6.19.4-rc1 review

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

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

From: Aleksander Jan Bajkowski <[email protected]>

[ Upstream commit 0ceeadc7b53a041d89d5843f6bf0ccb7c98b0b4f ]

EIP93 has an options register. This register indicates which crypto algorithms are implemented in silicon. Supported algorithms are registered on this basis. Unregister algorithms on the same basis. Currently, all algorithms are unregistered, even those not supported by HW. This results in panic on platforms that don't have all options implemented in silicon.

---
 .../crypto/inside-secure/eip93/eip93-main.c   | 92 +++++++++++--------
 1 file changed, 53 insertions(+), 39 deletions(-)

index 3cdc3308dcac8..b7fd9795062d4 100644
@@ -77,11 +77,44 @@ inline void eip93_irq_clear(struct eip93_device *eip93, u32 mask)
 	__raw_writel(mask, eip93->base + EIP93_REG_INT_CLR);
 }
 
-static void eip93_unregister_algs(unsigned int i)
+static int eip93_algo_is_supported(u32 alg_flags, u32 supported_algo_flags)
+{
+	if ((IS_DES(alg_flags) || IS_3DES(alg_flags)) &&
+	    !(supported_algo_flags & EIP93_PE_OPTION_TDES))
+		return 0;
+
+	if (IS_AES(alg_flags) &&
+	    !(supported_algo_flags & EIP93_PE_OPTION_AES))
+		return 0;
+
+	if (IS_HASH_MD5(alg_flags) &&
+	    !(supported_algo_flags & EIP93_PE_OPTION_MD5))
+		return 0;
+
+	if (IS_HASH_SHA1(alg_flags) &&
+	    !(supported_algo_flags & EIP93_PE_OPTION_SHA_1))
+		return 0;
+
+	if (IS_HASH_SHA224(alg_flags) &&
+	    !(supported_algo_flags & EIP93_PE_OPTION_SHA_224))
+		return 0;
+
+	if (IS_HASH_SHA256(alg_flags) &&
+	    !(supported_algo_flags & EIP93_PE_OPTION_SHA_256))
+		return 0;
+
+	return 1;
+}
+
+static void eip93_unregister_algs(u32 supported_algo_flags, unsigned int i)
 {
 	unsigned int j;
 
 	for (j = 0; j < i; j++) {
+		if (!eip93_algo_is_supported(eip93_algs[j]->flags,
+					     supported_algo_flags))
+			continue;
+
 		switch (eip93_algs[j]->type) {
 		case EIP93_ALG_TYPE_SKCIPHER:
 			crypto_unregister_skcipher(&eip93_algs[j]->alg.skcipher);

@@ -106,49 +139,27 @@ static int eip93_register_algs(struct eip93_device *eip93, u32 supported_algo_fl
 
 		eip93_algs[i]->eip93 = eip93;
 
-		if ((IS_DES(alg_flags) || IS_3DES(alg_flags)) &&
-		    !(supported_algo_flags & EIP93_PE_OPTION_TDES))
+		if (!eip93_algo_is_supported(alg_flags, supported_algo_flags))
 			continue;
 
-		if (IS_AES(alg_flags)) {
-			if (!(supported_algo_flags & EIP93_PE_OPTION_AES))
-				continue;
+		if (IS_AES(alg_flags) && !IS_HMAC(alg_flags)) {
+			if (supported_algo_flags & EIP93_PE_OPTION_AES_KEY128)
+				eip93_algs[i]->alg.skcipher.max_keysize =
+					AES_KEYSIZE_128;
 
-			if (!IS_HMAC(alg_flags)) {
-				if (supported_algo_flags & EIP93_PE_OPTION_AES_KEY128)
-					eip93_algs[i]->alg.skcipher.max_keysize =
-						AES_KEYSIZE_128;
+			if (supported_algo_flags & EIP93_PE_OPTION_AES_KEY192)
+				eip93_algs[i]->alg.skcipher.max_keysize =
+					AES_KEYSIZE_192;
 
-				if (supported_algo_flags & EIP93_PE_OPTION_AES_KEY192)
-					eip93_algs[i]->alg.skcipher.max_keysize =
-						AES_KEYSIZE_192;
+			if (supported_algo_flags & EIP93_PE_OPTION_AES_KEY256)
+				eip93_algs[i]->alg.skcipher.max_keysize =
+					AES_KEYSIZE_256;
 
-				if (supported_algo_flags & EIP93_PE_OPTION_AES_KEY256)
-					eip93_algs[i]->alg.skcipher.max_keysize =
-						AES_KEYSIZE_256;
-
-				if (IS_RFC3686(alg_flags))
-					eip93_algs[i]->alg.skcipher.max_keysize +=
-						CTR_RFC3686_NONCE_SIZE;
-			}
+			if (IS_RFC3686(alg_flags))
+				eip93_algs[i]->alg.skcipher.max_keysize +=
+					CTR_RFC3686_NONCE_SIZE;
 		}
 
-		if (IS_HASH_MD5(alg_flags) &&
-		    !(supported_algo_flags & EIP93_PE_OPTION_MD5))
-			continue;
-
-		if (IS_HASH_SHA1(alg_flags) &&
-		    !(supported_algo_flags & EIP93_PE_OPTION_SHA_1))
-			continue;
-
-		if (IS_HASH_SHA224(alg_flags) &&
-		    !(supported_algo_flags & EIP93_PE_OPTION_SHA_224))
-			continue;
-
-		if (IS_HASH_SHA256(alg_flags) &&
-		    !(supported_algo_flags & EIP93_PE_OPTION_SHA_256))
-			continue;
-
 		switch (eip93_algs[i]->type) {
 		case EIP93_ALG_TYPE_SKCIPHER:
 			ret = crypto_register_skcipher(&eip93_algs[i]->alg.skcipher);

@@ -167,7 +178,7 @@ static int eip93_register_algs(struct eip93_device *eip93, u32 supported_algo_fl
 	return 0;
 
 fail:
-	eip93_unregister_algs(i);
+	eip93_unregister_algs(supported_algo_flags, i);
 
 	return ret;
 }

@@ -469,8 +480,11 @@ static int eip93_crypto_probe(struct platform_device *pdev)
 static void eip93_crypto_remove(struct platform_device *pdev)
 {
 	struct eip93_device *eip93 = platform_get_drvdata(pdev);
+	u32 algo_flags;
+
+	algo_flags = readl(eip93->base + EIP93_REG_PE_OPTION_1);
 
-	eip93_unregister_algs(ARRAY_SIZE(eip93_algs));
+	eip93_unregister_algs(algo_flags, ARRAY_SIZE(eip93_algs));
 	eip93_cleanup(eip93);
 }
 

2.51.0

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

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

From: Jiri Olsa <[email protected]>

[ Upstream commit 8bc11700e0d23d4fdb7d8d5a73b2e95de427cabc ]

The previous change (Fixes commit) messed up the rsp register value, which is wrong because it's already adjusted with FRAME_SIZE, we need the original rsp value.

This change does not affect fprobe current kernel unwind, the !perf_hw_regs path perf_callchain_kernel:

        if (perf_hw_regs(regs)) {
                if (perf_callchain_store(entry, regs->ip))
                        return;
                unwind_start(&state, current, regs, NULL);
        } else {
                unwind_start(&state, current, NULL, (void *)regs->sp);
        }

which uses pt_regs.sp as first_frame boundary (FRAME_SIZE shift makes no difference, unwind stil stops at the right frame).

This change fixes the other path when we want to unwind directly from pt_regs sp/fp/ip state, which is coming in following change.

---
 arch/x86/kernel/ftrace_64.S | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

index a132608265f6c..62c1c93aa1c6a 100644
@@ -364,6 +364,9 @@ SYM_CODE_START(return_to_handler)
 	UNWIND_HINT_UNDEFINED
 	ANNOTATE_NOENDBR
 
+	/* Store original rsp for pt_regs.sp value. */
+	movq %rsp, %rdi
+
 	/* Restore return_to_handler value that got eaten by previous ret instruction. */
 	subq $8, %rsp
 	UNWIND_HINT_FUNC

@@ -374,7 +377,7 @@ SYM_CODE_START(return_to_handler)
 	movq %rax, RAX(%rsp)
 	movq %rdx, RDX(%rsp)
 	movq %rbp, RBP(%rsp)
-	movq %rsp, RSP(%rsp)
+	movq %rdi, RSP(%rsp)
 	movq %rsp, %rdi
 
 	call ftrace_return_to_handler

2.51.0

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

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

From: Jiri Olsa <[email protected]>

[ Upstream commit aea251799998aa1b78eacdfb308f18ea114ea5b3 ]

Mahe reported missing function from stack trace on top of kprobe multi program. The missing function is the very first one in the stacktrace, the one that the bpf program is attached to.

  # bpftrace -e 'kprobe:__x64_sys_newuname* { print(kstack)}'
  Attaching 1 probe...

        do_syscall_64+134
        entry_SYSCALL_64_after_hwframe+118

('*' is used for kprobe_multi attachment)

The reason is that the previous change (the Fixes commit) fixed stack unwind for tracepoint, but removed attached function address from the stack trace on top of kprobe multi programs, which I also overlooked in the related test (check following patch).

The tracepoint and kprobe_multi have different stack setup, but use same unwind path. I think it's better to keep the previous change, which fixed tracepoint unwind and instead change the kprobe multi unwind as explained below.

The bpf program stack unwind calls perf_callchain_kernel for kernel portion and it follows two unwind paths based on X86_EFLAGS_FIXED bit in pt_regs.flags.

When the bit set we unwind from stack represented by pt_regs argument, otherwise we unwind currently executed stack up to 'first_frame' boundary.

The 'first_frame' value is taken from regs.rsp value, but ftrace_caller and ftrace_regs_caller (ftrace trampoline) functions set the regs.rsp to the previous stack frame, so we skip the attached function entry.

If we switch kprobe_multi unwind to use the X86_EFLAGS_FIXED bit, we set the start of the unwind to the attached function address. As another benefit we also cut extra unwind cycles needed to reach the 'first_frame' boundary.

The speedup can be measured with trigger bench for kprobe_multi program and stacktrace support.

- trigger bench with stacktrace on current code:

        kprobe-multi   :     0.810 ± 0.001M/s
        kretprobe-multi:     0.808 ± 0.001M/s

- and with the fix:

        kprobe-multi   :     1.264 ± 0.001M/s
        kretprobe-multi:     1.401 ± 0.002M/s

With the fix, the entry probe stacktrace:

  # bpftrace -e 'kprobe:__x64_sys_newuname* { print(kstack)}'
  Attaching 1 probe...

        __x64_sys_newuname+9
        do_syscall_64+134
        entry_SYSCALL_64_after_hwframe+118

The return probe skips the attached function, because it's no longer on the stack at the point of the unwind and this way is the same how standard kretprobe works.

  # bpftrace -e 'kretprobe:__x64_sys_newuname* { print(kstack)}'
  Attaching 1 probe...

        do_syscall_64+134
        entry_SYSCALL_64_after_hwframe+118

---
 arch/x86/include/asm/ftrace.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index b08c95872eed9..c56e1e63b8932 100644
@@ -57,7 +57,7 @@ arch_ftrace_get_regs(struct ftrace_regs *fregs)
 }
 
 #define arch_ftrace_partial_regs(regs) do {	\
-	regs->flags &= ~X86_EFLAGS_FIXED;	\
+	regs->flags |= X86_EFLAGS_FIXED;	\
 	regs->cs = __KERNEL_CS;			\
 } while (0)
 

2.51.0

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

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

From: Jiri Olsa <[email protected]>

[ Upstream commit 0207f94971e72a13380e28022c86da147e8e090f ]

We now include the attached function in the stack trace, fixing the test accordingly.

---
 .../selftests/bpf/prog_tests/stacktrace_ips.c | 19 ++++++++++++++-----
 1 file changed, 14 insertions(+), 5 deletions(-)

index c9efdd2a5b18a..c93718dafd9b6 100644
@@ -74,11 +74,20 @@ static void test_stacktrace_ips_kprobe_multi(bool retprobe)
 
 	load_kallsyms();
 
-	check_stacktrace_ips(bpf_map__fd(skel->maps.stackmap), skel->bss->stack_key, 4,
-			     ksym_get_addr("bpf_testmod_stacktrace_test_3"),
-			     ksym_get_addr("bpf_testmod_stacktrace_test_2"),
-			     ksym_get_addr("bpf_testmod_stacktrace_test_1"),
-			     ksym_get_addr("bpf_testmod_test_read"));
+	if (retprobe) {
+		check_stacktrace_ips(bpf_map__fd(skel->maps.stackmap), skel->bss->stack_key, 4,
+				     ksym_get_addr("bpf_testmod_stacktrace_test_3"),
+				     ksym_get_addr("bpf_testmod_stacktrace_test_2"),
+				     ksym_get_addr("bpf_testmod_stacktrace_test_1"),
+				     ksym_get_addr("bpf_testmod_test_read"));
+	} else {
+		check_stacktrace_ips(bpf_map__fd(skel->maps.stackmap), skel->bss->stack_key, 5,
+				     ksym_get_addr("bpf_testmod_stacktrace_test"),
+				     ksym_get_addr("bpf_testmod_stacktrace_test_3"),
+				     ksym_get_addr("bpf_testmod_stacktrace_test_2"),
+				     ksym_get_addr("bpf_testmod_stacktrace_test_1"),
+				     ksym_get_addr("bpf_testmod_test_read"));
+	}
 
 cleanup:
 	stacktrace_ips__destroy(skel);

2.51.0

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

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

From: Weili Qian <[email protected]>

[ Upstream commit 3d3135057ff567d5c09fff4c9ef6391a684e8042 ]

Since the number of devices is limited, and the number of tfms may exceed the number of devices, to ensure that tfms can be successfully allocated, support tfms sharing the same device.

---
 drivers/crypto/hisilicon/trng/trng.c | 121 +++++++++++++++++++--------
 1 file changed, 86 insertions(+), 35 deletions(-)

index ac74df4a94712..5ca0b90859a81 100644
@@ -40,6 +40,7 @@
 #define SEED_SHIFT_24		24
 #define SEED_SHIFT_16		16
 #define SEED_SHIFT_8		8
+#define SW_MAX_RANDOM_BYTES	65520
 
 struct hisi_trng_list {
 	struct mutex lock;

@@ -53,8 +54,10 @@ struct hisi_trng {
 	struct list_head list;
 	struct hwrng rng;
 	u32 ver;
-	bool is_used;
-	struct mutex mutex;
+	u32 ctx_num;
+	/* The bytes of the random number generated since the last seeding. */
+	u32 random_bytes;
+	struct mutex lock;
 };
 
 struct hisi_trng_ctx {

@@ -63,10 +66,14 @@ struct hisi_trng_ctx {
 
 static atomic_t trng_active_devs;
 static struct hisi_trng_list trng_devices;
+static int hisi_trng_read(struct hwrng *rng, void *buf, size_t max, bool wait);
 
-static void hisi_trng_set_seed(struct hisi_trng *trng, const u8 *seed)
+static int hisi_trng_set_seed(struct hisi_trng *trng, const u8 *seed)
 {
 	u32 val, seed_reg, i;
+	int ret;
+
+	writel(0x0, trng->base + SW_DRBG_BLOCKS);
 
 	for (i = 0; i < SW_DRBG_SEED_SIZE;
 	     i += SW_DRBG_SEED_SIZE / SW_DRBG_SEED_REGS_NUM) {

@@ -78,6 +85,20 @@ static void hisi_trng_set_seed(struct hisi_trng *trng, const u8 *seed)
 		seed_reg = (i >> SW_DRBG_NUM_SHIFT) % SW_DRBG_SEED_REGS_NUM;
 		writel(val, trng->base + SW_DRBG_SEED(seed_reg));
 	}
+
+	writel(SW_DRBG_BLOCKS_NUM | (0x1 << SW_DRBG_ENABLE_SHIFT),
+	       trng->base + SW_DRBG_BLOCKS);
+	writel(0x1, trng->base + SW_DRBG_INIT);
+	ret = readl_relaxed_poll_timeout(trng->base + SW_DRBG_STATUS,
+					 val, val & BIT(0), SLEEP_US, TIMEOUT_US);
+	if (ret) {
+		pr_err("failed to init trng(%d)\n", ret);
+		return -EIO;
+	}
+
+	trng->random_bytes = 0;
+
+	return 0;
 }
 
 static int hisi_trng_seed(struct crypto_rng *tfm, const u8 *seed,

@@ -85,8 +106,7 @@ static int hisi_trng_seed(struct crypto_rng *tfm, const u8 *seed,
 {
 	struct hisi_trng_ctx *ctx = crypto_rng_ctx(tfm);
 	struct hisi_trng *trng = ctx->trng;
-	u32 val = 0;
-	int ret = 0;
+	int ret;
 
 	if (slen < SW_DRBG_SEED_SIZE) {
 		pr_err("slen(%u) is not matched with trng(%d)\n", slen,

