   path_getxattrat (fs/xattr.c:876)
   do_syscall_64 (arch/x86/entry/syscall_64.c:94)
  ...
  The buggy address is located 15 bytes inside of
   allocated 20-byte region, cache kmalloc-32

Validate the name/value layout in the ef->size branch too, bounding it against ef->size, as the zero-size branch validates it against the buffer.

Fixes be71b5cba2e6 ("fs/ntfs3: Add attrib operations")
Assisted-by Claude:claude-opus-4-8
[[email protected]: remove duplicate check of need > ea_size]
---
 fs/ntfs3/xattr.c | 2 ++
 1 file changed, 2 insertions(+)
fs/ntfs3/xattr.c +2 -0
index 078f4abb42ec2..f12eaaf1647e8 100644
@@ -167,6 +167,8 @@ static int ntfs_read_ea(struct ntfs_inode *ni, struct EA_FULL **ea,
 			ea_size = le32_to_cpu(ef->size);
 			if (ea_size > bytes || packed_size > ea_size)
 				goto out1;
+			if (bytes < offsetof(struct EA_FULL, name))
+				goto out1;
 			continue;
 		}
 
Signature · 1 line
2.53.0
author maintainer
#1544 lore raw

PATCH 1541/2214 fs/ntfs3: validate $SDH entry size in ntfs_insert_security

In reply to Greg Kroah-Hartman

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

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

From: Xiang Mei <[email protected]>

[ Upstream commit 5b76487f94dacc2884135ce9a7bda61ce7ffa209 ]

ntfs_insert_security() walks the $Secure::$SDH index and reads e->sec_hdr.size for every entry returned by indx_find()/indx_find_sort(). Those only guarantee the NTFS_DE envelope (e->size), so a crafted $SDH entry shorter than a full NTFS_DE_SDH at the tail of an index buffer makes the read of e->sec_hdr.size run past the allocation. Triggerable by an unprivileged setxattr(2) on a crafted NTFS image.

  BUG: KASAN: slab-out-of-bounds in ntfs_insert_security (fs/ntfs3/fsntfs.c:2148)
  Read of size 4 at addr ffff88800fa1de00 by task exploit/142
   ntfs_insert_security (fs/ntfs3/fsntfs.c:2148)
   ntfs_setxattr (fs/ntfs3/xattr.c:947)
   __vfs_setxattr (fs/xattr.c:223)
   __vfs_setxattr_noperm (fs/xattr.c:257)
   vfs_setxattr (fs/xattr.c:344)
   filename_setxattr (fs/xattr.c:690)
   path_setxattrat (fs/xattr.c:734)
   __x64_sys_setxattr (fs/xattr.c:764)
   [...]
  The buggy address is located 0 bytes to the right of
   allocated 512-byte region [ffff88800fa1dc00, ffff88800fa1de00)

Stop the scan when an entry is too small to hold a SECURITY_HDR.

Fixes 82cae269cfa9 ("fs/ntfs3: Add initialization of super block")
---
 fs/ntfs3/fsntfs.c | 3 +++
 1 file changed, 3 insertions(+)
fs/ntfs3/fsntfs.c +3 -0
index cab0679c978c3..31a014dc5b066 100644
@@ -2153,6 +2153,9 @@ int ntfs_insert_security(struct ntfs_sb_info *sbi,
 		goto out;
 
 	while (e) {
+		if (le16_to_cpu(e->de.size) < SIZEOF_SDH_DIRENTRY)
+			break;
+
 		if (le32_to_cpu(e->sec_hdr.size) == new_sec_size) {
 			err = ntfs_read_run_nb(sbi, &ni->file.run,
 					       le64_to_cpu(e->sec_hdr.off),
Signature · 1 line
2.53.0
author maintainer
#1545 lore raw

PATCH 1542/2214 pmdomain: imx: Propagate SCU provider registration failures

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit d3efba7921f3ebebd91fc481910a428c78414fd8 ]

imx_scu_init_pm_domains() ignores of_genpd_add_provider_onecell() and returns success even when none of the initialized domains are published to DT consumers.

Return the provider error and remove the initialized domains in reverse order before devm releases their backing storage on probe failure.

The issue was found by our static-analysis tool and manually reviewed.

Fixes c800cd7824bd ("firmware: imx: add SCU power domain driver")
Assisted-by gpt 5
---
 drivers/firmware/imx/scu-pd.c | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
drivers/firmware/imx/scu-pd.c +8 -2
index a0fc352a79936..f684168a14bd2 100644
@@ -348,7 +348,7 @@ static int imx_scu_init_pm_domains(struct device *dev,
 	struct genpd_onecell_data *pd_data;
 	struct imx_sc_pm_domain *sc_pd;
 	u32 count = 0;
-	int i, j;
+	int i, j, ret;
 
 	for (i = 0; i < pd_soc->num_ranges; i++)
 		count += pd_ranges[i].num;
@@ -377,7 +377,13 @@ static int imx_scu_init_pm_domains(struct device *dev,
 	pd_data->num_domains = count;
 	pd_data->xlate = imx_scu_pd_xlate;
 
-	of_genpd_add_provider_onecell(dev->of_node, pd_data);
+	ret = of_genpd_add_provider_onecell(dev->of_node, pd_data);
+	if (ret) {
+		while (count)
+			pm_genpd_remove(domains[--count]);
+
+		return ret;
+	}
 
 	return 0;
 }
Signature · 1 line
2.53.0
author maintainer
#1546 lore raw

PATCH 1543/2214 dt-bindings: vendor-prefixes: add vendor prefix for Nexperia B.V.

In reply to Greg Kroah-Hartman

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

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

From: Neil Armstrong <[email protected]>

[ Upstream commit 819a3e62d37a2fe73c1e7f241f2ac086042e7757 ]

Document the vendor prefix for the semiconductor company Nexperia B.V.

The company website is https://www.nexperia.com/

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index 2ca68ee8602cf..fd2d25d7fa51b 100644
@@ -802,6 +802,8 @@ patternProperties:
     description: Guangdong Neweast Optoelectronics CO., LTD
   "^nexbox,.*":
     description: Nexbox
+  "^nexperia,.*":
+    description: Nexperia B.V.
   "^nextthing,.*":
     description: Next Thing Co.
   "^newhaven,.*":
Signature · 1 line
2.53.0
author maintainer
#1547 lore raw

PATCH 1544/2214 pstore: dont leave an ERR_PTR in the ramoops zone pointer

In reply to Greg Kroah-Hartman

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

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

From: Andrea Parri <[email protected]>

[ Upstream commit 70fec428c845216da98041b1ae9f16a8f54b3d68 ]

ramoops_init_prz() stores the result of persistent_ram_new() in *prz and returns the error without clearing it when initialization fails. ramoops_probe() then cleans up through ramoops_free_przs(), which hands *prz to persistent_ram_free(); that treats the ERR_PTR as a valid zone and dereferences it, crashing the kernel.

Clear *prz before returning so that cleanup sees a NULL zone.

Fixes b5d38e9bf1b0c ("pstore/ram: Add console messages handling")
Assisted-by LLM
---
 fs/pstore/ram.c | 1 +
 1 file changed, 1 insertion(+)
fs/pstore/ram.c +1 -0
index e15b4631364a8..1e28b1fdcde16 100644
@@ -590,6 +590,7 @@ static int ramoops_init_prz(const char *name,
 	if (IS_ERR(*prz)) {
 		int err = PTR_ERR(*prz);
 
+		*prz = NULL;
 		dev_err(dev, "failed to request %s mem region (0x%zx@0x%llx): %d\n",
 			name, sz, (unsigned long long)*paddr, err);
 		return err;
Signature · 1 line
2.53.0
author maintainer
#1548 lore raw

PATCH 1545/2214 pstore: reject persistent RAM zones too small for the buffer header

In reply to Greg Kroah-Hartman

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

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

From: Andrea Parri <[email protected]>

[ Upstream commit 1b30df8662ac85a6759be4332f3bec7e3cb28efc ]

persistent_ram_buffer_map() computes

	prz->buffer_size = size - sizeof(struct persistent_ram_buffer);

without checking that size is large enough to hold the header. A ramoops zone configured with a tiny size (e.g. ramoops.pmsg_size=1) underflows buffer_size to a huge value, which defeats every bounds check in persistent_ram_write() and persistent_ram_write_user(); a write to /dev/pmsg0 then copies user data far past the end of the zone and corrupts kernel memory.

Reject zones that are not larger than the buffer header.

Fixes 404a6043385de ("staging: android: persistent_ram: handle reserving and mapping memory")
Assisted-by LLM
---
 fs/pstore/ram_core.c | 6 ++++++
 1 file changed, 6 insertions(+)
fs/pstore/ram_core.c +6 -0
index 8a86b9928503d..2d89725929d27 100644
@@ -507,6 +507,12 @@ static int persistent_ram_buffer_map(phys_addr_t start, phys_addr_t size,
 	prz->paddr = start;
 	prz->size = size;
 
+	if (size <= sizeof(struct persistent_ram_buffer)) {
+		pr_err("%s: persistent ram zone too small (%#llx bytes)\n",
+		       prz->label, (unsigned long long)size);
+		return -EINVAL;
+	}
+
 	if (pfn_valid(start >> PAGE_SHIFT))
 		prz->vaddr = persistent_ram_vmap(start, size, memtype);
 	else
Signature · 1 line
2.53.0
author maintainer
#1549 lore raw

PATCH 1546/2214 leds: pca9532: Check i2c_smbus_read_byte_data() return value

In reply to Greg Kroah-Hartman

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

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

From: Haibo Chen <[email protected]>

[ Upstream commit 1b75f038e5f71ed32ce6593b5763f9776ec58a25 ]

Coverity report INTEGER_OVERFLOW (tainted_data_return -> cast_underflow -> overflow) in pca9532_getled(). pca9532_getled() stored the return value of i2c_smbus_read_byte_data() into a char and used it directly to compute the LED state without checking for an error. i2c_smbus_read_byte_data() returns a negative error code on failure. On an I2C read error the negative value was assigned to an unsigned type, triggering an integer underflow/overflow, and a meaningless LED state was returned to the caller.

Read the value into an int, check for the error, and fall back to PCA9532_OFF while emitting a warning when the read fails.

Fixes 28c5fe99016d ("leds: pca9532: Extend pca9532 device tree support")
Assisted-by VeroCoder:vero-1 [cli]
---
 drivers/leds/leds-pca9532.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
drivers/leds/leds-pca9532.c +6 -3
index b7d0aeeecfcf8..977b4419df34c 100644
@@ -259,12 +259,15 @@ static enum pca9532_state pca9532_getled(struct pca9532_led *led)
 	struct i2c_client *client = led->client;
 	struct pca9532_data *data = i2c_get_clientdata(client);
 	u8 maxleds = data->chip_info->num_leds;
-	char reg;
-	enum pca9532_state ret;
+	int reg;
+	enum pca9532_state ret = PCA9532_OFF;
 
 	mutex_lock(&data->update_lock);
 	reg = i2c_smbus_read_byte_data(client, LED_REG(maxleds, led->id));
-	ret = (reg & LED_MASK(led->id)) >> LED_SHIFT(led->id);
+	if (reg >= 0)
+		ret = (reg & LED_MASK(led->id)) >> LED_SHIFT(led->id);
+	else
+		dev_warn(&client->dev, "failed to read LED register: %d\n", reg);
 	mutex_unlock(&data->update_lock);
 	return ret;
 }
Signature · 1 line
2.53.0
author maintainer
#1550 lore raw

PATCH 1547/2214 leds: lm3692x: Fix regulator/gpio leak and wrong enabled state on last write failure

In reply to Greg Kroah-Hartman

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

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

From: Cong Nguyen <[email protected]>

[ Upstream commit 145ca6bca2cc5fd8d9c7af216df58a9c75c0e84c ]

Every write in lm3692x_leds_enable() checks its return and jumps to
out: on failure -- except the last one, which falls through and sets led->enabled = true regardless of whether it actually succeeded. Leaks the regulator on a failing probe, and short-circuits later brightness_set() calls without retrying.

Route it through out: like every other write here, and only set enabled = true once it succeeds.

Fixes 260718b3a35d ("leds: lm3692x: Disable chip on brightness 0")
Assisted-by Claude:claude-opus-4
---
 drivers/leds/leds-lm3692x.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/leds/leds-lm3692x.c +3 -1
index afe6fb297855f..d67b58f0ea018 100644
@@ -274,9 +274,11 @@ static int lm3692x_leds_enable(struct lm3692x_led *led)
 
 	ret = regmap_update_bits(led->regmap, LM3692X_EN, LM3692X_ENABLE_MASK,
 				 enable_state | LM3692X_DEVICE_EN);
+	if (ret)
+		goto out;
 
 	led->enabled = true;
-	return ret;
+	return 0;
 out:
 	dev_err(&led->client->dev, "Fail writing initialization values\n");
 
Signature · 1 line
2.53.0
author maintainer
#1551 lore raw

PATCH 1548/2214 RDMA/core: Fix partial copy of IPv6 flow_lbl in LAG hash skb

In reply to Greg Kroah-Hartman

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

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

From: Li RongQing <[email protected]>

[ Upstream commit b63d21631f836528ef660d7b5a0d84890ff6dbd5 ]

rdma_build_skb() constructs a synthetic IPv6 header for LAG slave selection and copies the flow label from the AH attribute:

	memcpy(&ip6h->flow_lbl, &ah_attr->grh.flow_label,
	       sizeof(*ip6h->flow_lbl));

ipv6hdr.flow_lbl is __u8[3], so sizeof(*ip6h->flow_lbl) dereferences the array to a single __u8 and yields 1. The memcpy therefore copies only the first byte of the flow label, leaving flow_lbl[1] and flow_lbl[2] uninitialised in the skb buffer (alloc_skb() does not zero the data). The resulting LAG hash is computed over one byte of real flow-label data and two bytes of kmalloc residue, so IPv6 RoCEv2 traffic on a bonded interface can be steered to the wrong slave.

Use sizeof(ip6h->flow_lbl) (the array, 3 bytes) so the whole flow label is copied.

Fixes bd3920eac1331 ("RDMA/core: Add LAG functionality")
---
 drivers/infiniband/core/lag.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/infiniband/core/lag.c +1 -1
index 7063e41eaf261..fc1dd6b5bf5ef 100644
@@ -60,7 +60,7 @@ static struct sk_buff *rdma_build_skb(struct ib_device *device,
 		ip6h->version = 6;
 		ip6h->nexthdr = IPPROTO_UDP;
 		memcpy(&ip6h->flow_lbl, &ah_attr->grh.flow_label,
-		       sizeof(*ip6h->flow_lbl));
+		       sizeof(ip6h->flow_lbl));
 		memcpy(&ip6h->saddr, ah_attr->grh.sgid_attr->gid.raw,
 		       sizeof(struct in6_addr));
 		memcpy(&ip6h->daddr, ah_attr->grh.dgid.raw,
Signature · 1 line
2.53.0
author maintainer
#1552 lore raw

PATCH 1549/2214 leds: flash: Prevent using uninitialized variable in flash_fault_show()

In reply to Greg Kroah-Hartman

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

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

From: Dmitry Antipov <[email protected]>

[ Upstream commit d229695f2162a41bfa614f175798fa56a5f1c969 ]

In flash_fault_show(), address of uninitialized 'fault' variable is passed to device-specific .fault_get callbacks. This is not a problem if such a callback directly assigns the value like rt4505_fault_get(), but results in an undefined behavior if callback just performs bitwise ORs like as3645a_get_fault(). So just initialize 'fault' with zero.

---
 drivers/leds/led-class-flash.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/leds/led-class-flash.c +1 -1
index 185e17055317c..56a29ca6f4a0a 100644
@@ -188,7 +188,7 @@ static ssize_t flash_fault_show(struct device *dev,
 {
 	struct led_classdev *led_cdev = dev_get_drvdata(dev);
 	struct led_classdev_flash *fled_cdev = lcdev_to_flcdev(led_cdev);
-	u32 fault, mask = 0x1;
+	u32 fault = 0, mask = 0x1;
 	char *pbuf = buf;
 	int i, ret, buf_len;
 
Signature · 1 line
2.53.0
author maintainer
#1553 lore raw

PATCH 1550/2214 regulator: mp886x: check get_voltage_sel() return value

In reply to Greg Kroah-Hartman

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

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

From: Jisheng Zhang <[email protected]>

[ Upstream commit 87121c42fadc7683469aee96ad833bc560197a55 ]

The get_voltage_sel() can fail, resulting in negative error codes stored in unsigned integers and corrupted state logic.

---
 drivers/regulator/mp886x.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
drivers/regulator/mp886x.c +7 -1
index 7910896df51a5..f42572d0f94e5 100644
@@ -256,6 +256,7 @@ static int mp886x_regulator_register(struct mp886x_device_info *di,
 {
 	struct regulator_desc *rdesc = &di->desc;
 	struct regulator_dev *rdev;
+	int sel;
 
 	rdesc->name = "mp886x-reg";
 	rdesc->supply_name = "vin";
@@ -277,7 +278,12 @@ static int mp886x_regulator_register(struct mp886x_device_info *di,
 	rdev = devm_regulator_register(di->dev, &di->desc, config);
 	if (IS_ERR(rdev))
 		return PTR_ERR(rdev);
-	di->sel = rdesc->ops->get_voltage_sel(rdev);
+
+	sel = rdesc->ops->get_voltage_sel(rdev);
+	if (sel < 0)
+		return sel;
+	di->sel = sel;
+
 	return 0;
 }
 
Signature · 1 line
2.53.0
author maintainer
#1554 lore raw

PATCH 1551/2214 regulator: mp886x: fix pontential division by zero

In reply to Greg Kroah-Hartman

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

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

From: Jisheng Zhang <[email protected]>

[ Upstream commit d16ef5debccc4289c1f25309a2cccfbbe879601a ]

If the "mps,fb-voltage-divider" 2nd var is 0, there will be division by zero bug.

---
 drivers/regulator/mp886x.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/regulator/mp886x.c +3 -0
index f42572d0f94e5..ae45a8dbdb27b 100644
@@ -317,6 +317,9 @@ static int mp886x_i2c_probe(struct i2c_client *client)
 	if (ret)
 		return ret;
 
+	if (di->r[1] == 0)
+		return -EINVAL;
+
 	di->en_gpio = devm_gpiod_get(dev, "enable", GPIOD_OUT_HIGH);
 	if (IS_ERR(di->en_gpio))
 		return PTR_ERR(di->en_gpio);
Signature · 1 line
2.53.0
author maintainer
#1555 lore raw

PATCH 1552/2214 RDMA/irdma: Use kvzalloc for lvl2 leafmem allocation

In reply to Greg Kroah-Hartman

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

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

From: Jeremy Bongio <[email protected]>

[ Upstream commit 9c40f13a63a4b87177de3e8c318af25068fc6ab4 ]

During VM live migration or when registering large memory regions (e.g., guest memory >= 256 GB backed by 4 KiB pages), get_lvl2_pble() calculates total leaves exceeding ~131,072. The memory required for lvl2->leafmem.va exceeds 4 MiB (sizeof(*leaf) * total > 5.24 MiB).

Because kzalloc() requires physically contiguous pages, requests larger than MAX_PAGE_ORDER (4 MiB on x86_64) fail and trigger a warning:

  WARNING: CPU: 123 PID: 119906 at mm/page_alloc.c:5997 __alloc_frozen_pages_noprof+0x50c/0x1050
  Call Trace:
   <TASK>
   alloc_frozen_pages_noprof+0xc0/0x240
   ___kmalloc_large_node+0x3d/0xd0
   __kmalloc_large_node_noprof+0x1b/0x90
   __kmalloc_noprof+0x254/0x660
   get_lvl1_lvl2_pble+0xbc/0x270 [irdma]
   irdma_get_pble+0x44/0xc0 [irdma]
   irdma_setup_pbles+0x46/0x1d0 [irdma]
   irdma_reg_user_mr_type_mem+0x41/0x230 [irdma]
   irdma_reg_user_mr+0x194/0x1c0 [irdma]
   ib_uverbs_reg_mr+0x1b0/0x2c0 [ib_uverbs]

The leafmem array is purely host software tracking structures (struct irdma_pble_info) and does not require physically contiguous memory.

Switch the allocation from kzalloc() to kvzalloc() and free it with kvfree() so that requests exceeding MAX_PAGE_ORDER fall back to virtually contiguous memory via vmalloc.

Fixes e8c4dbc2fcac ("RDMA/irdma: Add PBLE resource manager")
Assisted-by LLM
---
 drivers/infiniband/hw/irdma/pble.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/infiniband/hw/irdma/pble.c +3 -3
index f4d5d1cee681f..965b84c0bc1ec 100644
@@ -328,7 +328,7 @@ static void free_lvl2(struct irdma_hmc_pble_rsrc *pble_rsrc,
 	if (root->addr)
 		irdma_prm_return_pbles(&pble_rsrc->pinfo, &root->chunkinfo);
 
-	kfree(lvl2->leafmem.va);
+	kvfree(lvl2->leafmem.va);
 	lvl2->leaf = NULL;
 }
 
@@ -357,7 +357,7 @@ get_lvl2_pble(struct irdma_hmc_pble_rsrc *pble_rsrc,
 	lvl2->leaf_cnt = total;
 
 	lvl2->leafmem.size = (sizeof(*leaf) * total);
-	lvl2->leafmem.va = kzalloc(lvl2->leafmem.size, GFP_KERNEL);
+	lvl2->leafmem.va = kvzalloc(lvl2->leafmem.size, GFP_KERNEL);
 	if (!lvl2->leafmem.va)
 		return IRDMA_ERR_NO_MEMORY;
 
@@ -366,7 +366,7 @@ get_lvl2_pble(struct irdma_hmc_pble_rsrc *pble_rsrc,
 	ret_code = irdma_prm_get_pbles(&pble_rsrc->pinfo, &root->chunkinfo,
 				       total << 3, &root->addr, &fpm_addr);
 	if (ret_code) {
-		kfree(lvl2->leafmem.va);
+		kvfree(lvl2->leafmem.va);
 		lvl2->leaf = NULL;
 		return IRDMA_ERR_NO_MEMORY;
 	}
Signature · 1 line
2.53.0
author maintainer
#1556 lore raw

PATCH 1553/2214 wifi: rtl8xxxu: Use dev_* instead of pr_info

In reply to Greg Kroah-Hartman

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

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

From: Bitterblue Smith <[email protected]>

[ Upstream commit bd954a7e4aa0739630884fc572d2d0e4dcd16e25 ]

Replace two instances of bare pr_info with dev_info and dev_warn.

