Skip to contentSkip to conversation

6.1.16-rc1 review

PATCH In review Greg Kroah-Hartman 6 participants 892 messages Mar 7, 2023 – Mar 8, 2023 Page 5 of 7

Subsystems (18)
lore
Share & tools
lore
author maintainer
#572 lore raw

PATCH 571/885 netfilter: nf_tables: NULL pointer dereference in nf_tables_updobj()

In reply to Greg Kroah-Hartman

From: Alok Tiwari <[email protected]>

[ Upstream commit dac7f50a45216d652887fb92d6cd3b7ca7f006ea ]

static analyzer detect null pointer dereference case for 'type' function __nft_obj_type_get() can return NULL value which require to handle if type is NULL pointer return -ENOENT.

This is a theoretical issue, since an existing object has a type, but better add this failsafe check.

---
 net/netfilter/nf_tables_api.c | 3 +++
 1 file changed, 3 insertions(+)
net/netfilter/nf_tables_api.c +3 -0
index 3ba8c291fcaa7..dca5352bdf3d7 100644
@@ -6951,6 +6951,9 @@ static int nf_tables_newobj(struct sk_buff *skb, const struct nfnl_info *info,
 			return -EOPNOTSUPP;
 
 		type = __nft_obj_type_get(objtype);
+		if (WARN_ON_ONCE(!type))
+			return -ENOENT;
+
 		nft_ctx_init(&ctx, net, skb, info->nlh, family, table, NULL, nla);
 
 		return nf_tables_updobj(&ctx, type, nla[NFTA_OBJ_DATA], obj);
Signature · 1 line
2.39.2
author maintainer
#573 lore raw

PATCH 572/885 can: isotp: check CAN address family in isotp_bind()

In reply to Greg Kroah-Hartman

From: Oliver Hartkopp <[email protected]>

[ Upstream commit c6adf659a8ba85913e16a571d5a9bcd17d3d1234 ]

Add missing check to block non-AF_CAN binds.

Syzbot created some code which matched the right sockaddr struct size but used AF_XDP (0x2C) instead of AF_CAN (0x1D) in the address family field:

bind$xdp(r2, &(0x7f0000000540)={0x2c, 0x0, r4, 0x0, r2}, 0x10)

                                ^^^^

This has no funtional impact but the userspace should be notified about the wrong address family field content.

---
 net/can/isotp.c | 3 +++
 1 file changed, 3 insertions(+)
net/can/isotp.c +3 -0
index fc81d77724a13..9bc344851704e 100644
@@ -1220,6 +1220,9 @@ static int isotp_bind(struct socket *sock, struct sockaddr *uaddr, int len)
 	if (len < ISOTP_MIN_NAMELEN)
 		return -EINVAL;
 
+	if (addr->can_family != AF_CAN)
+		return -EINVAL;
+
 	/* sanitize tx CAN identifier */
 	if (tx_id & CAN_EFF_FLAG)
 		tx_id &= (CAN_EFF_FLAG | CAN_EFF_MASK);
Signature · 1 line
2.39.2
author maintainer
#574 lore raw

PATCH 573/885 gcc-plugins: drop -std=gnu++11 to fix GCC 13 build

In reply to Greg Kroah-Hartman

From: Sam James <[email protected]>

[ Upstream commit 5a6b64adc18d9adfb497a529ff004d59b6df151f ]

The latest GCC 13 snapshot (13.0.1 20230129) gives the following:

```
cc1: error: cannot load plugin ./scripts/gcc-plugins/randomize_layout_plugin.so
 :./scripts/gcc-plugins/randomize_layout_plugin.so: undefined symbol: tree_code_type
```

This ends up being because of https://gcc.gnu.org/git/gitweb.cgi?p=gcc.git;h=b0241ce6e37031
upstream in GCC which changes the visibility of some types used by the kernel's plugin infrastructure like tree_code_type.

After discussion with the GCC folks, we found that the kernel needs to be building plugins with the same flags used to build GCC - and GCC defaults to gnu++17 right now. The minimum GCC version needed to build the kernel is GCC 5.1 and GCC 5.1 already defaults to gnu++14 anyway, so just drop the flag, as all GCCs that could be used to build GCC already default to an acceptable version which was >= the version we forced via flags until now.

---
 scripts/gcc-plugins/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
scripts/gcc-plugins/Makefile +1 -1
index b34d11e226366..320afd3cf8e82 100644
@@ -29,7 +29,7 @@ GCC_PLUGINS_DIR = $(shell $(CC) -print-file-name=plugin)
 plugin_cxxflags	= -Wp,-MMD,$(depfile) $(KBUILD_HOSTCXXFLAGS) -fPIC \
 		  -include $(srctree)/include/linux/compiler-version.h \
 		  -DPLUGIN_VERSION=$(call stringify,$(KERNELVERSION)) \
-		  -I $(GCC_PLUGINS_DIR)/include -I $(obj) -std=gnu++11 \
+		  -I $(GCC_PLUGINS_DIR)/include -I $(obj) \
 		  -fno-rtti -fno-exceptions -fasynchronous-unwind-tables \
 		  -ggdb -Wno-narrowing -Wno-unused-variable \
 		  -Wno-format-diag
Signature · 1 line
2.39.2
author maintainer
#575 lore raw

PATCH 574/885 tools/power/x86/intel-speed-select: Add Emerald Rapid quirk

In reply to Greg Kroah-Hartman

From: Zhang Rui <[email protected]>

[ Upstream commit 61f9fdcdcd01f9a996b6db4e7092fcdfe8414ad5 ]

Need memory frequency quirk as Sapphire Rapids in Emerald Rapids. So add Emerald Rapids CPU model check in is_spr_platform().

---
 tools/power/x86/intel-speed-select/isst-config.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
tools/power/x86/intel-speed-select/isst-config.c +1 -1
index a160bad291eb7..be3668d37d654 100644
@@ -110,7 +110,7 @@ int is_skx_based_platform(void)
 
 int is_spr_platform(void)
 {
-	if (cpu_model == 0x8F)
+	if (cpu_model == 0x8F || cpu_model == 0xCF)
 		return 1;
 
 	return 0;
Signature · 1 line
2.39.2
author maintainer
#576 lore raw

PATCH 575/885 wifi: mt76: dma: free rx_head in mt76_dma_rx_cleanup

In reply to Greg Kroah-Hartman

From: Lorenzo Bianconi <[email protected]>

[ Upstream commit 1b88b47e898edef0e56e3a2f4e49f052a136153d ]

Free rx_head skb in mt76_dma_rx_cleanup routine in order to avoid possible memory leak at module unload.

---
 drivers/net/wireless/mediatek/mt76/dma.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
drivers/net/wireless/mediatek/mt76/dma.c +7 -6
index 7378c4d1e1567..478bffb7418d9 100644
@@ -573,6 +573,7 @@ mt76_dma_rx_cleanup(struct mt76_dev *dev, struct mt76_queue *q)
 		return;
 
 	spin_lock_bh(&q->lock);
+
 	do {
 		buf = mt76_dma_dequeue(dev, q, true, NULL, NULL, &more);
 		if (!buf)
@@ -580,6 +581,12 @@ mt76_dma_rx_cleanup(struct mt76_dev *dev, struct mt76_queue *q)
 
 		skb_free_frag(buf);
 	} while (1);
+
+	if (q->rx_head) {
+		dev_kfree_skb(q->rx_head);
+		q->rx_head = NULL;
+	}
+
 	spin_unlock_bh(&q->lock);
 
 	if (!q->rx_page.va)
@@ -605,12 +612,6 @@ mt76_dma_rx_reset(struct mt76_dev *dev, enum mt76_rxq_id qid)
 	mt76_dma_rx_cleanup(dev, q);
 	mt76_dma_sync_idx(dev, q);
 	mt76_dma_rx_fill(dev, q);
-
-	if (!q->rx_head)
-		return;
-
-	dev_kfree_skb(q->rx_head);
-	q->rx_head = NULL;
 }
 
 static void
Signature · 1 line
2.39.2
author maintainer
#577 lore raw

PATCH 576/885 ACPI: video: Fix Lenovo Ideapad Z570 DMI match

In reply to Greg Kroah-Hartman

From: Hans de Goede <[email protected]>

[ Upstream commit 2d11eae42d52a131f06061015e49dc0f085c5bfc ]

Multiple Ideapad Z570 variants need acpi_backlight=native to force native use on these pre Windows 8 machines since acpi_video backlight control does not work here.

The original DMI quirk matches on a product_name of "102434U" but other variants may have different product_name-s such as e.g. "1024D9U".

Move to checking product_version instead as is more or less standard for Lenovo DMI quirks for similar reasons.

---
 drivers/acpi/video_detect.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/acpi/video_detect.c +1 -1
index b48f85c3791e9..7f0ed845cd6ad 100644
@@ -432,7 +432,7 @@ static const struct dmi_system_id video_detect_dmi_table[] = {
 	 /* Lenovo Ideapad Z570 */
 	 .matches = {
 		DMI_MATCH(DMI_SYS_VENDOR, "LENOVO"),
-		DMI_MATCH(DMI_PRODUCT_NAME, "102434U"),
+		DMI_MATCH(DMI_PRODUCT_VERSION, "Ideapad Z570"),
 		},
 	},
 	{
Signature · 1 line
2.39.2
author maintainer
#578 lore raw

PATCH 577/885 net/mlx5: fw_tracer: Fix debug print

In reply to Greg Kroah-Hartman

From: Shay Drory <[email protected]>

[ Upstream commit 988c2352273997a242f15c4fc3711773515006a2 ]

The debug message specify tdsn, but takes as an argument the tmsn. The correct argument is tmsn, hence, fix the print.

---
 drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/net/ethernet/mellanox/mlx5/core/diag/fw_tracer.c +1 -1
index 5b05b884b5fb3..d7b2ee5de1158 100644
@@ -603,7 +603,7 @@ static int mlx5_tracer_handle_string_trace(struct mlx5_fw_tracer *tracer,
 	} else {
 		cur_string = mlx5_tracer_message_get(tracer, tracer_event);
 		if (!cur_string) {
-			pr_debug("%s Got string event for unknown string tdsm: %d\n",
+			pr_debug("%s Got string event for unknown string tmsn: %d\n",
 				 __func__, tracer_event->string_event.tmsn);
 			return -1;
 		}
Signature · 1 line
2.39.2
author maintainer
#579 lore raw

PATCH 578/885 coda: Avoid partial allocation of sig_inputArgs

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

[ Upstream commit 48df133578c70185a95a49390d42df1996ddba2a ]

GCC does not like having a partially allocated object, since it cannot reason about it for bounds checking when it is passed to other code. Instead, fully allocate sig_inputArgs. (Alternatively, sig_inputArgs should be defined as a struct coda_in_hdr, if it is actually not using any other part of the union.) Seen under GCC 13:

../fs/coda/upcall.c: In function 'coda_upcall':
../fs/coda/upcall.c:801:22: warning: array subscript 'union inputArgs[0]' is partly outside array bounds of 'unsigned char[20]' [-Warray-bounds=]
801 | sig_inputArgs->ih.opcode = CODA_SIGNAL;
| ^~

---
 fs/coda/upcall.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/coda/upcall.c +1 -1
index 59f6cfd06f96a..cd6a3721f6f69 100644
@@ -791,7 +791,7 @@ static int coda_upcall(struct venus_comm *vcp,
 	sig_req = kmalloc(sizeof(struct upc_req), GFP_KERNEL);
 	if (!sig_req) goto exit;
 
-	sig_inputArgs = kvzalloc(sizeof(struct coda_in_hdr), GFP_KERNEL);
+	sig_inputArgs = kvzalloc(sizeof(*sig_inputArgs), GFP_KERNEL);
 	if (!sig_inputArgs) {
 		kfree(sig_req);
 		goto exit;
Signature · 1 line
2.39.2
author maintainer
#580 lore raw

PATCH 579/885 uaccess: Add minimum bounds check on kernel buffer size

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

[ Upstream commit 04ffde1319a715bd0550ded3580d4ea3bc003776 ]

While there is logic about the difference between ksize and usize, copy_struct_from_user() didn't check the size of the destination buffer (when it was known) against ksize. Add this check so there is an upper bounds check on the possible memset() call, otherwise lower bounds checks made by callers will trigger bounds warnings under -Warray-bounds. Seen under GCC 13:

In function 'copy_struct_from_user',
inlined from 'iommufd_fops_ioctl' at
../drivers/iommu/iommufd/main.c:333:8:
../include/linux/fortify-string.h:59:33: warning: '__builtin_memset' offset [57, 4294967294] is out of the bounds [0, 56] of object 'buf' with type 'union ucmd_buffer' [-Warray-bounds=]

   59 | #define __underlying_memset     __builtin_memset
      |                                 ^
../include/linux/fortify-string.h:453:9: note: in expansion of macro '__underlying_memset'
  453 |         __underlying_memset(p, c, __fortify_size); \
      |         ^~~~~~~~~~~~~~~~~~~
../include/linux/fortify-string.h:461:25: note: in expansion of macro '__fortify_memset_chk'
  461 | #define memset(p, c, s) __fortify_memset_chk(p, c, s, \
      |                         ^~~~~~~~~~~~~~~~~~~~
../include/linux/uaccess.h:334:17: note: in expansion of macro 'memset'
  334 |                 memset(dst + size, 0, rest);
      |                 ^~~~~~