@@ -94,43 +114,45 @@ static int hisi_trng_seed(struct crypto_rng *tfm, const u8 *seed,
 		return -EINVAL;
 	}
 
-	writel(0x0, trng->base + SW_DRBG_BLOCKS);
-	hisi_trng_set_seed(trng, seed);
+	mutex_lock(&trng->lock);
+	ret = hisi_trng_set_seed(trng, seed);
+	mutex_unlock(&trng->lock);
 
-	writel(SW_DRBG_BLOCKS_NUM | (0x1 << SW_DRBG_ENABLE_SHIFT),
-	       trng->base + SW_DRBG_BLOCKS);
-	writel(0x1, trng->base + SW_DRBG_INIT);
+	return ret;
+}
 
-	ret = readl_relaxed_poll_timeout(trng->base + SW_DRBG_STATUS,
-					val, val & BIT(0), SLEEP_US, TIMEOUT_US);
-	if (ret)
-		pr_err("fail to init trng(%d)\n", ret);
+static int hisi_trng_reseed(struct hisi_trng *trng)
+{
+	u8 seed[SW_DRBG_SEED_SIZE];
+	int size;
 
-	return ret;
+	if (!trng->random_bytes)
+		return 0;
+
+	size = hisi_trng_read(&trng->rng, seed, SW_DRBG_SEED_SIZE, false);
+	if (size != SW_DRBG_SEED_SIZE)
+		return -EIO;
+
+	return hisi_trng_set_seed(trng, seed);
 }
 
-static int hisi_trng_generate(struct crypto_rng *tfm, const u8 *src,
-			      unsigned int slen, u8 *dstn, unsigned int dlen)
+static int hisi_trng_get_bytes(struct hisi_trng *trng, u8 *dstn, unsigned int dlen)
 {
-	struct hisi_trng_ctx *ctx = crypto_rng_ctx(tfm);
-	struct hisi_trng *trng = ctx->trng;
 	u32 data[SW_DRBG_DATA_NUM];
 	u32 currsize = 0;
 	u32 val = 0;
 	int ret;
 	u32 i;
 
-	if (dlen > SW_DRBG_BLOCKS_NUM * SW_DRBG_BYTES || dlen == 0) {
-		pr_err("dlen(%u) exceeds limit(%d)!\n", dlen,
-			SW_DRBG_BLOCKS_NUM * SW_DRBG_BYTES);
-		return -EINVAL;
-	}
+	ret = hisi_trng_reseed(trng);
+	if (ret)
+		return ret;
 
 	do {
 		ret = readl_relaxed_poll_timeout(trng->base + SW_DRBG_STATUS,
-		     val, val & BIT(1), SLEEP_US, TIMEOUT_US);
+						 val, val & BIT(1), SLEEP_US, TIMEOUT_US);
 		if (ret) {
-			pr_err("fail to generate random number(%d)!\n", ret);
+			pr_err("failed to generate random number(%d)!\n", ret);
 			break;
 		}
 

@@ -145,30 +167,57 @@ static int hisi_trng_generate(struct crypto_rng *tfm, const u8 *src,
 			currsize = dlen;
 		}
 
+		trng->random_bytes += SW_DRBG_BYTES;
 		writel(0x1, trng->base + SW_DRBG_GEN);
 	} while (currsize < dlen);
 
 	return ret;
 }
 
+static int hisi_trng_generate(struct crypto_rng *tfm, const u8 *src,
+			      unsigned int slen, u8 *dstn, unsigned int dlen)
+{
+	struct hisi_trng_ctx *ctx = crypto_rng_ctx(tfm);
+	struct hisi_trng *trng = ctx->trng;
+	unsigned int currsize = 0;
+	unsigned int block_size;
+	int ret;
+
+	if (!dstn || !dlen) {
+		pr_err("output is error, dlen %u!\n", dlen);
+		return -EINVAL;
+	}
+
+	do {
+		block_size = min_t(unsigned int, dlen - currsize, SW_MAX_RANDOM_BYTES);
+		mutex_lock(&trng->lock);
+		ret = hisi_trng_get_bytes(trng, dstn + currsize, block_size);
+		mutex_unlock(&trng->lock);
+		if (ret)
+			return ret;
+		currsize += block_size;
+	} while (currsize < dlen);
+
+	return 0;
+}
+
 static int hisi_trng_init(struct crypto_tfm *tfm)
 {
 	struct hisi_trng_ctx *ctx = crypto_tfm_ctx(tfm);
 	struct hisi_trng *trng;
-	int ret = -EBUSY;
+	u32 ctx_num = ~0;
 
 	mutex_lock(&trng_devices.lock);
 	list_for_each_entry(trng, &trng_devices.list, list) {
-		if (!trng->is_used) {
-			trng->is_used = true;
+		if (trng->ctx_num < ctx_num) {
+			ctx_num = trng->ctx_num;
 			ctx->trng = trng;
-			ret = 0;
-			break;
 		}
 	}
+	ctx->trng->ctx_num++;
 	mutex_unlock(&trng_devices.lock);
 
-	return ret;
+	return 0;
 }
 
 static void hisi_trng_exit(struct crypto_tfm *tfm)

@@ -176,7 +225,7 @@ static void hisi_trng_exit(struct crypto_tfm *tfm)
 	struct hisi_trng_ctx *ctx = crypto_tfm_ctx(tfm);
 
 	mutex_lock(&trng_devices.lock);
-	ctx->trng->is_used = false;
+	ctx->trng->ctx_num--;
 	mutex_unlock(&trng_devices.lock);
 }
 

@@ -238,7 +287,7 @@ static int hisi_trng_del_from_list(struct hisi_trng *trng)
 	int ret = -EBUSY;
 
 	mutex_lock(&trng_devices.lock);
-	if (!trng->is_used) {
+	if (!trng->ctx_num) {
 		list_del(&trng->list);
 		ret = 0;
 	}

@@ -262,7 +311,9 @@ static int hisi_trng_probe(struct platform_device *pdev)
 	if (IS_ERR(trng->base))
 		return PTR_ERR(trng->base);
 
-	trng->is_used = false;
+	trng->ctx_num = 0;
+	trng->random_bytes = SW_MAX_RANDOM_BYTES;
+	mutex_init(&trng->lock);
 	trng->ver = readl(trng->base + HISI_TRNG_VERSION);
 	if (!trng_devices.is_init) {
 		INIT_LIST_HEAD(&trng_devices.list);

2.51.0

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

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

From: Jianpeng Chang <[email protected]>

[ Upstream commit 7d43252b3060b0ba4a192dce5dba85a3f39ffe39 ]

When commit 0e1a4d427f58 ("crypto: caam: Unembed net_dev structure in dpaa2") converted embedded net_device to dynamically allocated pointers, it added cleanup in dpaa2_dpseci_disable() but missed adding cleanup in dpaa2_dpseci_free() for error paths.

This causes memory leaks when dpaa2_dpseci_dpio_setup() fails during probe due to DPIO devices not being ready yet. The kernel's deferred probe mechanism handles the retry successfully, but the netdevs allocated during the failed probe attempt are never freed, resulting in kmemleak reports showing multiple leaked netdev-related allocations all traced back to dpaa2_caam_probe().

Fix this by preserving the CPU mask of allocated netdevs during setup and using it for cleanup in dpaa2_dpseci_free(). This approach ensures that only the CPUs that actually had netdevs allocated will be cleaned up, avoiding potential issues with CPU hotplug scenarios.

---
 drivers/crypto/caam/caamalg_qi2.c | 27 +++++++++++++++------------
 drivers/crypto/caam/caamalg_qi2.h |  2 ++
 2 files changed, 17 insertions(+), 12 deletions(-)

index 107ccb2ade420..c6117c23eb25b 100644
@@ -4814,7 +4814,8 @@ static void dpaa2_dpseci_free(struct dpaa2_caam_priv *priv)
 {
 	struct device *dev = priv->dev;
 	struct fsl_mc_device *ls_dev = to_fsl_mc_device(dev);
-	int err;
+	struct dpaa2_caam_priv_per_cpu *ppriv;
+	int i, err;
 
 	if (DPSECI_VER(priv->major_ver, priv->minor_ver) > DPSECI_VER(5, 3)) {
 		err = dpseci_reset(priv->mc_io, 0, ls_dev->mc_handle);

@@ -4822,6 +4823,12 @@ static void dpaa2_dpseci_free(struct dpaa2_caam_priv *priv)
 			dev_err(dev, "dpseci_reset() failed\n");
 	}
 
+	for_each_cpu(i, priv->clean_mask) {
+		ppriv = per_cpu_ptr(priv->ppriv, i);
+		free_netdev(ppriv->net_dev);
+	}
+	free_cpumask_var(priv->clean_mask);
+
 	dpaa2_dpseci_congestion_free(priv);
 	dpseci_close(priv->mc_io, 0, ls_dev->mc_handle);
 }

@@ -5007,16 +5014,15 @@ static int __cold dpaa2_dpseci_setup(struct fsl_mc_device *ls_dev)
 	struct device *dev = &ls_dev->dev;
 	struct dpaa2_caam_priv *priv;
 	struct dpaa2_caam_priv_per_cpu *ppriv;
-	cpumask_var_t clean_mask;
 	int err, cpu;
 	u8 i;
 
 	err = -ENOMEM;
-	if (!zalloc_cpumask_var(&clean_mask, GFP_KERNEL))
-		goto err_cpumask;
-
 	priv = dev_get_drvdata(dev);
 
+	if (!zalloc_cpumask_var(&priv->clean_mask, GFP_KERNEL))
+		goto err_cpumask;
+
 	priv->dev = dev;
 	priv->dpsec_id = ls_dev->obj_desc.id;
 

@@ -5118,7 +5124,7 @@ static int __cold dpaa2_dpseci_setup(struct fsl_mc_device *ls_dev)
 			err = -ENOMEM;
 			goto err_alloc_netdev;
 		}
-		cpumask_set_cpu(cpu, clean_mask);
+		cpumask_set_cpu(cpu, priv->clean_mask);
 		ppriv->net_dev->dev = *dev;
 
 		netif_napi_add_tx_weight(ppriv->net_dev, &ppriv->napi,

@@ -5126,18 +5132,16 @@ static int __cold dpaa2_dpseci_setup(struct fsl_mc_device *ls_dev)
 					 DPAA2_CAAM_NAPI_WEIGHT);
 	}
 
-	err = 0;
-	goto free_cpumask;
+	return 0;
 
 err_alloc_netdev:
-	free_dpaa2_pcpu_netdev(priv, clean_mask);
+	free_dpaa2_pcpu_netdev(priv, priv->clean_mask);
 err_get_rx_queue:
 	dpaa2_dpseci_congestion_free(priv);
 err_get_vers:
 	dpseci_close(priv->mc_io, 0, ls_dev->mc_handle);
 err_open:
-free_cpumask:
-	free_cpumask_var(clean_mask);
+	free_cpumask_var(priv->clean_mask);
 err_cpumask:
 	return err;
 }

@@ -5182,7 +5186,6 @@ static int __cold dpaa2_dpseci_disable(struct dpaa2_caam_priv *priv)
 		ppriv = per_cpu_ptr(priv->ppriv, i);
 		napi_disable(&ppriv->napi);
 		netif_napi_del(&ppriv->napi);
-		free_netdev(ppriv->net_dev);
 	}
 
 	return 0;

index 61d1219a202fc..8e65b4b28c7ba 100644
@@ -42,6 +42,7 @@
  * @mc_io: pointer to MC portal's I/O object
  * @domain: IOMMU domain
  * @ppriv: per CPU pointers to privata data
+ * @clean_mask: CPU mask of CPUs that have allocated netdevs
  */
 struct dpaa2_caam_priv {
 	int dpsec_id;

@@ -65,6 +66,7 @@ struct dpaa2_caam_priv {
 
 	struct dpaa2_caam_priv_per_cpu __percpu *ppriv;
 	struct dentry *dfs_root;
+	cpumask_var_t clean_mask;
 };
 
 /**

2.51.0

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

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

From: Paul Chaignon <[email protected]>

[ Upstream commit 6557f1565d779851c4db9c488c49c05a47a6e72f ]

While making some maps in Cilium read-only from the BPF side, we noticed that the bpf_xdp_store_bytes proto is incorrect. In particular, the verifier was throwing the following error:

; ret = ctx_store_bytes(ctx, l3_off + offsetof(struct iphdr, saddr),

                          &nat->address, 4, 0);
  635: (79) r1 = *(u64 *)(r10 -144)     ; R1=ctx() R10=fp0 fp-144=ctx()
  636: (b4) w2 = 26                     ; R2=26
  637: (b4) w4 = 4                      ; R4=4
  638: (b4) w5 = 0                      ; R5=0

639: (85) call bpf_xdp_store_bytes#190 write into map forbidden, value_size=6 off=0 size=4

nat comes from a BPF_F_RDONLY_PROG map, so R3 is a PTR_TO_MAP_VALUE. The verifier checks the helper's memory access to R3 in check_mem_size_reg, as it reaches ARG_CONST_SIZE argument. The third argument has expected type ARG_PTR_TO_UNINIT_MEM, which includes the MEM_WRITE flag. The verifier thus checks for a BPF_WRITE access on R3. Given R3 points to a read-only map, the check fails.

Conversely, ARG_PTR_TO_UNINIT_MEM can also lead to the helper reading from uninitialized memory.

This patch simply fixes the expected argument type to match that of bpf_skb_store_bytes.

---
 net/core/filter.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index b1f8e2930e1c4..51318cb40f778 100644
@@ -4137,7 +4137,7 @@ static const struct bpf_func_proto bpf_xdp_store_bytes_proto = {
 	.ret_type	= RET_INTEGER,
 	.arg1_type	= ARG_PTR_TO_CTX,
 	.arg2_type	= ARG_ANYTHING,
-	.arg3_type	= ARG_PTR_TO_UNINIT_MEM,
+	.arg3_type	= ARG_PTR_TO_MEM | MEM_RDONLY,
 	.arg4_type	= ARG_CONST_SIZE,
 };
 

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 943b052ded21feb84f293d40b06af3181cd0d0d7 ]

The flag IRQF_COND_ONESHOT was already force-added to request_irq() because the ACPI SCI interrupt handler is using the IRQF_ONESHOT flag which breaks all shared handlers.

devm_request_irq() needs the same change since some users, such as int0002_vgpio, are using this function instead.

Add IRQF_COND_ONESHOT to the flags passed to devm_request_irq().

---
 include/linux/interrupt.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 266f2b39213a0..b2bb878abd113 100644
@@ -228,7 +228,7 @@ static inline int __must_check
 devm_request_irq(struct device *dev, unsigned int irq, irq_handler_t handler,
 		 unsigned long irqflags, const char *devname, void *dev_id)
 {
-	return devm_request_threaded_irq(dev, irq, handler, NULL, irqflags,
+	return devm_request_threaded_irq(dev, irq, handler, NULL, irqflags | IRQF_COND_ONESHOT,
 					 devname, dev_id);
 }
 

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit f6bc712877f24dc89bdfd7bdbf1a32f3b9960b34 ]

Passing IRQF_ONESHOT ensures that the interrupt source is masked until the secondary (threaded) handler is done. If only a primary handler is used then the flag makes no sense because the interrupt cannot fire (again) while its handler is running.

The flag also prevents force-threading of the primary handler and the irq-core will warn about this.

The flag was added to match the flag on the shared handler which uses a threaded handler and therefore IRQF_ONESHOT. This is no longer needed because devm_request_irq() now passes IRQF_COND_ONESHOT for this case.