Also make their messages a little more informative.

Stable-dep-of: 9f7edf4e8584 ("wifi: rtl8xxxu: free RX skb when URB submission fails")

---
 drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c +4 -2
index 526491b1ec082..d95ac983177af 100644
@@ -1996,7 +1996,8 @@ static int rtl8xxxu_download_firmware(struct rtl8xxxu_priv *priv)
 
 	val8 = rtl8xxxu_read8(priv, REG_MCU_FW_DL);
 	if (val8 & MCU_FW_RAM_SEL) {
-		pr_info("do the RAM reset\n");
+		dev_info(&priv->udev->dev,
+			 "Firmware is already running, resetting the MCU.\n");
 		rtl8xxxu_write8(priv, REG_MCU_FW_DL, 0x00);
 		priv->fops->reset_8051(priv);
 	}
@@ -5228,7 +5229,8 @@ static void rtl8xxxu_rx_urb_work(struct work_struct *work)
 			rtl8xxxu_queue_rx_urb(priv, rx_urb);
 			break;
 		default:
-			pr_info("failed to requeue urb %i\n", ret);
+			dev_warn(&priv->udev->dev,
+				 "failed to requeue urb with error %i\n", ret);
 			skb = (struct sk_buff *)rx_urb->urb.context;
 			dev_kfree_skb(skb);
 			usb_free_urb(&rx_urb->urb);
Signature · 1 line
2.53.0
author maintainer
#1557 lore raw

PATCH 1554/2214 wifi: rtl8xxxu: free RX skb when URB submission fails

In reply to Greg Kroah-Hartman

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

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

From: Kim Wooseok <[email protected]>

[ Upstream commit 9f7edf4e8584c87e3c478c6dd87c08d51b5157c2 ]

When usb_submit_urb() fails, rtl8xxxu_submit_rx_urb() unanchors the URB but leaves the newly allocated skb in urb.context. For ENOMEM/EAGAIN, the RX worker puts the request back on the pending list. The next submission allocates another skb and overwrites that pointer, leaking the previous buffer. Stopping before the retry also leaks it, because pending-list cleanup frees only the URB.

Free the skb and clear urb.context in rtl8xxxu_submit_rx_urb() when submission fails. This keeps buffer allocation and failure cleanup in the same function, so a request returned for retry or teardown no longer owns an skb. Remove the corresponding cleanup from start and the RX worker; they only need to decide whether to retry or free the URB.

Fixes 26f1fad29ad9 ("New driver: rtl8xxxu (mac80211)")
Assisted-by GPT-6 Astra
---
 .../wireless/realtek/rtl8xxxu/rtl8xxxu_core.c    | 16 +++++-----------
 1 file changed, 5 insertions(+), 11 deletions(-)
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_core.c +5 -11
index d95ac983177af..841b3d7dd6a7b 100644
@@ -5199,7 +5199,6 @@ static void rtl8xxxu_rx_urb_work(struct work_struct *work)
 	struct rtl8xxxu_priv *priv;
 	struct rtl8xxxu_rx_urb *rx_urb, *tmp;
 	struct list_head local;
-	struct sk_buff *skb;
 	unsigned long flags;
 	int ret;
 
@@ -5231,8 +5230,6 @@ static void rtl8xxxu_rx_urb_work(struct work_struct *work)
 		default:
 			dev_warn(&priv->udev->dev,
 				 "failed to requeue urb with error %i\n", ret);
-			skb = (struct sk_buff *)rx_urb->urb.context;
-			dev_kfree_skb(skb);
 			usb_free_urb(&rx_urb->urb);
 		}
 	}
@@ -5811,8 +5808,11 @@ static int rtl8xxxu_submit_rx_urb(struct rtl8xxxu_priv *priv,
 			  skb_size, rtl8xxxu_rx_complete, skb);
 	usb_anchor_urb(&rx_urb->urb, &priv->rx_anchor);
 	ret = usb_submit_urb(&rx_urb->urb, GFP_ATOMIC);
-	if (ret)
+	if (ret) {
 		usb_unanchor_urb(&rx_urb->urb);
+		dev_kfree_skb(skb);
+		rx_urb->urb.context = NULL;
+	}
 	return ret;
 }
 
@@ -6372,7 +6372,6 @@ static int rtl8xxxu_start(struct ieee80211_hw *hw)
 	struct rtl8xxxu_priv *priv = hw->priv;
 	struct rtl8xxxu_rx_urb *rx_urb;
 	struct rtl8xxxu_tx_urb *tx_urb;
-	struct sk_buff *skb;
 	unsigned long flags;
 	int ret, i;
 
@@ -6423,13 +6422,8 @@ static int rtl8xxxu_start(struct ieee80211_hw *hw)
 		rx_urb->hw = hw;
 
 		ret = rtl8xxxu_submit_rx_urb(priv, rx_urb);
-		if (ret) {
-			if (ret != -ENOMEM) {
-				skb = (struct sk_buff *)rx_urb->urb.context;
-				dev_kfree_skb(skb);
-			}
+		if (ret)
 			rtl8xxxu_queue_rx_urb(priv, rx_urb);
-		}
 	}
 
 	schedule_delayed_work(&priv->ra_watchdog, 2 * HZ);
Signature · 1 line
2.53.0
author maintainer
#1558 lore raw

PATCH 1555/2214 RDMA/cxgb4: Fix neigh reference leak in rx_pkt()

In reply to Greg Kroah-Hartman

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

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

From: Wentao Liang <[email protected]>

[ Upstream commit 5fb3051aa4611ae18e304814fb1b74a30065fd85 ]

If ip_dev_find() fails on the loopback path the code jumps to free_dst without releasing the neighbour reference taken by dst_neigh_lookup_skb(), leaking one neigh reference.

Release the neighbour before jumping to free_dst.

Fixes ef42520240aa ("RDMA/cxgb4: add null-ptr-check after ip_dev_find()")
---
 drivers/infiniband/hw/cxgb4/cm.c | 1 +
 1 file changed, 1 insertion(+)
drivers/infiniband/hw/cxgb4/cm.c +1 -0
index 3efd06d5f7e70..2d651fce63f92 100644
@@ -4162,6 +4162,7 @@ static int rx_pkt(struct c4iw_dev *dev, struct sk_buff *skb)
 		pdev = ip_dev_find(&init_net, iph->daddr);
 		if (!pdev) {
 			pr_err("%s - failed to find device!\n", __func__);
+			neigh_release(neigh);
 			goto free_dst;
 		}
 		e = cxgb4_l2t_get(dev->rdev.lldi.l2t, neigh,
Signature · 1 line
2.53.0
author maintainer
#1559 lore raw

PATCH 1556/2214 RDMA/cxgb4: Fix skb reference leak in rx_pkt()

In reply to Greg Kroah-Hartman

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

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

From: Wentao Liang <[email protected]>

[ Upstream commit 5eb6ecb612841927b5ede554a0e9d88240934282 ]

rx_pkt() takes an extra reference on the received skb so it can be handed to the firmware completion path via req->cookie. When the work request skb allocation in send_fw_pass_open_req() fails the function returns without releasing that extra reference, leaking one skb reference.

Free the packet skb before returning when the allocation fails, matching the cxgb4_ofld_send() failure path.

Fixes 1cab775c3e75 ("RDMA/cxgb4: Fix LE hash collision bug for passive open connection")
---
 drivers/infiniband/hw/cxgb4/cm.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/infiniband/hw/cxgb4/cm.c +3 -1
index 2d651fce63f92..e5f0a8b779ed7 100644
@@ -4012,8 +4012,10 @@ static void send_fw_pass_open_req(struct c4iw_dev *dev, struct sk_buff *skb,
 	int ret;
 
 	req_skb = alloc_skb(sizeof(struct fw_ofld_connection_wr), GFP_KERNEL);
-	if (!req_skb)
+	if (!req_skb) {
+		kfree_skb(skb);
 		return;
+	}
 	req = __skb_put_zero(req_skb, sizeof(*req));
 	req->op_compl = htonl(WR_OP_V(FW_OFLD_CONNECTION_WR) | FW_WR_COMPL_F);
 	req->len16_pkd = htonl(FW_WR_LEN16_V(DIV_ROUND_UP(sizeof(*req), 16)));
Signature · 1 line
2.53.0
author maintainer
#1560 lore raw

PATCH 1557/2214 RDMA/mlx5: Fix RQ resource reference leak in mlx5_ib_wq_event()

In reply to Greg Kroah-Hartman

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

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

From: Wentao Liang <[email protected]>

[ Upstream commit 65f3eab431944415bbe5e3af1bbb2736d4bb1523 ]

Since commit 312b8f79eb05 ("RDMA/mlx: Calling qp event handler in workqueue context") rsc_event_notifier() expects the event handler to drop the resource reference taken by mlx5_get_rsc(). mlx5_ib_wq_event() was missed by that conversion, so every RQ event leaks one reference, which keeps destroy_resource_common() waiting on the completion.

Put the resource reference on all exits of the handler.

Fixes 312b8f79eb05 ("RDMA/mlx: Calling qp event handler in workqueue context")
---
 drivers/infiniband/hw/mlx5/qp.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/infiniband/hw/mlx5/qp.c +4 -1
index 3df6fb5edbf34..8caded8c3b324 100644
@@ -5121,11 +5121,14 @@ static void mlx5_ib_wq_event(struct mlx5_core_qp *core_qp, int type)
 			break;
 		default:
 			mlx5_ib_warn(dev, "Unexpected event type %d on WQ %06x\n", type, core_qp->qpn);
-			return;
+			goto out;
 		}
 
 		rwq->ibwq.event_handler(&event, rwq->ibwq.wq_context);
 	}
+
+out:
+	mlx5_core_res_put(&core_qp->common);
 }
 
 static int set_delay_drop(struct mlx5_ib_dev *dev)
Signature · 1 line
2.53.0
author maintainer
#1561 lore raw

PATCH 1558/2214 RDMA/siw: Fix qp use-after-free in siw_get_base_qp()

In reply to Greg Kroah-Hartman

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

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

From: Wentao Liang <[email protected]>

[ Upstream commit ef15197f81ffedf29560edf2221199d92fd83dc2 ]

The reference taken by siw_qp_id2obj() is dropped before the address of the embedded ib_qp is taken from the QP. If that reference was the last one the QP is freed and the pointer returned to the iwarp core is dangling; the core only pins the QP with iw_add_ref() after the call has returned.

Copy the pointer while the reference is still held.

Fixes 6c52fdc244b5 ("rdma/siw: connection management")
---
 drivers/infiniband/sw/siw/siw_main.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)
drivers/infiniband/sw/siw/siw_main.c +4 -1
index f853f3c23540a..dd4c2dcd377c2 100644
@@ -244,8 +244,11 @@ static struct ib_qp *siw_get_base_qp(struct ib_device *base_dev, int id)
 		/*
 		 * siw_qp_id2obj() increments object reference count
 		 */
+		struct ib_qp *base_qp = &qp->base_qp;
+
 		siw_qp_put(qp);
-		return &qp->base_qp;
+
+		return base_qp;
 	}
 	return NULL;
 }
Signature · 1 line
2.53.0
author maintainer
#1562 lore raw

PATCH 1559/2214 crypto: hisilicon/hpre - remove dma_free_coherent on NULL req->dst in ECDH

In reply to Greg Kroah-Hartman

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

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

From: Weili Qian <[email protected]>

[ Upstream commit 5b87f619cc5c0a55dcd85ce97974e3f5b9333b7b ]

In hpre_ecdh_hw_data_clr_all(), req->dst is always NULL because hpre_ecdh_dst_data_init() sets hpre_req->dst = NULL before mapping the user buffer with dma_map_single(). The dma_free_coherent() call on req->dst is therefore dead code. Remove it.

Fixes 05e7b906aa7c ("crypto: hisilicon/hpre - add 'ECDH' algorithm")
---
 drivers/crypto/hisilicon/hpre/hpre_crypto.c | 2 --
 1 file changed, 2 deletions(-)
drivers/crypto/hisilicon/hpre/hpre_crypto.c +0 -2
index d77056178d7ff..204d559a9be54 100644
@@ -1438,8 +1438,6 @@ static void hpre_ecdh_hw_data_clr_all(struct hpre_ctx *ctx,
 	if (unlikely(dma_mapping_error(dev, dma)))
 		return;
 
-	if (req->dst)
-		dma_free_coherent(dev, ctx->key_sz << 1, req->dst, dma);
 	if (dst)
 		dma_unmap_single(dev, dma, ctx->key_sz << 1, DMA_FROM_DEVICE);
 }
Signature · 1 line
2.53.0
author maintainer
#1563 lore raw

PATCH 1560/2214 crypto: ccp - kill v3 IRQ tasklet during teardown

In reply to Greg Kroah-Hartman

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

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

From: Jiale Yao <[email protected]>

[ Upstream commit b72a639bdee6e07d8091718ac3e451d9c5cba1b5 ]

On PCI devices using MSI-X, ccp_irq_handler() schedules irq_tasklet. The teardown paths release the IRQ but do not drain a tasklet that was already queued, so it can access ccp after the device-managed allocation has been released.

Kill the tasklet after releasing the IRQ in both the initialization error path and ccp_destroy().

Fixes 7b537b24e76a ("crypto: ccp - Change ISR handler method for a v3 CCP")
---
 drivers/crypto/ccp/ccp-dev-v3.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/crypto/ccp/ccp-dev-v3.c +4 -0
index 0d5576f6ad216..b3d12f46aedf2 100644
@@ -507,6 +507,8 @@ static int ccp_init(struct ccp_device *ccp)
 			kthread_stop(ccp->cmd_q[i].kthread);
 
 	sp_free_ccp_irq(ccp->sp, ccp);
+	if (ccp->use_tasklet)
+		tasklet_kill(&ccp->irq_tasklet);
 
 e_pool:
 	for (i = 0; i < ccp->cmd_q_count; i++)
@@ -546,6 +548,8 @@ static void ccp_destroy(struct ccp_device *ccp)
 			kthread_stop(ccp->cmd_q[i].kthread);
 
 	sp_free_ccp_irq(ccp->sp, ccp);
+	if (ccp->use_tasklet)
+		tasklet_kill(&ccp->irq_tasklet);
 
 	for (i = 0; i < ccp->cmd_q_count; i++)
 		dma_pool_destroy(ccp->cmd_q[i].dma_pool);
Signature · 1 line
2.53.0
author maintainer
#1564 lore raw

PATCH 1561/2214 crypto: ccp - kill v5 IRQ tasklet during teardown

In reply to Greg Kroah-Hartman

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

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

From: Jiale Yao <[email protected]>

[ Upstream commit 67aefeccc4101a93627f36abbd049f191b54f903 ]

On PCI devices using MSI-X, ccp5_irq_handler() schedules irq_tasklet. The teardown paths release the IRQ but do not drain a tasklet that was already queued, so it can access ccp after the device-managed allocation has been released.

Kill the tasklet after releasing the IRQ in both the initialization error path and ccp5_destroy().

Fixes 6263b51eb319 ("crypto: ccp - Change ISR handler method for a v5 CCP")
---
 drivers/crypto/ccp/ccp-dev-v5.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/crypto/ccp/ccp-dev-v5.c +4 -0
index 7838f63bab32c..e1966f43b9f7c 100644
@@ -996,6 +996,8 @@ static int ccp5_init(struct ccp_device *ccp)
 
 e_irq:
 	sp_free_ccp_irq(ccp->sp, ccp);
+	if (ccp->use_tasklet)
+		tasklet_kill(&ccp->irq_tasklet);
 
 e_pool:
 	for (i = 0; i < ccp->cmd_q_count; i++)
@@ -1047,6 +1049,8 @@ static void ccp5_destroy(struct ccp_device *ccp)
 			kthread_stop(ccp->cmd_q[i].kthread);
 
 	sp_free_ccp_irq(ccp->sp, ccp);
+	if (ccp->use_tasklet)
+		tasklet_kill(&ccp->irq_tasklet);
 
 	/* Flush the cmd and backlog queue */
 	while (!list_empty(&ccp->cmd)) {
Signature · 1 line
2.53.0
author maintainer
#1565 lore raw

PATCH 1562/2214 swim: Assert strobe with stable outputs

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit adcc410bfb41715fef198a0975db4ed594a49bb7 ]

Inside Macintosh says, "Be sure that you don't change CA0-CA2 or SEL while LSTRB is high". Unfortunately, those bits do change. The CA0-CA2 outputs become inputs when (LSTRB << 4) | LSTRB gets written to the phase register. Then LSTRB is driven low (with CA0-CA2 bits set).

This is a problem because the drive interprets a STEP command as an EJECT command when these pins float high. This occurs intermittently, perhaps because interrupts are disabled and the race condition happens to end well. However, when I add code to step the heads with interrupts enabled, the disk always ejects.

Keep the four phase pin directions set to output and hold their levels constant during LSTRB signalling. Introduce the PHASE_PIN_DIRECTION macro to separate the pin configuration from the logic level changes.

---
 drivers/block/swim.c | 62 ++++++++++++++++++++++----------------------
 1 file changed, 31 insertions(+), 31 deletions(-)
drivers/block/swim.c +31 -31
index 3911d0833e1b9..c4fca0ed99345 100644
@@ -87,37 +87,37 @@ struct iwm {
 #define iwm_write(base, reg, v) 	out_8(&(base)->reg, (v))
 #define iwm_read(base, reg)		in_8(&(base)->reg)
 
-/* bits in phase register */
+/* Bits in phase register */
 
-#define SEEK_POSITIVE	0x070
-#define SEEK_NEGATIVE	0x074
-#define STEP		0x071
-#define MOTOR_ON	0x072
-#define MOTOR_OFF	0x076
-#define INDEX		0x073
-#define EJECT		0x077
-#define SETMFM		0x171
-#define SETGCR		0x175
-
-#define RELAX		0x033
-#define LSTRB		0x008
-
-#define CA_MASK		0x077
+#define RELAX		0x03
+#define LSTRB		0x08
+#define CA_MASK		0x07
+#define PHASE_PIN_DIR	0xF0
 
 /* Select values for swim_select and swim_readbit */
 
-#define READ_DATA_0	0x074
-#define ONEMEG_DRIVE	0x075
-#define SINGLE_SIDED	0x076
-#define DRIVE_PRESENT	0x077
-#define DISK_IN		0x170
-#define WRITE_PROT	0x171
-#define TRACK_ZERO	0x172
-#define TACHO		0x173
-#define READ_DATA_1	0x174
-#define GCR_MODE	0x175
-#define SEEK_COMPLETE	0x176
-#define TWOMEG_MEDIA	0x177
+#define SEEK_POSITIVE	0x000
+#define SEEK_NEGATIVE	0x004
+#define STEP		0x001
+#define MOTOR_ON	0x002
+#define MOTOR_OFF	0x006
+#define INDEX		0x003
+#define EJECT		0x007
+#define SETMFM		0x101
+#define SETGCR		0x105
+
+#define READ_DATA_0	0x004
+#define ONEMEG_DRIVE	0x005
+#define SINGLE_SIDED	0x006
+#define DRIVE_PRESENT	0x007
+#define DISK_IN		0x100
+#define WRITE_PROT	0x101
+#define TRACK_ZERO	0x102
+#define TACHO		0x103
+#define READ_DATA_1	0x104
+#define GCR_MODE	0x105
+#define SEEK_COMPLETE	0x106
+#define TWOMEG_MEDIA	0x107
 
 /* Bits in handshake register */
 
@@ -268,11 +268,11 @@ static inline int get_swim_mode(struct swim __iomem *base)
 
 static inline void swim_select(struct swim __iomem *base, int sel)
 {
-	swim_write(base, phase, RELAX);
+	swim_write(base, phase, RELAX | PHASE_PIN_DIR);
 
 	via1_set_head(sel & 0x100);
 
-	swim_write(base, phase, sel & CA_MASK);
+	swim_write(base, phase, (sel & CA_MASK) | PHASE_PIN_DIR);
 }
 
 static inline void swim_action(struct swim __iomem *base, int action)
@@ -283,9 +283,9 @@ static inline void swim_action(struct swim __iomem *base, int action)
 
 	swim_select(base, action);
 	udelay(1);
-	swim_write(base, phase, (LSTRB<<4) | LSTRB);
+	swim_write(base, phase, LSTRB | action | PHASE_PIN_DIR);
 	udelay(1);
-	swim_write(base, phase, (LSTRB<<4) | ((~LSTRB) & 0x0F));
+	swim_write(base, phase, action | PHASE_PIN_DIR);
 	udelay(1);
 
 	local_irq_restore(flags);
Signature · 1 line
2.53.0
author maintainer
#1566 lore raw

PATCH 1563/2214 swim: Select appropriate drive once only

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit a2d22b3f25c388f9fbce7bf36a45a8900dce4aa3 ]

Before turning off the spindle motor, call swim_drive() to select the appropriate drive. Remove the swim_drive() call from swim_add_floppy() because it was already called by swim_floppy_init().