../drivers/iommu/iommufd/main.c: In function 'iommufd_fops_ioctl':
../drivers/iommu/iommufd/main.c:311:27: note: 'buf' declared here
311 | union ucmd_buffer buf;
| ^~~

Cc Christian Brauner <[email protected]>
Cc Rasmus Villemoes <[email protected]>
Cc Arnd Bergmann <[email protected]>
Cc Dinh Nguyen <[email protected]>
Cc Catalin Marinas <[email protected]>
Cc Andrew Morton <[email protected]>
Cc Geert Uytterhoeven <[email protected]>
Cc Alexander Potapenko <[email protected]>
---
 include/linux/uaccess.h | 4 ++++
 1 file changed, 4 insertions(+)
include/linux/uaccess.h +4 -0
index afb18f198843b..ab9728138ad67 100644
@@ -329,6 +329,10 @@ copy_struct_from_user(void *dst, size_t ksize, const void __user *src,
 	size_t size = min(ksize, usize);
 	size_t rest = max(ksize, usize) - size;
 
+	/* Double check if ksize is larger than a known object size. */
+	if (WARN_ON_ONCE(ksize > __builtin_object_size(dst, 1)))
+		return -E2BIG;
+
 	/* Deal with trailing bytes. */
 	if (usize < ksize) {
 		memset(dst + size, 0, rest);
Signature · 1 line
2.39.2
author maintainer
#581 lore raw

PATCH 580/885 s390/idle: mark arch_cpu_idle() noinstr

In reply to Greg Kroah-Hartman

From: Heiko Carstens <[email protected]>

[ Upstream commit a9cbc1b471d291c865907542394f1c483b93a811 ]

linux-next commit ("cpuidle: tracing: Warn about !rcu_is_watching()") adds a new warning which hits on s390's arch_cpu_idle() function:

RCU not on for: arch_cpu_idle+0x0/0x28

WARNING: CPU: 2 PID: 0 at include/linux/trace_recursion.h:162 arch_ftrace_ops_list_func+0x24c/0x258
Modules linked in:
CPU: 2 PID: 0 Comm: swapper/2 Not tainted 6.2.0-rc6-next-20230202 #4
Hardware name: IBM 8561 T01 703 (z/VM 7.3.0)
Krnl PSW : 0404d00180000000 00000000002b55c0 (arch_ftrace_ops_list_func+0x250/0x258)
           R:0 T:1 IO:0 EX:0 Key:0 M:1 W:0 P:0 AS:3 CC:1 PM:0 RI:0 EA:3
Krnl GPRS: c0000000ffffbfff 0000000080000002 0000000000000026 0000000000000000
           0000037ffffe3a28 0000037ffffe3a20 0000000000000000 0000000000000000
           0000000000000000 0000000000f4acf6 00000000001044f0 0000037ffffe3cb0
           0000000000000000 0000000000000000 00000000002b55bc 0000037ffffe3bb8
Krnl Code: 00000000002b55b0: c02000840051        larl    %r2,0000000001335652
           00000000002b55b6: c0e5fff512d1        brasl   %r14,0000000000157b58
          #00000000002b55bc: af000000            mc      0,0
          >00000000002b55c0: a7f4ffe7            brc     15,00000000002b558e
           00000000002b55c4: 0707                bcr     0,%r7
           00000000002b55c6: 0707                bcr     0,%r7
           00000000002b55c8: eb6ff0480024        stmg    %r6,%r15,72(%r15)
           00000000002b55ce: b90400ef            lgr     %r14,%r15
Call Trace:
 [<00000000002b55c0>] arch_ftrace_ops_list_func+0x250/0x258
([<00000000002b55bc>] arch_ftrace_ops_list_func+0x24c/0x258)
 [<0000000000f5f0fc>] ftrace_common+0x1c/0x20
 [<00000000001044f6>] arch_cpu_idle+0x6/0x28
 [<0000000000f4acf6>] default_idle_call+0x76/0x128
 [<00000000001cc374>] do_idle+0xf4/0x1b0
 [<00000000001cc6ce>] cpu_startup_entry+0x36/0x40
 [<0000000000119d00>] smp_start_secondary+0x140/0x150
 [<0000000000f5d2ae>] restart_int_handler+0x6e/0x90

Mark arch_cpu_idle() noinstr like all other architectures with CONFIG_ARCH_WANTS_NO_INSTR (should) have it to fix this.

---
 arch/s390/kernel/idle.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/s390/kernel/idle.c +1 -1
index 4bf1ee293f2b3..a0da049e73609 100644
@@ -44,7 +44,7 @@ void account_idle_time_irq(void)
 	S390_lowcore.last_update_timer = idle->timer_idle_exit;
 }
 
-void arch_cpu_idle(void)
+void noinstr arch_cpu_idle(void)
 {
 	struct s390_idle_data *idle = this_cpu_ptr(&s390_idle);
 	unsigned long idle_time;
Signature · 1 line
2.39.2
author maintainer
#582 lore raw

PATCH 581/885 time/debug: Fix memory leak with using debugfs_lookup()

In reply to Greg Kroah-Hartman

From: Greg Kroah-Hartman <[email protected]>

[ Upstream commit 5b268d8abaec6cbd4bd70d062e769098d96670aa ]

When calling debugfs_lookup() the result must have dput() called on it, otherwise the memory will leak over time. To make things simpler, just call debugfs_lookup_and_remove() instead which handles all of the logic at once.

---
 kernel/time/test_udelay.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/time/test_udelay.c +1 -1
index 13b11eb62685e..20d5df631570e 100644
@@ -149,7 +149,7 @@ module_init(udelay_test_init);
 static void __exit udelay_test_exit(void)
 {
 	mutex_lock(&udelay_test_lock);
-	debugfs_remove(debugfs_lookup(DEBUGFS_FILENAME, NULL));
+	debugfs_lookup_and_remove(DEBUGFS_FILENAME, NULL);
 	mutex_unlock(&udelay_test_lock);
 }
 
Signature · 1 line
2.39.2
author maintainer
#583 lore raw

PATCH 582/885 PM: domains: fix memory leak with using debugfs_lookup()

In reply to Greg Kroah-Hartman

From: Greg Kroah-Hartman <[email protected]>

[ Upstream commit 0b6200e1e9f53dabdc30d0f6c51af9a5f664d32b ]

When calling debugfs_lookup() the result must have dput() called on it, otherwise the memory will leak over time. To make things simpler, just call debugfs_lookup_and_remove() instead which handles all of the logic at once.

---
 drivers/base/power/domain.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
drivers/base/power/domain.c +1 -4
index 6471b559230e9..b411201f75bfb 100644
@@ -220,13 +220,10 @@ static void genpd_debug_add(struct generic_pm_domain *genpd);
 
 static void genpd_debug_remove(struct generic_pm_domain *genpd)
 {
-	struct dentry *d;
-
 	if (!genpd_debugfs_dir)
 		return;
 
-	d = debugfs_lookup(genpd->name, genpd_debugfs_dir);
-	debugfs_remove(d);
+	debugfs_lookup_and_remove(genpd->name, genpd_debugfs_dir);
 }
 
 static void genpd_update_accounting(struct generic_pm_domain *genpd)
Signature · 1 line
2.39.2
author maintainer
#584 lore raw

PATCH 583/885 PM: EM: fix memory leak with using debugfs_lookup()

In reply to Greg Kroah-Hartman

From: Greg Kroah-Hartman <[email protected]>

[ Upstream commit a0e8c13ccd6a9a636d27353da62c2410c4eca337 ]

When calling debugfs_lookup() the result must have dput() called on it, otherwise the memory will leak over time. To make things simpler, just call debugfs_lookup_and_remove() instead which handles all of the logic at once.

---
 kernel/power/energy_model.c | 5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
kernel/power/energy_model.c +1 -4
index f82111837b8d1..7b44f5b89fa15 100644
@@ -87,10 +87,7 @@ static void em_debug_create_pd(struct device *dev)
 
 static void em_debug_remove_pd(struct device *dev)
 {
-	struct dentry *debug_dir;
-
-	debug_dir = debugfs_lookup(dev_name(dev), rootdir);
-	debugfs_remove_recursive(debug_dir);
+	debugfs_lookup_and_remove(dev_name(dev), rootdir);
 }
 
 static int __init em_debug_init(void)
Signature · 1 line
2.39.2
author maintainer
#585 lore raw

PATCH 584/885 Bluetooth: Fix issue with Actions Semi ATS2851 based devices

In reply to Greg Kroah-Hartman

From: Marcel Holtmann <[email protected]>

[ Upstream commit 5d44f76fab0839799b19cbc88d13575da968dc08 ]

Their devices claim to support the erroneous data reporting, but don't actually support the required commands. So blacklist them and add a quirk.

< HCI Command: Read Default Erroneous Data Reporting (0x03|0x005a) plen 0

Quoted 1 line

HCI Event: Command Status (0x0f) plen 4

        Read Default Erroneous Data Reporting (0x03|0x005a) ncmd 1
          Status: Unknown HCI Command (0x01)
T:  Bus=02 Lev=02 Prnt=08 Port=02 Cnt=01 Dev#= 10 Spd=12   MxCh= 0
D:  Ver= 2.00 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=10d7 ProdID=b012 Rev=88.91
S:  Manufacturer=Actions
S:  Product=general adapter
S:  SerialNumber=ACTIONS1234
C:* #Ifs= 2 Cfg#= 1 Atr=c0 MxPwr=100mA
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  64 Ivl=1ms
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=01(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=01(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=01(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=01(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=01(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=01(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
---
 drivers/bluetooth/btusb.c | 9 +++++++++
 1 file changed, 9 insertions(+)
drivers/bluetooth/btusb.c +9 -0
index 93e9ae928e4e8..4394a0e40c733 100644
@@ -63,6 +63,7 @@ static struct usb_driver btusb_driver;
 #define BTUSB_INTEL_BROKEN_SHUTDOWN_LED	BIT(24)
 #define BTUSB_INTEL_BROKEN_INITIAL_NCMD BIT(25)
 #define BTUSB_INTEL_NO_WBS_SUPPORT	BIT(26)
+#define BTUSB_ACTIONS_SEMI		BIT(27)
 
 static const struct usb_device_id btusb_table[] = {
 	/* Generic Bluetooth USB device */
@@ -663,6 +664,9 @@ static const struct usb_device_id blacklist_table[] = {
 	{ USB_DEVICE(0x0cb5, 0xc547), .driver_info = BTUSB_REALTEK |
 						     BTUSB_WIDEBAND_SPEECH },
 
+	/* Actions Semiconductor ATS2851 based devices */
+	{ USB_DEVICE(0x10d7, 0xb012), .driver_info = BTUSB_ACTIONS_SEMI },
+
 	/* Silicon Wave based devices */
 	{ USB_DEVICE(0x0c10, 0x0000), .driver_info = BTUSB_SWAVE },
 
@@ -4012,6 +4016,11 @@ static int btusb_probe(struct usb_interface *intf,
 		set_bit(BTUSB_USE_ALT3_FOR_WBS, &data->flags);
 	}
 
+	if (id->driver_info & BTUSB_ACTIONS_SEMI) {
+		/* Support is advertised, but not implemented */
+		set_bit(HCI_QUIRK_BROKEN_ERR_DATA_REPORTING, &hdev->quirks);
+	}
+
 	if (!reset)
 		set_bit(HCI_QUIRK_RESET_ON_CLOSE, &hdev->quirks);
 
Signature · 1 line
2.39.2
author maintainer
#586 lore raw

PATCH 585/885 Bluetooth: btusb: Add new PID/VID 0489:e0f2 for MT7921

In reply to Greg Kroah-Hartman

From: Mario Limonciello <[email protected]>

[ Upstream commit 83458a5f272b303479e7d2f451600817a7350b6b ]

This bluetooth device is found in a combo WLAN/BT card for a MediaTek 7921e.

The device information:

T:  Bus=01 Lev=01 Prnt=01 Port=02 Cnt=01 Dev#=  2 Spd=480  MxCh= 0
D:  Ver= 2.10 Cls=ef(misc ) Sub=02 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=0489 ProdID=e0f2 Rev= 1.00
S:  Manufacturer=MediaTek Inc.
S:  Product=Wireless_Device
S:  SerialNumber=000000000
C:* #Ifs= 3 Cfg#= 1 Atr=e0 MxPwr=100mA
A:  FirstIf#= 0 IfCount= 3 Cls=e0(wlcon) Sub=01 Prot=01
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  16 Ivl=125us
E:  Ad=82(I) Atr=02(Bulk) MxPS= 512 Ivl=0ms
E:  Ad=02(O) Atr=02(Bulk) MxPS= 512 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
I:  If#= 1 Alt= 6 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=83(I) Atr=01(Isoc) MxPS=  63 Ivl=1ms
E:  Ad=03(O) Atr=01(Isoc) MxPS=  63 Ivl=1ms
I:* If#= 2 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=(none)
E:  Ad=8a(I) Atr=03(Int.) MxPS=  64 Ivl=125us
E:  Ad=0a(O) Atr=03(Int.) MxPS=  64 Ivl=125us
I:  If#= 2 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=(none)
E:  Ad=8a(I) Atr=03(Int.) MxPS= 512 Ivl=125us
E:  Ad=0a(O) Atr=03(Int.) MxPS= 512 Ivl=125us
---
 drivers/bluetooth/btusb.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/bluetooth/btusb.c +3 -0
index 4394a0e40c733..b8cd42d81d5ca 100644
@@ -558,6 +558,9 @@ static const struct usb_device_id blacklist_table[] = {
 	{ USB_DEVICE(0x0489, 0xe0e0), .driver_info = BTUSB_MEDIATEK |
 						     BTUSB_WIDEBAND_SPEECH |
 						     BTUSB_VALID_LE_STATES },
+	{ USB_DEVICE(0x0489, 0xe0f2), .driver_info = BTUSB_MEDIATEK |
+						     BTUSB_WIDEBAND_SPEECH |
+						     BTUSB_VALID_LE_STATES },
 	{ USB_DEVICE(0x04ca, 0x3802), .driver_info = BTUSB_MEDIATEK |
 						     BTUSB_WIDEBAND_SPEECH |
 						     BTUSB_VALID_LE_STATES },
Signature · 1 line
2.39.2
author maintainer
#587 lore raw

PATCH 586/885 Bluetooth: btusb: Add VID:PID 13d3:3529 for Realtek RTL8821CE

In reply to Greg Kroah-Hartman

From: Moises Cardona <[email protected]>

[ Upstream commit 1eec3b95b5ce7fb2cdd273ac4f8b24b1ed6776a1 ]

This patch adds VID:PID 13d3:3529 to the btusb.c file.