Revert adding IRQF_ONESHOT to irqflags.

---
 drivers/platform/x86/intel/int0002_vgpio.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 6f5629dc3f8db..562e880256436 100644
@@ -206,8 +206,8 @@ static int int0002_probe(struct platform_device *pdev)
 	 * FIXME: augment this if we managed to pull handling of shared
 	 * IRQs into gpiolib.
 	 */
-	ret = devm_request_irq(dev, irq, int0002_irq,
-			       IRQF_ONESHOT | IRQF_SHARED, "INT0002", chip);
+	ret = devm_request_irq(dev, irq, int0002_irq, IRQF_SHARED, "INT0002",
+			       chip);
 	if (ret) {
 		dev_err(dev, "Error requesting IRQ %d: %d\n", irq, ret);
 		return ret;

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 5bfcdccb4d18d3909b7f87942be67fd6bdc00c1d ]

request_threaded_irq() is invoked with a primary and a secondary handler and no flags are passed. The primary handler is the same as irq_default_primary_handler() so there is no need to have an identical copy.

The lack of the IRQF_ONESHOT can be dangerous because the interrupt source is not masked while the threaded handler is active. This means, especially on LEVEL typed interrupt lines, the interrupt can fire again before the threaded handler had a chance to run.

Use the default primary interrupt handler by specifying NULL and set IRQF_ONESHOT so the interrupt source is masked until the secondary handler is done.

---
 drivers/iommu/amd/amd_iommu.h |  1 -
 drivers/iommu/amd/init.c      | 12 ++++--------
 drivers/iommu/amd/iommu.c     |  5 -----
 3 files changed, 4 insertions(+), 14 deletions(-)

index b742ef1adb352..df1c238dc8885 100644
@@ -15,7 +15,6 @@ irqreturn_t amd_iommu_int_thread(int irq, void *data);
 irqreturn_t amd_iommu_int_thread_evtlog(int irq, void *data);
 irqreturn_t amd_iommu_int_thread_pprlog(int irq, void *data);
 irqreturn_t amd_iommu_int_thread_galog(int irq, void *data);
-irqreturn_t amd_iommu_int_handler(int irq, void *data);
 void amd_iommu_restart_log(struct amd_iommu *iommu, const char *evt_type,
 			   u8 cntrl_intr, u8 cntrl_log,
 			   u32 status_run_mask, u32 status_overflow_mask);

index 384c90b4f90a0..62a7a718acf8f 100644
@@ -2356,12 +2356,8 @@ static int iommu_setup_msi(struct amd_iommu *iommu)
 	if (r)
 		return r;
 
-	r = request_threaded_irq(iommu->dev->irq,
-				 amd_iommu_int_handler,
-				 amd_iommu_int_thread,
-				 0, "AMD-Vi",
-				 iommu);
-
+	r = request_threaded_irq(iommu->dev->irq, NULL, amd_iommu_int_thread,
+				 IRQF_ONESHOT, "AMD-Vi", iommu);
 	if (r) {
 		pci_disable_msi(iommu->dev);
 		return r;

@@ -2535,8 +2531,8 @@ static int __iommu_setup_intcapxt(struct amd_iommu *iommu, const char *devname,
 		return irq;
 	}
 
-	ret = request_threaded_irq(irq, amd_iommu_int_handler,
-				   thread_fn, 0, devname, iommu);
+	ret = request_threaded_irq(irq, NULL, thread_fn, IRQF_ONESHOT, devname,
+				   iommu);
 	if (ret) {
 		irq_domain_free_irqs(irq, 1);
 		irq_domain_remove(domain);

index 7c12be1b247f4..0f9045ce93af1 100644
@@ -1151,11 +1151,6 @@ irqreturn_t amd_iommu_int_thread(int irq, void *data)
 	return IRQ_HANDLED;
 }
 
-irqreturn_t amd_iommu_int_handler(int irq, void *data)
-{
-	return IRQ_WAKE_THREAD;
-}
-
 /****************************************************************************
  *
  * IOMMU command queuing functions

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 28abed6569c87eab9071ab56c64433c2f0d9ce51 ]

There is no added value in btintel_pcie_msix_isr() compared to irq_default_primary_handler().

Using a threaded interrupt without a dedicated primary handler mandates the IRQF_ONESHOT flag to mask the interrupt source while the threaded handler is active. Otherwise the interrupt can fire again before the threaded handler had a chance to run.

Use the default primary interrupt handler by specifying NULL and set IRQF_ONESHOT so the interrupt source is masked until the secondary handler is done.

---
 drivers/bluetooth/btintel_pcie.c | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)

index 2936b535479f2..704767b334b98 100644
@@ -1431,11 +1431,6 @@ static void btintel_pcie_msix_rx_handle(struct btintel_pcie_data *data)
 	}
 }
 
-static irqreturn_t btintel_pcie_msix_isr(int irq, void *data)
-{
-	return IRQ_WAKE_THREAD;
-}
-
 static inline bool btintel_pcie_is_rxq_empty(struct btintel_pcie_data *data)
 {
 	return data->ia.cr_hia[BTINTEL_PCIE_RXQ_NUM] == data->ia.cr_tia[BTINTEL_PCIE_RXQ_NUM];

@@ -1537,9 +1532,9 @@ static int btintel_pcie_setup_irq(struct btintel_pcie_data *data)
 
 		err = devm_request_threaded_irq(&data->pdev->dev,
 						msix_entry->vector,
-						btintel_pcie_msix_isr,
+						NULL,
 						btintel_pcie_irq_msix_handler,
-						IRQF_SHARED,
+						IRQF_ONESHOT | IRQF_SHARED,
 						KBUILD_MODNAME,
 						msix_entry);
 		if (err) {

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit bd81f07e9a27c341cd7e72be95eb0b7cf3910926 ]

There is no added value in efct_intr_msix() compared to irq_default_primary_handler().

Using a threaded interrupt without a dedicated primary handler mandates the IRQF_ONESHOT flag to mask the interrupt source while the threaded handler is active. Otherwise the interrupt can fire again before the threaded handler had a chance to run.

Use the default primary interrupt handler by specifying NULL and set IRQF_ONESHOT so the interrupt source is masked until the secondary handler is done.

---
 drivers/scsi/elx/efct/efct_driver.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

index 1bd42f7db1773..528399f725d42 100644
@@ -415,12 +415,6 @@ efct_intr_thread(int irq, void *handle)
 	return IRQ_HANDLED;
 }
 
-static irqreturn_t
-efct_intr_msix(int irq, void *handle)
-{
-	return IRQ_WAKE_THREAD;
-}
-
 static int
 efct_setup_msix(struct efct *efct, u32 num_intrs)
 {

@@ -450,7 +444,7 @@ efct_setup_msix(struct efct *efct, u32 num_intrs)
 		intr_ctx->index = i;
 
 		rc = request_threaded_irq(pci_irq_vector(efct->pci, i),
-					  efct_intr_msix, efct_intr_thread, 0,
+					  NULL, efct_intr_thread, IRQF_ONESHOT,
 					  EFCT_DRIVER_NAME, intr_ctx);
 		if (rc) {
 			dev_err(&efct->pci->dev,

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 5c858d6c66304b4c7579582ec5235f02d43578ea ]

Passing IRQF_ONESHOT ensures that the interrupt source is masked until the secondary (threaded) handler is done. If only a primary handler is used then the flag makes no sense because the interrupt can not fire (again) while its handler is running.

The flag also prevents force-threading of the primary handler and the irq-core will warn about this.

Remove IRQF_ONESHOT from irqflags.

---
 drivers/edac/altera_edac.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

index 0c5b94e64ea15..4edd2088c2db6 100644
@@ -1563,8 +1563,7 @@ static int altr_portb_setup(struct altr_edac_device_dev *device)
 		goto err_release_group_1;
 	}
 	rc = devm_request_irq(&altdev->ddev, altdev->sb_irq,
-			      prv->ecc_irq_handler,
-			      IRQF_ONESHOT | IRQF_TRIGGER_HIGH,
+			      prv->ecc_irq_handler, IRQF_TRIGGER_HIGH,
 			      ecc_name, altdev);
 	if (rc) {
 		edac_printk(KERN_ERR, EDAC_DEVICE, "PortB SBERR IRQ error\n");

@@ -1587,8 +1586,7 @@ static int altr_portb_setup(struct altr_edac_device_dev *device)
 		goto err_release_group_1;
 	}
 	rc = devm_request_irq(&altdev->ddev, altdev->db_irq,
-			      prv->ecc_irq_handler,
-			      IRQF_ONESHOT | IRQF_TRIGGER_HIGH,
+			      prv->ecc_irq_handler, IRQF_TRIGGER_HIGH,
 			      ecc_name, altdev);
 	if (rc) {
 		edac_printk(KERN_ERR, EDAC_DEVICE, "PortB DBERR IRQ error\n");

@@ -1970,8 +1968,7 @@ static int altr_edac_a10_device_add(struct altr_arria10_edac *edac,
 		goto err_release_group1;
 	}
 	rc = devm_request_irq(edac->dev, altdev->sb_irq, prv->ecc_irq_handler,
-			      IRQF_ONESHOT | IRQF_TRIGGER_HIGH,
-			      ecc_name, altdev);
+			      IRQF_TRIGGER_HIGH, ecc_name, altdev);
 	if (rc) {
 		edac_printk(KERN_ERR, EDAC_DEVICE, "No SBERR IRQ resource\n");
 		goto err_release_group1;

@@ -1993,7 +1990,7 @@ static int altr_edac_a10_device_add(struct altr_arria10_edac *edac,
 		goto err_release_group1;
 	}
 	rc = devm_request_irq(edac->dev, altdev->db_irq, prv->ecc_irq_handler,
-			      IRQF_ONESHOT | IRQF_TRIGGER_HIGH,
+			      IRQF_TRIGGER_HIGH,
 			      ecc_name, altdev);
 	if (rc) {
 		edac_printk(KERN_ERR, EDAC_DEVICE, "No DBERR IRQ resource\n");

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit a7fb84ea70aae9a92a842932206e70ed1b3c7007 ]

Passing IRQF_ONESHOT ensures that the interrupt source is masked until the secondary (threaded) handler is done. If only a primary handler is used then the flag makes no sense because the interrupt can not fire (again) while its handler is running.

The flag also prevents force-threading of the primary handler and the irq-core will warn about this.

Remove IRQF_ONESHOT from irqflags.

---
 drivers/usb/typec/tcpm/fusb302.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

index 870a71f953f6c..19ff8217818e7 100644
@@ -1756,8 +1756,7 @@ static int fusb302_probe(struct i2c_client *client)
 	}
 
 	ret = request_irq(chip->gpio_int_n_irq, fusb302_irq_intn,
-			  IRQF_ONESHOT | IRQF_TRIGGER_LOW,
-			  "fsc_interrupt_int_n", chip);
+			  IRQF_TRIGGER_LOW, "fsc_interrupt_int_n", chip);
 	if (ret < 0) {
 		dev_err(dev, "cannot request IRQ for GPIO Int_N, ret=%d", ret);
 		goto tcpm_unregister_port;

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 18d28446231390e4ea3634fb16200865df2c6506 ]

Passing IRQF_ONESHOT ensures that the interrupt source is masked until the secondary (threaded) handler is done. If only a primary handler is used then the flag makes no sense because the interrupt can not fire (again) while its handler is running.

The flag also prevents force-threading of the primary handler and the irq-core will warn about this.

Remove IRQF_ONESHOT from irqflags.

---
 drivers/rtc/rtc-amlogic-a4.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 123fb372fc9fe..50938c35af36a 100644
@@ -369,7 +369,7 @@ static int aml_rtc_probe(struct platform_device *pdev)
 		return PTR_ERR(rtc->rtc_dev);
 
 	ret = devm_request_irq(dev, rtc->irq, aml_rtc_handler,
-			       IRQF_ONESHOT, "aml-rtc alarm", rtc);
+			       0, "aml-rtc alarm", rtc);
 	if (ret) {
 		dev_err_probe(dev, ret, "IRQ%d request failed, ret = %d\n",
 			      rtc->irq, ret);

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit 553b4999cbe231b5011cb8db05a3092dec168aca ]

Using a threaded interrupt without a dedicated primary handler mandates the IRQF_ONESHOT flag to mask the interrupt source while the threaded handler is active. Otherwise the interrupt can fire again before the threaded handler had a chance to run.

Mark explained that this should not happen with this hardware since it is a slow irqchip which is behind an I2C/ SPI bus but the IRQ-core will refuse to accept such a handler.

Set IRQF_ONESHOT so the interrupt source is masked until the secondary handler is done.

---
 include/linux/mfd/wm8350/core.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 5f70d3b5d1b1a..097ef4dfcdac8 100644
@@ -667,7 +667,7 @@ static inline int wm8350_register_irq(struct wm8350 *wm8350, int irq,
 		return -ENODEV;
 
 	return request_threaded_irq(irq + wm8350->irq_base, NULL,
-				    handler, flags, name, data);
+				    handler, flags | IRQF_ONESHOT, name, data);
 }
 
 static inline void wm8350_free_irq(struct wm8350 *wm8350, int irq, void *data)

2.51.0

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

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

From: Sebastian Andrzej Siewior <[email protected]>

[ Upstream commit ef92b98f5f6758a049898b53aa30476010db04fa ]

The interrupt handler iio_trigger_generic_data_rdy_poll() will invoke other interrupt handlers and this supposed to happen from hard interrupt context.

Use IRQF_NO_THREAD to forbid forced-threading.

---
 drivers/media/pci/mgb4/mgb4_trigger.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 4f9a35904b418..70cad324df608 100644
@@ -115,7 +115,7 @@ static int probe_trigger(struct iio_dev *indio_dev, int irq)
 	if (!st->trig)
 		return -ENOMEM;
 
-	ret = request_irq(irq, &iio_trigger_generic_data_rdy_poll, 0,
+	ret = request_irq(irq, &iio_trigger_generic_data_rdy_poll, IRQF_NO_THREAD,
 			  "mgb4-trigger", st->trig);
 	if (ret)
 		goto error_free_trig;

2.51.0

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

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

From: Joel Fernandes <[email protected]>

[ Upstream commit 3cb3b27693bf30defb16aa096158a3b24583b8d2 ]

The defer params were not cleared in __dl_clear_params. Clear them.

Without this is some of my test cases are flaking and the DL timer is not starting correctly AFAICS.

---
 kernel/sched/deadline.c | 3 +++
 1 file changed, 3 insertions(+)

index 7bcde7114f1b6..e3a6b8ed1d6db 100644
@@ -3656,6 +3656,9 @@ static void __dl_clear_params(struct sched_dl_entity *dl_se)
 	dl_se->dl_non_contending	= 0;
 	dl_se->dl_overrun		= 0;
 	dl_se->dl_server		= 0;
+	dl_se->dl_defer			= 0;
+	dl_se->dl_defer_running		= 0;
+	dl_se->dl_defer_armed		= 0;
 
 #ifdef CONFIG_RT_MUTEXES
 	dl_se->pi_se			= dl_se;

2.51.0

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

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

From: Chen Jinghuang <[email protected]>

[ Upstream commit 94894c9c477e53bcea052e075c53f89df3d2a33e ]

CPU0 becomes overloaded when hosting a CPU-bound RT task, a non-CPU-bound RT task, and a CFS task stuck in kernel space. When other CPUs switch from RT to non-RT tasks, RT load balancing (LB) is triggered; with HAVE_RT_PUSH_IPI enabled, they send IPIs to CPU0 to drive the execution of rto_push_irq_work_func. During push_rt_task on CPU0,
if next_task->prio < rq->donor->prio, resched_curr() sets NEED_RESCHED and after the push operation completes, CPU0 calls rto_next_cpu(). Since only CPU0 is overloaded in this scenario, rto_next_cpu() should ideally return -1 (no further IPI needed).

However, multiple CPUs invoking tell_cpu_to_push() during LB increments rd->rto_loop_next. Even when rd->rto_cpu is set to -1, the mismatch between rd->rto_loop and rd->rto_loop_next forces rto_next_cpu() to restart its search from -1. With CPU0 remaining overloaded (satisfying rt_nr_migratory && rt_nr_total > 1), it gets reselected, causing CPU0 to queue irq_work to itself and send self-IPIs repeatedly. As long as CPU0 stays overloaded and other CPUs run pull_rt_tasks(), it falls into an infinite self-IPI loop, which triggers a CPU hardlockup due to continuous self-interrupts.