---
 drivers/block/swim.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/block/swim.c +3 -3
index c4fca0ed99345..77fe8e81dbb17 100644
@@ -680,8 +680,10 @@ static void floppy_release(struct gendisk *disk, fmode_t mode)
 	else if (fs->ref_count > 0)
 		--fs->ref_count;
 
-	if (fs->ref_count == 0)
+	if (fs->ref_count == 0) {
+		swim_drive(base, fs->location);
 		swim_motor(base, OFF);
+	}
 	mutex_unlock(&swim_mutex);
 }
 
@@ -753,8 +755,6 @@ static int swim_add_floppy(struct swim_priv *swd, enum drive_location location)
 
 	fs->location = location;
 
-	swim_drive(base, location);
-
 	swim_motor(base, OFF);
 
 	fs->type = HD_MEDIA;
Signature · 1 line
2.53.0
author maintainer
#1567 lore raw

PATCH 1564/2214 swim: Enable the drive when probing

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 57d88c8f4ea34a4dd7c5351bd026e225c7283f3f ]

According to the ISM spec, the MOTON bit "causes the Enable 1 and Enable 2 signals to be turned on to the drive." It doesn't actually turn on the motor. When selecting a drive, enable MOTON. Disable it upon eject or release. This fixes detection of the Sony FDHD drive.

---
 drivers/block/swim.c | 18 +++++++++++++++---
 1 file changed, 15 insertions(+), 3 deletions(-)
drivers/block/swim.c +15 -3
index 77fe8e81dbb17..f956a2e40f70b 100644
@@ -154,8 +154,9 @@ struct iwm {
 /*----------------------------------------------------------------------------*/
 
 enum drive_location {
-	INTERNAL_DRIVE = 0x02,
-	EXTERNAL_DRIVE = 0x04,
+	NO_DRIVE = 0,
+	INTERNAL_DRIVE = BIT(1),
+	EXTERNAL_DRIVE = BIT(2),
 };
 
 enum media_type {
@@ -310,9 +311,15 @@ static inline void swim_drive(struct swim __iomem *base,
 	if (location == INTERNAL_DRIVE) {
 		swim_write(base, mode0, EXTERNAL_DRIVE); /* clear drive 1 bit */
 		swim_write(base, mode1, INTERNAL_DRIVE); /* set drive 0 bit */
+		swim_write(base, mode1, MOTON);
 	} else if (location == EXTERNAL_DRIVE) {
 		swim_write(base, mode0, INTERNAL_DRIVE); /* clear drive 0 bit */
 		swim_write(base, mode1, EXTERNAL_DRIVE); /* set drive 1 bit */
+		swim_write(base, mode1, MOTON);
+	} else {
+		swim_write(base, mode0, INTERNAL_DRIVE);
+		swim_write(base, mode0, EXTERNAL_DRIVE);
+		swim_write(base, mode0, MOTON);
 	}
 }
 
@@ -446,6 +453,7 @@ static int floppy_eject(struct floppy_state *fs)
 	swim_drive(base, fs->location);
 	swim_motor(base, OFF);
 	swim_eject(base);
+	swim_drive(base, NO_DRIVE);
 
 	fs->disk_in = 0;
 	fs->ejected = 1;
@@ -653,8 +661,10 @@ static int floppy_open(struct block_device *bdev, fmode_t mode)
 	else if (fs->ref_count > 0)
 		--fs->ref_count;
 
-	if (fs->ref_count == 0)
+	if (fs->ref_count == 0) {
 		swim_motor(base, OFF);
+		swim_drive(base, NO_DRIVE);
+	}
 	return err;
 }
 
@@ -683,6 +693,7 @@ static void floppy_release(struct gendisk *disk, fmode_t mode)
 	if (fs->ref_count == 0) {
 		swim_drive(base, fs->location);
 		swim_motor(base, OFF);
+		swim_drive(base, NO_DRIVE);
 	}
 	mutex_unlock(&swim_mutex);
 }
@@ -788,6 +799,7 @@ static int swim_floppy_init(struct swim_priv *swd)
 	if (swim_readbit(base, DRIVE_PRESENT) &&
 	    !swim_readbit(base, ONEMEG_DRIVE))
 		swim_add_floppy(swd, EXTERNAL_DRIVE);
+	swim_drive(base, NO_DRIVE);
 
 	/* register floppy drives */
 
Signature · 1 line
2.53.0
author maintainer
#1568 lore raw

PATCH 1565/2214 swim: Dont disable drive after every sector

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 05a7c4b197279cfb3ec3df26146ce085a2438aec ]

This fixes multi-sector reads on a Sony drive. The Mitsubishi drive doesn't seem to care either way.

---
 drivers/block/swim.c | 4 ----
 1 file changed, 4 deletions(-)
drivers/block/swim.c +0 -4
index f956a2e40f70b..cb0bdf5c8ac30 100644
@@ -472,8 +472,6 @@ static inline int swim_read_sector(struct floppy_state *fs,
 	short i;
 
 	swim_track(fs, track);
-
-	swim_write(base, mode1, MOTON);
 	swim_head(base, side);
 	swim_write(base, mode0, side);
 
@@ -489,8 +487,6 @@ static inline int swim_read_sector(struct floppy_state *fs,
 	}
 	local_irq_restore(flags);
 
-	swim_write(base, mode0, MOTON);
-
 	if ((header.side != side)  || (header.track != track) ||
 	     (header.sector != sector))
 		return 0;
Signature · 1 line
2.53.0
author maintainer
#1569 lore raw

PATCH 1566/2214 swim: Perform ISM/IWM mode switching according to specs

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 52981bb358d1b5c38040b00f81e2cf973f9148ad ]

The SWIM spec says, "MOTOREN must be low to switch modes" and "after switching from ISM to IWM, the very first command must be a clear L7".

The ISM spec says, MOTOREN "must not be cleared until after the Action bit is cleared". Perform those operations in the correct sequence.

When switching to ISM mode, the Mode register has to be selected with a particular sequence of bit flips. Set q7 low then q6 low then mtrOff.

---
 drivers/block/swim.c | 10 +++++++---
 1 file changed, 7 insertions(+), 3 deletions(-)
drivers/block/swim.c +7 -3
index cb0bdf5c8ac30..c9b7b8a0ad62d 100644
@@ -149,6 +149,7 @@ struct iwm {
 #define ACTION		0x08
 #define WRITE_MODE	0x10
 #define HEDSEL		0x20
+#define ISM_SELECT	0x40
 #define MOTON		0x80
 
 /*----------------------------------------------------------------------------*/
@@ -222,18 +223,21 @@ extern int swim_read_sector_data(struct swim __iomem *base,
 static DEFINE_MUTEX(swim_mutex);
 static inline void set_swim_mode(struct swim __iomem *base, int enable)
 {
-	struct iwm __iomem *iwm_base;
+	struct iwm __iomem *iwm_base = (struct iwm __iomem *)base;
 	unsigned long flags;
 
 	if (!enable) {
-		swim_write(base, mode0, 0xf8);
+		swim_write(base, mode0, ACTION);
+		swim_write(base, mode0, ENBL1 | ENBL2 | MOTON);
+		swim_write(base, mode0, ISM_SELECT);
+		iwm_read(iwm_base, q7L);
 		return;
 	}
 
-	iwm_base = (struct iwm __iomem *)base;
 	local_irq_save(flags);
 
 	iwm_read(iwm_base, q7L);
+	iwm_read(iwm_base, q6L);
 	iwm_read(iwm_base, mtrOff);
 	iwm_read(iwm_base, q6H);
 
Signature · 1 line
2.53.0
author maintainer
#1570 lore raw

PATCH 1567/2214 swim: simplify using blk_cleanup_disk() on swim_remove()

In reply to Greg Kroah-Hartman

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

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

From: Luis Chamberlain <[email protected]>

[ Upstream commit b76a30c254d987b4ae7d47415081121d4c0a7423 ]

We can simplify swim_remove() by using one call instead of two, just as other drivers do. Use that pattern.

Stable-dep-of: 994a778d9665 ("swim: Configure parameter memory")

---
 drivers/block/swim.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/block/swim.c +1 -2
index c9b7b8a0ad62d..2934b7dc02bb2 100644
@@ -924,9 +924,8 @@ static int swim_remove(struct platform_device *dev)
 
 	for (drive = 0; drive < swd->floppy_count; drive++) {
 		del_gendisk(swd->unit[drive].disk);
-		blk_cleanup_queue(swd->unit[drive].disk->queue);
+		blk_cleanup_disk(swd->unit[drive].disk);
 		blk_mq_free_tag_set(&swd->unit[drive].tag_set);
-		put_disk(swd->unit[drive].disk);
 	}
 
 	unregister_blkdev(FLOPPY_MAJOR, "fd");
Signature · 1 line
2.53.0
author maintainer
#1571 lore raw

PATCH 1568/2214 swim: add helper for disk cleanup

In reply to Greg Kroah-Hartman

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

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

From: Luis Chamberlain <[email protected]>

[ Upstream commit 4e9abe72530a2baf5f80d60e8d0bcdb84964d2e4 ]

Disk cleanup can be shared between exit and bringup. Use a helper to do the work required. The only functional change at this point is we're being overly paraoid on exit to check for a null disk as well now, and this should be safe.

We'll later expand on this, this change just makes subsequent changes easier to read.

Stable-dep-of: 994a778d9665 ("swim: Configure parameter memory")

---
 drivers/block/swim.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)
drivers/block/swim.c +13 -8
index 2934b7dc02bb2..f32fb9b388c13 100644
@@ -783,6 +783,17 @@ static const struct blk_mq_ops swim_mq_ops = {
 	.queue_rq = swim_queue_rq,
 };
 
+static void swim_cleanup_floppy_disk(struct floppy_state *fs)
+{
+	struct gendisk *disk = fs->disk;
+
+	if (!disk)
+		return;
+
+	blk_cleanup_disk(disk);
+	blk_mq_free_tag_set(&fs->tag_set);
+}
+
 static int swim_floppy_init(struct swim_priv *swd)
 {
 	int err;
@@ -848,12 +859,7 @@ static int swim_floppy_init(struct swim_priv *swd)
 exit_put_disks:
 	unregister_blkdev(FLOPPY_MAJOR, "fd");
 	do {
-		struct gendisk *disk = swd->unit[drive].disk;
-
-		if (!disk)
-			continue;
-		blk_cleanup_disk(disk);
-		blk_mq_free_tag_set(&swd->unit[drive].tag_set);
+		swim_cleanup_floppy_disk(&swd->unit[drive]);
 	} while (drive--);
 	return err;
 }
@@ -924,8 +930,7 @@ static int swim_remove(struct platform_device *dev)
 
 	for (drive = 0; drive < swd->floppy_count; drive++) {
 		del_gendisk(swd->unit[drive].disk);
-		blk_cleanup_disk(swd->unit[drive].disk);
-		blk_mq_free_tag_set(&swd->unit[drive].tag_set);
+		swim_cleanup_floppy_disk(&swd->unit[drive]);
 	}
 
 	unregister_blkdev(FLOPPY_MAJOR, "fd");
Signature · 1 line
2.53.0
author maintainer
#1572 lore raw

PATCH 1569/2214 swim: Configure parameter memory

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 994a778d966592d0e211deaf6e436fcb21413935 ]

The original SWIM chip has a 16 byte memory to store parameters which control read/write timing. The SWIM 2 chip retains the last 4 parameters which control write pre-compensation, but omits the first 12. Hence SWIM 2 can perform reads without any parameter memory setup but SWIM 1 cannot.

Configure the SWIM parameter memory with the appropriate values so that SWIM 1 can read too. The parameters used here were observed in SWIM chip memory, using Macsbug, while MacOS was reading from an MFM floppy disk.

---
 drivers/block/swim.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/block/swim.c +13 -0
index f32fb9b388c13..ea64e386d2f16 100644
@@ -794,12 +794,25 @@ static void swim_cleanup_floppy_disk(struct floppy_state *fs)
 	blk_mq_free_tag_set(&fs->tag_set);
 }
 
+static void swim_set_parameters(struct swim __iomem *base)
+{
+	unsigned int i;
+	static const u8 mem[] = { 0x18, 0x41, 0x2e, 0x2e, 0x18, 0x18, 0x1b, 0x1b,
+				  0x2f, 0x2f, 0x19, 0x19, 0x97, 0x1b, 0x57, 0x3b, };
+
+	swim_write(base, mode0, 0); /* reset parameter memory index */
+	for (i = 0; i < 16; ++i)
+		swim_write(base, parameter, mem[i]);
+}
+
 static int swim_floppy_init(struct swim_priv *swd)
 {
 	int err;
 	int drive;
 	struct swim __iomem *base = swd->base;
 
+	swim_set_parameters(base);
+
 	/* scan floppy drives */
 
 	swim_drive(base, INTERNAL_DRIVE);
Signature · 1 line
2.53.0
author maintainer
#1573 lore raw

PATCH 1570/2214 swim: Dont start motor until medium is present

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit dbbaaae83cb277d76e70426c0f0ce289e867ed77 ]

The spindle motor should not be running when a disk is to be inserted. Don't start the motor while the drive is empty.

---
 drivers/block/swim.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/block/swim.c +4 -2
index ea64e386d2f16..ec2191ddda847 100644
@@ -632,8 +632,7 @@ static int floppy_open(struct block_device *bdev, fmode_t mode)
 	swim_write(base, setup, S_IBM_DRIVE  | S_FCLK_DIV2);
 	udelay(10);
 	swim_drive(base, fs->location);
-	swim_motor(base, ON);
-	swim_action(base, SETMFM);
+
 	if (fs->ejected)
 		setup_medium(fs);
 	if (!fs->disk_in) {
@@ -641,6 +640,9 @@ static int floppy_open(struct block_device *bdev, fmode_t mode)
 		goto out;
 	}
 
+	swim_motor(base, ON);
+	swim_action(base, SETMFM);
+
 	set_capacity(fs->disk, fs->total_secs);
 
 	if (mode & FMODE_NDELAY)
Signature · 1 line
2.53.0
author maintainer
#1574 lore raw

PATCH 1571/2214 swim: Recalibrate when drive is probed

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit ba087774a1493ade524f8a783461fa8be9c75039 ]

Track zero recalibration can be slow and is normally done only once i.e. during system POST or boot-up. Recalibrate once after the drive is probed rather than every time the device is opened. Don't register the drive if recalibration fails. Park the heads before ejecting.

---
 drivers/block/swim.c | 37 +++++++++++++++++--------------------
 1 file changed, 17 insertions(+), 20 deletions(-)
drivers/block/swim.c +17 -20
index ec2191ddda847..6fdfcfc5f05ef 100644
@@ -411,6 +411,7 @@ static inline int swim_track00(struct swim __iomem *base)
 	if (swim_readbit(base, TRACK_ZERO))
 		return 0;
 
+	pr_err("swim: track zero recalibration failed\n");
 	return -1;
 }
 
@@ -455,6 +456,7 @@ static int floppy_eject(struct floppy_state *fs)
 	struct swim __iomem *base = fs->swd->base;
 
 	swim_drive(base, fs->location);
+	swim_track(fs, 40);
 	swim_motor(base, OFF);
 	swim_eject(base);
 	swim_drive(base, NO_DRIVE);
@@ -595,13 +597,6 @@ static void setup_medium(struct floppy_state *fs)
 		struct floppy_struct *g;
 		fs->disk_in = 1;
 		fs->write_protected = swim_readbit(base, WRITE_PROT);
-
-		if (swim_track00(base))
-			printk(KERN_ERR
-				"SWIM: cannot move floppy head to track 0\n");
-
-		swim_track00(base);
-
 		fs->type = swim_readbit(base, TWOMEG_MEDIA) ?
 			HD_MEDIA : DD_MEDIA;
 		fs->head_number = swim_readbit(base, SINGLE_SIDED) ? 1 : 2;
@@ -609,7 +604,6 @@ static void setup_medium(struct floppy_state *fs)
 		fs->total_secs = g->size;
 		fs->secpercyl = g->head * g->sect;
 		fs->secpertrack = g->sect;
-		fs->track = 0;
 	} else {
 		fs->disk_in = 0;
 	}
@@ -761,24 +755,33 @@ static const struct block_device_operations floppy_fops = {
 	.check_events	 = floppy_check_events,
 };
 
-static int swim_add_floppy(struct swim_priv *swd, enum drive_location location)
+static void swim_add_floppy(struct swim_priv *swd, enum drive_location location)
 {
 	struct floppy_state *fs = &swd->unit[swd->floppy_count];
 	struct swim __iomem *base = swd->base;
 
-	fs->location = location;
+	swim_drive(base, location);
+	if (!swim_readbit(base, DRIVE_PRESENT) ||
+	    swim_readbit(base, ONEMEG_DRIVE))
+		goto out;
+	if (swim_readbit(base, DISK_IN))
+		swim_motor(base, ON);
+	if (swim_track00(base))
+		goto out;
 
-	swim_motor(base, OFF);
+	fs->location = location;
 
 	fs->type = HD_MEDIA;
 	fs->head_number = 2;
 
 	fs->ref_count = 0;
 	fs->ejected = 1;
+	fs->track = 0;
 
 	swd->floppy_count++;
 
-	return 0;
+out:
+	swim_motor(base, OFF);
 }
 
 static const struct blk_mq_ops swim_mq_ops = {
@@ -817,14 +820,8 @@ static int swim_floppy_init(struct swim_priv *swd)
 
 	/* scan floppy drives */
 
-	swim_drive(base, INTERNAL_DRIVE);
-	if (swim_readbit(base, DRIVE_PRESENT) &&
-	    !swim_readbit(base, ONEMEG_DRIVE))
-		swim_add_floppy(swd, INTERNAL_DRIVE);
-	swim_drive(base, EXTERNAL_DRIVE);
-	if (swim_readbit(base, DRIVE_PRESENT) &&
-	    !swim_readbit(base, ONEMEG_DRIVE))
-		swim_add_floppy(swd, EXTERNAL_DRIVE);
+	swim_add_floppy(swd, INTERNAL_DRIVE);
+	swim_add_floppy(swd, EXTERNAL_DRIVE);
 	swim_drive(base, NO_DRIVE);
 
 	/* register floppy drives */
Signature · 1 line
2.53.0
author maintainer
#1575 lore raw

PATCH 1572/2214 swim: Add track zero recalibration delay

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit b9c1095048537cc9563f12d0a24f2f0df129e4d5 ]

The UPD72070 spec indicates that the track zero sensor can take 3 ms to stabilize following a STEP command so add a call to msleep(). Remove the duplicate swim_readbit() call as there's no need for that once the sensor signal has stabilized.

---
 drivers/block/swim.c | 9 ++++-----
 1 file changed, 4 insertions(+), 5 deletions(-)
drivers/block/swim.c +4 -5
index 6fdfcfc5f05ef..066cceafd9d70 100644
@@ -401,16 +401,15 @@ static inline int swim_track00(struct swim __iomem *base)
 	for (try = 0; try < 100; try++) {
 
 		swim_select(base, RELAX);
+		msleep(3);
+
 		if (swim_readbit(base, TRACK_ZERO))
-			break;
+			return 0;
 
 		if (swim_step(base))
-			return -1;
+			break;
 	}
 
-	if (swim_readbit(base, TRACK_ZERO))
-		return 0;
-
 	pr_err("swim: track zero recalibration failed\n");
 	return -1;
 }
Signature · 1 line
2.53.0
author maintainer
#1576 lore raw

PATCH 1573/2214 swim: Handle FIFO timeout error

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 2e9bcbecee68431b689d16b23bc5529db686f8d4 ]

When polling the FIFO for a mark byte in the sector header, don't return zero if the timeout counter has expired, return an error code.

---
 drivers/block/swim_asm.S | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/block/swim_asm.S +1 -2
index 3d7a2d87595a5..208622ac4aeb9 100644
@@ -79,7 +79,7 @@ wait_addr_mark_byte:
 
 	tstb	%a2@
 	dbmi	%d2, wait_addr_mark_byte
-	bpl	header_exit
+	bpl	signal_nonyb
 
 	moveb	%a3@, %d3
 	cmpb	%a0@+, %d3
@@ -136,7 +136,6 @@ crc1:	tstb	%a2@
 
 	tstb	%a3@(read_error - read_mark)
 
-header_exit:
 	moveq	#0, %d0
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
 	rts
Signature · 1 line
2.53.0
author maintainer
#1577 lore raw

PATCH 1574/2214 swim: Check for CRC errors

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 01bff87152e6748fbac9ccd645d3d5f490d75ddc ]

After reading either the sector header or sector data, examine that flag in the handshake register which holds the result of the CRC calculation. CRC validation has to take place with the last byte still in the FIFO. This flag can't be checked by the caller because by then all bytes will have been retrieved from the FIFO. Return an error code when appropriate.

---
 drivers/block/swim_asm.S | 10 ++++++++++
 1 file changed, 10 insertions(+)
drivers/block/swim_asm.S +10 -0
index 208622ac4aeb9..250e408060e8b 100644
@@ -42,6 +42,8 @@
 	.equ	max_retry, 40
 	.equ	sector_size, 512
 
+	.equ	.Lhr_crc_error,		0x02
+
 	.global swim_read_sector_header
 swim_read_sector_header:
 	link	%a6, #0
@@ -134,6 +136,10 @@ crc1:	tstb	%a2@
 
 	moveb	%a3@, %a4@(o_crc1)
 
+	moveb	%a2@, %d5
+	andb	#.Lhr_crc_error, %d5
+	bne	signal_nonyb
+
 	tstb	%a3@(read_error - read_mark)
 
 	moveq	#0, %d0
@@ -226,6 +232,10 @@ data_crc1:
 
 	moveb	%a3@, %d5
 
+	moveb	%a2@, %d5
+	andb	#.Lhr_crc_error, %d5
+	bne	data_exit
+
 	tstb	%a3@(read_error - read_mark)
 
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
Signature · 1 line
2.53.0
author maintainer
#1578 lore raw

PATCH 1575/2214 swim: Simplify return value initialization

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 385a44e78702762e2fce796cf37b6f465d86a87f ]

Initialize the error result once only. Update the result only after a successful read.