This VID:PID is found in the Realtek RTL8821CE module (M.2 module AW-CB304NF on an ASUS E210MA laptop)

Output of /sys/kernel/debug/usb/devices:

T:  Bus=01 Lev=01 Prnt=01 Port=07 Cnt=02 Dev#=  3 Spd=12   MxCh= 0
D:  Ver= 1.10 Cls=e0(wlcon) Sub=01 Prot=01 MxPS=64 #Cfgs=  1
P:  Vendor=13d3 ProdID=3529 Rev= 1.10
S:  Manufacturer=Realtek
S:  Product=Bluetooth Radio
C:* #Ifs= 2 Cfg#= 1 Atr=e0 MxPwr=500mA
I:* If#= 0 Alt= 0 #EPs= 3 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=81(I) Atr=03(Int.) MxPS=  16 Ivl=1ms
E:  Ad=02(O) Atr=02(Bulk) MxPS=  64 Ivl=0ms
E:  Ad=82(I) Atr=02(Bulk) MxPS=  64 Ivl=0ms
I:* If#= 1 Alt= 0 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   0 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   0 Ivl=1ms
I:  If#= 1 Alt= 1 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=   9 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=   9 Ivl=1ms
I:  If#= 1 Alt= 2 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  17 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  17 Ivl=1ms
I:  If#= 1 Alt= 3 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  25 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  25 Ivl=1ms
I:  If#= 1 Alt= 4 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  33 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  33 Ivl=1ms
I:  If#= 1 Alt= 5 #EPs= 2 Cls=e0(wlcon) Sub=01 Prot=01 Driver=btusb
E:  Ad=03(O) Atr=01(Isoc) MxPS=  49 Ivl=1ms
E:  Ad=83(I) Atr=01(Isoc) MxPS=  49 Ivl=1ms
---
 drivers/bluetooth/btusb.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/bluetooth/btusb.c +4 -0
index b8cd42d81d5ca..952dc9d2404ed 100644
@@ -492,6 +492,10 @@ static const struct usb_device_id blacklist_table[] = {
 	{ USB_VENDOR_AND_INTERFACE_INFO(0x8087, 0xe0, 0x01, 0x01),
 	  .driver_info = BTUSB_IGNORE },
 
+	/* Realtek 8821CE Bluetooth devices */
+	{ USB_DEVICE(0x13d3, 0x3529), .driver_info = BTUSB_REALTEK |
+						     BTUSB_WIDEBAND_SPEECH },
+
 	/* Realtek 8822CE Bluetooth devices */
 	{ USB_DEVICE(0x0bda, 0xb00c), .driver_info = BTUSB_REALTEK |
 						     BTUSB_WIDEBAND_SPEECH },
Signature · 1 line
2.39.2
author maintainer
#588 lore raw

PATCH 587/885 wifi: rtw89: debug: avoid invalid access on RTW89_DBG_SEL_MAC_30

In reply to Greg Kroah-Hartman

From: Zong-Zhe Yang <[email protected]>

[ Upstream commit c074da21dd346e0cfef5d08b0715078d7aea7f8d ]

Only 8852C chip has valid pages on RTW89_DBG_SEL_MAC_30. To other chips, this section is an address hole. It will lead to crash if trying to access this section on chips except for 8852C. So, we avoid that.

---
 drivers/net/wireless/realtek/rtw89/debug.c | 7 +++++++
 1 file changed, 7 insertions(+)
drivers/net/wireless/realtek/rtw89/debug.c +7 -0
index 730e83d54257f..50701c55ed602 100644
@@ -594,6 +594,7 @@ rtw89_debug_priv_mac_reg_dump_select(struct file *filp,
 	struct seq_file *m = (struct seq_file *)filp->private_data;
 	struct rtw89_debugfs_priv *debugfs_priv = m->private;
 	struct rtw89_dev *rtwdev = debugfs_priv->rtwdev;
+	const struct rtw89_chip_info *chip = rtwdev->chip;
 	char buf[32];
 	size_t buf_size;
 	int sel;
@@ -613,6 +614,12 @@ rtw89_debug_priv_mac_reg_dump_select(struct file *filp,
 		return -EINVAL;
 	}
 
+	if (sel == RTW89_DBG_SEL_MAC_30 && chip->chip_id != RTL8852C) {
+		rtw89_info(rtwdev, "sel %d is address hole on chip %d\n", sel,
+			   chip->chip_id);
+		return -EINVAL;
+	}
+
 	debugfs_priv->cb_data = sel;
 	rtw89_info(rtwdev, "select mac page dump %d\n", debugfs_priv->cb_data);
 
Signature · 1 line
2.39.2
author maintainer
#589 lore raw

PATCH 588/885 hv_netvsc: Check status in SEND_RNDIS_PKT completion message

In reply to Greg Kroah-Hartman

From: Michael Kelley <[email protected]>

[ Upstream commit dca5161f9bd052e9e73be90716ffd57e8762c697 ]

Completion responses to SEND_RNDIS_PKT messages are currently processed regardless of the status in the response, so that resources associated with the request are freed. While this is appropriate, code bugs that cause sending a malformed message, or errors on the Hyper-V host, go undetected. Fix this by checking the status and outputting a rate-limited message if there is an error.

---
 drivers/net/hyperv/netvsc.c | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)
drivers/net/hyperv/netvsc.c +18 -0
index 79f4e13620a46..da737d959e81c 100644
@@ -851,6 +851,7 @@ static void netvsc_send_completion(struct net_device *ndev,
 	u32 msglen = hv_pkt_datalen(desc);
 	struct nvsp_message *pkt_rqst;
 	u64 cmd_rqst;
+	u32 status;
 
 	/* First check if this is a VMBUS completion without data payload */
 	if (!msglen) {
@@ -922,6 +923,23 @@ static void netvsc_send_completion(struct net_device *ndev,
 		break;
 
 	case NVSP_MSG1_TYPE_SEND_RNDIS_PKT_COMPLETE:
+		if (msglen < sizeof(struct nvsp_message_header) +
+		    sizeof(struct nvsp_1_message_send_rndis_packet_complete)) {
+			if (net_ratelimit())
+				netdev_err(ndev, "nvsp_rndis_pkt_complete length too small: %u\n",
+					   msglen);
+			return;
+		}
+
+		/* If status indicates an error, output a message so we know
+		 * there's a problem. But process the completion anyway so the
+		 * resources are released.
+		 */
+		status = nvsp_packet->msg.v1_msg.send_rndis_pkt_complete.status;
+		if (status != NVSP_STAT_SUCCESS && net_ratelimit())
+			netdev_err(ndev, "nvsp_rndis_pkt_complete error status: %x\n",
+				   status);
+
 		netvsc_send_tx_complete(ndev, net_device, incoming_channel,
 					desc, budget);
 		break;
Signature · 1 line
2.39.2
author maintainer
#590 lore raw

PATCH 589/885 s390/kfence: fix page fault reporting

In reply to Greg Kroah-Hartman

From: Heiko Carstens <[email protected]>

[ Upstream commit d9c2cf67b9cfd643ba85d51bc865a89a92e4f979 ]

Baoquan He reported lots of KFENCE reports when /proc/kcore is read, e.g. with crash or even simpler with dd:

 BUG: KFENCE: invalid read in copy_from_kernel_nofault+0x5e/0x120
 Invalid read at 0x00000000f4f5149f:
  copy_from_kernel_nofault+0x5e/0x120
  read_kcore+0x6b2/0x870
  proc_reg_read+0x9a/0xf0
  vfs_read+0x94/0x270
  ksys_read+0x70/0x100
  __do_syscall+0x1d0/0x200
  system_call+0x82/0xb0

The reason for this is that read_kcore() simply reads memory that might have been unmapped by KFENCE with copy_from_kernel_nofault(). Any fault due to pages being unmapped by KFENCE would be handled gracefully by the fault handler (exception table fixup).

However the s390 fault handler first reports the fault, and only afterwards would perform the exception table fixup. Most architectures have this in reversed order, which also avoids the false positive KFENCE reports when an unmapped page is accessed.

Therefore change the s390 fault handler so it handles exception table fixups before KFENCE page faults are reported.

---
 arch/s390/mm/fault.c | 49 +++++++++++++++++++++++++++++++-------------
 1 file changed, 35 insertions(+), 14 deletions(-)
arch/s390/mm/fault.c +35 -14
index 9649d9382e0ae..8e84ed2bb944e 100644
@@ -96,6 +96,20 @@ static enum fault_type get_fault_type(struct pt_regs *regs)
 	return KERNEL_FAULT;
 }
 
+static unsigned long get_fault_address(struct pt_regs *regs)
+{
+	unsigned long trans_exc_code = regs->int_parm_long;
+
+	return trans_exc_code & __FAIL_ADDR_MASK;
+}
+
+static bool fault_is_write(struct pt_regs *regs)
+{
+	unsigned long trans_exc_code = regs->int_parm_long;
+
+	return (trans_exc_code & store_indication) == 0x400;
+}
+
 static int bad_address(void *p)
 {
 	unsigned long dummy;
@@ -228,15 +242,26 @@ static noinline void do_sigsegv(struct pt_regs *regs, int si_code)
 			(void __user *)(regs->int_parm_long & __FAIL_ADDR_MASK));
 }
 
-static noinline void do_no_context(struct pt_regs *regs)
+static noinline void do_no_context(struct pt_regs *regs, vm_fault_t fault)
 {
+	enum fault_type fault_type;
+	unsigned long address;
+	bool is_write;
+
 	if (fixup_exception(regs))
 		return;
+	fault_type = get_fault_type(regs);
+	if ((fault_type == KERNEL_FAULT) && (fault == VM_FAULT_BADCONTEXT)) {
+		address = get_fault_address(regs);
+		is_write = fault_is_write(regs);
+		if (kfence_handle_page_fault(address, is_write, regs))
+			return;
+	}
 	/*
 	 * Oops. The kernel tried to access some bad page. We'll have to
 	 * terminate things with extreme prejudice.
 	 */
-	if (get_fault_type(regs) == KERNEL_FAULT)
+	if (fault_type == KERNEL_FAULT)
 		printk(KERN_ALERT "Unable to handle kernel pointer dereference"
 		       " in virtual kernel address space\n");
 	else
@@ -255,7 +280,7 @@ static noinline void do_low_address(struct pt_regs *regs)
 		die (regs, "Low-address protection");
 	}
 
-	do_no_context(regs);
+	do_no_context(regs, VM_FAULT_BADACCESS);
 }
 
 static noinline void do_sigbus(struct pt_regs *regs)
@@ -286,28 +311,28 @@ static noinline void do_fault_error(struct pt_regs *regs, vm_fault_t fault)
 		fallthrough;
 	case VM_FAULT_BADCONTEXT:
 	case VM_FAULT_PFAULT:
-		do_no_context(regs);
+		do_no_context(regs, fault);
 		break;
 	case VM_FAULT_SIGNAL:
 		if (!user_mode(regs))
-			do_no_context(regs);
+			do_no_context(regs, fault);
 		break;
 	default: /* fault & VM_FAULT_ERROR */
 		if (fault & VM_FAULT_OOM) {
 			if (!user_mode(regs))
-				do_no_context(regs);
+				do_no_context(regs, fault);
 			else
 				pagefault_out_of_memory();
 		} else if (fault & VM_FAULT_SIGSEGV) {
 			/* Kernel mode? Handle exceptions or die */
 			if (!user_mode(regs))
-				do_no_context(regs);
+				do_no_context(regs, fault);
 			else
 				do_sigsegv(regs, SEGV_MAPERR);
 		} else if (fault & VM_FAULT_SIGBUS) {
 			/* Kernel mode? Handle exceptions or die */
 			if (!user_mode(regs))
-				do_no_context(regs);
+				do_no_context(regs, fault);
 			else
 				do_sigbus(regs);
 		} else
@@ -334,7 +359,6 @@ static inline vm_fault_t do_exception(struct pt_regs *regs, int access)
 	struct mm_struct *mm;
 	struct vm_area_struct *vma;
 	enum fault_type type;
-	unsigned long trans_exc_code;
 	unsigned long address;
 	unsigned int flags;
 	vm_fault_t fault;
@@ -351,9 +375,8 @@ static inline vm_fault_t do_exception(struct pt_regs *regs, int access)
 		return 0;
 
 	mm = tsk->mm;
-	trans_exc_code = regs->int_parm_long;
-	address = trans_exc_code & __FAIL_ADDR_MASK;
-	is_write = (trans_exc_code & store_indication) == 0x400;
+	address = get_fault_address(regs);
+	is_write = fault_is_write(regs);
 
 	/*
 	 * Verify that the fault happened in user space, that
@@ -364,8 +387,6 @@ static inline vm_fault_t do_exception(struct pt_regs *regs, int access)
 	type = get_fault_type(regs);
 	switch (type) {
 	case KERNEL_FAULT:
-		if (kfence_handle_page_fault(address, is_write, regs))
-			return 0;
 		goto out;
 	case USER_FAULT:
 	case GMAP_FAULT:
Signature · 1 line
2.39.2
author maintainer
#591 lore raw

PATCH 590/885 devlink: Fix TP_STRUCT_entry in trace of devlink health report

In reply to Greg Kroah-Hartman

From: Moshe Shemesh <[email protected]>

[ Upstream commit d0ab772c1f1558af84f3293a52e9e886e08e0754 ]

Fix a bug in trace point definition for devlink health report, as TP_STRUCT_entry of reporter_name should get reporter_name and not msg.