The trigging scenario is as follows:

cpu0 cpu1 cpu2
pull_rt_task
tell_cpu_to_push
<------------irq_work_queue_on
rto_push_irq_work_func

       push_rt_task
    resched_curr(rq)                                   pull_rt_task
    rto_next_cpu                                     tell_cpu_to_push
                      <-------------------------- atomic_inc(rto_loop_next)

rd->rto_loop != next
rto_next_cpu
irq_work_queue_on
rto_push_irq_work_func

Fix redundant self-IPI by filtering the initiating CPU in rto_next_cpu(). This solution has been verified to effectively eliminate spurious self-IPIs and prevent CPU hardlockup scenarios.

---
 kernel/sched/rt.c | 5 +++++
 1 file changed, 5 insertions(+)

index f1867fe8e5c53..e0ff909050190 100644
@@ -2100,6 +2100,7 @@ static void push_rt_tasks(struct rq *rq)
  */
 static int rto_next_cpu(struct root_domain *rd)
 {
+	int this_cpu = smp_processor_id();
 	int next;
 	int cpu;
 

@@ -2123,6 +2124,10 @@ static int rto_next_cpu(struct root_domain *rd)
 
 		rd->rto_cpu = cpu;
 
+		/* Do not send IPI to self */
+		if (cpu == this_cpu)
+			continue;
+
 		if (cpu < nr_cpu_ids)
 			return cpu;
 

2.51.0

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

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

From: Zicheng Qu <[email protected]>

[ Upstream commit e34881c84c255bc300f24d9fe685324be20da3d1 ]

Consider the following sequence on a CPU configured with nohz_full:

1) A task P runs in cgroup A, and cgroup A becomes throttled due to CFS bandwidth control. The gse (cgroup A) where the task P attached is
dequeued and the CPU switches to idle.

2) Before cgroup A is unthrottled, task P is migrated from cgroup A to another cgroup B (not throttled).

During sched_move_task(), the task P is observed as queued but not
running, and therefore no resched_curr() is triggered.

3) Since the CPU is nohz_full, it remains in do_idle() waiting for an explicit scheduling event, i.e., resched_curr().

4) For kernel <= 5.10: Later, cgroup A is unthrottled. However, the task P has already been migrated out of cgroup A, so unthrottle_cfs_rq()
may observe load_weight == 0 and return early without resched_curr() called. For kernel >= 6.6: The unthrottling path normally triggers resched_curr() almost cases even when no runnable tasks remain in the unthrottled cgroup, preventing the idle stall described above. However, if cgroup A is removed before it gets unthrottled, the unthrottling path for cgroup A is never executed. In a result, no resched_curr() can be called.

5) At this point, the task P is runnable in cgroup B (not throttled), but the CPU remains in do_idle() with no pending reschedule point. The system stays in this state until an unrelated event (e.g. a new task wakeup or any cases) that can trigger a resched_curr() breaks the nohz_full idle state, and then the task P finally gets scheduled.

The root cause is that sched_move_task() may classify the task as only queued, not running, and therefore fails to trigger a resched_curr(), while the later unthrottling path no longer has visibility of the migrated task.

Preserve the existing behavior for running tasks by issuing resched_curr(), and explicitly invoke check_preempt_curr() for tasks that were queued at the time of migration. This ensures that runnable tasks are reconsidered for scheduling even when nohz_full suppresses periodic ticks.

---
 kernel/sched/core.c | 4 ++++
 1 file changed, 4 insertions(+)

index c3b6e123fa00e..dbf4e32a063f7 100644
@@ -9115,6 +9115,7 @@ void sched_move_task(struct task_struct *tsk, bool for_autogroup)
 {
 	unsigned int queue_flags = DEQUEUE_SAVE | DEQUEUE_MOVE;
 	bool resched = false;
+	bool queued = false;
 	struct rq *rq;
 
 	CLASS(task_rq_lock, rq_guard)(tsk);

@@ -9126,10 +9127,13 @@ void sched_move_task(struct task_struct *tsk, bool for_autogroup)
 			scx_cgroup_move_task(tsk);
 		if (scope->running)
 			resched = true;
+		queued = scope->queued;
 	}
 
 	if (resched)
 		resched_curr(rq);
+	else if (queued)
+		wakeup_preempt(rq, tsk, 0);
 
 	__balance_callbacks(rq, &rq_guard.rf);
 }

2.51.0

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

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

From: Titouan Ameline de Cadeville <[email protected]>

[ Upstream commit 46a03ea50b5f380bdb99178b8f90b39c6ba1f528 ]

Remove duplicate entries from the bprm_stack_limits KUnit test vector table. The duplicates do not add coverage and only increase test size.

---
 fs/tests/exec_kunit.c | 6 ------
 1 file changed, 6 deletions(-)

index 7c77d039680bb..f412d1a0f6bba 100644
@@ -87,9 +87,6 @@ static const struct bprm_stack_limits_result bprm_stack_limits_results[] = {
 	    .argc = 0, .envc = ARG_MAX / sizeof(void *) - 1 },
 	  .expected_argmin = ULONG_MAX - sizeof(void *) },
 	/* Raising rlim_stack / 4 to _STK_LIM / 4 * 3 will see more space. */
-	{ { .p = ULONG_MAX, .rlim_stack.rlim_cur = 4 * (_STK_LIM / 4 * 3),
-	    .argc = 0, .envc = 0 },
-	  .expected_argmin = ULONG_MAX - (_STK_LIM / 4 * 3) + sizeof(void *) },
 	{ { .p = ULONG_MAX, .rlim_stack.rlim_cur = 4 * (_STK_LIM / 4 * 3),
 	    .argc = 0, .envc = 0 },
 	  .expected_argmin = ULONG_MAX - (_STK_LIM / 4 * 3) + sizeof(void *) },

@@ -103,9 +100,6 @@ static const struct bprm_stack_limits_result bprm_stack_limits_results[] = {
 	{ { .p = ULONG_MAX, .rlim_stack.rlim_cur = 4 * _STK_LIM,
 	    .argc = 0, .envc = 0 },
 	  .expected_argmin = ULONG_MAX - (_STK_LIM / 4 * 3) + sizeof(void *) },
-	{ { .p = ULONG_MAX, .rlim_stack.rlim_cur = 4 * _STK_LIM,
-	    .argc = 0, .envc = 0 },
-	  .expected_argmin = ULONG_MAX - (_STK_LIM / 4 * 3) + sizeof(void *) },
 };
 
 static void exec_test_bprm_stack_limits(struct kunit *test)

2.51.0

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

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

From: Thomas Gleixner <[email protected]>

[ Upstream commit 42e025b719c128bdf8ff88584589a1e4a2448c81 ]

The driver is handling the number of hardware interrupts inconsistently.

The reason is that the firmware enumerates the maximum number of device interrupts, but the actual number of hardware interrupts is one more because hardware interrupt 0 is reserved.

There are two loop variants where this matters:

1) Iterating over the device interrupts

     for (irq = 1; irq < total_irqs; irq++)

2) Iterating over the number of interrupt register groups

     for (grp = 0; grp < irq_groups; grp++)

The current code stores the number of device interrupts and that requires to write the loops as:

1) for (irq = 1; irq <= device_irqs; irq++)

2) for (grp = 0; grp < DIV_ROUND_UP(device_irqs + 1); grp++)

But the code gets it wrong all over the place. Just fixing up the conditions and off by ones is not a sustainable solution as the next changes will reintroduce the same bugs over and over.

Sanitize it by storing the total number of hardware interrupts during probe and precalculating the number of groups. To future proof it mark priv::total_irqs __private, provide a correct iterator macro and adjust the code to this.

Marking it private allows sparse (C=1 build) to catch direct access to this member:

drivers/irqchip/irq-sifive-plic.c:270:9: warning: dereference of noderef expression

That should prevent at least the most obvious future damage in that area.

---
 drivers/irqchip/irq-sifive-plic.c | 82 +++++++++++++++++--------------
 1 file changed, 45 insertions(+), 37 deletions(-)

index 210a579596377..60fd8f91762b1 100644
@@ -68,15 +68,17 @@
 #define PLIC_QUIRK_CP100_CLAIM_REGISTER_ERRATUM	1
 
 struct plic_priv {
-	struct fwnode_handle *fwnode;
-	struct cpumask lmask;
-	struct irq_domain *irqdomain;
-	void __iomem *regs;
-	unsigned long plic_quirks;
-	unsigned int nr_irqs;
-	unsigned long *prio_save;
-	u32 gsi_base;
-	int acpi_plic_id;
+	struct fwnode_handle	*fwnode;
+	struct cpumask		lmask;
+	struct irq_domain	*irqdomain;
+	void __iomem		*regs;
+	unsigned long		plic_quirks;
+	/* device interrupts + 1 to compensate for the reserved hwirq 0 */
+	unsigned int __private	total_irqs;
+	unsigned int		irq_groups;
+	unsigned long		*prio_save;
+	u32			gsi_base;
+	int			acpi_plic_id;
 };
 
 struct plic_handler {

@@ -91,6 +93,12 @@ struct plic_handler {
 	u32			*enable_save;
 	struct plic_priv	*priv;
 };
+
+/*
+ * Macro to deal with the insanity of hardware interrupt 0 being reserved */
+#define for_each_device_irq(iter, priv)	\
+	for (unsigned int iter = 1; iter < ACCESS_PRIVATE(priv, total_irqs); iter++)
+
 static int plic_parent_irq __ro_after_init;
 static bool plic_global_setup_done __ro_after_init;
 static DEFINE_PER_CPU(struct plic_handler, plic_handlers);

@@ -257,14 +265,11 @@ static int plic_irq_set_type(struct irq_data *d, unsigned int type)
 
 static int plic_irq_suspend(void *data)
 {
-	struct plic_priv *priv;
-
-	priv = per_cpu_ptr(&plic_handlers, smp_processor_id())->priv;
+	struct plic_priv *priv = this_cpu_ptr(&plic_handlers)->priv;
 
-	/* irq ID 0 is reserved */
-	for (unsigned int i = 1; i < priv->nr_irqs; i++) {
-		__assign_bit(i, priv->prio_save,
-			     readl(priv->regs + PRIORITY_BASE + i * PRIORITY_PER_ID));
+	for_each_device_irq(irq, priv) {
+		__assign_bit(irq, priv->prio_save,
+			     readl(priv->regs + PRIORITY_BASE + irq * PRIORITY_PER_ID));
 	}
 
 	return 0;

@@ -272,18 +277,15 @@ static int plic_irq_suspend(void *data)
 
 static void plic_irq_resume(void *data)
 {
-	unsigned int i, index, cpu;
+	struct plic_priv *priv = this_cpu_ptr(&plic_handlers)->priv;
+	unsigned int index, cpu;
 	unsigned long flags;
 	u32 __iomem *reg;
-	struct plic_priv *priv;
-
-	priv = per_cpu_ptr(&plic_handlers, smp_processor_id())->priv;
 
-	/* irq ID 0 is reserved */
-	for (i = 1; i < priv->nr_irqs; i++) {
-		index = BIT_WORD(i);
-		writel((priv->prio_save[index] & BIT_MASK(i)) ? 1 : 0,
-		       priv->regs + PRIORITY_BASE + i * PRIORITY_PER_ID);
+	for_each_device_irq(irq, priv) {
+		index = BIT_WORD(irq);
+		writel((priv->prio_save[index] & BIT_MASK(irq)) ? 1 : 0,
+		       priv->regs + PRIORITY_BASE + irq * PRIORITY_PER_ID);
 	}
 
 	for_each_present_cpu(cpu) {

@@ -293,7 +295,7 @@ static void plic_irq_resume(void *data)
 			continue;
 
 		raw_spin_lock_irqsave(&handler->enable_lock, flags);
-		for (i = 0; i < DIV_ROUND_UP(priv->nr_irqs, 32); i++) {
+		for (unsigned int i = 0; i < priv->irq_groups; i++) {
 			reg = handler->enable_base + i * sizeof(u32);
 			writel(handler->enable_save[i], reg);
 		}

@@ -431,7 +433,7 @@ static u32 cp100_isolate_pending_irq(int nr_irq_groups, struct plic_handler *han
 
 static irq_hw_number_t cp100_get_hwirq(struct plic_handler *handler, void __iomem *claim)
 {
-	int nr_irq_groups = DIV_ROUND_UP(handler->priv->nr_irqs, 32);
+	int nr_irq_groups = handler->priv->irq_groups;
 	u32 __iomem *enable = handler->enable_base;
 	irq_hw_number_t hwirq = 0;
 	u32 iso_mask;

@@ -614,7 +616,6 @@ static int plic_probe(struct fwnode_handle *fwnode)
 	struct plic_handler *handler;
 	u32 nr_irqs, parent_hwirq;
 	struct plic_priv *priv;
-	irq_hw_number_t hwirq;
 	void __iomem *regs;
 	int id, context_id;
 	u32 gsi_base;

@@ -647,7 +648,16 @@ static int plic_probe(struct fwnode_handle *fwnode)
 
 	priv->fwnode = fwnode;
 	priv->plic_quirks = plic_quirks;
-	priv->nr_irqs = nr_irqs;
+	/*
+	 * The firmware provides the number of device interrupts. As
+	 * hardware interrupt 0 is reserved, the number of total interrupts
+	 * is nr_irqs + 1.
+	 */
+	nr_irqs++;
+	ACCESS_PRIVATE(priv, total_irqs) = nr_irqs;
+	/* Precalculate the number of register groups */
+	priv->irq_groups = DIV_ROUND_UP(nr_irqs, 32);
+
 	priv->regs = regs;
 	priv->gsi_base = gsi_base;
 	priv->acpi_plic_id = id;

@@ -686,7 +696,7 @@ static int plic_probe(struct fwnode_handle *fwnode)
 				u32 __iomem *enable_base = priv->regs +	CONTEXT_ENABLE_BASE +
 							   i * CONTEXT_ENABLE_SIZE;
 
-				for (int j = 0; j <= nr_irqs / 32; j++)
+				for (int j = 0; j < priv->irq_groups; j++)
 					writel(0, enable_base + j);
 			}
 			continue;

@@ -718,23 +728,21 @@ static int plic_probe(struct fwnode_handle *fwnode)
 			context_id * CONTEXT_ENABLE_SIZE;
 		handler->priv = priv;
 
-		handler->enable_save = kcalloc(DIV_ROUND_UP(nr_irqs, 32),
-					       sizeof(*handler->enable_save), GFP_KERNEL);
+		handler->enable_save = kcalloc(priv->irq_groups, sizeof(*handler->enable_save),
+					       GFP_KERNEL);
 		if (!handler->enable_save) {
 			error = -ENOMEM;
 			goto fail_cleanup_contexts;
 		}
 done:
-		for (hwirq = 1; hwirq <= nr_irqs; hwirq++) {
+		for_each_device_irq(hwirq, priv) {
 			plic_toggle(handler, hwirq, 0);
-			writel(1, priv->regs + PRIORITY_BASE +
-				  hwirq * PRIORITY_PER_ID);
+			writel(1, priv->regs + PRIORITY_BASE + hwirq * PRIORITY_PER_ID);
 		}
 		nr_handlers++;
 	}
 
-	priv->irqdomain = irq_domain_create_linear(fwnode, nr_irqs + 1,
-						   &plic_irqdomain_ops, priv);
+	priv->irqdomain = irq_domain_create_linear(fwnode, nr_irqs, &plic_irqdomain_ops, priv);
 	if (WARN_ON(!priv->irqdomain)) {
 		error = -ENOMEM;
 		goto fail_cleanup_contexts;

2.51.0

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

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

From: KP Singh <[email protected]>

[ Upstream commit ea1535e28bb3773fc0b3cbd1f3842b808016990c ]

Practical BPF signatures are significantly smaller than KMALLOC_MAX_CACHE_SIZE

Allowing larger sizes opens the door for abuse by passing excessive size values and forcing the kernel into expensive allocation paths (via kmalloc_large or vmalloc).