Stable-dep-of: d66bdef0a308 ("swim: Check error register during sector read")

---
 drivers/block/swim_asm.S | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)
drivers/block/swim_asm.S +3 -8
index 250e408060e8b..81c824923336c 100644
@@ -49,6 +49,7 @@ swim_read_sector_header:
 	link	%a6, #0
 	moveml	%d1-%d5/%a0-%a4,%sp@-
 	movel	%a6@(0x0c), %a4
+	moveq	#-1, %d0
 	bsr	mfm_read_addrmark
 	moveml	%sp@+, %d1-%d5/%a0-%a4
 	unlk	%a6
@@ -63,7 +64,6 @@ mfm_read_addrmark:
 	movel	%a6@(0x08), %a3
 	lea	%a3@(read_handshake), %a2
 	lea	%a3@(read_mark), %a3
-	moveq	#-1, %d0
 	movew	#seek_time, %d2
 
 wait_header_init:
@@ -143,10 +143,7 @@ crc1:	tstb	%a2@
 	tstb	%a3@(read_error - read_mark)
 
 	moveq	#0, %d0
-	moveb	#0x18, %a3@(write_mode0 - read_mark)
-	rts
 signal_nonyb:
-	moveq	#-1, %d0
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
 	rts
 
@@ -155,6 +152,7 @@ swim_read_sector_data:
 	link	%a6, #0
 	moveml	%d1-%d5/%a0-%a5,%sp@-
 	movel	%a6@(0x0c), %a4
+	moveq	#-1, %d0
 	bsr	mfm_read_data
 	moveml	%sp@+, %d1-%d5/%a0-%a5
 	unlk	%a6
@@ -238,15 +236,12 @@ data_crc1:
 
 	tstb	%a3@(read_error - read_mark)
 
-	moveb	#0x18, %a3@(write_mode0 - read_mark)
-
 	/* return number of bytes read */
 
 	movel	#sector_size, %d0
 	addw	#1, %d4
 	subl	%d4, %d0
-	rts
+
 data_exit:
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
-	moveq	#-1, %d0
 	rts
Signature · 1 line
2.53.0
author maintainer
#1579 lore raw

PATCH 1576/2214 swim: Check error register during sector read

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit d66bdef0a3083fed1a0fc4f0ee1c85b26476b21c ]

Clear the error register only once before a sector read operation. Don't clear it afterwards -- the caller needs it. Check the error register in swim_read_sector() and return the appropriate error when necessary. Fully validate the sector header. Don't terminate the search loop early just because an erroneous sector header showed up.

---
 drivers/block/swim.c     | 13 +++++++------
 drivers/block/swim_asm.S |  8 --------
 2 files changed, 7 insertions(+), 14 deletions(-)
drivers/block/swim.c +7 -6
index 066cceafd9d70..f4fbe195cbc09 100644
@@ -482,20 +482,21 @@ static inline int swim_read_sector(struct floppy_state *fs,
 
 	local_irq_save(flags);
 	for (i = 0; i < 36; i++) {
-		ret = swim_read_sector_header(base, &header);
-		if (!ret && (header.sector == sector)) {
+		if (swim_read_sector_header(base, &header) ||
+		    swim_read(base, error) || header.track != track ||
+		    header.side != side || header.size != 2)
+			continue;
+		if (header.sector == sector) {
 			/* found */
 
 			ret = swim_read_sector_data(base, buffer);
+			if (swim_read(base, error))
+				ret = -EIO;
 			break;
 		}
 	}
 	local_irq_restore(flags);
 
-	if ((header.side != side)  || (header.track != track) ||
-	     (header.sector != sector))
-		return 0;
-
 	return ret;
 }
 
drivers/block/swim_asm.S +0 -8
index 81c824923336c..fcf574ced85ec 100644
@@ -67,7 +67,6 @@ mfm_read_addrmark:
 	movew	#seek_time, %d2
 
 wait_header_init:
-	tstb	%a3@(read_error - read_mark)
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
 	moveb	#0x01, %a3@(write_mode1 - read_mark)
 	moveb	#0x01, %a3@(write_mode0 - read_mark)
@@ -140,8 +139,6 @@ crc1:	tstb	%a2@
 	andb	#.Lhr_crc_error, %d5
 	bne	signal_nonyb
 
-	tstb	%a3@(read_error - read_mark)
-
 	moveq	#0, %d0
 signal_nonyb:
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
@@ -166,7 +163,6 @@ mfm_read_data:
 	movew	#seek_time, %d2
 
 wait_data_init:
-	tstb	%a3@(read_error - read_mark)
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
 	moveb	#0x01, %a3@(write_mode1 - read_mark)
 	moveb	#0x01, %a3@(write_mode0 - read_mark)
@@ -191,8 +187,6 @@ wait_data_mark_byte:
 
 	/* read data */
 
-	tstb	%a3@(read_error - read_mark)
-
 	movel	#sector_size-1, %d4		/* sector size */
 read_new_data:
 	movew	#max_retry, %d2
@@ -234,8 +228,6 @@ data_crc1:
 	andb	#.Lhr_crc_error, %d5
 	bne	data_exit
 
-	tstb	%a3@(read_error - read_mark)
-
 	/* return number of bytes read */
 
 	movel	#sector_size, %d0
Signature · 1 line
2.53.0
author maintainer
#1580 lore raw

PATCH 1577/2214 swim: Dont use the mark register to read data

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit cb7923133ec18e67e2d6718e29426a4423af6b73 ]

If an unexpected mark byte were to be read from the data register, an error would be flagged. But no error gets flagged when such a byte is read from the mark register, which is misleading. Always use the data register except when a mark byte is expected.

---
 drivers/block/swim_asm.S | 21 +++++++++++----------
 1 file changed, 11 insertions(+), 10 deletions(-)
drivers/block/swim_asm.S +11 -10
index fcf574ced85ec..699f7c90dd1c0 100644
@@ -47,11 +47,11 @@
 	.global swim_read_sector_header
 swim_read_sector_header:
 	link	%a6, #0
-	moveml	%d1-%d5/%a0-%a4,%sp@-
+	moveml	%d1-%d5/%a0-%a5,%sp@-
 	movel	%a6@(0x0c), %a4
 	moveq	#-1, %d0
 	bsr	mfm_read_addrmark
-	moveml	%sp@+, %d1-%d5/%a0-%a4
+	moveml	%sp@+, %d1-%d5/%a0-%a5
 	unlk	%a6
 	rts
 
@@ -63,6 +63,7 @@ sector_data_mark:
 mfm_read_addrmark:
 	movel	%a6@(0x08), %a3
 	lea	%a3@(read_handshake), %a2
+	lea	%a3@(read_data), %a5
 	lea	%a3@(read_mark), %a3
 	movew	#seek_time, %d2
 
@@ -93,7 +94,7 @@ amark0:	tstb	%a2@
 	dbmi	%d2, amark0
 	bpl	signal_nonyb
 
-	moveb	%a3@, %a4@(o_track)
+	moveb	%a5@, %a4@(o_track)
 
 	moveq	#max_retry, %d2
 
@@ -101,7 +102,7 @@ amark1:	tstb	%a2@
 	dbmi	%d2, amark1
 	bpl	signal_nonyb
 
-	moveb	%a3@, %a4@(o_side)
+	moveb	%a5@, %a4@(o_side)
 
 	moveq	#max_retry, %d2
 
@@ -109,7 +110,7 @@ amark2:	tstb	%a2@
 	dbmi	%d2, amark2
 	bpl	signal_nonyb
 
-	moveb	%a3@, %a4@(o_sector)
+	moveb	%a5@, %a4@(o_sector)
 
 	moveq	#max_retry, %d2
 
@@ -117,7 +118,7 @@ amark3:	tstb	%a2@
 	dbmi	%d2, amark3
 	bpl	signal_nonyb
 
-	moveb	%a3@, %a4@(o_size)
+	moveb	%a5@, %a4@(o_size)
 
 	moveq	#max_retry, %d2
 
@@ -125,7 +126,7 @@ crc0:	tstb	%a2@
 	dbmi	%d2, crc0
 	bpl	signal_nonyb
 
-	moveb	%a3@, %a4@(o_crc0)
+	moveb	%a5@, %a4@(o_crc0)
 
 	moveq	#max_retry, %d2
 
@@ -133,7 +134,7 @@ crc1:	tstb	%a2@
 	dbmi	%d2, crc1
 	bpl	signal_nonyb
 
-	moveb	%a3@, %a4@(o_crc1)
+	moveb	%a5@, %a4@(o_crc1)
 
 	moveb	%a2@, %d5
 	andb	#.Lhr_crc_error, %d5
@@ -212,7 +213,7 @@ data_crc0:
 	dbmi	%d2, data_crc0
 	bpl	data_exit
 
-	moveb	%a3@, %d5
+	moveb	%a5@, %d5
 
 	moveq	#max_retry, %d2
 
@@ -222,7 +223,7 @@ data_crc1:
 	dbmi	%d2, data_crc1
 	bpl	data_exit
 
-	moveb	%a3@, %d5
+	moveb	%a5@, %d5
 
 	moveb	%a2@, %d5
 	andb	#.Lhr_crc_error, %d5
Signature · 1 line
2.53.0
author maintainer
#1581 lore raw

PATCH 1578/2214 swim: Fix buffer overflow

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit a54870a4ae4962acf341eb6ca010cae02af1a6ea ]

The effect of this bug can be observed as swim_read_sector_data() inexplicably returning -5, or an error flag indicating that a mark byte was read from the data register, or other odd behviour.

When copying bytes from the chip FIFO to the read buffer, the driver keeps count of the remaining buffer space using register %d4. A counter in register %d2 serves as a timeout. The driver polls (%a2), the handshake register, until flags indicate that byte(s) have arrived in the FIFO.

        movel   #sector_size-1, %d4
read_new_data:
        movew   #max_retry, %d2
read_data_loop:
        moveb   %a2@, %d5
        andb    #0xc0, %d5
        dbne    %d2, read_data_loop
        beq     data_exit
        moveb   %a5@, %a4@+
        andb    #0x40, %d5
        dbne    %d4, read_new_data
        beq     exit_loop

Note that the exit_loop branch depends upon a flag in the handshake register and not on the remaining buffer space. Hence there may be no branch to exit_loop after %d4 is decremented to -1 (i.e. full buffer).

moveb %a5@, %a4@+
dbra %d4, read_new_data
exit_loop:

Here is a second decrement of %d4 which can now reach -2. But the buffer bounds check is a comparison with -1, which is now ineffective. Hence the loop will continue copying until %d2 eventually reaches -1.

Fix this bug by terminating the loop as soon as %d4 or %d2 reach -1. Reset the timeout whenever a byte is copied.

---
 drivers/block/swim_asm.S | 16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)
drivers/block/swim_asm.S +9 -7
index 699f7c90dd1c0..e06aadb411a14 100644
@@ -43,6 +43,8 @@
 	.equ	sector_size, 512
 
 	.equ	.Lhr_crc_error,		0x02
+	.equ	.Lhr_fifo_2bytes,	0x40
+	.equ	.Lhr_fifo_1byte,	0x80
 
 	.global swim_read_sector_header
 swim_read_sector_header:
@@ -189,20 +191,20 @@ wait_data_mark_byte:
 	/* read data */
 
 	movel	#sector_size-1, %d4		/* sector size */
-read_new_data:
 	movew	#max_retry, %d2
 read_data_loop:
 	moveb	%a2@, %d5
-	andb	#0xc0, %d5
+	andb	#(.Lhr_fifo_1byte + .Lhr_fifo_2bytes), %d5
 	dbne	%d2, read_data_loop
 	beq	data_exit
+	moveq	#max_retry, %d2
 	moveb	%a5@, %a4@+
-	andb	#0x40, %d5
-	dbne	%d4, read_new_data
-	beq	exit_loop
+	dbra	%d4, 1f
+	bra	data_crc0
+1:	andb	#.Lhr_fifo_2bytes, %d5
+	beq	read_data_loop
 	moveb	%a5@, %a4@+
-	dbra	%d4, read_new_data
-exit_loop:
+	dbra	%d4, read_data_loop
 
 	/* read CRC */
 
Signature · 1 line
2.53.0
author maintainer
#1582 lore raw

PATCH 1579/2214 swim: Remove redundant RELAX actions

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 04cb277e6841b55b950b42063fe11501b5d1676a ]

Wherever we have a swim_select() or swim_readbit() call there is an implicit RELAX. That means the caller doesn't have to do it. Remove the redundant code.

BTW, Inside Macintosh says, "Be sure [...] that CA0 and CA1 are set high before changing SEL." Hence the RELAX found in swim_select(). The SwimIII driver in mkLinux also has that. But the swim3.c driver in Linux is odd:
it scatters RELAX actions around as though SEL was not actually under its control... In anycase, swim.c really does control SEL so there's no need for that here.

Stable-dep-of: b736189c93c9 ("swim: Deduplicate polling loops")

---
 drivers/block/swim.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)
drivers/block/swim.c +1 -7
index f4fbe195cbc09..adde8dc537259 100644
@@ -336,7 +336,6 @@ static inline void swim_motor(struct swim __iomem *base,
 		swim_action(base, MOTOR_ON);
 
 		for (i = 0; i < 2*HZ; i++) {
-			swim_select(base, RELAX);
 			if (swim_readbit(base, MOTOR_ON))
 				break;
 			set_current_state(TASK_INTERRUPTIBLE);
@@ -344,7 +343,7 @@ static inline void swim_motor(struct swim __iomem *base,
 		}
 	} else if (action == OFF) {
 		swim_action(base, MOTOR_OFF);
-		swim_select(base, RELAX);
+		swim_write(base, phase, RELAX | PHASE_PIN_DIR);
 	}
 }
 
@@ -355,13 +354,11 @@ static inline void swim_eject(struct swim __iomem *base)
 	swim_action(base, EJECT);
 
 	for (i = 0; i < 2*HZ; i++) {
-		swim_select(base, RELAX);
 		if (!swim_readbit(base, DISK_IN))
 			break;
 		set_current_state(TASK_INTERRUPTIBLE);
 		schedule_timeout(1);
 	}
-	swim_select(base, RELAX);
 }
 
 static inline void swim_head(struct swim __iomem *base, enum head head)
@@ -385,7 +382,6 @@ static inline int swim_step(struct swim __iomem *base)
 		set_current_state(TASK_INTERRUPTIBLE);
 		schedule_timeout(1);
 
-		swim_select(base, RELAX);
 		if (!swim_readbit(base, STEP))
 			return 0;
 	}
@@ -399,8 +395,6 @@ static inline int swim_track00(struct swim __iomem *base)
 	swim_action(base, SEEK_NEGATIVE);
 
 	for (try = 0; try < 100; try++) {
-
-		swim_select(base, RELAX);
 		msleep(3);
 
 		if (swim_readbit(base, TRACK_ZERO))
Signature · 1 line
2.53.0
author maintainer
#1583 lore raw

PATCH 1580/2214 swim: Deduplicate polling loops

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit b736189c93c9ad1056521a4229ec2f252b714b5a ]

Replace duplicated polling loops with poll_timeout_us(). Change the interruptible sleep to uninterruptible because signal delivery shouldn't be allowed to shorten delays required by the drive hardware.

Change the timeout for the !STEP transition to 20 ms in accordance with the maximum interval required by the UPD72070 spec. The existing 1 second timeout is impractical considering the number of steps in a typical seek.

Change the return type of swim_readbit() to bool because that way the bit names make sense i.e. the reader doesn't have to remember to invert the active-low logic used for drive signals.

---
 drivers/block/swim.c | 39 ++++++++-------------------------------
 1 file changed, 8 insertions(+), 31 deletions(-)
drivers/block/swim.c +8 -31
index adde8dc537259..795058cdf4343 100644
@@ -19,6 +19,7 @@
 #include <linux/major.h>
 #include <linux/mutex.h>
 #include <linux/hdreg.h>
+#include <linux/iopoll.h>
 #include <linux/kernel.h>
 #include <linux/delay.h>
 #include <linux/platform_device.h>
@@ -296,7 +297,7 @@ static inline void swim_action(struct swim __iomem *base, int action)
 	local_irq_restore(flags);
 }
 
-static inline int swim_readbit(struct swim __iomem *base, int bit)
+static inline bool swim_readbit(struct swim __iomem *base, int bit)
 {
 	int stat;
 
@@ -309,6 +310,9 @@ static inline int swim_readbit(struct swim __iomem *base, int bit)
 	return (stat & SENSE) == 0;
 }
 
+#define swim_readbit_timeout(base, bit, val, timeout_us) \
+	poll_timeout_us(, swim_readbit(base, bit) == val, 1000, timeout_us, false)
+
 static inline void swim_drive(struct swim __iomem *base,
 			      enum drive_location location)
 {
@@ -331,16 +335,8 @@ static inline void swim_motor(struct swim __iomem *base,
 			      enum motor_action action)
 {
 	if (action == ON) {
-		int i;
-
 		swim_action(base, MOTOR_ON);
-
-		for (i = 0; i < 2*HZ; i++) {
-			if (swim_readbit(base, MOTOR_ON))
-				break;
-			set_current_state(TASK_INTERRUPTIBLE);
-			schedule_timeout(1);
-		}
+		swim_readbit_timeout(base, MOTOR_ON, true, 2000 * 1000);
 	} else if (action == OFF) {
 		swim_action(base, MOTOR_OFF);
 		swim_write(base, phase, RELAX | PHASE_PIN_DIR);
@@ -349,16 +345,8 @@ static inline void swim_motor(struct swim __iomem *base,
 
 static inline void swim_eject(struct swim __iomem *base)
 {
-	int i;
-
 	swim_action(base, EJECT);
-
-	for (i = 0; i < 2*HZ; i++) {
-		if (!swim_readbit(base, DISK_IN))
-			break;
-		set_current_state(TASK_INTERRUPTIBLE);
-		schedule_timeout(1);
-	}
+	swim_readbit_timeout(base, DISK_IN, false, 2000 * 1000);
 }
 
 static inline void swim_head(struct swim __iomem *base, enum head head)
@@ -373,19 +361,8 @@ static inline void swim_head(struct swim __iomem *base, enum head head)
 
 static inline int swim_step(struct swim __iomem *base)
 {
-	int wait;
-
 	swim_action(base, STEP);
-
-	for (wait = 0; wait < HZ; wait++) {
-
-		set_current_state(TASK_INTERRUPTIBLE);
-		schedule_timeout(1);
-
-		if (!swim_readbit(base, STEP))
-			return 0;
-	}
-	return -1;
+	return swim_readbit_timeout(base, STEP, false, 20 * 1000);
 }
 
 static inline int swim_track00(struct swim __iomem *base)
Signature · 1 line
2.53.0
author maintainer
#1584 lore raw

PATCH 1581/2214 swim: Check drive ready bit

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit abae04f5bed924590e98fe1519566a6e0144e6b5 ]

The drive provides a readiness signal that has to be tested before certain commands are issued to the drive. Rename the SEEK_COMPLETE flag as READY because that's how it's known in the documentation as well as the mkLinux source code.

Poll for that signal after stepping the heads and also after switching to MFM mode, as that's what mkLinux does. Check for readiness when stepping because testing shows that some drives require this.

---
 drivers/block/swim.c | 29 ++++++++++++++++++++++-------
 1 file changed, 22 insertions(+), 7 deletions(-)
drivers/block/swim.c +22 -7
index 795058cdf4343..6e9fe07a8440e 100644
@@ -117,7 +117,7 @@ struct iwm {
 #define TACHO		0x103
 #define READ_DATA_1	0x104
 #define GCR_MODE	0x105
-#define SEEK_COMPLETE	0x106
+#define READY		0x106
 #define TWOMEG_MEDIA	0x107
 
 /* Bits in handshake register */
@@ -313,6 +313,14 @@ static inline bool swim_readbit(struct swim __iomem *base, int bit)
 #define swim_readbit_timeout(base, bit, val, timeout_us) \
 	poll_timeout_us(, swim_readbit(base, bit) == val, 1000, timeout_us, false)
 
+#define swim_READY_timeout(base) \
+({ \
+	int ret = swim_readbit_timeout(base, READY, true, 1000 * 1000); \
+	if (ret) \
+		printk(KERN_DEBUG "%s: drive not ready\n", __func__); \
+	ret; \
+})
+
 static inline void swim_drive(struct swim __iomem *base,
 			      enum drive_location location)
 {
@@ -351,8 +359,6 @@ static inline void swim_eject(struct swim __iomem *base)
 
 static inline void swim_head(struct swim __iomem *base, enum head head)
 {
-	/* wait drive is ready */
-
 	if (head == UPPER_HEAD)
 		swim_select(base, READ_DATA_1);
 	else if (head == LOWER_HEAD)
@@ -387,20 +393,25 @@ static inline int swim_track00(struct swim __iomem *base)
 
 static inline int swim_seek(struct swim __iomem *base, int step)
 {
-	if (step == 0)
-		return 0;
-
 	if (step < 0) {
 		swim_action(base, SEEK_NEGATIVE);
 		step = -step;
-	} else
+	} else if (step > 0)
 		swim_action(base, SEEK_POSITIVE);
 
+	swim_READY_timeout(base);
+
+	if (step == 0)
+		return 0;
+
 	for ( ; step > 0; step--) {
 		if (swim_step(base))
 			return -1;
 	}
 
+	msleep(30);
+	swim_READY_timeout(base);
+
 	return 0;
 }
 
@@ -482,6 +493,8 @@ static blk_status_t floppy_read_sectors(struct floppy_state *fs,
 
 
 	swim_drive(base, fs->location);
+	swim_READY_timeout(base);
+
 	for (i = req_sector; i < req_sector + sectors_nb; i++) {
 		int x;
 		track = i / fs->secpercyl;
@@ -607,6 +620,8 @@ static int floppy_open(struct block_device *bdev, fmode_t mode)
 
 	swim_motor(base, ON);
 	swim_action(base, SETMFM);
+	msleep(30);
+	swim_READY_timeout(base);
 
 	set_capacity(fs->disk, fs->total_secs);
 
Signature · 1 line
2.53.0
author maintainer
#1585 lore raw

PATCH 1582/2214 swim: Revisit delays

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit bbcd3e38a4f355d6b0320f3f6916df96c109f2ad ]

AFAIK, timing requirements for the various FDHD drive mechanisms aren't well documented. But we do have the UPD72070 spec and secondary sources like swim3.c and mkLinux source code. This patch is needed to satisfy the requirements in the UPD72070 spec and follows mkLinux.