Note no fixes tag as this is a harmless bug as both reporter_name and msg are strings and TP_fast_assign for this entry is correct.

---
 include/trace/events/devlink.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/trace/events/devlink.h +1 -1
index 24969184c5348..77ff7cfc6049a 100644
@@ -88,7 +88,7 @@ TRACE_EVENT(devlink_health_report,
 		__string(bus_name, devlink_to_dev(devlink)->bus->name)
 		__string(dev_name, dev_name(devlink_to_dev(devlink)))
 		__string(driver_name, devlink_to_dev(devlink)->driver->name)
-		__string(reporter_name, msg)
+		__string(reporter_name, reporter_name)
 		__string(msg, msg)
 	),
 
Signature · 1 line
2.39.2
author maintainer
#592 lore raw

PATCH 591/885 scm: add user copy checks to put_cmsg()

In reply to Greg Kroah-Hartman

From: Eric Dumazet <[email protected]>

[ Upstream commit 5f1eb1ff58ea122e24adf0bc940f268ed2227462 ]

This is a followup of commit 2558b8039d05 ("net: use a bounce buffer for copying skb->mark")

x86 and powerpc define user_access_begin, meaning that they are not able to perform user copy checks when using user_write_access_begin() / unsafe_copy_to_user() and friends [1]

Instead of waiting bugs to trigger on other arches, add a check_object_size() in put_cmsg() to make sure that new code tested on x86 with CONFIG_HARDENED_USERCOPY=y will perform more security checks.

[1] We can not generically call check_object_size() from unsafe_copy_to_user() because UACCESS is enabled at this point.

---
 net/core/scm.c | 2 ++
 1 file changed, 2 insertions(+)
net/core/scm.c +2 -0
index 5c356f0dee30c..acb7d776fa6ec 100644
@@ -229,6 +229,8 @@ int put_cmsg(struct msghdr * msg, int level, int type, int len, void *data)
 	if (msg->msg_control_is_user) {
 		struct cmsghdr __user *cm = msg->msg_control_user;
 
+		check_object_size(data, cmlen - sizeof(*cm), true);
+
 		if (!user_write_access_begin(cm, cmlen))
 			goto efault;
 
Signature · 1 line
2.39.2
author maintainer
#593 lore raw

PATCH 592/885 drm: panel-orientation-quirks: Add quirk for Lenovo Yoga Tab 3 X90F

In reply to Greg Kroah-Hartman

From: Hans de Goede <[email protected]>

[ Upstream commit 8a238d7f7eea7592e0764bc3b9e79e7c6354b04c ]

The Lenovo Yoga Tab 3 X90F has a portrait 1600x2560 LCD used in landscape mode, add a quirk for this.

---
 drivers/gpu/drm/drm_panel_orientation_quirks.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)
drivers/gpu/drm/drm_panel_orientation_quirks.c +13 -0
index 3659f0465a724..23d63a4d42d9c 100644
@@ -127,6 +127,12 @@ static const struct drm_dmi_panel_orientation_data lcd1600x2560_leftside_up = {
 	.orientation = DRM_MODE_PANEL_ORIENTATION_LEFT_UP,
 };
 
+static const struct drm_dmi_panel_orientation_data lcd1600x2560_rightside_up = {
+	.width = 1600,
+	.height = 2560,
+	.orientation = DRM_MODE_PANEL_ORIENTATION_RIGHT_UP,
+};
+
 static const struct dmi_system_id orientation_data[] = {
 	{	/* Acer One 10 (S1003) */
 		.matches = {
@@ -331,6 +337,13 @@ static const struct dmi_system_id orientation_data[] = {
 		 DMI_MATCH(DMI_BIOS_VERSION, "BLADE_21"),
 		},
 		.driver_data = (void *)&lcd1200x1920_rightside_up,
+	}, {	/* Lenovo Yoga Tab 3 X90F */
+		.matches = {
+		 DMI_MATCH(DMI_SYS_VENDOR, "Intel Corporation"),
+		 DMI_MATCH(DMI_PRODUCT_NAME, "CHERRYVIEW D1 PLATFORM"),
+		 DMI_MATCH(DMI_PRODUCT_VERSION, "Blade3-10A-001"),
+		},
+		.driver_data = (void *)&lcd1600x2560_rightside_up,
 	}, {	/* Nanote UMPC-01 */
 		.matches = {
 		 DMI_MATCH(DMI_SYS_VENDOR, "RWC CO.,LTD"),
Signature · 1 line
2.39.2
author maintainer
#594 lore raw

PATCH 593/885 drm: panel-orientation-quirks: Add quirk for DynaBook K50

In reply to Greg Kroah-Hartman

From: Allen Ballway <[email protected]>

[ Upstream commit a3caf7ea0c3d5872ed0f2c51f5476aee0c47a73a ]

Like the ASUS T100HAN for which there is already a quirk, the DynaBook K50 has a 800x1280 portrait screen mounted in the tablet part of a landscape oriented 2-in-1. Update the quirk to be more generic and apply to this device.

---
 .../gpu/drm/drm_panel_orientation_quirks.c    | 20 ++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)
drivers/gpu/drm/drm_panel_orientation_quirks.c +13 -7
index 23d63a4d42d9c..b409fe256fd0a 100644
@@ -30,12 +30,6 @@ struct drm_dmi_panel_orientation_data {
 	int orientation;
 };
 
-static const struct drm_dmi_panel_orientation_data asus_t100ha = {
-	.width = 800,
-	.height = 1280,
-	.orientation = DRM_MODE_PANEL_ORIENTATION_LEFT_UP,
-};
-
 static const struct drm_dmi_panel_orientation_data gpd_micropc = {
 	.width = 720,
 	.height = 1280,
@@ -97,6 +91,12 @@ static const struct drm_dmi_panel_orientation_data lcd720x1280_rightside_up = {
 	.orientation = DRM_MODE_PANEL_ORIENTATION_RIGHT_UP,
 };
 
+static const struct drm_dmi_panel_orientation_data lcd800x1280_leftside_up = {
+	.width = 800,
+	.height = 1280,
+	.orientation = DRM_MODE_PANEL_ORIENTATION_LEFT_UP,
+};
+
 static const struct drm_dmi_panel_orientation_data lcd800x1280_rightside_up = {
 	.width = 800,
 	.height = 1280,
@@ -157,7 +157,7 @@ static const struct dmi_system_id orientation_data[] = {
 		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "ASUSTeK COMPUTER INC."),
 		  DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "T100HAN"),
 		},
-		.driver_data = (void *)&asus_t100ha,
+		.driver_data = (void *)&lcd800x1280_leftside_up,
 	}, {	/* Asus T101HA */
 		.matches = {
 		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "ASUSTeK COMPUTER INC."),
@@ -202,6 +202,12 @@ static const struct dmi_system_id orientation_data[] = {
 		  DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "Hi10 pro tablet"),
 		},
 		.driver_data = (void *)&lcd1200x1920_rightside_up,
+	}, {	/* Dynabook K50 */
+		.matches = {
+		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Dynabook Inc."),
+		  DMI_EXACT_MATCH(DMI_PRODUCT_NAME, "dynabook K50/FR"),
+		},
+		.driver_data = (void *)&lcd800x1280_leftside_up,
 	}, {	/* GPD MicroPC (generic strings, also match on bios date) */
 		.matches = {
 		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "Default string"),
Signature · 1 line
2.39.2
author maintainer
#595 lore raw

PATCH 594/885 drm/amd/display: Reduce expected sdp bandwidth for dcn321

In reply to Greg Kroah-Hartman

From: Dillon Varone <[email protected]>

[ Upstream commit 6b81090d6d4cc0fd818c9ec9dbb6906f921ad396 ]

[Description]
Modify soc BB to reduce expected sdp bandwidth and align with measurements to fix underflow issues.

---
 drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dml/dcn321/dcn321_fpu.c +1 -1
index f4b176599be7a..0ea406145c1d7 100644
@@ -136,7 +136,7 @@ struct _vcs_dpi_soc_bounding_box_st dcn3_21_soc = {
 	.urgent_out_of_order_return_per_channel_pixel_only_bytes = 4096,
 	.urgent_out_of_order_return_per_channel_pixel_and_vm_bytes = 4096,
 	.urgent_out_of_order_return_per_channel_vm_only_bytes = 4096,
-	.pct_ideal_sdp_bw_after_urgent = 100.0,
+	.pct_ideal_sdp_bw_after_urgent = 90.0,
 	.pct_ideal_fabric_bw_after_urgent = 67.0,
 	.pct_ideal_dram_sdp_bw_after_urgent_pixel_only = 20.0,
 	.pct_ideal_dram_sdp_bw_after_urgent_pixel_and_vm = 60.0, // N/A, for now keep as is until DML implemented
Signature · 1 line
2.39.2
author maintainer
#596 lore raw

PATCH 595/885 drm/amd/display: Revert Reduce delay when sink device not able to ACK 00340h write

In reply to Greg Kroah-Hartman

From: Ian Chen <[email protected]>

[ Upstream commit 639f6ad6df7f47db48b59956b469a6917a136afb ]

[WHY]
It causes regression AMD source will not write DPCD 340.

---
 drivers/gpu/drm/amd/display/dc/core/dc_link.c    |  6 ------
 drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c | 14 +++-----------
 drivers/gpu/drm/amd/display/dc/dc_dp_types.h     |  1 -
 3 files changed, 3 insertions(+), 18 deletions(-)
drivers/gpu/drm/amd/display/dc/core/dc_link.c +0 -6
index 40b9d2ce08e66..328c5e33cc66b 100644
@@ -1916,12 +1916,6 @@ struct dc_link *link_create(const struct link_init_data *init_params)
 	if (false == dc_link_construct(link, init_params))
 		goto construct_fail;
 
-	/*
-	 * Must use preferred_link_setting, not reported_link_cap or verified_link_cap,
-	 * since struct preferred_link_setting won't be reset after S3.
-	 */
-	link->preferred_link_setting.dpcd_source_device_specific_field_support = true;
-
 	return link;
 
 construct_fail:
drivers/gpu/drm/amd/display/dc/core/dc_link_dp.c +3 -11
index 1254d38f1778a..24f1aba4ae133 100644
@@ -6591,18 +6591,10 @@ void dpcd_set_source_specific_data(struct dc_link *link)
 
 			uint8_t hblank_size = (uint8_t)link->dc->caps.min_horizontal_blanking_period;
 
-			if (link->preferred_link_setting.dpcd_source_device_specific_field_support) {
-				result_write_min_hblank = core_link_write_dpcd(link,
-					DP_SOURCE_MINIMUM_HBLANK_SUPPORTED, (uint8_t *)(&hblank_size),
-					sizeof(hblank_size));
-
-				if (result_write_min_hblank == DC_ERROR_UNEXPECTED)
-					link->preferred_link_setting.dpcd_source_device_specific_field_support = false;
-			} else {
-				DC_LOG_DC("Sink device does not support 00340h DPCD write. Skipping on purpose.\n");
-			}
+			result_write_min_hblank = core_link_write_dpcd(link,
+				DP_SOURCE_MINIMUM_HBLANK_SUPPORTED, (uint8_t *)(&hblank_size),
+				sizeof(hblank_size));
 		}
-
 		DC_TRACE_LEVEL_MESSAGE(DAL_TRACE_LEVEL_INFORMATION,
 							WPP_BIT_FLAG_DC_DETECTION_DP_CAPS,
 							"result=%u link_index=%u enum dce_version=%d DPCD=0x%04X min_hblank=%u branch_dev_id=0x%x branch_dev_name='%c%c%c%c%c%c'",
drivers/gpu/drm/amd/display/dc/dc_dp_types.h +0 -1
index 2c54b6e0498bf..296793d8b2bf2 100644
@@ -149,7 +149,6 @@ struct dc_link_settings {
 	enum dc_link_spread link_spread;
 	bool use_link_rate_set;
 	uint8_t link_rate_set;
-	bool dpcd_source_device_specific_field_support;
 };
 
 union dc_dp_ffe_preset {
Signature · 1 line
2.39.2
author maintainer
#597 lore raw

PATCH 596/885 drm/amd/display: Fix potential null-deref in dm_resume

In reply to Greg Kroah-Hartman

From: Roman Li <[email protected]>

[ Upstream commit 7a7175a2cd84b7874bebbf8e59f134557a34161b ]

[Why]
Fixing smatch error:
dm_resume() error: we previously assumed 'aconnector->dc_link' could be null

[How]
Check if dc_link null at the beginning of the loop, so further checks can be dropped.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +4 -2
index a930b1873f2a4..25ffc51ae6fa7 100644
@@ -2744,12 +2744,14 @@ static int dm_resume(void *handle)
 	drm_for_each_connector_iter(connector, &iter) {
 		aconnector = to_amdgpu_dm_connector(connector);
 
+		if (!aconnector->dc_link)
+			continue;
+
 		/*
 		 * this is the case when traversing through already created
 		 * MST connectors, should be skipped
 		 */
-		if (aconnector->dc_link &&
-		    aconnector->dc_link->type == dc_connection_mst_branch)
+		if (aconnector->dc_link->type == dc_connection_mst_branch)
 			continue;
 
 		mutex_lock(&aconnector->hpd_lock);
Signature · 1 line
2.39.2
author maintainer
#598 lore raw

PATCH 597/885 drm/omap: dsi: Fix excessive stack usage

In reply to Greg Kroah-Hartman

From: Tomi Valkeinen <[email protected]>

[ Upstream commit cfca78971b9233aef0891507a98fba62046d4542 ]

dsi_dump_dsi_irqs(), a function used for debugfs prints, has a large struct in its frame, which can result in:

drivers/gpu/drm/omapdrm/dss/dsi.c:1126:1: warning: the frame size of 1060 bytes is larger than 1024 bytes [-Wframe-larger-than=]

As the performance of the function is of no concern, let's allocate the struct with kmalloc instead.