---
 kernel/bpf/syscall.c | 7 +++++++
 1 file changed, 7 insertions(+)

index 763868d327b4a..f89aa142f71b8 100644
@@ -2815,6 +2815,13 @@ static int bpf_prog_verify_signature(struct bpf_prog *prog, union bpf_attr *attr
 	void *sig;
 	int err = 0;
 
+	/*
+	 * Don't attempt to use kmalloc_large or vmalloc for signatures.
+	 * Practical signature for BPF program should be below this limit.
+	 */
+	if (attr->signature_size > KMALLOC_MAX_CACHE_SIZE)
+		return -EINVAL;
+
 	if (system_keyring_id_check(attr->keyring_id) == 0)
 		key = bpf_lookup_system_key(attr->keyring_id);
 	else

2.51.0

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

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

From: KP Singh <[email protected]>

[ Upstream commit a2c86aa621c22f2a7e26c654f936d65cfff0aa91 ]

Currently, bpf_map_get_info_by_fd calculates and caches the hash of the map regardless of the map's frozen state.

This leads to a TOCTOU bug where userspace can call BPF_OBJ_GET_INFO_BY_FD to cache the hash and then modify the map contents before freezing.

Therefore, a trusted loader can be tricked into verifying the stale hash while loading the modified contents.

Fix this by returning -EPERM if the map is not frozen when the hash is requested. This ensures the hash is only generated for the final, immutable state of the map.

---
 kernel/bpf/syscall.c | 3 +++
 1 file changed, 3 insertions(+)

index f89aa142f71b8..ce7db2f3be6f6 100644
@@ -5314,6 +5314,9 @@ static int bpf_map_get_info_by_fd(struct file *file,
 		if (info.hash_size != SHA256_DIGEST_SIZE)
 			return -EINVAL;
 
+		if (!READ_ONCE(map->frozen))
+			return -EPERM;
+
 		err = map->ops->map_get_hash(map, SHA256_DIGEST_SIZE, map->sha);
 		if (err != 0)
 			return err;

2.51.0

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

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

From: Zilin Guan <[email protected]>

[ Upstream commit ccb679fdae2e62ed92fd9acb25ed809c0226fcc6 ]

The starfive_aes_aead_do_one_req() function allocates rctx->adata with kzalloc() but fails to free it if sg_copy_to_buffer() or starfive_aes_hw_init() fails, which lead to memory leaks.

Since rctx->adata is unconditionally freed after the write_adata operations, ensure consistent cleanup by freeing the allocation in these earlier error paths as well.

Compile tested only. Issue found using a prototype static analysis tool and code review.

---
 drivers/crypto/starfive/jh7110-aes.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)

index 426b24889af85..01195664cc7cd 100644
@@ -669,8 +669,10 @@ static int starfive_aes_aead_do_one_req(struct crypto_engine *engine, void *areq
 			return -ENOMEM;
 
 		if (sg_copy_to_buffer(req->src, sg_nents_for_len(req->src, cryp->assoclen),
-				      rctx->adata, cryp->assoclen) != cryp->assoclen)
+				      rctx->adata, cryp->assoclen) != cryp->assoclen) {
+			kfree(rctx->adata);
 			return -EINVAL;
+		}
 	}
 
 	if (cryp->total_in)

@@ -681,8 +683,11 @@ static int starfive_aes_aead_do_one_req(struct crypto_engine *engine, void *areq
 	ctx->rctx = rctx;
 
 	ret = starfive_aes_hw_init(ctx);
-	if (ret)
+	if (ret) {
+		if (cryp->assoclen)
+			kfree(rctx->adata);
 		return ret;
+	}
 
 	if (!cryp->assoclen)
 		goto write_text;

2.51.0

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

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

From: Lianjie Wang <[email protected]>

[ Upstream commit cc2f39d6ac48e6e3cb2d6240bc0d6df839dd0828 ]

Currently, hwrng_fill is not cleared until the hwrng_fillfn() thread exits. Since hwrng_unregister() reads hwrng_fill outside the rng_mutex lock, a concurrent hwrng_unregister() may call kthread_stop() again on the same task.

Additionally, if hwrng_unregister() is called immediately after hwrng_register(), the stopped thread may have never been executed. Thus, hwrng_fill remains dirty even after hwrng_unregister() returns. In this case, subsequent calls to hwrng_register() will fail to start new threads, and hwrng_unregister() will call kthread_stop() on the same freed task. In both cases, a use-after-free occurs:

refcount_t: addition on 0; use-after-free.
WARNING: ... at lib/refcount.c:25 refcount_warn_saturate+0xec/0x1c0

Call Trace:
 kthread_stop+0x181/0x360
 hwrng_unregister+0x288/0x380
 virtrng_remove+0xe3/0x200

This patch fixes the race by protecting the global hwrng_fill pointer inside the rng_mutex lock, so that hwrng_fillfn() thread is stopped only once, and calls to kthread_run() and kthread_stop() are serialized with the lock held.

To avoid deadlock in hwrng_fillfn() while being stopped with the lock held, we convert current_rng to RCU, so that get_current_rng() can read current_rng without holding the lock. To remove the lock from put_rng(), we also delay the actual cleanup into a work_struct.

Since get_current_rng() no longer returns ERR_PTR values, the IS_ERR() checks are removed from its callers.

With hwrng_fill protected by the rng_mutex lock, hwrng_fillfn() can no longer clear hwrng_fill itself. Therefore, if hwrng_fillfn() returns directly after current_rng is dropped, kthread_stop() would be called on a freed task_struct later. To fix this, hwrng_fillfn() calls schedule() now to keep the task alive until being stopped. The kthread_stop() call is also moved from hwrng_unregister() to drop_current_rng(), ensuring kthread_stop() is called on all possible paths where current_rng becomes NULL, so that the thread would not wait forever.

---
 drivers/char/hw_random/core.c | 168 +++++++++++++++++++++-------------
 include/linux/hw_random.h     |   2 +
 2 files changed, 107 insertions(+), 63 deletions(-)

index 96d7fe41b373d..aba92d777f726 100644
@@ -20,23 +20,25 @@
 #include <linux/miscdevice.h>
 #include <linux/module.h>
 #include <linux/random.h>
+#include <linux/rcupdate.h>
 #include <linux/sched.h>
 #include <linux/sched/signal.h>
 #include <linux/slab.h>
 #include <linux/string.h>
 #include <linux/uaccess.h>
+#include <linux/workqueue.h>
 
 #define RNG_MODULE_NAME		"hw_random"
 
 #define RNG_BUFFER_SIZE (SMP_CACHE_BYTES < 32 ? 32 : SMP_CACHE_BYTES)
 
-static struct hwrng *current_rng;
+static struct hwrng __rcu *current_rng;
 /* the current rng has been explicitly chosen by user via sysfs */
 static int cur_rng_set_by_user;
 static struct task_struct *hwrng_fill;
 /* list of registered rngs */
 static LIST_HEAD(rng_list);
-/* Protects rng_list and current_rng */
+/* Protects rng_list, hwrng_fill and updating on current_rng */
 static DEFINE_MUTEX(rng_mutex);
 /* Protects rng read functions, data_avail, rng_buffer and rng_fillbuf */
 static DEFINE_MUTEX(reading_mutex);

@@ -64,18 +66,39 @@ static size_t rng_buffer_size(void)
 	return RNG_BUFFER_SIZE;
 }
 
-static inline void cleanup_rng(struct kref *kref)
+static void cleanup_rng_work(struct work_struct *work)
 {
-	struct hwrng *rng = container_of(kref, struct hwrng, ref);
+	struct hwrng *rng = container_of(work, struct hwrng, cleanup_work);
+
+	/*
+	 * Hold rng_mutex here so we serialize in case they set_current_rng
+	 * on rng again immediately.
+	 */
+	mutex_lock(&rng_mutex);
+
+	/* Skip if rng has been reinitialized. */
+	if (kref_read(&rng->ref)) {
+		mutex_unlock(&rng_mutex);
+		return;
+	}
 
 	if (rng->cleanup)
 		rng->cleanup(rng);
 
 	complete(&rng->cleanup_done);
+	mutex_unlock(&rng_mutex);
+}
+
+static inline void cleanup_rng(struct kref *kref)
+{
+	struct hwrng *rng = container_of(kref, struct hwrng, ref);
+
+	schedule_work(&rng->cleanup_work);
 }
 
 static int set_current_rng(struct hwrng *rng)
 {
+	struct hwrng *old_rng;
 	int err;
 
 	BUG_ON(!mutex_is_locked(&rng_mutex));

@@ -84,8 +107,14 @@ static int set_current_rng(struct hwrng *rng)
 	if (err)
 		return err;
 
-	drop_current_rng();
-	current_rng = rng;
+	old_rng = rcu_dereference_protected(current_rng,
+					    lockdep_is_held(&rng_mutex));
+	rcu_assign_pointer(current_rng, rng);
+
+	if (old_rng) {
+		synchronize_rcu();
+		kref_put(&old_rng->ref, cleanup_rng);
+	}
 
 	/* if necessary, start hwrng thread */
 	if (!hwrng_fill) {

@@ -101,47 +130,56 @@ static int set_current_rng(struct hwrng *rng)
 
 static void drop_current_rng(void)
 {
-	BUG_ON(!mutex_is_locked(&rng_mutex));
-	if (!current_rng)
+	struct hwrng *rng;
+
+	rng = rcu_dereference_protected(current_rng,
+					lockdep_is_held(&rng_mutex));
+	if (!rng)
 		return;
 
+	RCU_INIT_POINTER(current_rng, NULL);
+	synchronize_rcu();
+
+	if (hwrng_fill) {
+		kthread_stop(hwrng_fill);
+		hwrng_fill = NULL;
+	}
+
 	/* decrease last reference for triggering the cleanup */
-	kref_put(&current_rng->ref, cleanup_rng);
-	current_rng = NULL;
+	kref_put(&rng->ref, cleanup_rng);
 }
 
-/* Returns ERR_PTR(), NULL or refcounted hwrng */
+/* Returns NULL or refcounted hwrng */
 static struct hwrng *get_current_rng_nolock(void)
 {
-	if (current_rng)
-		kref_get(&current_rng->ref);
+	struct hwrng *rng;
+
+	rng = rcu_dereference_protected(current_rng,
+					lockdep_is_held(&rng_mutex));
+	if (rng)
+		kref_get(&rng->ref);
 
-	return current_rng;
+	return rng;
 }
 
 static struct hwrng *get_current_rng(void)
 {
 	struct hwrng *rng;
 
-	if (mutex_lock_interruptible(&rng_mutex))
-		return ERR_PTR(-ERESTARTSYS);
+	rcu_read_lock();
+	rng = rcu_dereference(current_rng);
+	if (rng)
+		kref_get(&rng->ref);
 
-	rng = get_current_rng_nolock();
+	rcu_read_unlock();
 
-	mutex_unlock(&rng_mutex);
 	return rng;
 }
 
 static void put_rng(struct hwrng *rng)
 {
-	/*
-	 * Hold rng_mutex here so we serialize in case they set_current_rng
-	 * on rng again immediately.
-	 */
-	mutex_lock(&rng_mutex);
 	if (rng)
 		kref_put(&rng->ref, cleanup_rng);
-	mutex_unlock(&rng_mutex);
 }
 
 static int hwrng_init(struct hwrng *rng)

@@ -213,10 +251,6 @@ static ssize_t rng_dev_read(struct file *filp, char __user *buf,
 
 	while (size) {
 		rng = get_current_rng();
-		if (IS_ERR(rng)) {
-			err = PTR_ERR(rng);
-			goto out;
-		}
 		if (!rng) {
 			err = -ENODEV;
 			goto out;

@@ -303,7 +337,7 @@ static struct miscdevice rng_miscdev = {
 
 static int enable_best_rng(void)
 {
-	struct hwrng *rng, *new_rng = NULL;
+	struct hwrng *rng, *cur_rng, *new_rng = NULL;
 	int ret = -ENODEV;
 
 	BUG_ON(!mutex_is_locked(&rng_mutex));

@@ -321,7 +355,9 @@ static int enable_best_rng(void)
 			new_rng = rng;
 	}
 
-	ret = ((new_rng == current_rng) ? 0 : set_current_rng(new_rng));
+	cur_rng = rcu_dereference_protected(current_rng,
+					    lockdep_is_held(&rng_mutex));
+	ret = ((new_rng == cur_rng) ? 0 : set_current_rng(new_rng));
 	if (!ret)
 		cur_rng_set_by_user = 0;
 

@@ -371,8 +407,6 @@ static ssize_t rng_current_show(struct device *dev,
 	struct hwrng *rng;
 
 	rng = get_current_rng();
-	if (IS_ERR(rng))
-		return PTR_ERR(rng);
 
 	ret = sysfs_emit(buf, "%s\n", rng ? rng->name : "none");
 	put_rng(rng);

@@ -416,8 +450,6 @@ static ssize_t rng_quality_show(struct device *dev,
 	struct hwrng *rng;
 
 	rng = get_current_rng();
-	if (IS_ERR(rng))
-		return PTR_ERR(rng);
 
 	if (!rng) /* no need to put_rng */
 		return -ENODEV;

@@ -432,6 +464,7 @@ static ssize_t rng_quality_store(struct device *dev,
 				 struct device_attribute *attr,
 				 const char *buf, size_t len)
 {
+	struct hwrng *rng;
 	u16 quality;
 	int ret = -EINVAL;
 

@@ -448,12 +481,13 @@ static ssize_t rng_quality_store(struct device *dev,
 		goto out;
 	}
 
-	if (!current_rng) {
+	rng = rcu_dereference_protected(current_rng, lockdep_is_held(&rng_mutex));
+	if (!rng) {
 		ret = -ENODEV;
 		goto out;
 	}
 
-	current_rng->quality = quality;
+	rng->quality = quality;
 	current_quality = quality; /* obsolete */
 
 	/* the best available RNG may have changed */

@@ -489,8 +523,20 @@ static int hwrng_fillfn(void *unused)
 		struct hwrng *rng;
 
 		rng = get_current_rng();
-		if (IS_ERR(rng) || !rng)
+		if (!rng) {
+			/*
+			 * Keep the task_struct alive until kthread_stop()
+			 * is called to avoid UAF in drop_current_rng().
+			 */
+			while (!kthread_should_stop()) {
+				set_current_state(TASK_INTERRUPTIBLE);
+				if (!kthread_should_stop())
+					schedule();
+			}
+			set_current_state(TASK_RUNNING);
 			break;
+		}
+
 		mutex_lock(&reading_mutex);
 		rc = rng_get_data(rng, rng_fillbuf,
 				  rng_buffer_size(), 1);

@@ -518,14 +564,13 @@ static int hwrng_fillfn(void *unused)
 		add_hwgenerator_randomness((void *)rng_fillbuf, rc,
 					   entropy >> 10, true);
 	}
-	hwrng_fill = NULL;
 	return 0;
 }
 
 int hwrng_register(struct hwrng *rng)
 {
 	int err = -EINVAL;
-	struct hwrng *tmp;
+	struct hwrng *cur_rng, *tmp;
 
 	if (!rng->name || (!rng->data_read && !rng->read))
 		goto out;

@@ -540,6 +585,7 @@ int hwrng_register(struct hwrng *rng)
 	}
 	list_add_tail(&rng->list, &rng_list);
 
+	INIT_WORK(&rng->cleanup_work, cleanup_rng_work);
 	init_completion(&rng->cleanup_done);
 	complete(&rng->cleanup_done);
 	init_completion(&rng->dying);

@@ -547,16 +593,19 @@ int hwrng_register(struct hwrng *rng)
 	/* Adjust quality field to always have a proper value */
 	rng->quality = min3(default_quality, 1024, rng->quality ?: 1024);
 
-	if (!cur_rng_set_by_user &&
-	    (!current_rng || rng->quality > current_rng->quality)) {
-		/*
-		 * Set new rng as current as the new rng source
-		 * provides better entropy quality and was not
-		 * chosen by userspace.
-		 */
-		err = set_current_rng(rng);
-		if (err)
-			goto out_unlock;
+	if (!cur_rng_set_by_user) {
+		cur_rng = rcu_dereference_protected(current_rng,
+						    lockdep_is_held(&rng_mutex));
+		if (!cur_rng || rng->quality > cur_rng->quality) {
+			/*
+			 * Set new rng as current as the new rng source
+			 * provides better entropy quality and was not
+			 * chosen by userspace.
+			 */
+			err = set_current_rng(rng);
+			if (err)
+				goto out_unlock;
+		}
 	}
 	mutex_unlock(&rng_mutex);
 	return 0;