Change the LSTRB pulse to 2 microseconds, because this is what mkLinux does. Inside Macintosh says, "Hold LSTRB high for at least one usec but not more than one msec".

When a disk is ejected, pause before de-asserting /ENBL. Wait 150 us after the STEP command for valid signalling. Pause for 1 us after setting the step direction before sending the STEP command.

---
 drivers/block/swim.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/block/swim.c +5 -1
index 6e9fe07a8440e..155ff90599b61 100644
@@ -290,7 +290,7 @@ static inline void swim_action(struct swim __iomem *base, int action)
 	swim_select(base, action);
 	udelay(1);
 	swim_write(base, phase, LSTRB | action | PHASE_PIN_DIR);
-	udelay(1);
+	udelay(2);
 	swim_write(base, phase, action | PHASE_PIN_DIR);
 	udelay(1);
 
@@ -337,6 +337,7 @@ static inline void swim_drive(struct swim __iomem *base,
 		swim_write(base, mode0, EXTERNAL_DRIVE);
 		swim_write(base, mode0, MOTON);
 	}
+	udelay(1);
 }
 
 static inline void swim_motor(struct swim __iomem *base,
@@ -355,6 +356,7 @@ static inline void swim_eject(struct swim __iomem *base)
 {
 	swim_action(base, EJECT);
 	swim_readbit_timeout(base, DISK_IN, false, 2000 * 1000);
+	msleep(1);
 }
 
 static inline void swim_head(struct swim __iomem *base, enum head head)
@@ -368,6 +370,7 @@ static inline void swim_head(struct swim __iomem *base, enum head head)
 static inline int swim_step(struct swim __iomem *base)
 {
 	swim_action(base, STEP);
+	udelay(150);
 	return swim_readbit_timeout(base, STEP, false, 20 * 1000);
 }
 
@@ -398,6 +401,7 @@ static inline int swim_seek(struct swim __iomem *base, int step)
 		step = -step;
 	} else if (step > 0)
 		swim_action(base, SEEK_POSITIVE);
+	udelay(1);
 
 	swim_READY_timeout(base);
 
Signature · 1 line
2.53.0
author maintainer
#1586 lore raw

PATCH 1583/2214 swim: Dont search beyond the first data mark

In reply to Greg Kroah-Hartman

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

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

From: Finn Thain <[email protected]>

[ Upstream commit 79eafa4ccb74be023fd2cae87d75425d6eb903bb ]

The ISM chip does an automatic MFM gap/sync search when the Action bit is first set. That search may stop at any of a) post-index gap, b) address field gap or c) data field gap. To find the next sector header, the driver need not search at all. It only has to validate the mark bytes.

Once the sector address mark has been validated, swim_read_sector_data() is called to read the sector contents. Between the sector address and data fields lies an intra-sector gap followed by a data field mark. After this mark is validated, the 512-byte data area is read into the IO request buffer.

Problem is, if any byte in the data field mark is mis-read, the driver searches the whole sector and then reaches the data field mark in the following sector. The wrong sector is then read into the buffer, and swim_read_sector_data() returns success. The request is silently corrupted.

The existing limit on polling loop iterations does constrain the search distance but is inherently tied to CPU speed. This is probably the reason why corruption was only observed on a 68030 system.

Discontinue the mark search when the mark bytes fail validation.

---
 drivers/block/swim_asm.S | 73 ++++++++++++++++++++++------------------
 1 file changed, 40 insertions(+), 33 deletions(-)
drivers/block/swim_asm.S +40 -33
index e06aadb411a14..4d6404234bf09 100644
@@ -41,18 +41,45 @@
 	.equ	seek_time, 30000
 	.equ	max_retry, 40
 	.equ	sector_size, 512
+	.equ	.Lmark_sequence_len,	4
 
 	.equ	.Lhr_crc_error,		0x02
 	.equ	.Lhr_fifo_2bytes,	0x40
 	.equ	.Lhr_fifo_1byte,	0x80
 
+.Lmfm_mark_check:
+	/*
+	 * This subroutine reads and validates a mark byte sequence.
+	 * On entry, %a1 and %d4 shall hold the location and length (resp.)
+	 * of the mark byte array.
+	 * %a2 and %a3 shall hold the locations of the handshake and mark
+	 * registers.
+	 * Returns zero in %d1 for success.
+	 */
+
+	moveq	#-1, %d1
+	subq	#1, %d4
+	movew	#seek_time, %d2
+
+5:	tstb	%a2@
+	dbmi	%d2, 5b
+	bpl	6f
+
+	moveb	%a3@, %d3
+	cmpb	%a1@+, %d3
+	dbne	%d4, 5b
+	bne	6f
+
+	moveq	#0, %d1
+6:	rts
+
 	.global swim_read_sector_header
 swim_read_sector_header:
 	link	%a6, #0
 	moveml	%d1-%d5/%a0-%a5,%sp@-
 	movel	%a6@(0x0c), %a4
 	moveq	#-1, %d0
-	bsr	mfm_read_addrmark
+	bsr	.Lmfm_read_header
 	moveml	%sp@+, %d1-%d5/%a0-%a5
 	unlk	%a6
 	rts
@@ -62,33 +89,25 @@ sector_address_mark:
 sector_data_mark:
 	.byte	0xa1, 0xa1, 0xa1, 0xfb
 
-mfm_read_addrmark:
+.Lmfm_read_header:
 	movel	%a6@(0x08), %a3
 	lea	%a3@(read_handshake), %a2
 	lea	%a3@(read_data), %a5
 	lea	%a3@(read_mark), %a3
-	movew	#seek_time, %d2
 
-wait_header_init:
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
 	moveb	#0x01, %a3@(write_mode1 - read_mark)
 	moveb	#0x01, %a3@(write_mode0 - read_mark)
 	tstb	%a3@(read_error - read_mark)
 	moveb	#0x08, %a3@(write_mode1 - read_mark)
 
-	lea	sector_address_mark, %a0
-	moveq	#3, %d1
-
-wait_addr_mark_byte:
-
-	tstb	%a2@
-	dbmi	%d2, wait_addr_mark_byte
-	bpl	signal_nonyb
+	lea	sector_address_mark, %a1
+	moveq	#.Lmark_sequence_len, %d4
+	bsr	.Lmfm_mark_check
+	tstl	%d1
+	bne	signal_nonyb
 
-	moveb	%a3@, %d3
-	cmpb	%a0@+, %d3
-	dbne	%d1, wait_addr_mark_byte
-	bne	wait_header_init
+	/* read header */
 
 	moveq	#max_retry, %d2
 
@@ -163,30 +182,18 @@ mfm_read_data:
 	lea	%a3@(read_handshake), %a2
 	lea	%a3@(read_data), %a5
 	lea	%a3@(read_mark), %a3
-	movew	#seek_time, %d2
 
-wait_data_init:
 	moveb	#0x18, %a3@(write_mode0 - read_mark)
 	moveb	#0x01, %a3@(write_mode1 - read_mark)
 	moveb	#0x01, %a3@(write_mode0 - read_mark)
 	tstb	%a3@(read_error - read_mark)
 	moveb	#0x08, %a3@(write_mode1 - read_mark)
 
-	lea	sector_data_mark, %a0
-	moveq	#3, %d1
-
-	/* wait data address mark */
-
-wait_data_mark_byte:
-
-	tstb	%a2@
-	dbmi	%d2, wait_data_mark_byte
-	bpl	data_exit
-
-	moveb	%a3@, %d3
-	cmpb	%a0@+, %d3
-	dbne	%d1, wait_data_mark_byte
-	bne	wait_data_init
+	lea	sector_data_mark, %a1
+	moveq	#.Lmark_sequence_len, %d4
+	bsr	.Lmfm_mark_check
+	tstl	%d1
+	bne	data_exit
 
 	/* read data */
 
Signature · 1 line
2.53.0
author maintainer
#1587 lore raw

PATCH 1584/2214 media: i2c: ov5648: Fix a data race on state->streaming

In reply to Greg Kroah-Hartman

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

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

From: Ginger Li <[email protected]>

[ Upstream commit 76c8ffbe1b9bc0ee4c977180e3f498352de0d01e ]

ov5648_s_stream() updates state->streaming after it has released sensor->mutex, while ov5648_resume() reads that field with sensor->mutex held. A concurrent resume can therefore observe a stale streaming state and leave the sensor in standby while the software believes that it is streaming, or the other way around.

Keep the update inside the sensor->mutex protected section that already programs the standby mode.

Fixes e43ccb0a045f ("media: i2c: Add support for the OV5648 image sensor")
---
 drivers/media/i2c/ov5648.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/media/i2c/ov5648.c +2 -2
index bb3666fc56183..741f0dcb801eb 100644
@@ -2145,13 +2145,13 @@ static int ov5648_s_stream(struct v4l2_subdev *subdev, int enable)
 
 	mutex_lock(&sensor->mutex);
 	ret = ov5648_sw_standby(sensor, !enable);
+	if (!ret)
+		state->streaming = !!enable;
 	mutex_unlock(&sensor->mutex);
 
 	if (ret)
 		return ret;
 
-	state->streaming = !!enable;
-
 	if (!enable)
 		pm_runtime_put(sensor->dev);
 
Signature · 1 line
2.53.0
author maintainer
#1588 lore raw

PATCH 1585/2214 serial: sifive: fix off-by-one in console port bounds check

In reply to Greg Kroah-Hartman

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

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

From: Zongmin Zhou <[email protected]>

[ Upstream commit bf961847813d9edcb5a9c089714069428d141fa9 ]

sifive_serial_console_ports[] is sized SIFIVE_SERIAL_MAX_PORTS (8), so valid indices are 0..7. The bounds check in sifive_serial_probe() uses "id > SIFIVE_SERIAL_MAX_PORTS", which lets id == 8 pass and causes an out-of-bounds write in __ssp_add_console_port() through sifive_serial_console_ports[ssp->port.line].

Fix the check to use >= instead, matching the check already used in sifive_serial_console_setup().

Fixes 45c054d0815b ("tty: serial: add driver for the SiFive UART")
---
 drivers/tty/serial/sifive.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/tty/serial/sifive.c +1 -1
index 493bde7b8e398..a3ab203304bfb 100644
@@ -975,7 +975,7 @@ static int sifive_serial_probe(struct platform_device *pdev)
 	}
 
 #ifdef CONFIG_SERIAL_SIFIVE_CONSOLE
-	if (id > SIFIVE_SERIAL_MAX_PORTS) {
+	if (id >= SIFIVE_SERIAL_MAX_PORTS) {
 		dev_err(&pdev->dev, "too many UARTs (%d)\n", id);
 		return -EINVAL;
 	}
Signature · 1 line
2.53.0
author maintainer
#1589 lore raw

PATCH 1586/2214 dt-bindings: vendor-prefixes: add Genesys Logic

In reply to Greg Kroah-Hartman

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

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

From: Icenowy Zheng <[email protected]>

[ Upstream commit 38cea8e31e9ef143187135d714aed4d7bd18463c ]

Genesys Logic, Inc. is a manufacturer for interface chips, especially USB hubs.

https://www.genesyslogic.com.tw/

Stable-dep-of: aa906abf611c ("dt-bindings: vendor-prefixes: add Free Mobile")

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index fd2d25d7fa51b..f32408840be63 100644
@@ -445,6 +445,8 @@ patternProperties:
     description: GE Fanuc Intelligent Platforms Embedded Systems, Inc.
   "^gemei,.*":
     description: Gemei Digital Technology Co., Ltd.
+  "^genesys,.*":
+    description: Genesys Logic, Inc.
   "^geniatech,.*":
     description: Geniatech, Inc.
   "^giantec,.*":
Signature · 1 line
2.53.0
author maintainer
#1590 lore raw

PATCH 1587/2214 dt-bindings: vendor-prefixes: Add GPLUS

In reply to Greg Kroah-Hartman

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

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

From: Lin, Meng-Bo <[email protected]>

[ Upstream commit d7a2d86837b9de8d378f040cfc9184905d82bd78 ]

Add vendor prefix for GPLUS.
https://www.gplus.com.tw

Stable-dep-of: aa906abf611c ("dt-bindings: vendor-prefixes: add Free Mobile")

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index f32408840be63..7dc90c24d5261 100644
@@ -463,6 +463,8 @@ patternProperties:
     description: Shenzhen Huiding Technology Co., Ltd.
   "^google,.*":
     description: Google, Inc.
+  "^gplus,.*":
+    description: GPLUS
   "^grinn,.*":
     description: Grinn
   "^grmn,.*":
Signature · 1 line
2.53.0
author maintainer
#1591 lore raw

PATCH 1588/2214 dt-bindings: Add vendor prefixes

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit 714ff69c331c005c9cd05d69d3c6fe9da20ef687 ]

These vendor prefixes are used by some routers supported by e.g. OpenWrt.

- ADI Engineering is a US telecom equipment company.

- Arcom Controllers is a US manufacturer of repeaters.

- Freecom Gmbh is a german telecom equipment company.

- Gemtek Technology is a Taiwan telecom company.

- Gateway Communications was a telecommunication company, now acquired by HKT Limited/PCCW.

- Goramo Gorecki is a privately owned Polish telecom company.

- U.S. Robotics Corporation, known through their brand name USRobotics is generally referred to as "USR" so use this prefix for the company's device tree bindings.

Stable-dep-of: aa906abf611c ("dt-bindings: vendor-prefixes: add Free Mobile")

---
 .../devicetree/bindings/vendor-prefixes.yaml       | 14 ++++++++++++++
 1 file changed, 14 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +14 -0
index 7dc90c24d5261..d68bfde3b28d0 100644
@@ -53,6 +53,8 @@ patternProperties:
     description: AD Holdings Plc.
   "^adi,.*":
     description: Analog Devices, Inc.
+  "^adieng,.*":
+    description: ADI Engineering, Inc.
   "^advantech,.*":
     description: Advantech Corporation
   "^aeroflexgaisler,.*":
@@ -113,6 +115,8 @@ patternProperties:
     description: Arasan Chip Systems
   "^archermind,.*":
     description: ArcherMind Technology (Nanjing) Co., Ltd.
+  "^arcom,.*":
+    description: Arcom Controllers
   "^arctic,.*":
     description: Arctic Sand
   "^arcx,.*":
@@ -421,6 +425,8 @@ patternProperties:
     description: FocalTech Systems Co.,Ltd
   "^forlinx,.*":
     description: Baoding Forlinx Embedded Technology Co., Ltd.
+  "^freecom,.*":
+    description: Freecom Gmbh
   "^frida,.*":
     description: Shenzhen Frida LCD Co., Ltd.
   "^friendlyarm,.*":
@@ -431,6 +437,8 @@ patternProperties:
     description: Fujitsu Ltd.
   "^gardena,.*":
     description: GARDENA GmbH
+  "^gateway,.*":
+    description: Gateway Communications
   "^gateworks,.*":
     description: Gateworks Corporation
   "^gcw,.*":
@@ -445,6 +453,8 @@ patternProperties:
     description: GE Fanuc Intelligent Platforms Embedded Systems, Inc.
   "^gemei,.*":
     description: Gemei Digital Technology Co., Ltd.
+  "^gemtek,.*":
+    description: Gemtek Technology Co., Ltd.
   "^genesys,.*":
     description: Genesys Logic, Inc.
   "^geniatech,.*":
@@ -463,6 +473,8 @@ patternProperties:
     description: Shenzhen Huiding Technology Co., Ltd.
   "^google,.*":
     description: Google, Inc.
+  "^goramo,.*":
+    description: Goramo Gorecki
   "^gplus,.*":
     description: GPLUS
   "^grinn,.*":
@@ -1248,6 +1260,8 @@ patternProperties:
     description: United Radiant Technology Corporation
   "^usi,.*":
     description: Universal Scientific Industrial Co., Ltd.
+  "^usr,.*":
+    description: U.S. Robotics Corporation
   "^utoo,.*":
     description: Aigo Digital Technology Co., Ltd.
   "^v3,.*":
Signature · 1 line
2.53.0
author maintainer
#1592 lore raw

PATCH 1589/2214 dt-bindings: vendor-prefixes: add Free Mobile

In reply to Greg Kroah-Hartman

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

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

From: Vincent Jardin <[email protected]>

[ Upstream commit aa906abf611c7f37c0fe9171c1d33c40270aa5fe ]

Free Mobile is a French mobile network operator. Add its vendor prefix, used by the Nodebox v3 CPU Module (freemobile,nbxv3).

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index d68bfde3b28d0..9de9c419f043f 100644
@@ -427,6 +427,8 @@ patternProperties:
     description: Baoding Forlinx Embedded Technology Co., Ltd.
   "^freecom,.*":
     description: Freecom Gmbh
+  "^freemobile,.*":
+    description: Free Mobile
   "^frida,.*":
     description: Shenzhen Frida LCD Co., Ltd.
   "^friendlyarm,.*":
Signature · 1 line
2.53.0
author maintainer
#1593 lore raw

PATCH 1590/2214 mfd: intel_quark_i2c_gpio: Manage the fixed-rate clock

In reply to Greg Kroah-Hartman

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

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

From: Myeonghun Pak <[email protected]>

[ Upstream commit 14a7cf1a3730c7c2393ef5a226e90baf61aa8a9f ]

The I2C clock is allocated by clk_register_fixed_rate(), but cleanup uses clk_unregister(), leaving the fixed-rate provider allocation behind.

Manage the provider with devm_clk_hw_register_fixed_rate() and its lookup with devm_clk_hw_register_clkdev(). Managed resources release the lookup before the provider, after the MFD children have been removed.

Remove the manual clock cleanup and the private structure that only stored the clock and lookup pointers.

This issue was identified during our ongoing static-analysis research while reviewing kernel code.

Fixes 60ae5b9f5cdd ("mfd: intel_quark_i2c_gpio: Add Intel Quark X1000 I2C-GPIO MFD Driver")
Assisted-by OpenAI:GPT-5.6
---
 drivers/mfd/intel_quark_i2c_gpio.c | 52 ++++--------------------------
 1 file changed, 6 insertions(+), 46 deletions(-)
drivers/mfd/intel_quark_i2c_gpio.c +6 -46
index 9b9c76bd067b8..443a1a0cf7778 100644
@@ -35,11 +35,6 @@
 /* The Quark I2C controller source clock */
 #define INTEL_QUARK_I2C_CLK_HZ	33000000
 
-struct intel_quark_mfd {
-	struct clk		*i2c_clk;
-	struct clk_lookup	*i2c_clk_lookup;
-};
-
 static const struct property_entry intel_quark_i2c_controller_standard_properties[] = {
 	PROPERTY_ENTRY_U32("clock-frequency", I2C_MAX_STANDARD_MODE_FREQ),
 	{ }
@@ -160,37 +155,12 @@ MODULE_DEVICE_TABLE(pci, intel_quark_mfd_ids);
 
 static int intel_quark_register_i2c_clk(struct device *dev)
 {
-	struct intel_quark_mfd *quark_mfd = dev_get_drvdata(dev);
-	struct clk *i2c_clk;
-
-	i2c_clk = clk_register_fixed_rate(dev,
-					  INTEL_QUARK_I2C_CONTROLLER_CLK, NULL,
-					  0, INTEL_QUARK_I2C_CLK_HZ);
-	if (IS_ERR(i2c_clk))
-		return PTR_ERR(i2c_clk);
-
-	quark_mfd->i2c_clk = i2c_clk;
-	quark_mfd->i2c_clk_lookup = clkdev_create(i2c_clk, NULL,
-						INTEL_QUARK_I2C_CONTROLLER_CLK);
-
-	if (!quark_mfd->i2c_clk_lookup) {
-		clk_unregister(quark_mfd->i2c_clk);
-		dev_err(dev, "Fixed clk register failed\n");
-		return -ENOMEM;
-	}
-
-	return 0;
-}
-
-static void intel_quark_unregister_i2c_clk(struct device *dev)
-{
-	struct intel_quark_mfd *quark_mfd = dev_get_drvdata(dev);
-
-	if (!quark_mfd->i2c_clk_lookup)
-		return;
+	struct clk_hw *hw;
 
-	clkdev_drop(quark_mfd->i2c_clk_lookup);
-	clk_unregister(quark_mfd->i2c_clk);
+	hw = devm_clk_hw_register_fixed_rate(dev, INTEL_QUARK_I2C_CONTROLLER_CLK,
+					     NULL, 0, INTEL_QUARK_I2C_CLK_HZ);
+	return devm_clk_hw_register_clkdev(dev, hw, NULL,
+					 INTEL_QUARK_I2C_CONTROLLER_CLK);
 }
 
 static int intel_quark_i2c_setup(struct pci_dev *pdev)
@@ -238,19 +208,12 @@ static int intel_quark_gpio_setup(struct pci_dev *pdev)
 static int intel_quark_mfd_probe(struct pci_dev *pdev,
 				 const struct pci_device_id *id)
 {
-	struct intel_quark_mfd *quark_mfd;
 	int ret;
 
 	ret = pcim_enable_device(pdev);
 	if (ret)
 		return ret;
 
-	quark_mfd = devm_kzalloc(&pdev->dev, sizeof(*quark_mfd), GFP_KERNEL);
-	if (!quark_mfd)
-		return -ENOMEM;
-
-	dev_set_drvdata(&pdev->dev, quark_mfd);
-
 	ret = intel_quark_register_i2c_clk(&pdev->dev);
 	if (ret)
 		return ret;
@@ -260,7 +223,7 @@ static int intel_quark_mfd_probe(struct pci_dev *pdev,
 	/* This driver only requires 1 IRQ vector */
 	ret = pci_alloc_irq_vectors(pdev, 1, 1, PCI_IRQ_ALL_TYPES);
 	if (ret < 0)
-		goto err_unregister_i2c_clk;
+		return ret;
 
 	ret = intel_quark_i2c_setup(pdev);
 	if (ret)
@@ -282,8 +245,6 @@ static int intel_quark_mfd_probe(struct pci_dev *pdev,
 	software_node_unregister_node_group(intel_quark_gpio_node_group);
 err_free_irq_vectors:
 	pci_free_irq_vectors(pdev);
-err_unregister_i2c_clk:
-	intel_quark_unregister_i2c_clk(&pdev->dev);
 	return ret;
 }
 
@@ -292,7 +253,6 @@ static void intel_quark_mfd_remove(struct pci_dev *pdev)
 	mfd_remove_devices(&pdev->dev);
 	software_node_unregister_node_group(intel_quark_gpio_node_group);
 	pci_free_irq_vectors(pdev);
-	intel_quark_unregister_i2c_clk(&pdev->dev);
 }
 
 static struct pci_driver intel_quark_mfd_driver = {
Signature · 1 line
2.53.0
author maintainer
#1594 lore raw

PATCH 1591/2214 mfd: da9150: Balance IRQ wake on teardown

In reply to Greg Kroah-Hartman

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

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

From: Myeonghun Pak <[email protected]>

[ Upstream commit ef464f82f3aecad4071af71074b2ceb77ffbb609 ]

DA9150 enables IRQ wake after registering its regmap IRQ chip, but does not disable it on a later probe failure or driver removal. This leaves the wake depth elevated after the handler is removed, and repeated bind attempts can accumulate the imbalance.