Compile-tested only.

---
 drivers/gpu/drm/omapdrm/dss/dsi.c | 26 ++++++++++++++++----------
 1 file changed, 16 insertions(+), 10 deletions(-)
drivers/gpu/drm/omapdrm/dss/dsi.c +16 -10
index a6845856cbce4..4c1084eb01759 100644
@@ -1039,22 +1039,26 @@ static int dsi_dump_dsi_irqs(struct seq_file *s, void *p)
 {
 	struct dsi_data *dsi = s->private;
 	unsigned long flags;
-	struct dsi_irq_stats stats;
+	struct dsi_irq_stats *stats;
+
+	stats = kmalloc(sizeof(*stats), GFP_KERNEL);
+	if (!stats)
+		return -ENOMEM;
 
 	spin_lock_irqsave(&dsi->irq_stats_lock, flags);
 
-	stats = dsi->irq_stats;
+	*stats = dsi->irq_stats;
 	memset(&dsi->irq_stats, 0, sizeof(dsi->irq_stats));
 	dsi->irq_stats.last_reset = jiffies;
 
 	spin_unlock_irqrestore(&dsi->irq_stats_lock, flags);
 
 	seq_printf(s, "period %u ms\n",
-			jiffies_to_msecs(jiffies - stats.last_reset));
+			jiffies_to_msecs(jiffies - stats->last_reset));
 
-	seq_printf(s, "irqs %d\n", stats.irq_count);
+	seq_printf(s, "irqs %d\n", stats->irq_count);
 #define PIS(x) \
-	seq_printf(s, "%-20s %10d\n", #x, stats.dsi_irqs[ffs(DSI_IRQ_##x)-1]);
+	seq_printf(s, "%-20s %10d\n", #x, stats->dsi_irqs[ffs(DSI_IRQ_##x)-1]);
 
 	seq_printf(s, "-- DSI%d interrupts --\n", dsi->module_id + 1);
 	PIS(VC0);
@@ -1078,10 +1082,10 @@ static int dsi_dump_dsi_irqs(struct seq_file *s, void *p)
 
 #define PIS(x) \
 	seq_printf(s, "%-20s %10d %10d %10d %10d\n", #x, \
-			stats.vc_irqs[0][ffs(DSI_VC_IRQ_##x)-1], \
-			stats.vc_irqs[1][ffs(DSI_VC_IRQ_##x)-1], \
-			stats.vc_irqs[2][ffs(DSI_VC_IRQ_##x)-1], \
-			stats.vc_irqs[3][ffs(DSI_VC_IRQ_##x)-1]);
+			stats->vc_irqs[0][ffs(DSI_VC_IRQ_##x)-1], \
+			stats->vc_irqs[1][ffs(DSI_VC_IRQ_##x)-1], \
+			stats->vc_irqs[2][ffs(DSI_VC_IRQ_##x)-1], \
+			stats->vc_irqs[3][ffs(DSI_VC_IRQ_##x)-1]);
 
 	seq_printf(s, "-- VC interrupts --\n");
 	PIS(CS);
@@ -1097,7 +1101,7 @@ static int dsi_dump_dsi_irqs(struct seq_file *s, void *p)
 
 #define PIS(x) \
 	seq_printf(s, "%-20s %10d\n", #x, \
-			stats.cio_irqs[ffs(DSI_CIO_IRQ_##x)-1]);
+			stats->cio_irqs[ffs(DSI_CIO_IRQ_##x)-1]);
 
 	seq_printf(s, "-- CIO interrupts --\n");
 	PIS(ERRSYNCESC1);
@@ -1122,6 +1126,8 @@ static int dsi_dump_dsi_irqs(struct seq_file *s, void *p)
 	PIS(ULPSACTIVENOT_ALL1);
 #undef PIS
 
+	kfree(stats);
+
 	return 0;
 }
 #endif
Signature · 1 line
2.39.2
author maintainer
#599 lore raw

PATCH 598/885 HID: Add Mapping for System Microphone Mute

In reply to Greg Kroah-Hartman

From: Jingyuan Liang <[email protected]>

[ Upstream commit 2d60f9f4f26785a00273cb81fe60eff129ebd449 ]

HUTRR110 added a new usage code for a key that is supposed to mute/unmute microphone system-wide.

Map the new usage code(0x01 0xa9) to keycode KEY_MICMUTE. Additionally hid-debug is adjusted to recognize this keycode as well.

---
 drivers/hid/hid-debug.c | 1 +
 drivers/hid/hid-input.c | 8 ++++++++
 2 files changed, 9 insertions(+)
drivers/hid/hid-debug.c +1 -0
index 2ca6ab600bc9f..15e35702773cd 100644
@@ -972,6 +972,7 @@ static const char *keys[KEY_MAX + 1] = {
 	[KEY_KBD_LAYOUT_NEXT] = "KbdLayoutNext",
 	[KEY_EMOJI_PICKER] = "EmojiPicker",
 	[KEY_DICTATE] = "Dictate",
+	[KEY_MICMUTE] = "MicrophoneMute",
 	[KEY_BRIGHTNESS_MIN] = "BrightnessMin",
 	[KEY_BRIGHTNESS_MAX] = "BrightnessMax",
 	[KEY_BRIGHTNESS_AUTO] = "BrightnessAuto",
drivers/hid/hid-input.c +8 -0
index 7e94ca1822afb..8797ec7b29322 100644
@@ -793,6 +793,14 @@ static void hidinput_configure_usage(struct hid_input *hidinput, struct hid_fiel
 			break;
 		}
 
+		if ((usage->hid & 0xf0) == 0xa0) {	/* SystemControl */
+			switch (usage->hid & 0xf) {
+			case 0x9: map_key_clear(KEY_MICMUTE); break;
+			default: goto ignore;
+			}
+			break;
+		}
+
 		if ((usage->hid & 0xf0) == 0xb0) {	/* SC - Display */
 			switch (usage->hid & 0xf) {
 			case 0x05: map_key_clear(KEY_SWITCHVIDEOMODE); break;
Signature · 1 line
2.39.2
author maintainer
#600 lore raw

PATCH 599/885 drm/tiny: ili9486: Do not assume 8-bit only SPI controllers

In reply to Greg Kroah-Hartman

From: Carlo Caione <[email protected]>

[ Upstream commit 77772e607522daa61f3af74df018559db75c43d6 ]

The pixel data for the ILI9486 is always 16-bits wide and it must be sent over the SPI bus. When the controller is only able to deal with 8-bit transfers, this 16-bits data needs to be swapped before the sending to account for the big endian bus, this is on the contrary not needed when the SPI controller already supports 16-bits transfers.

The decision about swapping the pixel data or not is taken in the MIPI DBI code by probing the controller capabilities: if the controller only suppors 8-bit transfers the data is swapped, otherwise it is not.

This swapping/non-swapping is relying on the assumption that when the controller does support 16-bit transactions then the data is sent unswapped in 16-bits-per-word over SPI.

The problem with the ILI9486 driver is that it is forcing 8-bit transactions also for controllers supporting 16-bits, violating the assumption and corrupting the pixel data.

Align the driver to what is done in the MIPI DBI code by adjusting the transfer size to the maximum allowed by the SPI controller.

---
 drivers/gpu/drm/tiny/ili9486.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)
drivers/gpu/drm/tiny/ili9486.c +9 -4
index c80028bb1d110..7b3048a3d9086 100644
@@ -43,6 +43,7 @@ static int waveshare_command(struct mipi_dbi *mipi, u8 *cmd, u8 *par,
 			     size_t num)
 {
 	struct spi_device *spi = mipi->spi;
+	unsigned int bpw = 8;
 	void *data = par;
 	u32 speed_hz;
 	int i, ret;
@@ -56,8 +57,6 @@ static int waveshare_command(struct mipi_dbi *mipi, u8 *cmd, u8 *par,
 	 * The displays are Raspberry Pi HATs and connected to the 8-bit only
 	 * SPI controller, so 16-bit command and parameters need byte swapping
 	 * before being transferred as 8-bit on the big endian SPI bus.
-	 * Pixel data bytes have already been swapped before this function is
-	 * called.
 	 */
 	buf[0] = cpu_to_be16(*cmd);
 	gpiod_set_value_cansleep(mipi->dc, 0);
@@ -71,12 +70,18 @@ static int waveshare_command(struct mipi_dbi *mipi, u8 *cmd, u8 *par,
 		for (i = 0; i < num; i++)
 			buf[i] = cpu_to_be16(par[i]);
 		num *= 2;
-		speed_hz = mipi_dbi_spi_cmd_max_speed(spi, num);
 		data = buf;
 	}
 
+	/*
+	 * Check whether pixel data bytes needs to be swapped or not
+	 */
+	if (*cmd == MIPI_DCS_WRITE_MEMORY_START && !mipi->swap_bytes)
+		bpw = 16;
+
 	gpiod_set_value_cansleep(mipi->dc, 1);
-	ret = mipi_dbi_spi_transfer(spi, speed_hz, 8, data, num);
+	speed_hz = mipi_dbi_spi_cmd_max_speed(spi, num);
+	ret = mipi_dbi_spi_transfer(spi, speed_hz, bpw, data, num);
  free:
 	kfree(buf);
 
Signature · 1 line
2.39.2
author maintainer
#601 lore raw

PATCH 600/885 drm/amd/display: Defer DIG FIFO disable after VID stream enable

In reply to Greg Kroah-Hartman

From: Nicholas Kazlauskas <[email protected]>

[ Upstream commit 2d90a1c054831338d57b39aec4d273cf3e867590 ]

[Why]
On some monitors we see a brief flash of corruption during the monitor disable sequence caused by FIFO being disabled in the middle of an active DP stream.

[How]
Wait until DP vid stream is disabled before turning off the FIFO.

The FIFO reset on DP unblank should take care of clearing any FIFO error, if any.

---
 .../drm/amd/display/dc/dcn314/dcn314_dio_stream_encoder.c   | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/display/dc/dcn314/dcn314_dio_stream_encoder.c +3 -3
index 38842f938bed0..0926db0183383 100644
@@ -278,10 +278,10 @@ static void enc314_stream_encoder_dp_blank(
 	struct dc_link *link,
 	struct stream_encoder *enc)
 {
-	/* New to DCN314 - disable the FIFO before VID stream disable. */
-	enc314_disable_fifo(enc);
-
 	enc1_stream_encoder_dp_blank(link, enc);
+
+	/* Disable FIFO after the DP vid stream is disabled to avoid corruption. */
+	enc314_disable_fifo(enc);
 }
 
 static void enc314_stream_encoder_dp_unblank(
Signature · 1 line
2.39.2
author maintainer
#602 lore raw

PATCH 601/885 drm/radeon: free iio for atombios when driver shutdown

In reply to Greg Kroah-Hartman

From: Liwei Song <[email protected]>

[ Upstream commit 4773fadedca918faec443daaca5e4ea1c0ced144 ]

Fix below kmemleak when unload radeon driver:

unreferenced object 0xffff9f8608ede200 (size 512):
comm "systemd-udevd", pid 326, jiffies 4294682822 (age 716.338s) hex dump (first 32 bytes):

    00 00 00 00 c4 aa ec aa 14 ab 00 00 00 00 00 00  ................
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
    [<0000000062fadebe>] kmem_cache_alloc_trace+0x2f1/0x500
    [<00000000b6883cea>] atom_parse+0x117/0x230 [radeon]
    [<00000000158c23fd>] radeon_atombios_init+0xab/0x170 [radeon]
    [<00000000683f672e>] si_init+0x57/0x750 [radeon]
    [<00000000566cc31f>] radeon_device_init+0x559/0x9c0 [radeon]
    [<0000000046efabb3>] radeon_driver_load_kms+0xc1/0x1a0 [radeon]
    [<00000000b5155064>] drm_dev_register+0xdd/0x1d0
    [<0000000045fec835>] radeon_pci_probe+0xbd/0x100 [radeon]
    [<00000000e69ecca3>] pci_device_probe+0xe1/0x160
    [<0000000019484b76>] really_probe.part.0+0xc1/0x2c0
    [<000000003f2649da>] __driver_probe_device+0x96/0x130
    [<00000000231c5bb1>] driver_probe_device+0x24/0xf0
    [<0000000000a42377>] __driver_attach+0x77/0x190
    [<00000000d7574da6>] bus_for_each_dev+0x7f/0xd0
    [<00000000633166d2>] driver_attach+0x1e/0x30
    [<00000000313b05b8>] bus_add_driver+0x12c/0x1e0

iio was allocated in atom_index_iio() called by atom_parse(), but it doesn't got released when the dirver is shutdown. Fix this kmemleak by free it in radeon_atombios_fini().