@@ -569,14 +618,17 @@ EXPORT_SYMBOL_GPL(hwrng_register);
 
 void hwrng_unregister(struct hwrng *rng)
 {
-	struct hwrng *new_rng;
+	struct hwrng *cur_rng;
 	int err;
 
 	mutex_lock(&rng_mutex);
 
 	list_del(&rng->list);
 	complete_all(&rng->dying);
-	if (current_rng == rng) {
+
+	cur_rng = rcu_dereference_protected(current_rng,
+					    lockdep_is_held(&rng_mutex));
+	if (cur_rng == rng) {
 		err = enable_best_rng();
 		if (err) {
 			drop_current_rng();

@@ -584,17 +636,7 @@ void hwrng_unregister(struct hwrng *rng)
 		}
 	}
 
-	new_rng = get_current_rng_nolock();
-	if (list_empty(&rng_list)) {
-		mutex_unlock(&rng_mutex);
-		if (hwrng_fill)
-			kthread_stop(hwrng_fill);
-	} else
-		mutex_unlock(&rng_mutex);
-
-	if (new_rng)
-		put_rng(new_rng);
-
+	mutex_unlock(&rng_mutex);
 	wait_for_completion(&rng->cleanup_done);
 }
 EXPORT_SYMBOL_GPL(hwrng_unregister);

@@ -682,7 +724,7 @@ static int __init hwrng_modinit(void)
 static void __exit hwrng_modexit(void)
 {
 	mutex_lock(&rng_mutex);
-	BUG_ON(current_rng);
+	WARN_ON(rcu_access_pointer(current_rng));
 	kfree(rng_buffer);
 	kfree(rng_fillbuf);
 	mutex_unlock(&rng_mutex);

index b424555753b11..b77bc55a4cf35 100644
@@ -15,6 +15,7 @@
 #include <linux/completion.h>
 #include <linux/kref.h>
 #include <linux/types.h>
+#include <linux/workqueue_types.h>
 
 /**
  * struct hwrng - Hardware Random Number Generator driver

@@ -48,6 +49,7 @@ struct hwrng {
 	/* internal. */
 	struct list_head list;
 	struct kref ref;
+	struct work_struct cleanup_work;
 	struct completion cleanup_done;
 	struct completion dying;
 };

2.51.0

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

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

From: Larysa Zaremba <[email protected]>

[ Upstream commit 42e41b2a0afa04ca49ee2725aadf90ccb058ed28 ]

Referenced commit reduced the scope of the variable pkt, so now it has to be reinitialized via pkt_stream_get_next_rx_pkt(), which also increments some counters. When the packet is interrupted by the batch ending, pkt stream therefore proceeds to the next packet, while xsk ring still contains the previous one, this results in a pkt_nb mismatch.

Decrement the affected counters when packet is interrupted.

---
 tools/testing/selftests/bpf/prog_tests/test_xsk.c | 2 ++
 1 file changed, 2 insertions(+)

index 5af28f359cfda..69a5a9a5189b0 100644
@@ -1090,6 +1090,8 @@ static int __receive_pkts(struct test_spec *test, struct xsk_socket_info *xsk)
 			xsk_ring_prod__cancel(&umem->fq, nb_frags);
 		}
 		frags_processed -= nb_frags;
+		pkt_stream_cancel(pkt_stream);
+		pkts_sent--;
 	}
 
 	if (ifobj->use_fill_ring)

2.51.0

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

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

From: Larysa Zaremba <[email protected]>

[ Upstream commit 88af9fefed412e4bea9a1a771cbe6fe347fa3507 ]

The issue occurs in TOO_MANY_FRAGS test case when xdp_zc_max_segs is set to an odd number.

TOO_MANY_FRAGS test case contains an invalid packet consisting of (xdp_zc_max_segs) frags. Every frag, even the last one has XDP_PKT_CONTD flag set. This packet is expected to be dropped. After that, there is a valid linear packet, which is expected to be received back.

Once (xdp_zc_max_segs) is an odd number, the last packet cannot be received, if packet forwarding between Rx and Tx interfaces relies on the ethernet header, e.g. checks for ETH_P_LOOPBACK. Packet is malformed, if all traffic is looped.

Turns out, sending function processes multiple invalid frags as if they were in 2-frag packets. So once the invalid mbuf packet contains an odd number of those, the valid packet after gets paired with the previous invalid descriptor, and hence does not get an ethernet header generated, so it is either dropped or malformed.

Make invalid packets in verbatim mode always have only a single frag. For such packets, number of frags is otherwise meaningless, as descriptor flags are pre-configured in verbatim mode and packet data is not generated for invalid descriptors.

---
 tools/testing/selftests/bpf/prog_tests/test_xsk.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 69a5a9a5189b0..bab4a31621c75 100644
@@ -433,7 +433,7 @@ static u32 pkt_nb_frags(u32 frame_size, struct pkt_stream *pkt_stream, struct pk
 	}
 
 	/* Search for the end of the packet in verbatim mode */
-	if (!pkt_continues(pkt->options))
+	if (!pkt_continues(pkt->options) || !pkt->valid)
 		return nb_frags;
 
 	next_frag = pkt_stream->current_pkt_nb;

2.51.0

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

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

From: Sai Ritvik Tanksalkar <[email protected]>

[ Upstream commit 5669645c052f235726a85f443769b6fc02f66762 ]

persistent_ram_save_old() can be called multiple times for the same persistent_ram_zone (e.g., via ramoops_pstore_read -> ramoops_get_next_prz for PSTORE_TYPE_DMESG records).

Currently, the function only allocates prz->old_log when it is NULL, but it unconditionally updates prz->old_log_size to the current buffer size and then performs memcpy_fromio() using this new size. If the buffer size has grown since the first allocation (which can happen across different kernel boot cycles), this leads to:

1. A heap buffer overflow (OOB write) in the memcpy_fromio() calls
2. A subsequent OOB read when ramoops_pstore_read() accesses the buffer using the incorrect (larger) old_log_size

The KASAN splat would look similar to:

  BUG: KASAN: slab-out-of-bounds in ramoops_pstore_read+0x...
  Read of size N at addr ... by task ...

The conditions are likely extremely hard to hit:

0. Crash with a ramoops write of less-than-record-max-size bytes.
1. Reboot: ramoops registers, pstore_get_records(0) reads old crash, allocates old_log with size X
2. Crash handler registered, timer started (if pstore_update_ms >= 0)
3. Oops happens (non-fatal, system continues)
4. pstore_dump() writes oops via ramoops_pstore_write() size Y (>X)
5. pstore_new_entry = 1, pstore_timer_kick() called
6. System continues running (not a panic oops)
7. Timer fires after pstore_update_ms milliseconds
8. pstore_timefunc() → schedule_work() → pstore_dowork() → pstore_get_records(1)
9. ramoops_get_next_prz() → persistent_ram_save_old()
10. buffer_size() returns Y, but old_log is X bytes
11. Y > X: memcpy_fromio() overflows heap

Requirements:
- a prior crash record exists that did not fill the record size (almost impossible since the crash handler writes as much as it can possibly fit into the record, capped by max record size and the kmsg buffer almost always exceeds the max record size)
- pstore_update_ms >= 0 (disabled by default)
- Non-fatal oops (system survives)

Free and reallocate the buffer when the new size differs from the previously allocated size. This ensures old_log always has sufficient space for the data being copied.

---
 fs/pstore/ram_core.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

index f1848cdd6d348..c9eaacdec37e4 100644
@@ -298,6 +298,17 @@ void persistent_ram_save_old(struct persistent_ram_zone *prz)
 	if (!size)
 		return;
 
+	/*
+	 * If the existing buffer is differently sized, free it so a new
+	 * one is allocated. This can happen when persistent_ram_save_old()
+	 * is called early in boot and later for a timer-triggered
+	 * survivable crash when the crash dumps don't match in size
+	 * (which would be extremely unlikely given kmsg buffers usually
+	 * exceed prz buffer sizes).
+	 */
+	if (prz->old_log && prz->old_log_size != size)
+		persistent_ram_free_old(prz);
+
 	if (!prz->old_log) {
 		persistent_ram_ecc_old(prz);
 		prz->old_log = kvzalloc(size, GFP_KERNEL);

2.51.0

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

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

From: Francesco Dolcini <[email protected]>

[ Upstream commit 36ee9f8b1ae07fe82885a7a3553a5be347d3f978 ]

Change USB-C Sink PDO and the amount of power that the device can sink to zero to maximize compatibility with other USB peers (the Aquila Development Board is not sinking any current, it is self powered).

---
 arch/arm64/boot/dts/ti/k3-am69-aquila-dev.dts | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index c7ce804eac703..f48601ae38b7c 100644
@@ -399,8 +399,8 @@ connector {
 			try-power-role = "sink";
 			self-powered;
 			source-pdos = <PDO_FIXED(5000, 3000, PDO_FIXED_USB_COMM)>;
-			sink-pdos = <PDO_FIXED(5000, 3000, PDO_FIXED_USB_COMM)>;
-			op-sink-microwatt = <1000000>;
+			sink-pdos = <PDO_FIXED(5000, 0, PDO_FIXED_USB_COMM)>;
+			op-sink-microwatt = <0>;
 
 			ports {
 				#address-cells = <1>;

2.51.0

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

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

From: Francesco Dolcini <[email protected]>

[ Upstream commit b548f3949937b55ee19ab418343f05700fdf7009 ]

Change USB-C Sink PDO and the amount of power that the device can sink to zero to maximize compatibility with other USB peers (the Aquila Clover Board is not sinking any current, it is self powered).

---
 arch/arm64/boot/dts/ti/k3-am69-aquila-clover.dts | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 55fd214a82e44..c816ba3bfbdf7 100644
@@ -280,8 +280,8 @@ connector {
 			try-power-role = "sink";
 			self-powered;
 			source-pdos = <PDO_FIXED(5000, 3000, PDO_FIXED_USB_COMM)>;
-			sink-pdos = <PDO_FIXED(5000, 3000, PDO_FIXED_USB_COMM)>;
-			op-sink-microwatt = <1000000>;
+			sink-pdos = <PDO_FIXED(5000, 0, PDO_FIXED_USB_COMM)>;
+			op-sink-microwatt = <0>;
 
 			ports {
 				#address-cells = <1>;

2.51.0

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

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

From: Jorge Ramirez-Ortiz <[email protected]>

[ Upstream commit 0fe01a7955f4fef97e7cc6d14bfc5931c660402b ]

Fail the driver probe if the region can't be mapped

---
 drivers/soc/qcom/smem.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

index c18a0c946f762..d5c94b47f431f 100644
@@ -1219,7 +1219,9 @@ static int qcom_smem_probe(struct platform_device *pdev)
 		smem->item_count = qcom_smem_get_item_count(smem);
 		break;
 	case SMEM_GLOBAL_HEAP_VERSION:
-		qcom_smem_map_global(smem, size);
+		ret = qcom_smem_map_global(smem, size);
+		if (ret < 0)
+			return ret;
 		smem->item_count = SMEM_ITEM_COUNT;
 		break;
 	default:

2.51.0

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

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

From: Dan Carpenter <[email protected]>

[ Upstream commit 7b5c7e83ac405ff9ecbdd92b37a477f4288f8814 ]

The snprintf() can't really overflow because we're writing a max of 42 bytes to a PAGE_SIZE buffer. But the limit calculation doesn't take the first 11 bytes that we wrote into consideration so the limit is not correct. Just fix it for correctness even though it doesn't affect runtime.

---
 drivers/edac/i5000_edac.c | 1 +
 1 file changed, 1 insertion(+)

index 4a1bebc1ff14d..471b8540d18b0 100644
@@ -1111,6 +1111,7 @@ static void calculate_dimm_size(struct i5000_pvt *pvt)
 
 	n = snprintf(p, space, "           ");
 	p += n;
+	space -= n;
 	for (branch = 0; branch < MAX_BRANCHES; branch++) {
 		n = snprintf(p, space, "       branch %d       | ", branch);
 		p += n;

2.51.0

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

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

From: Dan Carpenter <[email protected]>

[ Upstream commit 72f12683611344853ab030fe7d19b23970ed2bd8 ]

The snprintf() can't really overflow because we're writing a max of 42 bytes to a PAGE_SIZE buffer. But my static checker complains because the limit calculation doesn't take the first 11 space characters that we wrote into the buffer into consideration. Fix this for the sake of correctness even though it doesn't affect runtime.

Also delete an earlier "space -= n;" which was not used.

---
 drivers/edac/i5400_edac.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index b5cf25905b059..fb49a1d1df112 100644
@@ -1026,13 +1026,13 @@ static void calculate_dimm_size(struct i5400_pvt *pvt)
 		space -= n;
 	}
 
-	space -= n;
 	edac_dbg(2, "%s\n", mem_buffer);
 	p = mem_buffer;
 	space = PAGE_SIZE;
 
 	n = snprintf(p, space, "           ");
 	p += n;
+	space -= n;
 	for (branch = 0; branch < MAX_BRANCHES; branch++) {
 		n = snprintf(p, space, "       branch %d       | ", branch);
 		p += n;

2.51.0

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

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

From: Sudeep Holla <[email protected]>

[ Upstream commit be4d4543f78074fbebd530ba5109d39a2a34e668 ]

The FF-A specification allows NOTIFICATION_INFO_GET to return either a 64-bit (FFA_FN64_SUCCESS) or a 32-bit (FFA_SUCCESS) response, depending on whether the firmware chooses the SMC64 or SMC32 calling convention.

The driver previously detected the response format by checking ret.a0, but still interpreted the returned ID lists (x3..x17 or w3..w7) as if they always followed the 64-bit SMC64 layout. In the SMC32 case, the upper 32 bits of each argument register are undefined by the calling convention, meaning the driver could read stale or garbage values when parsing notification IDs.

This resulted in incorrectly decoded partition/VCPU IDs whenever the FF-A firmware used an SMC32 return path.

Fix the issue by:

- Introducing logic to map list indices to the correct u16 offsets, depending on whether the response width matches the kernel word size or is a 32-bit response on a 64-bit kernel.
- Ensuring that the packed ID list is parsed using the proper layout, avoiding reads from undefined upper halves in the SMC32 case.

With this change, NOTIFICATION_INFO_GET now correctly interprets ID list entries regardless of the response width, aligning the driver with the FF-A specification.