Remember whether enabling IRQ wake succeeded and balance only a successful call. Remove MFD children first so their nested IRQ users are gone, then disable wake before removing the regmap IRQ chip. Preserve the existing non-fatal behavior when IRQ wake cannot be enabled.

This issue was identified during our ongoing static-analysis research while reviewing kernel code.

Fixes b8fce55c09d3 ("mfd: Add support for DA9150 combined charger & fuel-gauge device")
Assisted-by LLM
---
 drivers/mfd/da9150-core.c       | 8 ++++++--
 include/linux/mfd/da9150/core.h | 1 +
 2 files changed, 7 insertions(+), 2 deletions(-)
drivers/mfd/da9150-core.c +6 -2
index 58009c8cb8705..b05ac49dbd6c5 100644
@@ -451,7 +451,7 @@ static int da9150_probe(struct i2c_client *client,
 
 	da9150->irq_base = regmap_irq_chip_get_base(da9150->regmap_irq_data);
 
-	enable_irq_wake(da9150->irq);
+	da9150->irq_wake_enabled = !enable_irq_wake(da9150->irq);
 
 	ret = mfd_add_devices(da9150->dev, -1, da9150_devs,
 			      ARRAY_SIZE(da9150_devs), NULL,
@@ -464,6 +464,8 @@ static int da9150_probe(struct i2c_client *client,
 	return 0;
 
 mfd_fail:
+	if (da9150->irq_wake_enabled)
+		disable_irq_wake(da9150->irq);
 	regmap_del_irq_chip(da9150->irq, da9150->regmap_irq_data);
 regmap_irq_fail:
 	i2c_unregister_device(da9150->core_qif);
@@ -475,8 +477,10 @@ static int da9150_remove(struct i2c_client *client)
 {
 	struct da9150 *da9150 = i2c_get_clientdata(client);
 
-	regmap_del_irq_chip(da9150->irq, da9150->regmap_irq_data);
 	mfd_remove_devices(da9150->dev);
+	if (da9150->irq_wake_enabled)
+		disable_irq_wake(da9150->irq);
+	regmap_del_irq_chip(da9150->irq, da9150->regmap_irq_data);
 	i2c_unregister_device(da9150->core_qif);
 
 	return 0;
include/linux/mfd/da9150/core.h +1 -0
index d116d5f3ef561..369698036d1aa 100644
@@ -65,6 +65,7 @@ struct da9150 {
 	struct regmap_irq_chip_data *regmap_irq_data;
 	int irq;
 	int irq_base;
+	bool irq_wake_enabled;
 };
 
 /* Device I/O - Query Interface for FG and standard register access */
Signature · 1 line
2.53.0
author maintainer
#1595 lore raw

PATCH 1592/2214 kconfig: warn about malformed KCONFIG_PROBABILITY values

In reply to Greg Kroah-Hartman

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

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

From: Dmitrii Tulnov <[email protected]>

[ Upstream commit 9f01cd3a324390b235c67ae4ca3d35085369ebdd ]

randconfig checks the numeric range of each probability but does not validate where strtol() stops. For example, KCONFIG_PROBABILITY=50% is accepted as 50:0:0: the '%' is parsed repeatedly as zero. For the documented value 50, tristate y/m/n probabilities are 25%/25%/50%. With 50%, both y/m probabilities silently become zero, reducing the coverage of random configuration builds. Empty fields and extra fields are also accepted.

Warn when the value does not follow the documented decimal format. For malformed inputs whose parsed probabilities are in range, preserve the existing interpretation unless KCONFIG_WERROR is set. In that case, exit with an error before writing the configuration. Leading whitespace and signs are accepted by strtol(), but also trigger the warning because they are outside the documented format.

Keep the strtol() result as long until the range check. This rejects out-of-range values that narrowing to int previously made valid, such as 4294967296 becoming zero on a 64-bit host, regardless of KCONFIG_WERROR.

Add regression tests for one warning per malformed input, preserved configurations, KCONFIG_WERROR, out-of-range values, the supported probability formats, and the documented empty-value default.

Fixes e43956e60769 ("kconfig: implement KCONFIG_PROBABILITY for randconfig")
Assisted-by LLM
---
 scripts/kconfig/conf.c                        |  16 ++-
 scripts/kconfig/tests/conftest.py             |   6 +
 .../tests/randconfig_probability/Kconfig      |  12 ++
 .../tests/randconfig_probability/__init__.py  | 130 ++++++++++++++++++
 4 files changed, 163 insertions(+), 1 deletion(-)
 create mode 100644 scripts/kconfig/tests/randconfig_probability/Kconfig
 create mode 100644 scripts/kconfig/tests/randconfig_probability/__init__.py
scripts/kconfig/conf.c +15 -1
index 49bbfef4d51b3..5cd43e4becf65 100644
@@ -184,12 +184,23 @@ static bool conf_set_all_new_symbols(enum conf_def_mode mode)
 
 	if (mode == def_random) {
 		int n, p[3];
+		bool warned = false;
 		char *env = getenv("KCONFIG_PROBABILITY");
 
 		n = 0;
 		while (env && *env) {
 			char *endp;
-			int tmp = strtol(env, &endp, 10);
+			long tmp = strtol(env, &endp, 10);
+
+			if (!isdigit((unsigned char)*env) ||
+			    (*endp && *endp != ':') ||
+			    (*endp == ':' && (!endp[1] || n == 2))) {
+				if (!warned) {
+					fprintf(stderr,
+						"warning: KCONFIG_PROBABILITY has malformed format\n");
+					warned = true;
+				}
+			}
 
 			if (tmp >= 0 && tmp <= 100) {
 				p[n++] = tmp;
@@ -225,6 +236,9 @@ static bool conf_set_all_new_symbols(enum conf_def_mode mode)
 			perror("KCONFIG_PROBABILITY");
 			exit(1);
 		}
+
+		if (warned && getenv("KCONFIG_WERROR"))
+			exit(1);
 	}
 
 	for_all_symbols(i, sym) {
scripts/kconfig/tests/conftest.py +6 -0
index af8774a5697c8..3d75025437a70 100644
@@ -289,6 +289,12 @@ class Conf:
         return self._matches('stderr', expected)
 
 
[email protected](autouse=True)
+def clear_werror(monkeypatch):
+    # extra_env can set strict mode, but cannot remove an inherited flag.
+    monkeypatch.delenv('KCONFIG_WERROR', raising=False)
+
+
 @pytest.fixture(scope="module")
 def conf(request):
     """Create a Conf instance and provide it to test functions."""
scripts/kconfig/tests/randconfig_probability/Kconfig new file +12 -0
new file mode 100644
index 0000000000000..84f4e5fcc51e4
@@ -0,0 +1,12 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+config MODULES
+	bool
+	default y
+	modules
+
+config BOOL
+	bool "Bool"
+
+config TRI
+	tristate "Tristate"
scripts/kconfig/tests/randconfig_probability/__init__.py new file +130 -0

130 lines not on this page · raw message

Signature · 1 line
2.53.0
author maintainer
#1596 lore raw

PATCH 1593/2214 ASoC: fsl_asrc: use regmap_write_bits to update clock source registers

In reply to Greg Kroah-Hartman

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

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

From: Shengjiu Wang <[email protected]>

[ Upstream commit 92871abc1302b852b88d86114115f8fd6b50b9b9 ]

After a warm reboot the REG_ASRCSR register may not be in its hardware reset state. regmap caches the register value as 0 (the assumed default), so when the desired clock source index is also 0, regmap_update_bits() skips the actual hardware write because it believes the register already holds the correct value. This leaves a stale clock source in the hardware.

Replace regmap_update_bits() with regmap_write_bits() so that the masked fields are always written to hardware regardless of the cached value.

Fixes 3117bb3109dc ("ASoC: fsl_asrc: Add ASRC ASoC CPU DAI and platform drivers")
---
 sound/soc/fsl/fsl_asrc.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)
sound/soc/fsl/fsl_asrc.c +4 -4
index 3f36e3d2d9a20..a98cf1bef1117 100644
@@ -517,10 +517,10 @@ static int fsl_asrc_config_pair(struct fsl_asrc_pair *pair, bool use_ideal_rate)
 			   ASRCTR_USR(index));
 
 	/* Set the input and output clock sources */
-	regmap_update_bits(asrc->regmap, REG_ASRCSR,
-			   ASRCSR_AICSi_MASK(index) | ASRCSR_AOCSi_MASK(index),
-			   ASRCSR_AICS(index, clk_index[IN]) |
-			   ASRCSR_AOCS(index, clk_index[OUT]));
+	regmap_write_bits(asrc->regmap, REG_ASRCSR,
+			  ASRCSR_AICSi_MASK(index) | ASRCSR_AOCSi_MASK(index),
+			  ASRCSR_AICS(index, clk_index[IN]) |
+			  ASRCSR_AOCS(index, clk_index[OUT]));
 
 	/* Calculate the input clock divisors */
 	indiv = fsl_asrc_cal_asrck_divisor(pair, div[IN]);
Signature · 1 line
2.53.0
author maintainer
#1597 lore raw

PATCH 1594/2214 ASoC: fsl_asrc: check the second front-end DMA channel

In reply to Greg Kroah-Hartman

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

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

From: Slavin Liu <[email protected]>

[ Upstream commit b4136c0d69ead737197af741e20edd1c3574f6b4 ]

The temporary Front-End DMA request can fail independently of the persistent channel acquired earlier in fsl_asrc_dma_hw_params(). The returned NULL pointer is immediately used to read its private data.

Check the temporary channel and release the previously acquired persistent Front-End channel on failure. Clear its slot so a later hw_free cannot release it twice. ASoC marks a component's hw_params only after success and skips unmarked components during rollback, so returning an error alone would leak the earlier channel.

Detected by static analysis and reviewed with AI-assisted source auditing.

Fixes 3117bb3109dc ("ASoC: fsl_asrc: Add ASRC ASoC CPU DAI and platform drivers")
Assisted-by LLM
---
 sound/soc/fsl/fsl_asrc_dma.c | 5 +++++
 1 file changed, 5 insertions(+)
sound/soc/fsl/fsl_asrc_dma.c +5 -0
index 79dd5a9b8f481..e547e2e10cab2 100644
@@ -231,6 +231,11 @@ static int fsl_asrc_dma_hw_params(struct snd_soc_component *component,
 
 		/* Get DMA request of Front-End */
 		tmp_chan = asrc->get_dma_channel(pair, dir);
+		if (!tmp_chan) {
+			dma_release_channel(pair->dma_chan[!dir]);
+			pair->dma_chan[!dir] = NULL;
+			return -EINVAL;
+		}
 		tmp_data = tmp_chan->private;
 		pair->dma_data.dma_request2 = tmp_data->dma_request;
 		pair->dma_data.peripheral_type = tmp_data->peripheral_type;
Signature · 1 line
2.53.0
author maintainer
#1598 lore raw

PATCH 1595/2214 ASoC: adau17x1: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 8ca84987c11d11e0ba60f202e3199b3b80c87721 ]

adau17x1_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The optional switch-mode callback remains ordered before the cache replay.

The issue was found by a tool we developed.

Fixes 4101866c743a ("ASoC: Add ADAU1X61 and ADAU1X81 CODECs common code")
Assisted-by gpt 5
---
 sound/soc/codecs/adau17x1.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/adau17x1.c +1 -3
index 8aae7ab74091f..ba85fedcd50f8 100644
@@ -978,9 +978,7 @@ int adau17x1_resume(struct snd_soc_component *component)
 	if (adau->switch_mode)
 		adau->switch_mode(component->dev);
 
-	regcache_sync(adau->regmap);
-
-	return 0;
+	return regcache_sync(adau->regmap);
 }
 EXPORT_SYMBOL_GPL(adau17x1_resume);
 
Signature · 1 line
2.53.0
author maintainer
#1599 lore raw

PATCH 1596/2214 ASoC: ak4642: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 7142c34ed92fddca7038787b8695d22c661dc6ec ]

ak4642_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes a3a83d9a7cb0 ("ASoC: Add ak4642/ak4643 codec support")
Assisted-by gpt 5
---
 sound/soc/codecs/ak4642.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
sound/soc/codecs/ak4642.c +1 -2
index c49c58eeb4763..df1d68a3cabbe 100644
@@ -533,8 +533,7 @@ static int ak4642_resume(struct snd_soc_component *component)
 	struct regmap *regmap = dev_get_regmap(component->dev, NULL);
 
 	regcache_cache_only(regmap, false);
-	regcache_sync(regmap);
-	return 0;
+	return regcache_sync(regmap);
 }
 static int ak4642_probe(struct snd_soc_component *component)
 {
Signature · 1 line
2.53.0
author maintainer
#1600 lore raw

PATCH 1597/2214 ASoC: nau8540: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 0bf2b3411537e8e4d9c03702646a9fc87198d4b1 ]

nau8540_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes c1644e3de45d ("ASoC: nau8540: new codec driver")
Assisted-by gpt 5
---
 sound/soc/codecs/nau8540.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/nau8540.c +1 -3
index ace96995fedc4..5c5f62b477332 100644
@@ -786,9 +786,7 @@ static int __maybe_unused nau8540_resume(struct snd_soc_component *component)
 	struct nau8540 *nau8540 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(nau8540->regmap, false);
-	regcache_sync(nau8540->regmap);
-
-	return 0;
+	return regcache_sync(nau8540->regmap);
 }
 
 static const struct snd_soc_component_driver nau8540_component_driver = {
Signature · 1 line
2.53.0
author maintainer
#1601 lore raw

PATCH 1598/2214 ASoC: rt1011: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 432734a00bfae17baeff1a83b61bfc9c2ab20c12 ]

rt1011_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes d6e65bb7ff0d ("ASoC: rt1011: Add RT1011 amplifier driver")
Assisted-by gpt 5
---
 sound/soc/codecs/rt1011.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/rt1011.c +1 -3
index faff2b558687d..58495058178ea 100644
@@ -2048,9 +2048,7 @@ static int rt1011_resume(struct snd_soc_component *component)
 	struct rt1011_priv *rt1011 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(rt1011->regmap, false);
-	regcache_sync(rt1011->regmap);
-
-	return 0;
+	return regcache_sync(rt1011->regmap);
 }
 #else
 #define rt1011_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1602 lore raw

PATCH 1599/2214 ASoC: rt1016: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 98ace05793c7b80a9f49d92b78fcc88e0970f8e4 ]

rt1016_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes 79a4b670b4b4 ("ASoC: rt1016: Add the rt1016 support")
Assisted-by gpt 5
---
 sound/soc/codecs/rt1016.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/rt1016.c +1 -3
index 7561d202274cf..1f31517a076ca 100644
@@ -571,9 +571,7 @@ static int rt1016_resume(struct snd_soc_component *component)
 	struct rt1016_priv *rt1016 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(rt1016->regmap, false);
-	regcache_sync(rt1016->regmap);
-
-	return 0;
+	return regcache_sync(rt1016->regmap);
 }
 #else
 #define rt1016_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1603 lore raw

PATCH 1600/2214 ASoC: rt1305: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 30858325aa90380c6bcdde605214f1ebdbec568c ]

rt1305_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes 29bc643ddd7e ("ASoC: rt1305: Add RT1305/RT1306 amplifier driver")
Assisted-by gpt 5
---
 sound/soc/codecs/rt1305.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/rt1305.c +1 -3
index 7a0094578e467..f0207d97a07e5 100644
@@ -898,9 +898,7 @@ static int rt1305_resume(struct snd_soc_component *component)
 	struct rt1305_priv *rt1305 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(rt1305->regmap, false);
-	regcache_sync(rt1305->regmap);
-
-	return 0;
+	return regcache_sync(rt1305->regmap);
 }
 #else
 #define rt1305_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1604 lore raw

PATCH 1601/2214 ASoC: rt1308: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 3db3289178b94056753f7a8c6ee4ab4b232ee6a5 ]

rt1308_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes 2b9def8c0d92 ("ASoC: rt1308: Add RT1308 amplifier driver")
Assisted-by gpt 5
---
 sound/soc/codecs/rt1308.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/rt1308.c +1 -3
index b4e5546e2e217..54cea885515f4 100644
@@ -717,9 +717,7 @@ static int rt1308_resume(struct snd_soc_component *component)
 	struct rt1308_priv *rt1308 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(rt1308->regmap, false);
-	regcache_sync(rt1308->regmap);
-
-	return 0;
+	return regcache_sync(rt1308->regmap);
 }
 #else
 #define rt1308_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1605 lore raw

PATCH 1602/2214 ASoC: rt5616: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 8fc9d353d5371a946d9a9952278b96c09070374c ]

rt5616_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes b1d15059957d ("ASoC: rt5616: add rt5616 codec driver")
Assisted-by gpt 5
---
 sound/soc/codecs/rt5616.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
sound/soc/codecs/rt5616.c +1 -2
index fd0d3a08e9dda..52408f24a494b 100644
@@ -1250,8 +1250,7 @@ static int rt5616_resume(struct snd_soc_component *component)
 	struct rt5616_priv *rt5616 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(rt5616->regmap, false);
-	regcache_sync(rt5616->regmap);
-	return 0;
+	return regcache_sync(rt5616->regmap);
 }
 #else
 #define rt5616_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1606 lore raw

PATCH 1603/2214 ASoC: rt5659: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 3a9b5bce85a1baff95463a2bb021efac64bcf83f ]

rt5659_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes d3cb2de2479b ("ASoC: rt5659: add rt5659 codec driver")
Assisted-by gpt 5
---
 sound/soc/codecs/rt5659.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/rt5659.c +1 -3
index 4a50b169fe032..6a85157456d70 100644
@@ -3705,9 +3705,7 @@ static int rt5659_resume(struct snd_soc_component *component)
 	struct rt5659_priv *rt5659 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(rt5659->regmap, false);
-	regcache_sync(rt5659->regmap);
-
-	return 0;
+	return regcache_sync(rt5659->regmap);
 }
 #else
 #define rt5659_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1607 lore raw

PATCH 1604/2214 ASoC: rt5660: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 70fe8997b03078ff8253d985a161fef310f7c346 ]

rt5660_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The power-off resume delay remains ordered before the cache replay.

The issue was found by a tool we developed.

Fixes 2b26dd4c1fc5 ("ASoC: rt5660: add rt5660 codec driver")
Assisted-by gpt 5
---
 sound/soc/codecs/rt5660.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/rt5660.c +1 -3
index 33ff9156358b4..596b3b8acc342 100644
@@ -1152,9 +1152,7 @@ static int rt5660_resume(struct snd_soc_component *component)
 		msleep(350);
 
 	regcache_cache_only(rt5660->regmap, false);
-	regcache_sync(rt5660->regmap);
-
-	return 0;
+	return regcache_sync(rt5660->regmap);
 }
 #else
 #define rt5660_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1608 lore raw

PATCH 1605/2214 ASoC: rt5665: return cache sync errors from resume

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit 9f2933d9a437302d5ae32eb9022efef049d8ab22 ]

rt5665_resume() restores cached registers during component resume but discards the regcache_sync() result and returns success. A failed replay can therefore leave the hardware register state stale without reaching the component error reporting path.

Return the cache synchronization result. The ASoC component wrapper reports a negative callback result and retains the established best-effort resume state, so this adds observability without introducing rollback or retry semantics.

The successful resume path is unchanged.

The issue was found by a tool we developed.

Fixes 33ada14a26c8 ("ASoC: add rt5665 codec driver")
Assisted-by gpt 5
---
 sound/soc/codecs/rt5665.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
sound/soc/codecs/rt5665.c +1 -3
index 5e00aca0c418a..fb9ef278f1953 100644
@@ -4491,9 +4491,7 @@ static int rt5665_resume(struct snd_soc_component *component)
 	struct rt5665_priv *rt5665 = snd_soc_component_get_drvdata(component);
 
 	regcache_cache_only(rt5665->regmap, false);
-	regcache_sync(rt5665->regmap);
-
-	return 0;
+	return regcache_sync(rt5665->regmap);
 }
 #else
 #define rt5665_suspend NULL
Signature · 1 line
2.53.0
author maintainer
#1609 lore raw

PATCH 1606/2214 iio: accel: bma220: publish the SPI driver ACPI alias

In reply to Greg Kroah-Hartman

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

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

From: Pengpeng Hou <[email protected]>

[ Upstream commit cdc7d3dd101251a6373332726b8d6fad35e83b8a ]

The SPI driver advertises an ACPI match for BMA0220 but does not export that table. A modular build therefore lacks the ACPI alias used to load the driver when the device is enumerated through ACPI.