---
 drivers/gpu/drm/radeon/radeon_device.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/radeon/radeon_device.c +1 -0
index a556b6be11374..e1f3ab607e4f4 100644
@@ -1023,6 +1023,7 @@ void radeon_atombios_fini(struct radeon_device *rdev)
 {
 	if (rdev->mode_info.atom_context) {
 		kfree(rdev->mode_info.atom_context->scratch);
+		kfree(rdev->mode_info.atom_context->iio);
 	}
 	kfree(rdev->mode_info.atom_context);
 	rdev->mode_info.atom_context = NULL;
Signature · 1 line
2.39.2
author maintainer
#603 lore raw

PATCH 602/885 drm/amd: Avoid BUG() for case of SRIOV missing IP version

In reply to Greg Kroah-Hartman

From: Mario Limonciello <[email protected]>

[ Upstream commit 93fec4f8c158584065134b4d45e875499bf517c8 ]

No need to crash the kernel. AMDGPU will now fail to probe.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_psp.c +1 -1
index 712dd72f3ccf2..087147f09933a 100644
@@ -354,7 +354,7 @@ static int psp_init_sriov_microcode(struct psp_context *psp)
 		adev->virt.autoload_ucode_id = AMDGPU_UCODE_ID_CP_MES1_DATA;
 		break;
 	default:
-		BUG();
+		ret = -EINVAL;
 		break;
 	}
 	return ret;
Signature · 1 line
2.39.2
author maintainer
#604 lore raw

PATCH 603/885 drm/amdkfd: Page aligned memory reserve size

In reply to Greg Kroah-Hartman

From: Philip Yang <[email protected]>

[ Upstream commit 0c2dece8fb541ab07b68c3312a1065fa9c927a81 ]

Use page aligned size to reserve memory usage because page aligned TTM BO size is used to unreserve memory usage, otherwise no page aligned size causes memory usage accounting unbalanced.

Change vram_used definition type to int64_t to be able to trigger WARN_ONCE(adev && adev->kfd.vram_used < 0, "..."), to help debug the accounting issue with warning and backtrace.

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h       |  2 +-
 drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c | 12 +++++++-----
 drivers/gpu/drm/amd/amdkfd/kfd_chardev.c         |  9 +++++++--
 3 files changed, 15 insertions(+), 8 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd.h +1 -1
index 30f145dc8724e..dbc842590b253 100644
@@ -95,7 +95,7 @@ struct amdgpu_amdkfd_fence {
 
 struct amdgpu_kfd_dev {
 	struct kfd_dev *dev;
-	uint64_t vram_used;
+	int64_t vram_used;
 	uint64_t vram_used_aligned;
 	bool init_complete;
 	struct work_struct reset_work;
drivers/gpu/drm/amd/amdgpu/amdgpu_amdkfd_gpuvm.c +7 -5
index 404c839683b1c..da01c1424b4ad 100644
@@ -1653,6 +1653,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	struct amdgpu_bo *bo;
 	struct drm_gem_object *gobj = NULL;
 	u32 domain, alloc_domain;
+	uint64_t aligned_size;
 	u64 alloc_flags;
 	int ret;
 
@@ -1703,22 +1704,23 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	 * the memory.
 	 */
 	if ((*mem)->aql_queue)
-		size = size >> 1;
+		size >>= 1;
+	aligned_size = PAGE_ALIGN(size);
 
 	(*mem)->alloc_flags = flags;
 
 	amdgpu_sync_create(&(*mem)->sync);
 
-	ret = amdgpu_amdkfd_reserve_mem_limit(adev, size, flags);
+	ret = amdgpu_amdkfd_reserve_mem_limit(adev, aligned_size, flags);
 	if (ret) {
 		pr_debug("Insufficient memory\n");
 		goto err_reserve_limit;
 	}
 
 	pr_debug("\tcreate BO VA 0x%llx size 0x%llx domain %s\n",
-			va, size, domain_string(alloc_domain));
+			va, (*mem)->aql_queue ? size << 1 : size, domain_string(alloc_domain));
 
-	ret = amdgpu_gem_object_create(adev, size, 1, alloc_domain, alloc_flags,
+	ret = amdgpu_gem_object_create(adev, aligned_size, 1, alloc_domain, alloc_flags,
 				       bo_type, NULL, &gobj);
 	if (ret) {
 		pr_debug("Failed to create BO on domain %s. ret %d\n",
@@ -1775,7 +1777,7 @@ int amdgpu_amdkfd_gpuvm_alloc_memory_of_gpu(
 	/* Don't unreserve system mem limit twice */
 	goto err_reserve_limit;
 err_bo_create:
-	amdgpu_amdkfd_unreserve_mem_limit(adev, size, flags);
+	amdgpu_amdkfd_unreserve_mem_limit(adev, aligned_size, flags);
 err_reserve_limit:
 	mutex_destroy(&(*mem)->lock);
 	if (gobj)
drivers/gpu/drm/amd/amdkfd/kfd_chardev.c +7 -2
index 6d291aa6386bd..f79b8e964140e 100644
@@ -1127,8 +1127,13 @@ static int kfd_ioctl_alloc_memory_of_gpu(struct file *filep,
 	}
 
 	/* Update the VRAM usage count */
-	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM)
-		WRITE_ONCE(pdd->vram_usage, pdd->vram_usage + args->size);
+	if (flags & KFD_IOC_ALLOC_MEM_FLAGS_VRAM) {
+		uint64_t size = args->size;
+
+		if (flags & KFD_IOC_ALLOC_MEM_FLAGS_AQL_QUEUE_MEM)
+			size >>= 1;
+		WRITE_ONCE(pdd->vram_usage, pdd->vram_usage + PAGE_ALIGN(size));
+	}
 
 	mutex_unlock(&p->mutex);
 
Signature · 1 line
2.39.2
author maintainer
#605 lore raw

PATCH 604/885 scsi: lpfc: Fix use-after-free KFENCE violation during sysfs firmware write

In reply to Greg Kroah-Hartman

From: Justin Tee <[email protected]>

[ Upstream commit 21681b81b9ae548c5dae7ae00d931197a27f480c ]

During the sysfs firmware write process, a use-after-free read warning is logged from the lpfc_wr_object() routine:

  BUG: KFENCE: use-after-free read in lpfc_wr_object+0x235/0x310 [lpfc]
  Use-after-free read at 0x0000000000cf164d (in kfence-#111):
  lpfc_wr_object+0x235/0x310 [lpfc]
  lpfc_write_firmware.cold+0x206/0x30d [lpfc]
  lpfc_sli4_request_firmware_update+0xa6/0x100 [lpfc]
  lpfc_request_firmware_upgrade_store+0x66/0xb0 [lpfc]
  kernfs_fop_write_iter+0x121/0x1b0
  new_sync_write+0x11c/0x1b0
  vfs_write+0x1ef/0x280
  ksys_write+0x5f/0xe0
  do_syscall_64+0x59/0x90
  entry_SYSCALL_64_after_hwframe+0x63/0xcd

The driver accessed wr_object pointer data, which was initialized into mailbox payload memory, after the mailbox object was released back to the mailbox pool.

Fix by moving the mailbox free calls to the end of the routine ensuring that we don't reference internal mailbox memory after release.

---
 drivers/scsi/lpfc/lpfc_sli.c | 19 +++++++++++++------
 1 file changed, 13 insertions(+), 6 deletions(-)
drivers/scsi/lpfc/lpfc_sli.c +13 -6
index 21c52154626f1..b93c948c4fcc4 100644
@@ -20802,6 +20802,7 @@ lpfc_wr_object(struct lpfc_hba *phba, struct list_head *dmabuf_list,
 	struct lpfc_mbx_wr_object *wr_object;
 	LPFC_MBOXQ_t *mbox;
 	int rc = 0, i = 0;
+	int mbox_status = 0;
 	uint32_t shdr_status, shdr_add_status, shdr_add_status_2;
 	uint32_t shdr_change_status = 0, shdr_csf = 0;
 	uint32_t mbox_tmo;
@@ -20847,11 +20848,15 @@ lpfc_wr_object(struct lpfc_hba *phba, struct list_head *dmabuf_list,
 	wr_object->u.request.bde_count = i;
 	bf_set(lpfc_wr_object_write_length, &wr_object->u.request, written);
 	if (!phba->sli4_hba.intr_enable)
-		rc = lpfc_sli_issue_mbox(phba, mbox, MBX_POLL);
+		mbox_status = lpfc_sli_issue_mbox(phba, mbox, MBX_POLL);
 	else {
 		mbox_tmo = lpfc_mbox_tmo_val(phba, mbox);
-		rc = lpfc_sli_issue_mbox_wait(phba, mbox, mbox_tmo);
+		mbox_status = lpfc_sli_issue_mbox_wait(phba, mbox, mbox_tmo);
 	}
+
+	/* The mbox status needs to be maintained to detect MBOX_TIMEOUT. */
+	rc = mbox_status;
+
 	/* The IOCTL status is embedded in the mailbox subheader. */
 	shdr_status = bf_get(lpfc_mbox_hdr_status,
 			     &wr_object->header.cfg_shdr.response);
@@ -20866,10 +20871,6 @@ lpfc_wr_object(struct lpfc_hba *phba, struct list_head *dmabuf_list,
 				  &wr_object->u.response);
 	}
 
-	if (!phba->sli4_hba.intr_enable)
-		mempool_free(mbox, phba->mbox_mem_pool);
-	else if (rc != MBX_TIMEOUT)
-		mempool_free(mbox, phba->mbox_mem_pool);
 	if (shdr_status || shdr_add_status || shdr_add_status_2 || rc) {
 		lpfc_printf_log(phba, KERN_ERR, LOG_TRACE_EVENT,
 				"3025 Write Object mailbox failed with "
@@ -20887,6 +20888,12 @@ lpfc_wr_object(struct lpfc_hba *phba, struct list_head *dmabuf_list,
 		lpfc_log_fw_write_cmpl(phba, shdr_status, shdr_add_status,
 				       shdr_add_status_2, shdr_change_status,
 				       shdr_csf);
+
+	if (!phba->sli4_hba.intr_enable)
+		mempool_free(mbox, phba->mbox_mem_pool);
+	else if (mbox_status != MBX_TIMEOUT)
+		mempool_free(mbox, phba->mbox_mem_pool);
+
 	return rc;
 }
 
Signature · 1 line
2.39.2
author maintainer
#606 lore raw

PATCH 605/885 Revert "fbcon: dont lose the console font across generic->chip driver switch"

In reply to Greg Kroah-Hartman

From: Thomas Zimmermann <[email protected]>

[ Upstream commit 12d5796d55f9fd9e4b621003127c99e176665064 ]

This reverts commit ae1287865f5361fa138d4d3b1b6277908b54eac9.

Always free the console font when deinitializing the framebuffer console. Subsequent framebuffer consoles will then use the default font. Rely on userspace to load any user-configured font for these consoles.

Commit ae1287865f53 ("fbcon: don't lose the console font across generic->chip driver switch") was introduced to work around losing the font during graphics-device handover. [1][2] It kept a dangling pointer with the font data between loading the two consoles, which is fairly adventurous hack. It also never covered cases when the other consoles, such as VGA text mode, where involved.

The problem has meanwhile been solved in userspace. Systemd comes with a udev rule that re-installs the configured font when a console comes up. [3] So the kernel workaround can be removed.

This also removes one of the two special cases triggered by setting FBINFO_MISC_FIRMWARE in an fbdev driver.

Tested during device handover from efifb and simpledrm to radeon. Udev reloads the configured console font for the new driver's terminal.