Message-Id: <[email protected]>

---
 drivers/firmware/arm_ffa/driver.c | 33 +++++++++++++++++++++++++++----
 1 file changed, 29 insertions(+), 4 deletions(-)

index c72ee47565856..c501c3104b3a4 100644
@@ -981,10 +981,27 @@ static void __do_sched_recv_cb(u16 part_id, u16 vcpu, bool is_per_vcpu)
 	}
 }
 
+/*
+ * Map logical ID index to the u16 index within the packed ID list.
+ *
+ * For native responses (FF-A width == kernel word size), IDs are
+ * tightly packed: idx -> idx.
+ *
+ * For 32-bit responses on a 64-bit kernel, each 64-bit register
+ * contributes 4 x u16 values but only the lower 2 are defined; the
+ * upper 2 are garbage. This mapping skips those upper halves:
+ *   0,1,2,3,4,5,... -> 0,1,4,5,8,9,...
+ */
+static int list_idx_to_u16_idx(int idx, bool is_native_resp)
+{
+	return is_native_resp ? idx : idx + 2 * (idx >> 1);
+}
+
 static void ffa_notification_info_get(void)
 {
-	int idx, list, max_ids, lists_cnt, ids_processed, ids_count[MAX_IDS_64];
-	bool is_64b_resp;
+	int ids_processed, ids_count[MAX_IDS_64];
+	int idx, list, max_ids, lists_cnt;
+	bool is_64b_resp, is_native_resp;
 	ffa_value_t ret;
 	u64 id_list;
 

@@ -1001,6 +1018,7 @@ static void ffa_notification_info_get(void)
 		}
 
 		is_64b_resp = (ret.a0 == FFA_FN64_SUCCESS);
+		is_native_resp = (ret.a0 == FFA_FN_NATIVE(SUCCESS));
 
 		ids_processed = 0;
 		lists_cnt = FIELD_GET(NOTIFICATION_INFO_GET_ID_COUNT, ret.a2);

@@ -1017,12 +1035,16 @@ static void ffa_notification_info_get(void)
 
 		/* Process IDs */
 		for (list = 0; list < lists_cnt; list++) {
+			int u16_idx;
 			u16 vcpu_id, part_id, *packed_id_list = (u16 *)&ret.a3;
 
 			if (ids_processed >= max_ids - 1)
 				break;
 
-			part_id = packed_id_list[ids_processed++];
+			u16_idx = list_idx_to_u16_idx(ids_processed,
+						      is_native_resp);
+			part_id = packed_id_list[u16_idx];
+			ids_processed++;
 
 			if (ids_count[list] == 1) { /* Global Notification */
 				__do_sched_recv_cb(part_id, 0, false);

@@ -1034,7 +1056,10 @@ static void ffa_notification_info_get(void)
 				if (ids_processed >= max_ids - 1)
 					break;
 
-				vcpu_id = packed_id_list[ids_processed++];
+				u16_idx = list_idx_to_u16_idx(ids_processed,
+							      is_native_resp);
+				vcpu_id = packed_id_list[u16_idx];
+				ids_processed++;
 
 				__do_sched_recv_cb(part_id, vcpu_id, true);
 			}

2.51.0

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

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

From: Anton D. Stavinskii <[email protected]>

[ Upstream commit 03ea8676919af21b99bea01f18ef1a271d19f92f ]

I've tested the current dwc2 FIFO configuration and found that USB device mode breaks in ECM mode when transmitting frames larger than 128 bytes. For example, large ICMP packets or iperf3 traffic cause the USB link to hang and eventually disconnect without any messages in dmesg.

After switching to more conservative FIFO sizes, ECM becomes stable and no longer drops the connection. iperf3 now shows ~130 Mbit/s RX and ~100 Mbit/s TX on SG2002 (MilkV Duo 256M).

Fix the FIFO sizes accordingly.

---
 arch/riscv/boot/dts/sophgo/cv180x.dtsi | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 1b2b1969a6484..06b0ce5a2db7a 100644
@@ -438,8 +438,8 @@ usb: usb@4340000 {
 			clocks = <&clk CLK_AXI4_USB>, <&clk CLK_APB_USB>;
 			clock-names = "otg", "utmi";
 			g-np-tx-fifo-size = <32>;
-			g-rx-fifo-size = <536>;
-			g-tx-fifo-size = <768 512 512 384 128 128>;
+			g-rx-fifo-size = <1536>;
+			g-tx-fifo-size = <128 128 64 64 64 64 32 32>;
 			interrupts = <SOC_PERIPHERAL_IRQ(14) IRQ_TYPE_LEVEL_HIGH>;
 			phys = <&usbphy>;
 			phy-names = "usb2-phy";

2.51.0

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

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

From: Alexander Stein <[email protected]>

[ Upstream commit 8401527abb5e3a00c867b6597b8e1b29c80c9824 ]

As per datasheet of the HDMI protection IC the CEC_IC pin has been configured as open-drain.

---
 arch/arm64/boot/dts/freescale/imx8mp-tqma8mpql-mba8mpxl.dts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 59642a8a2c445..c73d40fb789f6 100644
@@ -867,7 +867,7 @@ pinctrl_hdmi: hdmigrp {
 		fsl,pins = <MX8MP_IOMUXC_HDMI_DDC_SCL__HDMIMIX_HDMI_SCL	0x400001c2>,
 			   <MX8MP_IOMUXC_HDMI_DDC_SDA__HDMIMIX_HDMI_SDA	0x400001c2>,
 			   <MX8MP_IOMUXC_HDMI_HPD__HDMIMIX_HDMI_HPD	0x40000010>,
-			   <MX8MP_IOMUXC_HDMI_CEC__HDMIMIX_HDMI_CEC	0x40000010>;
+			   <MX8MP_IOMUXC_HDMI_CEC__HDMIMIX_HDMI_CEC	0x40000030>;
 	};
 
 	pinctrl_hoggpio2: hoggpio2grp {

2.51.0

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

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

From: Alexander Stein <[email protected]>

[ Upstream commit 53a5c1d98d1155ece4c9446c0fea55e17d08774a ]

As per datasheet of the HDMI protection IC the CEC_IC pin has been configured as open-drain.

---
 .../arm64/boot/dts/freescale/imx8mp-tqma8mpql-mba8mp-ras314.dts | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index f7346b3d35fe5..a122f2ed5f531 100644
@@ -704,7 +704,7 @@ pinctrl_hdmi: hdmigrp {
 		fsl,pins = <MX8MP_IOMUXC_HDMI_DDC_SCL__HDMIMIX_HDMI_SCL	0x400001c2>,
 			   <MX8MP_IOMUXC_HDMI_DDC_SDA__HDMIMIX_HDMI_SDA	0x400001c2>,
 			   <MX8MP_IOMUXC_HDMI_HPD__HDMIMIX_HDMI_HPD	0x40000010>,
-			   <MX8MP_IOMUXC_HDMI_CEC__HDMIMIX_HDMI_CEC	0x40000154>;
+			   <MX8MP_IOMUXC_HDMI_CEC__HDMIMIX_HDMI_CEC	0x40000030>;
 	};
 
 	pinctrl_gpt1: gpt1grp {

2.51.0

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

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

From: Arnd Bergmann <[email protected]>

[ Upstream commit c816ba1dcd931b9db8d66c71e1ae34ddcdbf968f ]

Using a variable as a format string causes a (default-disabled) warning:

drivers/edac/amd64_edac.c: In function 'per_family_init':
drivers/edac/amd64_edac.c:3914:17: error: format not a string literal and no format arguments [-Werror=format-security]

   3914 |                 scnprintf(pvt->ctl_name, sizeof(pvt->ctl_name), tmp_name);
        |                 ^~~~~~~~~

The code here is safe, but in order to enable the warning by default in the future, change this instance to pass the name indirectly.

---
 drivers/edac/amd64_edac.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 2391f3469961a..63fca0ee2c23b 100644
@@ -3911,7 +3911,7 @@ static int per_family_init(struct amd64_pvt *pvt)
 	}
 
 	if (tmp_name)
-		scnprintf(pvt->ctl_name, sizeof(pvt->ctl_name), tmp_name);
+		scnprintf(pvt->ctl_name, sizeof(pvt->ctl_name), "%s", tmp_name);
 	else
 		scnprintf(pvt->ctl_name, sizeof(pvt->ctl_name), "F%02Xh_M%02Xh",
 			  pvt->fam, pvt->model);

2.51.0

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

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

From: Haotian Zhang <[email protected]>

[ Upstream commit 1e07ebe744fb522983bd52a4a6148601675330c7 ]

When devm_platform_ioremap_resource() fails, it returns various error codes. Returning a hardcoded -ENOMEM masks the actual failure reason.

Use PTR_ERR() to propagate the actual error code returned by devm_platform_ioremap_resource() instead of -ENOMEM.

---
 drivers/clk/qcom/common.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 1215918867741..eec369d2173b5 100644
@@ -454,7 +454,7 @@ int qcom_cc_probe_by_index(struct platform_device *pdev, int index,
 
 	base = devm_platform_ioremap_resource(pdev, index);
 	if (IS_ERR(base))
-		return -ENOMEM;
+		return PTR_ERR(base);
 
 	regmap = devm_regmap_init_mmio(&pdev->dev, base, desc->config);
 	if (IS_ERR(regmap))

2.51.0

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

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

From: Dmitry Baryshkov <[email protected]>

[ Upstream commit e814796dfcae8905682ac3ac2dd57f512a9f6726 ]

Historically sdm630.dtsi has used 1 byte length for the gpu_speed_bin cell, although it spans two bytes (offset 5, size 7 bits). It was being accepted by the kernel because before the commit 7a06ef751077 ("nvmem:
core: fix bit offsets of more than one byte") the kernel didn't have length check. After this commit nvmem core rejects QFPROM on sdm630 / sdm660, making GPU and USB unusable on those platforms.

Set the size of the gpu_speed_bin cell to 2 bytes, fixing the parsing error. While we are at it, update the length to 8 bits as pointed out by Alexey Minnekhanov.

---
 arch/arm64/boot/dts/qcom/sdm630.dtsi | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index 8b1a45a4e56ed..b383e480a394d 100644
@@ -598,8 +598,8 @@ qusb2_hstx_trim: hstx-trim@240 {
 			};
 
 			gpu_speed_bin: gpu-speed-bin@41a0 {
-				reg = <0x41a2 0x1>;
-				bits = <5 7>;
+				reg = <0x41a2 0x2>;
+				bits = <5 8>;
 			};
 		};
 

2.51.0

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

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

From: Dmitry Baryshkov <[email protected]>

[ Upstream commit d43019ef200d567454a8f68e60a5b2df01d8c706 ]

The DT file has GPU node enabled, but doesn't specify the file name of the ZAP firmware, which means using a default file name. Specify the name to the ZAP shader firmware, pointing to the file in the linux-firmware repo.

---
 arch/arm64/boot/dts/qcom/sm8150-hdk.dts | 4 ++++
 arch/arm64/boot/dts/qcom/sm8150-mtp.dts | 4 ++++
 2 files changed, 8 insertions(+)

index 0339a572f34d0..1eea9c5c66847 100644
@@ -387,6 +387,10 @@ &gpu {
 	status = "okay";
 };
 
+&gpu_zap_shader {
+	firmware-name = "qcom/sm8150/a640_zap.mbn";
+};
+
 &i2c4 {
 	clock-frequency = <100000>;
 

index 12e8e1ada6d8b..0f2d511624a8b 100644
@@ -358,6 +358,10 @@ &gpu {
 	status = "okay";
 };
 
+&gpu_zap_shader {
+	firmware-name = "qcom/sm8150/a640_zap.mbn";
+};
+
 &pon {
 	mode-bootloader = <0x2>;
 	mode-recovery = <0x1>;

2.51.0

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

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

From: Dmitry Baryshkov <[email protected]>

[ Upstream commit 95c121244a5d46435559bc74dbc7b5519394db08 ]

The DT file has GPU node enabled, but doesn't specify the file name of the ZAP firmware, which means using a default file name. Specify the name to the ZAP shader firmware, pointing to the file in the linux-firmware repo.

---
 arch/arm64/boot/dts/qcom/sm8250-hdk.dts | 4 ++++
 1 file changed, 4 insertions(+)

index f5c193c6c5f9b..3ea9d2b1a7d58 100644
@@ -373,6 +373,10 @@ &gpu {
 	status = "okay";
 };
 
+&gpu_zap_shader {
+	firmware-name = "qcom/sm8250/a650_zap.mbn";
+};
+
 &pon {
 	mode-bootloader = <0x2>;
 	mode-recovery = <0x1>;

2.51.0

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

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

From: Jingzhou Zhu <[email protected]>

[ Upstream commit a499c40ccd8e748ef363e2d13fb7a5c0ed6a788a ]

The adsp_mem and slpi_mem defined in sdm845.dtsi work well. Remove these nodes here to avoid redefinition.

Stable-dep-of: bc11f6f1d247 ("arm64: dts: qcom: sdm850-huawei-matebook-e-2019: Correct ipa_fw_mem for the driver to load successfully")

---
 .../boot/dts/qcom/sdm850-huawei-matebook-e-2019.dts   | 11 -----------
 1 file changed, 11 deletions(-)

index 0ef9ea38a424a..a5f025ae7dbe6 100644
@@ -30,9 +30,7 @@
 /delete-node/ &ipa_fw_mem;
 /delete-node/ &ipa_gsi_mem;
 /delete-node/ &gpu_mem;
-/delete-node/ &adsp_mem;
 /delete-node/ &wlan_msa_mem;
-/delete-node/ &slpi_mem;
 
 / {
 	model = "Huawei MateBook E 2019";

@@ -145,20 +143,11 @@ wlan_msa_mem: wlan-msa@8c400000 {
 			no-map;
 		};
 
-		adsp_mem: adsp@8c500000 {
-			reg = <0 0x8c500000 0 0x1a00000>;
-			no-map;
-		};
-
 		ipa_fw_mem: ipa-fw@8df00000 {
 			reg = <0 0x8df00000 0 0x100000>;
 			no-map;
 		};
 
-		slpi_mem: slpi@96700000 {
-			reg = <0 0x96700000 0 0x1200000>;
-		};
-
 		gpu_mem: gpu@97900000 {
 			reg = <0 0x97900000 0 0x5000>;
 			no-map;

2.51.0

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

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

From: Jingzhou Zhu <[email protected]>

[ Upstream commit bc11f6f1d2470fa59846be077555f9d4b7c2c0d3 ]

The ipa driver refuses to load with the old ipa_fw_mem in newer kernels. Shrinking its size to 0x5a000 fixes the problem.

---
 arch/arm64/boot/dts/qcom/sdm850-huawei-matebook-e-2019.dts | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

index a5f025ae7dbe6..f048653818702 100644
@@ -144,12 +144,12 @@ wlan_msa_mem: wlan-msa@8c400000 {
 		};
 
 		ipa_fw_mem: ipa-fw@8df00000 {
-			reg = <0 0x8df00000 0 0x100000>;
+			reg = <0 0x8df00000 0 0x5a000>;
 			no-map;
 		};
 
-		gpu_mem: gpu@97900000 {
-			reg = <0 0x97900000 0 0x5000>;
+		gpu_mem: gpu@8df5a000 {
+			reg = <0 0x8df5a000 0 0x5000>;
 			no-map;
 		};
 

2.51.0

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

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

From: Casey Connolly <[email protected]>

[ Upstream commit c9b98b9dad9749bf2eb7336a6fca31a6af1039d7 ]

The touchscreen isn't enabled by bootloader and doesn't need to be enabled at boot, only when the driver probes, thus remove the regulator-boot-on property.

---
 arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi | 1 -
 1 file changed, 1 deletion(-)

index db6dd04c51bb5..ee62adfa6af0b 100644
@@ -148,7 +148,6 @@ ts_1p8_supply: ts-1p8-regulator {
 
 		gpio = <&tlmm 88 0>;
 		enable-active-high;
-		regulator-boot-on;
 	};
 
 	panel_vci_3v3: panel-vci-3v3-regulator {

2.51.0

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

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

From: Casey Connolly <[email protected]>

[ Upstream commit 45d1f42d3e84b5880cf9fab1eb24a7818320eeb7 ]

The panel regulator doesn't need to be always on, so remove this property.

---
 arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi | 1 -
 1 file changed, 1 deletion(-)

index ee62adfa6af0b..5845bc3bb80b2 100644
@@ -272,7 +272,6 @@ vreg_l14a_1p88: ldo14 {
 			regulator-min-microvolt = <1800000>;
 			regulator-max-microvolt = <1800000>;
 			regulator-initial-mode = <RPMH_REGULATOR_MODE_HPM>;
-			regulator-always-on;
 		};
 
 		vreg_l17a_1p3: ldo17 {

2.51.0

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

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

From: Casey Connolly <[email protected]>

[ Upstream commit ad33ee060be46794a03d033894c9db3a9d6c1a0f ]

This regulator is used only for the display, which is enabled by the bootloader and left on for continuous splash. Mark it as such.

---
 arch/arm64/boot/dts/qcom/sdm845-oneplus-common.dtsi | 1 +
 1 file changed, 1 insertion(+)

index 5845bc3bb80b2..8251f5a2f9475 100644
@@ -272,6 +272,7 @@ vreg_l14a_1p88: ldo14 {
 			regulator-min-microvolt = <1800000>;
 			regulator-max-microvolt = <1800000>;
 			regulator-initial-mode = <RPMH_REGULATOR_MODE_HPM>;
+			regulator-boot-on;
 		};
 
 		vreg_l17a_1p3: ldo17 {

2.51.0

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

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

From: Abel Vesa <[email protected]>

[ Upstream commit 3af51501e2b8c87564b5cda43b0e5c316cf54717 ]

It seems the USB combo SS1 and SS2 ref clocks have another gate, unlike the SS0. These gates are part of the TCSR clock controller.