Export the ACPI table. Move the existing SPI export next to its SPI ID table so each export is adjacent to the table it describes.

The issue was found by our static-analysis tool.

Fixes bf2a5600a3eb ("iio: accel: Add support for Bosch BMA220")
Assisted-by LLM
---
 drivers/iio/accel/bma220_spi.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/iio/accel/bma220_spi.c +2 -1
index bc4c626e454d3..4497e80ef0199 100644
@@ -308,12 +308,13 @@ static const struct spi_device_id bma220_spi_id[] = {
 	{"bma220", 0},
 	{}
 };
+MODULE_DEVICE_TABLE(spi, bma220_spi_id);
 
 static const struct acpi_device_id bma220_acpi_id[] = {
 	{"BMA0220", 0},
 	{}
 };
-MODULE_DEVICE_TABLE(spi, bma220_spi_id);
+MODULE_DEVICE_TABLE(acpi, bma220_acpi_id);
 
 static struct spi_driver bma220_driver = {
 	.driver = {
Signature · 1 line
2.53.0
author maintainer
#1610 lore raw

PATCH 1607/2214 dt-bindings: vendor-prefixes: Add General Mobile

In reply to Greg Kroah-Hartman

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

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

From: Muzaffer Kadir <[email protected]>

[ Upstream commit 7034bec693b752bcd7ebb4fa00dc3b429b4dc4f7 ]

Add vendor prefix for General Mobile.

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index 9de9c419f043f..c5fabb4692902 100644
@@ -457,6 +457,8 @@ patternProperties:
     description: Gemei Digital Technology Co., Ltd.
   "^gemtek,.*":
     description: Gemtek Technology Co., Ltd.
+  "^generalmobile,.*":
+    description: General Mobile Fzc.
   "^genesys,.*":
     description: Genesys Logic, Inc.
   "^geniatech,.*":
Signature · 1 line
2.53.0
author maintainer
#1611 lore raw

PATCH 1608/2214 perf c2c: Fix documented default coalesce fields

In reply to Greg Kroah-Hartman

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

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

From: Jiebin Sun <[email protected]>

[ Upstream commit b36a6f180baad35a962677ae181681ca60dda35b ]

The default coalesce fields dropped pid in favor of iaddr, but the man page still documents the old pid,iaddr default. Update it to match the command.

Fixes 423701a0c8d754d9 ("perf c2c: Change the default coalesce setup")
Cc Dapeng Mi <[email protected]>
Cc Ian Rogers <[email protected]>
Cc James Clark <[email protected]>
Cc Jiri Olsa <[email protected]>
Cc Thomas Falcon <[email protected]>
---
 tools/perf/Documentation/perf-c2c.txt | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/perf/Documentation/perf-c2c.txt +1 -1
index 3b6a2c84ea027..4db7ed63f06b8 100644
@@ -269,7 +269,7 @@ output fields set for cacheline offsets output:
              Code address, Code symbol, Shared Object, Source line
   dso   - coalesced by shared object
 
-By default the coalescing is setup with 'pid,iaddr'.
+By default, coalescing uses `iaddr`.
 
 STDIO OUTPUT
 ------------
Signature · 1 line
2.53.0
author maintainer
#1612 lore raw

PATCH 1609/2214 mtd: block: prevent reclaim I/O during request processing

In reply to Greg Kroah-Hartman

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

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

From: Qingfang Deng <[email protected]>

[ Upstream commit d498b2ddda99eed28cde177b1620dbcea246a1c4 ]

The blktrans request callbacks run under dev->lock and may allocate memory with GFP_KERNEL. For example, mtdblock_writesect() uses vmalloc() to allocate its eraseblock cache. Direct reclaim can then recurse into block I/O and deadlock on resources held by the request being processed. syzbot reports a circular locking dependency involving the device mutex and fs_reclaim.

Commit d5ba1c8ffd0b ("mtd: don't use PF_MEMALLOC") removed PF_MEMALLOC from the MTD request thread. That flag had prevented direct reclaim, so its removal made allocations during request processing eligible for reclaim without excluding I/O.

Wrap request processing in mtd_queue_rq() in a NOIO scope. This covers both the request and background callbacks, including allocations made by lower MTD drivers, and prevents reclaim from initiating filesystem or block I/O.

Fixes d5ba1c8ffd0b ("mtd: don't use PF_MEMALLOC")
Assisted-by Codex:gpt-6-astra
Reported-by [email protected]
---
 drivers/mtd/mtd_blkdevs.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/mtd/mtd_blkdevs.c +5 -0
index 5288a6aaf1828..9d43eb86bb742 100644
@@ -18,6 +18,7 @@
 #include <linux/spinlock.h>
 #include <linux/hdreg.h>
 #include <linux/mutex.h>
+#include <linux/sched/mm.h>
 #include <linux/uaccess.h>
 
 #include "mtdcore.h"
@@ -166,6 +167,7 @@ static blk_status_t mtd_queue_rq(struct blk_mq_hw_ctx *hctx,
 				 const struct blk_mq_queue_data *bd)
 {
 	struct mtd_blktrans_dev *dev;
+	unsigned int noio_flags;
 
 	dev = hctx->queue->queuedata;
 	if (!dev) {
@@ -173,10 +175,13 @@ static blk_status_t mtd_queue_rq(struct blk_mq_hw_ctx *hctx,
 		return BLK_STS_IOERR;
 	}
 
+	/* Reclaim must not recurse into I/O while processing requests. */
+	noio_flags = memalloc_noio_save();
 	spin_lock_irq(&dev->queue_lock);
 	list_add_tail(&bd->rq->queuelist, &dev->rq_list);
 	mtd_blktrans_work(dev);
 	spin_unlock_irq(&dev->queue_lock);
+	memalloc_noio_restore(noio_flags);
 
 	return BLK_STS_OK;
 }
Signature · 1 line
2.53.0
author maintainer
#1613 lore raw

PATCH 1610/2214 writeback: size foreign flushes by target wb dirty pages

In reply to Greg Kroah-Hartman

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

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

From: Xin Yin <[email protected]>

[ Upstream commit 168a8c13159c6e3f0f08da6f8fa2f633a91ba9fd ]

Foreign dirty tracking records the bdi and wb memcg IDs of recently dirtied foreign inodes. When the source memcg needs foreign writeback, mem_cgroup_flush_foreign() queues WB_REASON_FOREIGN_FLUSH to the recorded target bdi_writeback.

cgroup_writeback_by_id() currently sizes this best-effort work from the target memcg's NR_FILE_DIRTY counter. The work is scoped to one target wb, so a memcg-wide dirty count can over-size a single-bdi flush when the target memcg has dirty pages on other devices. It can also under-size the flush when the target wb has dirty pages charged to other memcgs.

Over-sizing keeps the target wb busy longer and can delay later writeback work. Under-sizing can finish before enough pages are written back for the source memcg.

Use the target wb's WB_RECLAIMABLE counter and keep the existing 25% headroom. This keeps the budget aligned with the writeback object being queued.

---
 fs/fs-writeback.c | 17 +++++++----------
 1 file changed, 7 insertions(+), 10 deletions(-)
fs/fs-writeback.c +7 -10
index acc639682c7d5..35cf8e71f1411 100644
@@ -1084,7 +1084,7 @@ int cgroup_writeback_by_id(u64 bdi_id, int memcg_id,
 	struct cgroup_subsys_state *memcg_css;
 	struct bdi_writeback *wb;
 	struct wb_writeback_work *work;
-	unsigned long dirty;
+	long dirty;
 	int ret;
 
 	/* lookup bdi and memcg */
@@ -1113,16 +1113,13 @@ int cgroup_writeback_by_id(u64 bdi_id, int memcg_id,
 	}
 
 	/*
-	 * The caller is attempting to write out most of
-	 * the currently dirty pages.  Let's take the current dirty page
-	 * count and inflate it by 25% which should be large enough to
-	 * flush out most dirty pages while avoiding getting livelocked by
-	 * concurrent dirtiers.
-	 *
-	 * BTW the memcg stats are flushed periodically and this is best-effort
-	 * estimation, so some potential error is ok.
+	 * The caller is attempting to write out most of the target wb's
+	 * currently dirty pages.  Size the work from the wb's reclaimable pages
+	 * and inflate the count by 25%, which should be large enough to flush
+	 * out most dirty pages while avoiding getting livelocked by concurrent
+	 * dirtiers.
 	 */
-	dirty = memcg_page_state(mem_cgroup_from_css(memcg_css), NR_FILE_DIRTY);
+	dirty = wb_stat_sum(wb, WB_RECLAIMABLE);
 	dirty = dirty * 10 / 8;
 
 	/* issue the writeback work */
Signature · 1 line
2.53.0
author maintainer
#1614 lore raw

PATCH 1611/2214 ACPICA: Fix memory leak in acpi_ds_create_field()

In reply to Greg Kroah-Hartman

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

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

From: Breno Leitao <[email protected]>

[ Upstream commit dbb3d99ee95c32711950850ddd33eb22a6042eb2 ]

acpi_ds_create_field() allocates internal_pcc_buffer for operation regions in the PCC address space.

The function runs once per field, so a second field on the same region overwrites the pointer and orphans the previous buffer.

The matching ACPI_FREE() in acpi_ut_delete_internal_obj() only ever sees the last one.

An arm64 platform whose SSDT declares regions with multiple fields each leaks two buffers at boot, as detected by kmemleak:

unreferenced object 0xffff000051e65400 (size 192):
comm "swapper/0", pid 1, jiffies 4294668436 backtrace (crc 0):
__kmalloc_noprof
acpi_ds_create_field acpi_ds_load2_end_op acpi_ds_exec_end_op
acpi_ps_parse_loop
acpi_ps_parse_aml
acpi_ns_load_table
acpi_load_tables
acpi_init

Do not allocate a region if it was previously allocated before, preserving the current design, but, not leaking memory.

Fixes aa6ec56b574d ("ACPICA: ACPI 6.3: add PCC operation region support for AML interpreter")
[ rjw: Subject tweak, fixed the Fixes: tag ]
---
 drivers/acpi/acpica/dsfield.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)
drivers/acpi/acpica/dsfield.c +2 -1
index a168177679691..d53368871e537 100644
@@ -522,7 +522,8 @@ acpi_ds_create_field(union acpi_parse_object *op,
 	}
 
 	if (info.region_node->object->region.space_id ==
-	    ACPI_ADR_SPACE_PLATFORM_COMM) {
+	    ACPI_ADR_SPACE_PLATFORM_COMM &&
+	    !region_node->object->field.internal_pcc_buffer) {
 		region_node->object->field.internal_pcc_buffer =
 		    ACPI_ALLOCATE_ZEROED(info.region_node->object->region.
 					 length);
Signature · 1 line
2.53.0
author maintainer
#1615 lore raw

PATCH 1612/2214 ACPICA: Fix PCC OperationRegion command offsets

In reply to Greg Kroah-Hartman

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

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

From: Sudeep Holla <[email protected]>

[ Upstream commit 888dab4b04ea8725168c6d50d0bfb8057fccca18 ]

ACPI 6.3, section 5.5.2.4.7.3, states that the PCC Operation Region is associated with the region of shared memory that follows the PCC signature.

The generic and extended PCC shared memory layouts include the 4-byte signature at offset 0, so their raw shared memory COMMAND fields are at offsets 4 and 12 respectively. Since AML field offsets for the PCC OperationRegion are relative to the region after that signature, ACPICA must look for those COMMAND fields at OperationRegion offsets 0 and 8.

Adjust the generic and master subspace command checks to use those OperationRegion-relative offsets. Otherwise writes to the COMMAND field can fail to invoke the PCC address space handler at the offset described by the PCC OperationRegion definition.

Fixes aa6ec56b574d ("ACPICA: ACPI 6.3: add PCC operation region support for AML interpreter")
[ rjw: Fixed the Fixes: tag ]
---
 drivers/acpi/acpica/exfield.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)
drivers/acpi/acpica/exfield.c +6 -5
index 8618500f23b39..5f3a19e74252d 100644
@@ -45,12 +45,13 @@ static const u8 acpi_protocol_lengths[] = {
 
 /*
  * The following macros determine a given offset is a COMD field.
- * According to the specification, generic subspaces (types 0-2) contains a
- * 2-byte COMD field at offset 4 and master subspaces (type 3) contains a 4-byte
- * COMD field starting at offset 12.
+ * According to the specification, the PCC OperationRegion begins after
+ * the PCC signature. The raw shared memory COMD offsets of 4 for generic
+ * subspaces (types 0-2) and 12 for master subspaces (type 3) therefore
+ * appear at OperationRegion offsets 0 and 8.
  */
-#define GENERIC_SUBSPACE_COMMAND(a)     (4 == a || a == 5)
-#define MASTER_SUBSPACE_COMMAND(a)      (12 <= a && a <= 15)
+#define GENERIC_SUBSPACE_COMMAND(a)     (0 == a || a == 1)
+#define MASTER_SUBSPACE_COMMAND(a)      (8 <= a && a <= 11)
 
 /*******************************************************************************
  *
Signature · 1 line
2.53.0
author maintainer
#1616 lore raw

PATCH 1613/2214 ASoC: mediatek: mt8183: Fix wrong clock cleanup on clk_set_parent() failure

In reply to Greg Kroah-Hartman

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

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

From: bui duc phuc <[email protected]>

[ Upstream commit 7f05596bd4047e4623eeb1f3c6921d6f71481c3b ]

In mt8183_afe_enable_clock(), when clk_set_parent() fails, the current error path incorrectly cleans up the previously enabled clock instead of the clock used by clk_set_parent().

Fix the error path to clean up the correct clock when clk_set_parent() fails.

Fixes a94aec035a12 ("ASoC: mediatek: mt8183: add platform driver")
---
 sound/soc/mediatek/mt8183/mt8183-afe-clk.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
sound/soc/mediatek/mt8183/mt8183-afe-clk.c +2 -2
index cc4f8f4d3dab9..6ab0734ad1365 100644
@@ -137,7 +137,7 @@ int mt8183_afe_enable_clock(struct mtk_base_afe *afe)
 		dev_err(afe->dev, "%s(), clk_set_parent %s-%s fail %d\n",
 			__func__, aud_clks[CLK_MUX_AUDIO],
 			aud_clks[CLK_CLK26M], ret);
-		goto CLK_MUX_AUDIO_ERR;
+		goto CLK_MUX_AUDIO_INTBUS_ERR;
 	}
 
 	ret = clk_prepare_enable(afe_priv->clk[CLK_MUX_AUDIOINTBUS]);
@@ -153,7 +153,7 @@ int mt8183_afe_enable_clock(struct mtk_base_afe *afe)
 		dev_err(afe->dev, "%s(), clk_set_parent %s-%s fail %d\n",
 			__func__, aud_clks[CLK_MUX_AUDIOINTBUS],
 			aud_clks[CLK_TOP_SYSPLL_D2_D4], ret);
-		goto CLK_MUX_AUDIO_INTBUS_ERR;
+		goto CLK_AFE_ERR;
 	}
 
 	ret = clk_prepare_enable(afe_priv->clk[CLK_AFE]);
Signature · 1 line
2.53.0
author maintainer
#1617 lore raw

PATCH 1614/2214 ASoC: mediatek: mt8183: Fix clock handling in mux disable path

In reply to Greg Kroah-Hartman

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

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

From: bui duc phuc <[email protected]>

[ Upstream commit 4a0f9566ae0c3ad32620254342b266f37f4757de ]

In the disable path of the appl*_mux_setting() functions, clk_disable_unprepare() is called after clk_set_parent(). If clk_set_parent() fails, clk_disable_unprepare() is skipped, potentially leaving one of the clocks enabled.

Remove the goto exit paths so that all teardown steps are attempted and any errors are only reported.

---
 sound/soc/mediatek/mt8183/mt8183-afe-clk.c | 6 ------
 1 file changed, 6 deletions(-)
sound/soc/mediatek/mt8183/mt8183-afe-clk.c +0 -6
index 6ab0734ad1365..dd1a5925af96d 100644
@@ -271,7 +271,6 @@ static int apll1_mux_setting(struct mtk_base_afe *afe, bool enable)
 			dev_err(afe->dev, "%s clk_set_parent %s-%s fail %d\n",
 				__func__, aud_clks[CLK_TOP_MUX_AUD_ENG1],
 				aud_clks[CLK_CLK26M], ret);
-			goto EXIT;
 		}
 		clk_disable_unprepare(afe_priv->clk[CLK_TOP_MUX_AUD_ENG1]);
 
@@ -281,7 +280,6 @@ static int apll1_mux_setting(struct mtk_base_afe *afe, bool enable)
 			dev_err(afe->dev, "%s clk_set_parent %s-%s fail %d\n",
 				__func__, aud_clks[CLK_TOP_MUX_AUD_1],
 				aud_clks[CLK_CLK26M], ret);
-			goto EXIT;
 		}
 		clk_disable_unprepare(afe_priv->clk[CLK_TOP_MUX_AUD_1]);
 	}
@@ -298,7 +296,6 @@ static int apll1_mux_setting(struct mtk_base_afe *afe, bool enable)
 		       afe_priv->clk[CLK_CLK26M]);
 	clk_disable_unprepare(afe_priv->clk[CLK_TOP_MUX_AUD_1]);
 ERR_ENABLE_CLK_TOP_MUX_AUD_1:
-EXIT:
 	return ret;
 }
 
@@ -345,7 +342,6 @@ static int apll2_mux_setting(struct mtk_base_afe *afe, bool enable)
 			dev_err(afe->dev, "%s clk_set_parent %s-%s fail %d\n",
 				__func__, aud_clks[CLK_TOP_MUX_AUD_ENG2],
 				aud_clks[CLK_CLK26M], ret);
-			goto EXIT;
 		}
 		clk_disable_unprepare(afe_priv->clk[CLK_TOP_MUX_AUD_ENG2]);
 
@@ -355,7 +351,6 @@ static int apll2_mux_setting(struct mtk_base_afe *afe, bool enable)
 			dev_err(afe->dev, "%s clk_set_parent %s-%s fail %d\n",
 				__func__, aud_clks[CLK_TOP_MUX_AUD_2],
 				aud_clks[CLK_CLK26M], ret);
-			goto EXIT;
 		}
 		clk_disable_unprepare(afe_priv->clk[CLK_TOP_MUX_AUD_2]);
 	}
@@ -372,7 +367,6 @@ static int apll2_mux_setting(struct mtk_base_afe *afe, bool enable)
 		       afe_priv->clk[CLK_CLK26M]);
 	clk_disable_unprepare(afe_priv->clk[CLK_TOP_MUX_AUD_2]);
 ERR_ENABLE_CLK_TOP_MUX_AUD_2:
-EXIT:
 	return ret;
 }
 
Signature · 1 line
2.53.0
author maintainer
#1618 lore raw

PATCH 1615/2214 ASoC: mediatek: mt8183: Fix APLL enable error handling

In reply to Greg Kroah-Hartman

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

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

From: bui duc phuc <[email protected]>

[ Upstream commit 30cf48383006579d7cd35380cc6e766260b72bff ]

Currently, the mt8183_apll*_enable() functions call mux_setting(afe, true) but do not check its return value to handle failures.

In addition, the cleanup paths of mt8183_apll*_enable() do not call mux_setting(afe, false) when the enable operation fails, while the mt8183_apll*_disable() functions do.

Add error handling for apll*_mux_setting() and call mux_setting(afe, false) in the cleanup paths when mt8183_apll*_enable() fails.

Fixes a94aec035a12 ("ASoC: mediatek: mt8183: add platform driver")
---
 sound/soc/mediatek/mt8183/mt8183-afe-clk.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)
sound/soc/mediatek/mt8183/mt8183-afe-clk.c +10 -2
index dd1a5925af96d..06556689976db 100644
@@ -376,7 +376,9 @@ int mt8183_apll1_enable(struct mtk_base_afe *afe)
 	int ret;
 
 	/* setting for APLL */
-	apll1_mux_setting(afe, true);
+	ret = apll1_mux_setting(afe, true);
+	if (ret)
+		return ret;
 
 	ret = clk_prepare_enable(afe_priv->clk[CLK_APLL22M]);
 	if (ret) {
@@ -405,6 +407,8 @@ int mt8183_apll1_enable(struct mtk_base_afe *afe)
 ERR_CLK_APLL1_TUNER:
 	clk_disable_unprepare(afe_priv->clk[CLK_APLL22M]);
 ERR_CLK_APLL22M:
+	apll1_mux_setting(afe, false);
+
 	return ret;
 }
 
@@ -430,7 +434,9 @@ int mt8183_apll2_enable(struct mtk_base_afe *afe)
 	int ret;
 
 	/* setting for APLL */
-	apll2_mux_setting(afe, true);
+	ret = apll2_mux_setting(afe, true);
+	if (ret)
+		return ret;
 
 	ret = clk_prepare_enable(afe_priv->clk[CLK_APLL24M]);
 	if (ret) {
@@ -459,6 +465,8 @@ int mt8183_apll2_enable(struct mtk_base_afe *afe)
 ERR_CLK_APLL2_TUNER:
 	clk_disable_unprepare(afe_priv->clk[CLK_APLL24M]);
 ERR_CLK_APLL24M:
+	apll2_mux_setting(afe, false);
+
 	return ret;
 }
 
Signature · 1 line
2.53.0
author maintainer
#1619 lore raw

PATCH 1616/2214 dt-bindings: vendor-prefixes: Document Chongzhou

In reply to Greg Kroah-Hartman

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

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

From: Jagan Teki <[email protected]>

[ Upstream commit 90f86d0c617d9461cb00f4d8e861eda28011d46e ]

Chongzhou is a professional is a manufacturer of LCD panels from Shenzhen.

Add vendor prefix for it.