---
 drivers/video/fbdev/core/fbcon.c | 17 ++++++-----------
 1 file changed, 6 insertions(+), 11 deletions(-)
drivers/video/fbdev/core/fbcon.c +6 -11
index 1b14c21af2b74..2bc8baa90c0f2 100644
@@ -958,7 +958,7 @@ static const char *fbcon_startup(void)
 	set_blitting_type(vc, info);
 
 	/* Setup default font */
-	if (!p->fontdata && !vc->vc_font.data) {
+	if (!p->fontdata) {
 		if (!fontname[0] || !(font = find_font(fontname)))
 			font = get_default_font(info->var.xres,
 						info->var.yres,
@@ -968,8 +968,6 @@ static const char *fbcon_startup(void)
 		vc->vc_font.height = font->height;
 		vc->vc_font.data = (void *)(p->fontdata = font->data);
 		vc->vc_font.charcount = font->charcount;
-	} else {
-		p->fontdata = vc->vc_font.data;
 	}
 
 	cols = FBCON_SWAP(ops->rotate, info->var.xres, info->var.yres);
@@ -1135,9 +1133,9 @@ static void fbcon_init(struct vc_data *vc, int init)
 	ops->p = &fb_display[fg_console];
 }
 
-static void fbcon_free_font(struct fbcon_display *p, bool freefont)
+static void fbcon_free_font(struct fbcon_display *p)
 {
-	if (freefont && p->userfont && p->fontdata && (--REFCOUNT(p->fontdata) == 0))
+	if (p->userfont && p->fontdata && (--REFCOUNT(p->fontdata) == 0))
 		kfree(p->fontdata - FONT_EXTRA_WORDS * sizeof(int));
 	p->fontdata = NULL;
 	p->userfont = 0;
@@ -1172,8 +1170,8 @@ static void fbcon_deinit(struct vc_data *vc)
 	struct fb_info *info;
 	struct fbcon_ops *ops;
 	int idx;
-	bool free_font = true;
 
+	fbcon_free_font(p);
 	idx = con2fb_map[vc->vc_num];
 
 	if (idx == -1)
@@ -1184,8 +1182,6 @@ static void fbcon_deinit(struct vc_data *vc)
 	if (!info)
 		goto finished;
 
-	if (info->flags & FBINFO_MISC_FIRMWARE)
-		free_font = false;
 	ops = info->fbcon_par;
 
 	if (!ops)
@@ -1197,9 +1193,8 @@ static void fbcon_deinit(struct vc_data *vc)
 	ops->initialized = false;
 finished:
 
-	fbcon_free_font(p, free_font);
-	if (free_font)
-		vc->vc_font.data = NULL;
+	fbcon_free_font(p);
+	vc->vc_font.data = NULL;
 
 	if (vc->vc_hi_font_mask && vc->vc_screenbuf)
 		set_vc_hi_font(vc, false);
Signature · 1 line
2.39.2
author maintainer
#607 lore raw

PATCH 606/885 drm/amd: Avoid ASSERT for some message failures

In reply to Greg Kroah-Hartman

From: Mario Limonciello <[email protected]>

[ Upstream commit 3e5019ee67760cd61b2a5fd605e1289c2f92d983 ]

On DCN314 when resuming from s0i3 an ASSERT is shown indicating that VBIOSSMC_MSG_SetHardMinDcfclkByFreq returned VBIOSSMC_Result_Failed.

This isn't a driver bug; it's a BIOS/configuration bug. To make this easier to triage, add an explicit warning when this issue happens.

This matches the behavior utilized for failures with VBIOSSMC_MSG_TransferTableDram2Smu configuration.

---
 drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/amd/display/dc/clk_mgr/dcn314/dcn314_smu.c +3 -0
index 2db595672a469..aa264c600408d 100644
@@ -146,6 +146,9 @@ static int dcn314_smu_send_msg_with_param(struct clk_mgr_internal *clk_mgr,
 		if (msg_id == VBIOSSMC_MSG_TransferTableDram2Smu &&
 		    param == TABLE_WATERMARKS)
 			DC_LOG_WARNING("Watermarks table not configured properly by SMU");
+		else if (msg_id == VBIOSSMC_MSG_SetHardMinDcfclkByFreq ||
+			 msg_id == VBIOSSMC_MSG_SetMinDeepSleepDcfclk)
+			DC_LOG_WARNING("DCFCLK_DPM is not enabled by BIOS");
 		else
 			ASSERT(0);
 		REG_WRITE(MP1_SMN_C2PMSG_91, VBIOSSMC_Result_OK);
Signature · 1 line
2.39.2
author maintainer
#608 lore raw

PATCH 607/885 drm: amd: display: Fix memory leakage

In reply to Greg Kroah-Hartman

From: Konstantin Meskhidze <[email protected]>

[ Upstream commit 6b8701be1f66064ca72733c5f6e13748cdbf8397 ]

This commit fixes memory leakage in dc_construct_ctx() function.

---
 drivers/gpu/drm/amd/display/dc/core/dc.c | 1 +
 1 file changed, 1 insertion(+)
drivers/gpu/drm/amd/display/dc/core/dc.c +1 -0
index 5260ad6de8038..24015f8cac75a 100644
@@ -878,6 +878,7 @@ static bool dc_construct_ctx(struct dc *dc,
 
 	dc_ctx->perf_trace = dc_perf_trace_create();
 	if (!dc_ctx->perf_trace) {
+		kfree(dc_ctx);
 		ASSERT_CRITICAL(false);
 		return false;
 	}
Signature · 1 line
2.39.2
author maintainer
#609 lore raw

PATCH 608/885 drm/amd/display: fix mapping to non-allocated address

In reply to Greg Kroah-Hartman

From: Brandon Syu <[email protected]>

[ Upstream commit 9190d4a263264eabf715f5fc1827da45e3fdc247 ]

[Why]
There is an issue mapping non-allocated location of memory. It would allocate gpio registers from an array out of bounds.

[How]
Patch correct numbers of bounds for using.

---
 .../gpu/drm/amd/display/dc/gpio/dcn20/hw_factory_dcn20.c   | 6 ++++--
 .../gpu/drm/amd/display/dc/gpio/dcn30/hw_factory_dcn30.c   | 6 ++++--
 .../gpu/drm/amd/display/dc/gpio/dcn32/hw_factory_dcn32.c   | 6 ++++--
 drivers/gpu/drm/amd/display/dc/gpio/ddc_regs.h             | 7 +++++++
 4 files changed, 19 insertions(+), 6 deletions(-)
drivers/gpu/drm/amd/display/dc/gpio/dcn20/hw_factory_dcn20.c +4 -2
index 9b63c6c0cc844..e0bd0c722e006 100644
@@ -138,7 +138,8 @@ static const struct ddc_sh_mask ddc_shift[] = {
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 3),
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 4),
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 5),
-	DDC_MASK_SH_LIST_DCN2(__SHIFT, 6)
+	DDC_MASK_SH_LIST_DCN2(__SHIFT, 6),
+	DDC_MASK_SH_LIST_DCN2_VGA(__SHIFT)
 };
 
 static const struct ddc_sh_mask ddc_mask[] = {
@@ -147,7 +148,8 @@ static const struct ddc_sh_mask ddc_mask[] = {
 	DDC_MASK_SH_LIST_DCN2(_MASK, 3),
 	DDC_MASK_SH_LIST_DCN2(_MASK, 4),
 	DDC_MASK_SH_LIST_DCN2(_MASK, 5),
-	DDC_MASK_SH_LIST_DCN2(_MASK, 6)
+	DDC_MASK_SH_LIST_DCN2(_MASK, 6),
+	DDC_MASK_SH_LIST_DCN2_VGA(_MASK)
 };
 
 #include "../generic_regs.h"
drivers/gpu/drm/amd/display/dc/gpio/dcn30/hw_factory_dcn30.c +4 -2
index 687d4f128480e..36a5736c58c92 100644
@@ -145,7 +145,8 @@ static const struct ddc_sh_mask ddc_shift[] = {
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 3),
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 4),
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 5),
-	DDC_MASK_SH_LIST_DCN2(__SHIFT, 6)
+	DDC_MASK_SH_LIST_DCN2(__SHIFT, 6),
+	DDC_MASK_SH_LIST_DCN2_VGA(__SHIFT)
 };
 
 static const struct ddc_sh_mask ddc_mask[] = {
@@ -154,7 +155,8 @@ static const struct ddc_sh_mask ddc_mask[] = {
 	DDC_MASK_SH_LIST_DCN2(_MASK, 3),
 	DDC_MASK_SH_LIST_DCN2(_MASK, 4),
 	DDC_MASK_SH_LIST_DCN2(_MASK, 5),
-	DDC_MASK_SH_LIST_DCN2(_MASK, 6)
+	DDC_MASK_SH_LIST_DCN2(_MASK, 6),
+	DDC_MASK_SH_LIST_DCN2_VGA(_MASK)
 };
 
 #include "../generic_regs.h"
drivers/gpu/drm/amd/display/dc/gpio/dcn32/hw_factory_dcn32.c +4 -2
index 0ea52ba5ac827..9f6872ae40203 100644
@@ -149,7 +149,8 @@ static const struct ddc_sh_mask ddc_shift[] = {
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 3),
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 4),
 	DDC_MASK_SH_LIST_DCN2(__SHIFT, 5),
-	DDC_MASK_SH_LIST_DCN2(__SHIFT, 6)
+	DDC_MASK_SH_LIST_DCN2(__SHIFT, 6),
+	DDC_MASK_SH_LIST_DCN2_VGA(__SHIFT)
 };
 
 static const struct ddc_sh_mask ddc_mask[] = {
@@ -158,7 +159,8 @@ static const struct ddc_sh_mask ddc_mask[] = {
 	DDC_MASK_SH_LIST_DCN2(_MASK, 3),
 	DDC_MASK_SH_LIST_DCN2(_MASK, 4),
 	DDC_MASK_SH_LIST_DCN2(_MASK, 5),
-	DDC_MASK_SH_LIST_DCN2(_MASK, 6)
+	DDC_MASK_SH_LIST_DCN2(_MASK, 6),
+	DDC_MASK_SH_LIST_DCN2_VGA(_MASK)
 };
 
 #include "../generic_regs.h"
drivers/gpu/drm/amd/display/dc/gpio/ddc_regs.h +7 -0
index 308a543178a56..59884ef651b39 100644
@@ -113,6 +113,13 @@
 	(PHY_AUX_CNTL__AUX## cd ##_PAD_RXSEL## mask_sh),\
 	(DC_GPIO_AUX_CTRL_5__DDC_PAD## cd ##_I2CMODE## mask_sh)}
 
+#define DDC_MASK_SH_LIST_DCN2_VGA(mask_sh) \
+	{DDC_MASK_SH_LIST_COMMON(mask_sh),\
+	0,\
+	0,\
+	0,\
+	0}
+
 struct ddc_registers {
 	struct gpio_registers gpio;
 	uint32_t ddc_setup;
Signature · 1 line
2.39.2
author maintainer
#610 lore raw

PATCH 609/885 HID: uclogic: Add frame type quirk

In reply to Greg Kroah-Hartman

From: José Expósito <[email protected]>

[ Upstream commit 14b71e6ad8ca59dd734c7fa9676f3d60bddee2a9 ]

The report descriptor used to get information about UGEE v2 devices is incorrect in the XP-PEN Deco Pro SW. It indicates that the device frame is of type UCLOGIC_PARAMS_FRAME_BUTTONS but the device has a frame of type UCLOGIC_PARAMS_FRAME_MOUSE.

Here is the original report descriptor:

0x0e 0x03 0xc8 0xb3 0x34 0x65 0x08 0x00 0xff 0x1f 0xd8 0x13 0x00 0x00
^ This byte should be 2

Add a quirk to be able to fix the reported frame type.

---
 drivers/hid/hid-uclogic-core.c   | 20 +-------------------
 drivers/hid/hid-uclogic-params.c |  5 +++++
 drivers/hid/hid-uclogic-params.h | 23 +++++++++++++++++++++++
 3 files changed, 29 insertions(+), 19 deletions(-)
drivers/hid/hid-uclogic-core.c +1 -19
index cfbbc39807a69..739984b8fa1b8 100644
@@ -22,25 +22,6 @@
 
 #include "hid-ids.h"
 
-/* Driver data */
-struct uclogic_drvdata {
-	/* Interface parameters */
-	struct uclogic_params params;
-	/* Pointer to the replacement report descriptor. NULL if none. */
-	__u8 *desc_ptr;
-	/*
-	 * Size of the replacement report descriptor.
-	 * Only valid if desc_ptr is not NULL
-	 */
-	unsigned int desc_size;
-	/* Pen input device */
-	struct input_dev *pen_input;
-	/* In-range timer */
-	struct timer_list inrange_timer;
-	/* Last rotary encoder state, or U8_MAX for none */
-	u8 re_state;
-};
-
 /**
  * uclogic_inrange_timeout - handle pen in-range state timeout.
  * Emulate input events normally generated when pen goes out of range for
@@ -202,6 +183,7 @@ static int uclogic_probe(struct hid_device *hdev,
 	}
 	timer_setup(&drvdata->inrange_timer, uclogic_inrange_timeout, 0);
 	drvdata->re_state = U8_MAX;
+	drvdata->quirks = id->driver_data;
 	hid_set_drvdata(hdev, drvdata);
 
 	/* Initialize the device and retrieve interface parameters */
drivers/hid/hid-uclogic-params.c +5 -0
index 3c5eea3df3288..e052538a62fb3 100644
@@ -1298,6 +1298,7 @@ static int uclogic_params_ugee_v2_init(struct uclogic_params *params,
 				       struct hid_device *hdev)
 {
 	int rc = 0;
+	struct uclogic_drvdata *drvdata;
 	struct usb_interface *iface;
 	__u8 bInterfaceNumber;
 	const int str_desc_len = 12;
@@ -1316,6 +1317,7 @@ static int uclogic_params_ugee_v2_init(struct uclogic_params *params,
 		goto cleanup;
 	}
 