At least on Dell XPS 13 (9345), if the ref clock provided by the TCSR clock controller for SS1 PHY is disabled on the clk_disable_unused late initcall, the PHY fails to initialize. It doesn't happen on the SS0 PHY and the SS2 is not used on this device.

This doesn't seem to be a problem on CRD though. It might be that the RPMh has a vote for it from some other consumer and does not actually disable it when ther kernel drops its vote.

Either way, these TCSR provided clocks seem to be the correct ones for the SS1 and SS2, so use them instead.

---
 arch/arm64/boot/dts/qcom/hamoa.dtsi | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

index a17900eacb203..9c9e567731556 100644
@@ -2937,7 +2937,7 @@ usb_1_ss1_qmpphy: phy@fda000 {
 			reg = <0 0x00fda000 0 0x4000>;
 
 			clocks = <&gcc GCC_USB3_SEC_PHY_AUX_CLK>,
-				 <&rpmhcc RPMH_CXO_CLK>,
+				 <&tcsr TCSR_USB4_1_CLKREF_EN>,
 				 <&gcc GCC_USB3_SEC_PHY_COM_AUX_CLK>,
 				 <&gcc GCC_USB3_SEC_PHY_PIPE_CLK>;
 			clock-names = "aux",

@@ -3008,7 +3008,7 @@ usb_1_ss2_qmpphy: phy@fdf000 {
 			reg = <0 0x00fdf000 0 0x4000>;
 
 			clocks = <&gcc GCC_USB3_TERT_PHY_AUX_CLK>,
-				 <&rpmhcc RPMH_CXO_CLK>,
+				 <&tcsr TCSR_USB4_2_CLKREF_EN>,
 				 <&gcc GCC_USB3_TERT_PHY_COM_AUX_CLK>,
 				 <&gcc GCC_USB3_TERT_PHY_PIPE_CLK>;
 			clock-names = "aux",

2.51.0

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

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

From: Geert Uytterhoeven <[email protected]>

[ Upstream commit 44cfd102ce28e749a07bb0f1668cf932077b1175 ]

SW_LCD_EN is defined twice.

---
 arch/arm64/boot/dts/renesas/r9a09g047e57-smarc.dts | 1 -
 1 file changed, 1 deletion(-)

index 08e814c03fa85..ed6fcdc337a0b 100644
@@ -8,7 +8,6 @@
 /dts-v1/;
 
 /* Switch selection settings */
-#define SW_LCD_EN		0
 #define SW_GPIO8_CAN0_STB	0
 #define SW_GPIO9_CAN1_STB	0
 #define SW_LCD_EN		0

2.51.0

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

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

From: Vladimir Zapolskiy <[email protected]>

[ Upstream commit 7db5fbe508deedec6c183d5056cf3c504c027f40 ]

Trivial change, Analog Devices vendor prefix is "adi", but there is a valid "ad" vendor prefix of another company, this may explain why the issue hasn't been discovered by the automatic tests.

A problem of not described compatible value is out of this change scope.

---
 arch/arm64/boot/dts/qcom/msm8994-msft-lumia-octagon.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

index 4c983b10dd925..7ace3540ef0a0 100644
@@ -378,7 +378,7 @@ &blsp2_i2c1 {
 	status = "okay";
 
 	sideinteraction: touch@2c {
-		compatible = "ad,ad7147_captouch";
+		compatible = "adi,ad7147_captouch";
 		reg = <0x2c>;
 
 		pinctrl-names = "default", "sleep";

2.51.0

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

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

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

[ Upstream commit 0b2761eb1287bd9f62367cccf6626eb3107cef6f ]

The P66's device tree includes the reference design dtsi files, which defines a node and properties for the touchpanel in the common design. The P66 dts file then overrides all the properties to match its own design, but as the touchpanel model is different, a different schema is matched. This other schema uses a different name for the GPIO.

The original submission added the correct GPIO property, but did not delete the one inherited from the reference design, causing validation errors.

Explicitly delete the incorrect GPIO property.

---
 arch/arm/boot/dts/allwinner/sun5i-a13-utoo-p66.dts | 1 +
 1 file changed, 1 insertion(+)

index be486d28d04fa..428cab5a0e906 100644
@@ -102,6 +102,7 @@ &touchscreen {
 	/* The P66 uses a different EINT then the reference design */
 	interrupts = <6 9 IRQ_TYPE_EDGE_FALLING>; /* EINT9 (PG9) */
 	/* The icn8318 binding expects wake-gpios instead of power-gpios */
+	/delete-property/ power-gpios;
 	wake-gpios = <&pio 1 3 GPIO_ACTIVE_HIGH>; /* PB3 */
 	touchscreen-size-x = <800>;
 	touchscreen-size-y = <480>;

2.51.0

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

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

From: Christophe Leroy <[email protected]>

[ Upstream commit 5fbc09eb0b4f4b1a4b33abebacbeee0d29f195e9 ]

Commit 74e19ef0ff80 ("uaccess: Add speculation barrier to copy_from_user()") added a redundant barrier_nospec() in copy_from_user(), because powerpc is already calling barrier_nospec() in allow_read_from_user() and allow_read_write_user(). But on other architectures that call to barrier_nospec() was missing. So change powerpc instead of reverting the above commit and having to fix other architectures one by one. This is now possible because barrier_nospec() has also been added in copy_from_user_iter().

Move barrier_nospec() out of allow_read_from_user() and allow_read_write_user(). This will also allow reuse of those functions when implementing masked user access which doesn't require barrier_nospec().

Don't add it back in raw_copy_from_user() as it is already called by copy_from_user() and copy_from_user_iter().

---
 arch/powerpc/include/asm/kup.h     | 2 --
 arch/powerpc/include/asm/uaccess.h | 4 ++++
 2 files changed, 4 insertions(+), 2 deletions(-)

index dab63b82a8d4f..f2009d7c8cfa7 100644
@@ -134,7 +134,6 @@ static __always_inline void kuap_assert_locked(void)
 
 static __always_inline void allow_read_from_user(const void __user *from, unsigned long size)
 {
-	barrier_nospec();
 	allow_user_access(NULL, from, size, KUAP_READ);
 }
 

@@ -146,7 +145,6 @@ static __always_inline void allow_write_to_user(void __user *to, unsigned long s
 static __always_inline void allow_read_write_user(void __user *to, const void __user *from,
 						  unsigned long size)
 {
-	barrier_nospec();
 	allow_user_access(to, from, size, KUAP_READ_WRITE);
 }
 

index 784a00e681fa3..3e622e647d622 100644
@@ -301,6 +301,7 @@ do {								\
 	__typeof__(sizeof(*(ptr))) __gu_size = sizeof(*(ptr));	\
 								\
 	might_fault();					\
+	barrier_nospec();					\
 	allow_read_from_user(__gu_addr, __gu_size);		\
 	__get_user_size_allowed(__gu_val, __gu_addr, __gu_size, __gu_err);	\
 	prevent_read_from_user(__gu_addr, __gu_size);		\

@@ -329,6 +330,7 @@ raw_copy_in_user(void __user *to, const void __user *from, unsigned long n)
 {
 	unsigned long ret;
 
+	barrier_nospec();
 	allow_read_write_user(to, from, n);
 	ret = __copy_tofrom_user(to, from, n);
 	prevent_read_write_user(to, from, n);

@@ -415,6 +417,7 @@ static __must_check __always_inline bool user_access_begin(const void __user *pt
 
 	might_fault();
 
+	barrier_nospec();
 	allow_read_write_user((void __user *)ptr, ptr, len);
 	return true;
 }

@@ -431,6 +434,7 @@ user_read_access_begin(const void __user *ptr, size_t len)
 
 	might_fault();
 
+	barrier_nospec();
 	allow_read_from_user(ptr, len);
 	return true;
 }

2.51.0

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

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

From: Haotian Zhang <[email protected]>

[ Upstream commit 0da7824734d8d83e6a844dd0207f071cb0c50cf4 ]

If cmd_db_magic_matches() fails after memremap() succeeds, the function returns -EINVAL without unmapping the memory region, causing a potential resource leak.

Switch to devm_memremap to automatically manage the map resource.

---
 drivers/soc/qcom/cmd-db.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

index ae66c2623d250..84a75d8c4b702 100644
@@ -349,15 +349,16 @@ static int cmd_db_dev_probe(struct platform_device *pdev)
 		return -EINVAL;
 	}
 
-	cmd_db_header = memremap(rmem->base, rmem->size, MEMREMAP_WC);
-	if (!cmd_db_header) {
-		ret = -ENOMEM;
+	cmd_db_header = devm_memremap(&pdev->dev, rmem->base, rmem->size, MEMREMAP_WC);
+	if (IS_ERR(cmd_db_header)) {
+		ret = PTR_ERR(cmd_db_header);
 		cmd_db_header = NULL;
 		return ret;
 	}
 
 	if (!cmd_db_magic_matches(cmd_db_header)) {
 		dev_err(&pdev->dev, "Invalid Command DB Magic\n");
+		cmd_db_header = NULL;
 		return -EINVAL;
 	}
 

2.51.0

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

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

From: Zilin Guan <[email protected]>

[ Upstream commit 6259094ee806fb813ca95894c65fb80e2ec98bf1 ]

In svs_enable_debug_write(), the buf allocated by memdup_user_nul() is leaked if kstrtoint() fails.

Fix this by using __free(kfree) to automatically free buf, eliminating the need for explicit kfree() calls and preventing leaks.

---
 drivers/soc/mediatek/mtk-svs.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

index f45537546553e..99edecb204f25 100644
@@ -9,6 +9,7 @@
 #include <linux/bits.h>
 #include <linux/clk.h>
 #include <linux/completion.h>
+#include <linux/cleanup.h>
 #include <linux/cpu.h>
 #include <linux/cpuidle.h>
 #include <linux/debugfs.h>

@@ -789,7 +790,7 @@ static ssize_t svs_enable_debug_write(struct file *filp,
 	struct svs_bank *svsb = file_inode(filp)->i_private;
 	struct svs_platform *svsp = dev_get_drvdata(svsb->dev);
 	int enabled, ret;
-	char *buf = NULL;
+	char *buf __free(kfree) = NULL;
 
 	if (count >= PAGE_SIZE)
 		return -EINVAL;

@@ -807,8 +808,6 @@ static ssize_t svs_enable_debug_write(struct file *filp,
 		svsb->mode_support = SVSB_MODE_ALL_DISABLE;
 	}
 
-	kfree(buf);
-
 	return count;
 }
 

2.51.0

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

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

From: Narayana Murty N <[email protected]>

[ Upstream commit 815a8d2feb5615ae7f0b5befd206af0b0160614c ]

The recent commit 1010b4c012b0 ("powerpc/eeh: Make EEH driver device hotplug safe") restructured the EEH driver to improve synchronization with the PCI hotplug layer.

However, it inadvertently moved pci_lock_rescan_remove() outside its intended scope in eeh_handle_normal_event(), leading to broken PCI error reporting and improper EEH event triggering. Specifically, eeh_handle_normal_event() acquired pci_lock_rescan_remove() before calling eeh_pe_bus_get(), but eeh_pe_bus_get() itself attempts to acquire the same lock internally, causing nested locking and disrupting normal EEH event handling paths.

This patch adds a boolean parameter do_lock to _eeh_pe_bus_get(), with two public wrappers:
eeh_pe_bus_get() with locking enabled. eeh_pe_bus_get_nolock() that skips locking.

Callers that already hold pci_lock_rescan_remove() now use eeh_pe_bus_get_nolock() to avoid recursive lock acquisition.

Additionally, pci_lock_rescan_remove() calls are restored to the correct position—after eeh_pe_bus_get() and immediately before iterating affected PEs and devices. This ensures EEH-triggered PCI removes occur under proper bus rescan locking without recursive lock contention.

The eeh_pe_loc_get() function has been split into two functions:
eeh_pe_loc_get(struct eeh_pe *pe) which retrieves the loc for given PE. eeh_pe_loc_get_bus(struct pci_bus *bus) which retrieves the location code for given bus.

This resolves lockdep warnings such as:
<snip>

[   84.964298] [    T928] ============================================
[   84.964304] [    T928] WARNING: possible recursive locking detected
[   84.964311] [    T928] 6.18.0-rc3 #51 Not tainted
[   84.964315] [    T928] --------------------------------------------
[   84.964320] [    T928] eehd/928 is trying to acquire lock:
[   84.964324] [    T928] c000000003b29d58 (pci_rescan_remove_lock){+.+.}-{3:3}, at: pci_lock_rescan_remove+0x28/0x40
[   84.964342] [    T928]
                       but task is already holding lock:
[   84.964347] [    T928] c000000003b29d58 (pci_rescan_remove_lock){+.+.}-{3:3}, at: pci_lock_rescan_remove+0x28/0x40
[   84.964357] [    T928]
                       other info that might help us debug this:
[   84.964363] [    T928]  Possible unsafe locking scenario:

[   84.964367] [    T928]        CPU0
[   84.964370] [    T928]        ----
[   84.964373] [    T928]   lock(pci_rescan_remove_lock);
[   84.964378] [    T928]   lock(pci_rescan_remove_lock);
[   84.964383] [    T928]
                       *** DEADLOCK ***

[   84.964388] [    T928]  May be due to missing lock nesting notation

[   84.964393] [    T928] 1 lock held by eehd/928:
[   84.964397] [    T928]  #0: c000000003b29d58 (pci_rescan_remove_lock){+.+.}-{3:3}, at: pci_lock_rescan_remove+0x28/0x40
[   84.964408] [    T928]
                       stack backtrace:
[   84.964414] [    T928] CPU: 2 UID: 0 PID: 928 Comm: eehd Not tainted 6.18.0-rc3 #51 VOLUNTARY
[   84.964417] [    T928] Hardware name: IBM,9080-HEX POWER10 (architected) 0x800200 0xf000006 of:IBM,FW1060.00 (NH1060_022) hv:phyp pSeries
[   84.964419] [    T928] Call Trace:
[   84.964420] [    T928] [c0000011a7157990] [c000000001705de4] dump_stack_lvl+0xc8/0x130 (unreliable)
[   84.964424] [    T928] [c0000011a71579d0] [c0000000002f66e0] print_deadlock_bug+0x430/0x440
[   84.964428] [    T928] [c0000011a7157a70] [c0000000002fd0c0] __lock_acquire+0x1530/0x2d80
[   84.964431] [    T928] [c0000011a7157ba0] [c0000000002fea54] lock_acquire+0x144/0x410
[   84.964433] [    T928] [c0000011a7157cb0] [c0000011a7157cb0] __mutex_lock+0xf4/0x1050
[   84.964436] [    T928] [c0000011a7157e00] [c000000000de21d8] pci_lock_rescan_remove+0x28/0x40
[   84.964439] [    T928] [c0000011a7157e20] [c00000000004ed98] eeh_pe_bus_get+0x48/0xc0
[   84.964442] [    T928] [c0000011a7157e50] [c000000000050434] eeh_handle_normal_event+0x64/0xa60
[   84.964446] [    T928] [c0000011a7157f30] [c000000000051de8] eeh_event_handler+0xf8/0x190
[   84.964450] [    T928] [c0000011a7157f90] [c0000000002747ac] kthread+0x16c/0x180
[   84.964453] [    T928] [c0000011a7157fe0] [c00000000000ded8] start_kernel_thread+0x14/0x18

</snip>

---
 arch/powerpc/include/asm/eeh.h   |  2 +
 arch/powerpc/kernel/eeh_driver.c | 11 ++---
 arch/powerpc/kernel/eeh_pe.c     | 74 ++++++++++++++++++++++++++++++--
 3 files changed, 78 insertions(+), 9 deletions(-)

index 5e34611de9ef4..b7ebb4ac2c710 100644
@@ -289,6 +289,8 @@ void eeh_pe_dev_traverse(struct eeh_pe *root,
 void eeh_pe_restore_bars(struct eeh_pe *pe);
 const char *eeh_pe_loc_get(struct eeh_pe *pe);
 struct pci_bus *eeh_pe_bus_get(struct eeh_pe *pe);
+const char *eeh_pe_loc_get_bus(struct pci_bus *bus);
+struct pci_bus *eeh_pe_bus_get_nolock(struct eeh_pe *pe);
 
 void eeh_show_enabled(void);
 int __init eeh_init(struct eeh_ops *ops);