Stable-dep-of: 058ee00216a1 ("dt-bindings: vendor-prefixes: Add Chip Wealth Technology")

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index c5fabb4692902..9cb4c3b95bd1b 100644
@@ -229,6 +229,8 @@ patternProperties:
     description: ChipOne
   "^chipspark,.*":
     description: ChipSPARK
+  "^chongzhou,.*":
+    description: Shenzhen Chongzhou Electronic Technology Co., Ltd
   "^chrontel,.*":
     description: Chrontel, Inc.
   "^chrp,.*":
Signature · 1 line
2.53.0
author maintainer
#1620 lore raw

PATCH 1617/2214 dt-bindings: vendor-prefixes: Add Chip Wealth Technology

In reply to Greg Kroah-Hartman

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

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

From: Aaron Kling <[email protected]>

[ Upstream commit 058ee00216a1356e710446b59a800065d2abbd80 ]

Add an entry for Chip Wealth Technology (https://www.chipwealth.com/)

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index 9cb4c3b95bd1b..aa06eff78e45b 100644
@@ -229,6 +229,8 @@ patternProperties:
     description: ChipOne
   "^chipspark,.*":
     description: ChipSPARK
+  "^chipwealth,.*":
+    description: Chip Wealth Technology
   "^chongzhou,.*":
     description: Shenzhen Chongzhou Electronic Technology Co., Ltd
   "^chrontel,.*":
Signature · 1 line
2.53.0
author maintainer
#1621 lore raw

PATCH 1618/2214 hfs/hfsplus: serialize B-tree close against folio release

In reply to Greg Kroah-Hartman

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

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

From: Bruno Produit <[email protected]>

[ Upstream commit dce0e0248205bb3dc5b4c6793cd2ecb453152259 ]

B-tree nodes with a zero reference count remain in the node hash until folio reclaim or tree teardown frees them. The folio release callbacks remove nodes while holding hash_lock, but hfs_btree_close() walks and frees the same hash without that lock. Reclaim can therefore unhash and free a node after close has loaded its pointer, causing a use-after-free or double-free.

The following syzkaller crashes seem to be the same UAF in {hfs,hfsplus}_btree_close() (or {hfs,hfsplus}_bnode_unhash()) and follow the same pattern, but do not contain a reproducer to confirm.

Detach each node with hfs_bnode_unhash() while holding hash_lock before inspecting and freeing it. Drop the lock before hfs_bnode_free() so a large tree is not freed while holding a spinlock. Apply the same fix to the matching HFS+ implementation.

---
 fs/hfs/btree.c     | 7 +++++--
 fs/hfsplus/btree.c | 7 +++++--
 2 files changed, 10 insertions(+), 4 deletions(-)
fs/hfs/btree.c +5 -2
index f9ccfe0fc5748..5553de925448f 100644
@@ -143,15 +143,18 @@ void hfs_btree_close(struct hfs_btree *tree)
 		return;
 
 	for (i = 0; i < NODE_HASH_SIZE; i++) {
+		spin_lock(&tree->hash_lock);
 		while ((node = tree->node_hash[i])) {
-			tree->node_hash[i] = node->next_hash;
+			hfs_bnode_unhash(node);
+			spin_unlock(&tree->hash_lock);
 			if (atomic_read(&node->refcnt))
 				pr_err("node %d:%d still has %d user(s)!\n",
 				       node->tree->cnid, node->this,
 				       atomic_read(&node->refcnt));
 			hfs_bnode_free(node);
-			tree->node_hash_cnt--;
+			spin_lock(&tree->hash_lock);
 		}
+		spin_unlock(&tree->hash_lock);
 	}
 	iput(tree->inode);
 	kfree(tree);
fs/hfsplus/btree.c +5 -2
index d4b2591046474..926398b135929 100644
@@ -264,16 +264,19 @@ void hfs_btree_close(struct hfs_btree *tree)
 		return;
 
 	for (i = 0; i < NODE_HASH_SIZE; i++) {
+		spin_lock(&tree->hash_lock);
 		while ((node = tree->node_hash[i])) {
-			tree->node_hash[i] = node->next_hash;
+			hfs_bnode_unhash(node);
+			spin_unlock(&tree->hash_lock);
 			if (atomic_read(&node->refcnt))
 				pr_crit("node %d:%d "
 						"still has %d user(s)!\n",
 					node->tree->cnid, node->this,
 					atomic_read(&node->refcnt));
 			hfs_bnode_free(node);
-			tree->node_hash_cnt--;
+			spin_lock(&tree->hash_lock);
 		}
+		spin_unlock(&tree->hash_lock);
 	}
 	iput(tree->inode);
 	kfree(tree);
Signature · 1 line
2.53.0
author maintainer
#1622 lore raw

PATCH 1619/2214 dt-bindings: vendor-prefixes: Add LeEco

In reply to Greg Kroah-Hartman

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

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

From: Adam Skladowski <[email protected]>

[ Upstream commit 1658af4d25bd3133d1b3403ccdda6214d40c0ebb ]

Add vendor prefix for now dead ecosystem company LeEco

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index aa06eff78e45b..2b4e259fd816a 100644
@@ -651,6 +651,8 @@ patternProperties:
     description: Lattice Semiconductor
   "^leadtek,.*":
     description: Shenzhen Leadtek Technology Co., Ltd.
+  "^leeco,.*":
+    description: Leshi Holding Beijing.
   "^leez,.*":
     description: Leez
   "^lego,.*":
Signature · 1 line
2.53.0
author maintainer
#1623 lore raw

PATCH 1620/2214 dt-bindings: vendor-prefixes: Add Retronix Technology Inc.

In reply to Greg Kroah-Hartman

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

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

From: Marek Vasut <[email protected]>

[ Upstream commit da076d5562d2fc99138ff8ab04ff82eee7a4a3d8 ]

Add vendor prefix for Retronix Technology Inc.
https://www.retronix.com.tw/en/about.html

Stable-dep-of: ea55bba4c68a ("dt-bindings: vendor-prefixes: Add Retroid Pocket")

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index 2b4e259fd816a..e70b0c59a0806 100644
@@ -1004,6 +1004,8 @@ patternProperties:
     description: iMX6 Rex Project
   "^rervision,.*":
     description: Shenzhen Rervision Technology Co., Ltd.
+  "^retronix,.*":
+    description: Retronix Technology Inc.
   "^revotics,.*":
     description: Revolution Robotics, Inc. (Revotics)
   "^richtek,.*":
Signature · 1 line
2.53.0
author maintainer
#1624 lore raw

PATCH 1621/2214 dt-bindings: vendor-prefixes: Add Retroid Pocket

In reply to Greg Kroah-Hartman

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

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

From: Aaron Kling <[email protected]>

[ Upstream commit ea55bba4c68a49fcb527da6c71dd48879b705fd1 ]

Add an entry for Retroid Pocket (https://www.goretroid.com/)

---
 Documentation/devicetree/bindings/vendor-prefixes.yaml | 2 ++
 1 file changed, 2 insertions(+)
Documentation/devicetree/bindings/vendor-prefixes.yaml +2 -0
index e70b0c59a0806..afe41204f4a56 100644
@@ -1004,6 +1004,8 @@ patternProperties:
     description: iMX6 Rex Project
   "^rervision,.*":
     description: Shenzhen Rervision Technology Co., Ltd.
+  "^retroidpocket,.*":
+    description: Retroid Pocket
   "^retronix,.*":
     description: Retronix Technology Inc.
   "^revotics,.*":
Signature · 1 line
2.53.0
author maintainer
#1625 lore raw

PATCH 1622/2214 MIPS: Octeon: size the swiotlb for a modular OHCI platform driver

In reply to Greg Kroah-Hartman

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

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

From: Orgad Shaneh <[email protected]>

[ Upstream commit f1dd1f0c8d595a019be503d40060d25721313b4a ]

plat_swiotlb_setup() raises the swiotlb to 64 MB on OCTEON II with memory above 4 GB, because that OHCI can only DMA to 32 bits. The test is "#ifdef CONFIG_USB_OHCI_HCD_PLATFORM", which is false when the driver is a module - the usual way to build it - so the bump is silently skipped.

What is left depends on the rest of the function: with PCI enabled the "memory past the BAR1 hole" branch happens to ask for the same 64 MB, but with CONFIG_PCI off - which is normal on these SoCs when no PCI devices are wired - swiotlbsize stays at the PAGE_SIZE it was initialised to. A full- or low-speed device on an OCTEON II board with more than 4 GB of RAM then has no bounce buffer to speak of.

Use IS_ENABLED() so a modular OHCI gets the same sizing as a built-in one. Seen on a CN6635 board with 8 GB of RAM and CONFIG_USB_OHCI_HCD_PLATFORM=m.

Fixes a8667d706dfa ("MIPS: OCTEON: dma-octeon: fix OHCI USB config check")
Assisted-by Claude:claude-opus-5
---
 arch/mips/cavium-octeon/dma-octeon.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/mips/cavium-octeon/dma-octeon.c +1 -1
index df70308db0e69..8a9bf803c84ae 100644
@@ -231,7 +231,7 @@ void __init plat_swiotlb_setup(void)
 		swiotlbsize = 64 * (1<<20);
 	}
 #endif
-#ifdef CONFIG_USB_OHCI_HCD_PLATFORM
+#if IS_ENABLED(CONFIG_USB_OHCI_HCD_PLATFORM)
 	/* OCTEON II ohci is only 32-bit. */
 	if (OCTEON_IS_OCTEON2() && max_addr >= 0x100000000ul)
 		swiotlbsize = 64 * (1<<20);
Signature · 1 line
2.53.0
author maintainer
#1626 lore raw

PATCH 1623/2214 MIPS: Only check -msym32 when need-compiler

In reply to Greg Kroah-Hartman

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

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

From: WangYuli <[email protected]>

[ Upstream commit 7a8f286c235f5d308208cfdcb025b2a0e0c84742 ]

During 'make modules_install', the need-compiler variable is null, so Makefile.compiler isn't included. This results in call cc-option-yn returning nothing and consequently an error is issued:

*** CONFIG_CPU_DADDI_WORKAROUNDS unsupported without -msym32. Stop.

and the invocation terminated for configurations that set the Kconfig option named.

Previously commit a79a404e6c22 ("MIPS: Fix CONFIG_CPU_DADDI_WORKAROUNDS `modules_install' regression") addressed the same issue for GCC versions now supported for Linux compilation, but caused a build error with Clang because it doesn't support '-msym32'. Then commit 18ca63a2e23c ("MIPS:
Probe toolchain support of -msym32") fixed the issue with Clang, but reintroduced the 'make modules_install' error listed above.

Wrap the handling of the '-msym32' option then with ifdef need-compiler, fixing said error for 'make modules_install'.

For more technical details on why need-compiler is null during 'make modules_install' and why no compiler invocation is actually needed at this point, please refer to commit 4fe4a6374c4d ("MIPS: Only fiddle with CHECKFLAGS if need-compiler") and commit 805b2e1d427a ("kbuild: include Makefile.compiler only when compiler is needed").

NB leading 8 spaces rather than tabs required due to `make' syntax.

---
 arch/mips/Makefile | 22 ++++++++++++----------
 1 file changed, 12 insertions(+), 10 deletions(-)
arch/mips/Makefile +12 -10
index 8d1b1d7b8e025..20cdff5f457b8 100644
@@ -272,18 +272,20 @@ drivers-$(CONFIG_PCI)		+= arch/mips/pci/
 # We can always force a build with a 64-bits symbol format by
 # passing 'KBUILD_SYM32=no' option to the make's command line.
 #
-ifdef CONFIG_64BIT
-  ifndef KBUILD_SYM32
-    ifeq ($(shell expr $(load-y) \< 0xffffffff80000000), 0)
-      KBUILD_SYM32 = $(call cc-option-yn, -msym32)
+ifdef need-compiler
+  ifdef CONFIG_64BIT
+    ifndef KBUILD_SYM32
+      ifeq ($(shell expr $(load-y) \< 0xffffffff80000000), 0)
+        KBUILD_SYM32 = $(call cc-option-yn, -msym32)
+      endif
     endif
-  endif
 
-  ifeq ($(KBUILD_SYM32), y)
-    cflags-$(KBUILD_SYM32) += -msym32 -DKBUILD_64BIT_SYM32
-  else
-    ifeq ($(CONFIG_CPU_DADDI_WORKAROUNDS), y)
-      $(error CONFIG_CPU_DADDI_WORKAROUNDS unsupported without -msym32)
+    ifeq ($(KBUILD_SYM32), y)
+      cflags-$(KBUILD_SYM32) += -msym32 -DKBUILD_64BIT_SYM32
+    else
+      ifeq ($(CONFIG_CPU_DADDI_WORKAROUNDS), y)
+        $(error CONFIG_CPU_DADDI_WORKAROUNDS unsupported without -msym32)
+      endif
     endif
   endif
 endif
Signature · 1 line
2.53.0
author maintainer
#1627 lore raw

PATCH 1624/2214 m68k: apollo: Fix interrupt setup

In reply to Greg Kroah-Hartman

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

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

From: Daniel Palmer <[email protected]>

[ Upstream commit 6cdbe63badcb524c88f2c8b3cee08f2ffc099a1b ]

The user interrupt vector needs to be installed before the interrupts are unmasked otherwise the timer tick isn't handled and boot up gets stuck in calibrate_delay().

Assisted-by Claude:claude-fable-5 # Wrote Apollo DN3000 emulator using QEMU
Fixes 0aa78106c661cf6c ("[PATCH] m68k: convert apollo irq code")
---
 arch/m68k/apollo/dn_ints.c | 6 ++++++
 1 file changed, 6 insertions(+)
arch/m68k/apollo/dn_ints.c +6 -0
index 02cff7efc8342..e62c96c5b0dd6 100644
@@ -2,6 +2,7 @@
 #include <linux/interrupt.h>
 #include <linux/irq.h>
 
+#include <asm/irq.h>
 #include <asm/traps.h>
 #include <asm/apollohw.h>
 
@@ -9,10 +10,13 @@ unsigned int apollo_irq_startup(struct irq_data *data)
 {
 	unsigned int irq = data->irq;
 
+	m68k_irq_startup(data);
+
 	if (irq < 8)
 		*(volatile unsigned char *)(pica+1) &= ~(1 << irq);
 	else
 		*(volatile unsigned char *)(picb+1) &= ~(1 << (irq - 8));
+
 	return 0;
 }
 
@@ -24,6 +28,8 @@ void apollo_irq_shutdown(struct irq_data *data)
 		*(volatile unsigned char *)(pica+1) |= (1 << irq);
 	else
 		*(volatile unsigned char *)(picb+1) |= (1 << (irq - 8));
+
+	m68k_irq_shutdown(data);
 }
 
 void apollo_irq_eoi(struct irq_data *data)
Signature · 1 line
2.53.0
author maintainer
#1628 lore raw

PATCH 1625/2214 drm/msm/a6xx: Fix secondary rail vote in ARC votes

In reply to Greg Kroah-Hartman

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

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

From: Akhil P Oommen <[email protected]>

[ Upstream commit 128a0edde50780ed9eb93ab5d2e18ca24c1f1c43 ]

Using the requested level can pick a secondary corner lower than the one required to support the chosen primary corner, under-voting the dependent rail. To fix this, match against the selected primary voltage instead.

Fixes 4b565ca5a2cb ("drm/msm: Add A6XX device support")

Patchwork: https://patchwork.freedesktop.org/patch/755592/
Message-ID: <[email protected]>

---
 drivers/gpu/drm/msm/adreno/a6xx_gmu.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/gpu/drm/msm/adreno/a6xx_gmu.c +5 -1
index 8fa2b9e051002..a5aa7750a280d 100644
@@ -1304,10 +1304,14 @@ static int a6xx_gmu_rpmh_arc_votes_init(struct device *dev, u32 *votes,
 		/*
 		 * Look for a level in in the secondary list that matches. If
 		 * nothing fits, use the maximum non zero vote
+		 *
+		 * The secondary rail depends on the primary rail, so match it
+		 * against the quantized primary voltage (which is >= the
+		 * requested level), not the requested level itself.
 		 */
 
 		for (j = 0; j < sec_count; j++) {
-			if (sec[j] >= level) {
+			if (sec[j] >= pri[pindex]) {
 				sindex = j;
 				break;
 			} else if (sec[j]) {
Signature · 1 line
2.53.0
author maintainer
#1629 lore raw

PATCH 1626/2214 netfilter: synproxy: fix reset of ct seqadj when reopening a connection

In reply to Greg Kroah-Hartman

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

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

From: Fernando Fernandez Mancera <[email protected]>

[ Upstream commit a8708aa9fe1095ca7859cb3f102b6a3fa0fb98bd ]

SYNPROXY is resetting conntrack seqadj when a closed connection is re-opened, but it was using nf_ct_seqadj_init() which is a no-op for a zero offset.

This patch introduces nf_ct_seqadj_reset() which sets the offset values directly to zero and avoid setting IPS_SEQ_ADJUST_BIT flag, it changes SYNPROXY code to use it when needed.

Fixes 48b1de4c110a ("netfilter: add SYNPROXY core/target")
---
 include/net/netfilter/nf_conntrack_seqadj.h |  1 +
 net/netfilter/nf_conntrack_seqadj.c         | 17 +++++++++++++++++
 net/netfilter/nf_synproxy_core.c            |  4 ++--
 3 files changed, 20 insertions(+), 2 deletions(-)
include/net/netfilter/nf_conntrack_seqadj.h +1 -0
index 0a10b50537aee..df90bec7b7670 100644
@@ -33,6 +33,7 @@ static inline struct nf_conn_seqadj *nfct_seqadj_ext_add(struct nf_conn *ct)
 
 int nf_ct_seqadj_init(struct nf_conn *ct, enum ip_conntrack_info ctinfo,
 		      s32 off);
+void nf_ct_seqadj_reset(struct nf_conn *ct, enum ip_conntrack_info ctinfo);
 int nf_ct_seqadj_set(struct nf_conn *ct, enum ip_conntrack_info ctinfo,
 		     __be32 seq, s32 off);
 void nf_ct_tcp_seqadj_set(struct sk_buff *skb, struct nf_conn *ct,
net/netfilter/nf_conntrack_seqadj.c +17 -0
index bcb4adf9879dc..f570060035f71 100644
@@ -29,6 +29,23 @@ int nf_ct_seqadj_init(struct nf_conn *ct, enum ip_conntrack_info ctinfo,
 }
 EXPORT_SYMBOL_GPL(nf_ct_seqadj_init);
 
+void nf_ct_seqadj_reset(struct nf_conn *ct, enum ip_conntrack_info ctinfo)
+{
+	struct nf_conn_seqadj *seqadj = nfct_seqadj(ct);
+	enum ip_conntrack_dir dir = CTINFO2DIR(ctinfo);
+	struct nf_ct_seqadj *this_way;
+
+	if (unlikely(!seqadj))
+		return;
+
+	spin_lock_bh(&ct->lock);
+	this_way = &seqadj->seq[dir];
+	this_way->offset_before	 = 0;
+	this_way->offset_after	 = 0;
+	spin_unlock_bh(&ct->lock);
+}
+EXPORT_SYMBOL_GPL(nf_ct_seqadj_reset);
+
 int nf_ct_seqadj_set(struct nf_conn *ct, enum ip_conntrack_info ctinfo,
 		     __be32 seq, s32 off)
 {
net/netfilter/nf_synproxy_core.c +2 -2
index 91d3e3eea1be0..379f26277ca4f 100644
@@ -709,7 +709,7 @@ ipv4_synproxy_hook(void *priv, struct sk_buff *skb,
 		 * adjustments, they will get initialized once the connection is
 		 * reestablished.
 		 */
-		nf_ct_seqadj_init(ct, ctinfo, 0);
+		nf_ct_seqadj_reset(ct, ctinfo);
 		synproxy->tsoff = 0;
 		this_cpu_inc(snet->stats->conn_reopened);
 		fallthrough;
@@ -1137,7 +1137,7 @@ ipv6_synproxy_hook(void *priv, struct sk_buff *skb,
 		 * adjustments, they will get initialized once the connection is
 		 * reestablished.
 		 */
-		nf_ct_seqadj_init(ct, ctinfo, 0);
+		nf_ct_seqadj_reset(ct, ctinfo);
 		synproxy->tsoff = 0;
 		this_cpu_inc(snet->stats->conn_reopened);
 		fallthrough;
Signature · 1 line
2.53.0
author maintainer
#1630 lore raw

PATCH 1627/2214 ARM: dts: gemini: Use a level-high interrupt for TVE200

In reply to Greg Kroah-Hartman

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

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

From: Linus Walleij <[email protected]>

[ Upstream commit 5bf1bb010712bba28eda926ed453db3cab91381a ]

TVE200's interrupt is configured as rising-edge. On the D-Link DIR-685, this can stop interrupt delivery after the first vertical interrupt, causing vblank and page-flip timeouts and delaying framebuffer console startup.

The vertical interrupt can remain pending after acknowledgment. Use level-high triggering so the pending condition is delivered again without requiring another edge.

Fixes d2b85241a9f5 ("ARM: dts: Add TVE200 to the Gemini SoC DTSI")
Assisted-by LLM

Message-ID: <[email protected]>

---
 arch/arm/boot/dts/gemini.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/gemini.dtsi +1 -1
index cc053af3c3472..8fe76b520b03f 100644
@@ -411,7 +411,7 @@ dma-controller@67000000 {
 		display-controller@6a000000 {
 			compatible = "cortina,gemini-tvc", "faraday,tve200";
 			reg = <0x6a000000 0x1000>;
-			interrupts = <13 IRQ_TYPE_EDGE_RISING>;
+			interrupts = <13 IRQ_TYPE_LEVEL_HIGH>;
 			resets = <&syscon GEMINI_RESET_TVC>;
 			clocks = <&syscon GEMINI_CLK_GATE_TVC>,
 				 <&syscon GEMINI_CLK_TVC>;
Signature · 1 line
2.53.0