+	drvdata = hid_get_drvdata(hdev);
 	iface = to_usb_interface(hdev->dev.parent);
 	bInterfaceNumber = iface->cur_altsetting->desc.bInterfaceNumber;
 
@@ -1382,6 +1384,9 @@ static int uclogic_params_ugee_v2_init(struct uclogic_params *params,
 	p.pen.subreport_list[0].id = UCLOGIC_RDESC_V1_FRAME_ID;
 
 	/* Initialize the frame interface */
+	if (drvdata->quirks & UCLOGIC_MOUSE_FRAME_QUIRK)
+		frame_type = UCLOGIC_PARAMS_FRAME_MOUSE;
+
 	switch (frame_type) {
 	case UCLOGIC_PARAMS_FRAME_DIAL:
 	case UCLOGIC_PARAMS_FRAME_MOUSE:
drivers/hid/hid-uclogic-params.h +23 -0
index a97477c02ff82..10a05c7fd9398 100644
@@ -19,6 +19,8 @@
 #include <linux/usb.h>
 #include <linux/hid.h>
 
+#define UCLOGIC_MOUSE_FRAME_QUIRK	BIT(0)
+
 /* Types of pen in-range reporting */
 enum uclogic_params_pen_inrange {
 	/* Normal reports: zero - out of proximity, one - in proximity */
@@ -215,6 +217,27 @@ struct uclogic_params {
 	struct uclogic_params_frame frame_list[3];
 };
 
+/* Driver data */
+struct uclogic_drvdata {
+	/* Interface parameters */
+	struct uclogic_params params;
+	/* Pointer to the replacement report descriptor. NULL if none. */
+	__u8 *desc_ptr;
+	/*
+	 * Size of the replacement report descriptor.
+	 * Only valid if desc_ptr is not NULL
+	 */
+	unsigned int desc_size;
+	/* Pen input device */
+	struct input_dev *pen_input;
+	/* In-range timer */
+	struct timer_list inrange_timer;
+	/* Last rotary encoder state, or U8_MAX for none */
+	u8 re_state;
+	/* Device quirks */
+	unsigned long quirks;
+};
+
 /* Initialize a tablet interface and discover its parameters */
 extern int uclogic_params_init(struct uclogic_params *params,
 				struct hid_device *hdev);
Signature · 1 line
2.39.2
author maintainer
#611 lore raw

PATCH 610/885 HID: uclogic: Add battery quirk

In reply to Greg Kroah-Hartman

From: José Expósito <[email protected]>

[ Upstream commit f60c377f52de37f8705c5fc6d57737fdaf309ff9 ]

Some UGEE v2 tablets have a wireless version with an internal battery and their firmware is able to report their battery level.

However, there was not found a field on their descriptor indicating whether the tablet has battery or not, making it mandatory to classify such devices through the UCLOGIC_BATTERY_QUIRK quirk.

---
 drivers/hid/hid-uclogic-params.c | 5 +++++
 drivers/hid/hid-uclogic-params.h | 1 +
 2 files changed, 6 insertions(+)
drivers/hid/hid-uclogic-params.c +5 -0
index e052538a62fb3..23624d5b07b5a 100644
@@ -1222,6 +1222,11 @@ static int uclogic_params_ugee_v2_init_frame_mouse(struct uclogic_params *p)
  */
 static bool uclogic_params_ugee_v2_has_battery(struct hid_device *hdev)
 {
+	struct uclogic_drvdata *drvdata = hid_get_drvdata(hdev);
+
+	if (drvdata->quirks & UCLOGIC_BATTERY_QUIRK)
+		return true;
+
 	/* The XP-PEN Deco LW vendor, product and version are identical to the
 	 * Deco L. The only difference reported by their firmware is the product
 	 * name. Add a quirk to support battery reporting on the wireless
drivers/hid/hid-uclogic-params.h +1 -0
index 10a05c7fd9398..b0e7f3807939b 100644
@@ -20,6 +20,7 @@
 #include <linux/hid.h>
 
 #define UCLOGIC_MOUSE_FRAME_QUIRK	BIT(0)
+#define UCLOGIC_BATTERY_QUIRK		BIT(1)
 
 /* Types of pen in-range reporting */
 enum uclogic_params_pen_inrange {
Signature · 1 line
2.39.2
author maintainer
#612 lore raw

PATCH 611/885 HID: uclogic: Add support for XP-PEN Deco Pro SW

In reply to Greg Kroah-Hartman

From: José Expósito <[email protected]>

[ Upstream commit 7744ca571af55b794595cff2da9d51a26904998f ]

The XP-PEN Deco Pro SW is a UGEE v2 device with a frame with 8 buttons, a bitmap dial and a mouse; however, the UCLOGIC_MOUSE_FRAME_QUIRK is required because it reports an incorrect frame type. Its pen has 2 buttons, supports tilt and pressure.

It can be connected using a USB cable or, to use it in wireless mode, using a USB Bluetooth dongle. When it is connected in wireless mode the device battery is used to power it.

All the pieces to support it are already in place. Add its ID and quirks in order to support the device.

---
 drivers/hid/hid-ids.h            | 1 +
 drivers/hid/hid-input.c          | 2 ++
 drivers/hid/hid-uclogic-core.c   | 3 +++
 drivers/hid/hid-uclogic-params.c | 2 ++
 4 files changed, 8 insertions(+)
drivers/hid/hid-ids.h +1 -0
index 9e36b4cd905ee..c662994d73381 100644
@@ -1300,6 +1300,7 @@
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO01_V2	0x0905
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L	0x0935
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_S	0x0909
+#define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_SW	0x0933
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_STAR06	0x0078
 #define USB_DEVICE_ID_UGEE_TABLET_G5		0x0074
 #define USB_DEVICE_ID_UGEE_TABLET_EX07S		0x0071
drivers/hid/hid-input.c +2 -0
index 8797ec7b29322..fd8fe2594a61f 100644
@@ -378,6 +378,8 @@ static const struct hid_device_id hid_battery_quirks[] = {
 	  HID_BATTERY_QUIRK_IGNORE },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE, USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L),
 	  HID_BATTERY_QUIRK_AVOID_QUERY },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE, USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_SW),
+	  HID_BATTERY_QUIRK_AVOID_QUERY },
 	{ HID_I2C_DEVICE(USB_VENDOR_ID_ELAN, I2C_DEVICE_ID_HP_ENVY_X360_15),
 	  HID_BATTERY_QUIRK_IGNORE },
 	{ HID_I2C_DEVICE(USB_VENDOR_ID_ELAN, I2C_DEVICE_ID_HP_ENVY_X360_15T_DR100),
drivers/hid/hid-uclogic-core.c +3 -0
index 739984b8fa1b8..7c05d38d3afad 100644
@@ -513,6 +513,9 @@ static const struct hid_device_id uclogic_devices[] = {
 				USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE,
 				USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_S) },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE,
+				USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_SW),
+		.driver_data = UCLOGIC_MOUSE_FRAME_QUIRK | UCLOGIC_BATTERY_QUIRK },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE,
 				USB_DEVICE_ID_UGEE_XPPEN_TABLET_STAR06) },
 	{ }
drivers/hid/hid-uclogic-params.c +2 -0
index 23624d5b07b5a..492a30f83575a 100644
@@ -1671,6 +1671,8 @@ int uclogic_params_init(struct uclogic_params *params,
 		     USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L):
 	case VID_PID(USB_VENDOR_ID_UGEE,
 		     USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_S):
+	case VID_PID(USB_VENDOR_ID_UGEE,
+		     USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_SW):
 		rc = uclogic_params_ugee_v2_init(&p, hdev);
 		if (rc != 0)
 			goto cleanup;
Signature · 1 line
2.39.2
author maintainer
#613 lore raw

PATCH 612/885 HID: uclogic: Add support for XP-PEN Deco Pro MW

In reply to Greg Kroah-Hartman

From: José Expósito <[email protected]>

[ Upstream commit 9266a88156d1fbb8e50d6eeff7bac44ad4eaecc2 ]

The XP-PEN Deco Pro MW is a UGEE v2 device with a frame with 8 buttons, a bitmap dial and a mouse. Its pen has 2 buttons, supports tilt and pressure.

It can be connected using a USB cable or, to use it in wireless mode, using a USB Bluetooth dongle. When it is connected in wireless mode the device battery is used to power it.

All the pieces to support it are already in place. Add its ID and quirks in order to support the device.

---
 drivers/hid/hid-ids.h            | 1 +
 drivers/hid/hid-input.c          | 2 ++
 drivers/hid/hid-uclogic-core.c   | 3 +++
 drivers/hid/hid-uclogic-params.c | 2 ++
 4 files changed, 8 insertions(+)
drivers/hid/hid-ids.h +1 -0
index c662994d73381..2235d78784b1b 100644
@@ -1299,6 +1299,7 @@
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO01	0x0042
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO01_V2	0x0905
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L	0x0935
+#define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_MW	0x0934
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_S	0x0909
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_SW	0x0933
 #define USB_DEVICE_ID_UGEE_XPPEN_TABLET_STAR06	0x0078
drivers/hid/hid-input.c +2 -0
index fd8fe2594a61f..c3f80b516f398 100644
@@ -378,6 +378,8 @@ static const struct hid_device_id hid_battery_quirks[] = {
 	  HID_BATTERY_QUIRK_IGNORE },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE, USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L),
 	  HID_BATTERY_QUIRK_AVOID_QUERY },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE, USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_MW),
+	  HID_BATTERY_QUIRK_AVOID_QUERY },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE, USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_SW),
 	  HID_BATTERY_QUIRK_AVOID_QUERY },
 	{ HID_I2C_DEVICE(USB_VENDOR_ID_ELAN, I2C_DEVICE_ID_HP_ENVY_X360_15),
drivers/hid/hid-uclogic-core.c +3 -0
index 7c05d38d3afad..bfbb51f8b5beb 100644
@@ -511,6 +511,9 @@ static const struct hid_device_id uclogic_devices[] = {
 				USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO01_V2) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE,
 				USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L) },
+	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE,
+				USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_MW),
+		.driver_data = UCLOGIC_MOUSE_FRAME_QUIRK | UCLOGIC_BATTERY_QUIRK },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE,
 				USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_S) },
 	{ HID_USB_DEVICE(USB_VENDOR_ID_UGEE,
drivers/hid/hid-uclogic-params.c +2 -0
index 492a30f83575a..0cc03c11ecc22 100644
@@ -1669,6 +1669,8 @@ int uclogic_params_init(struct uclogic_params *params,
 		     USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO01_V2):
 	case VID_PID(USB_VENDOR_ID_UGEE,
 		     USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_L):
+	case VID_PID(USB_VENDOR_ID_UGEE,
+		     USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_MW):
 	case VID_PID(USB_VENDOR_ID_UGEE,
 		     USB_DEVICE_ID_UGEE_XPPEN_TABLET_DECO_PRO_S):
 	case VID_PID(USB_VENDOR_ID_UGEE,
Signature · 1 line
2.39.2
author maintainer
#614 lore raw

PATCH 613/885 drm/msm/dsi: Add missing check for alloc_ordered_workqueue

In reply to Greg Kroah-Hartman

From: Jiasheng Jiang <[email protected]>

[ Upstream commit 115906ca7b535afb1fe7b5406c566ccd3873f82b ]

Add check for the return value of alloc_ordered_workqueue as it may return NULL pointer and cause NULL pointer dereference.

Patchwork: https://patchwork.freedesktop.org/patch/517646/

---
 drivers/gpu/drm/msm/dsi/dsi_host.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/msm/dsi/dsi_host.c +3 -0
index 89aadd3b3202b..f167a45f1fbdd 100644
@@ -1977,6 +1977,9 @@ int msm_dsi_host_init(struct msm_dsi *msm_dsi)
 
 	/* setup workqueue */
 	msm_host->workqueue = alloc_ordered_workqueue("dsi_drm_work", 0);
+	if (!msm_host->workqueue)
+		return -ENOMEM;
+
 	INIT_WORK(&msm_host->err_work, dsi_err_worker);
 
 	msm_dsi->id = msm_host->id;
Signature · 1 line
2.39.2
author maintainer
#615 lore raw

PATCH 614/885 drm: rcar-du: Add quirk for H3 ES1.x pclk workaround

In reply to Greg Kroah-Hartman

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 4f548bc48a2b4c4e54eecfddb6f7d24aa1b98768 ]

rcar_du_crtc.c does a soc_device_match() in rcar_du_crtc_set_display_timing() to find out if the SoC is H3 ES1.x, and if so, apply a workaround.

We will need another H3 ES1.x check in the following patch, so rather than adding more soc_device_match() calls, let's add a rcar_du_device_info entry for the ES1, and a quirk flag, RCAR_DU_QUIRK_H3_ES1_PCLK_STABILITY, for the workaround.

---
 drivers/gpu/drm/rcar-du/rcar_du_crtc.c |  8 +----
 drivers/gpu/drm/rcar-du/rcar_du_drv.c  | 48 ++++++++++++++++++++++++++
 drivers/gpu/drm/rcar-du/rcar_du_drv.h  |  1 +
 3 files changed, 50 insertions(+), 7 deletions(-)
drivers/gpu/drm/rcar-du/rcar_du_crtc.c +1 -7
index 3619e1ddeb620..f2d3266509cc1 100644
@@ -10,7 +10,6 @@
 #include <linux/clk.h>
 #include <linux/mutex.h>
 #include <linux/platform_device.h>
-#include <linux/sys_soc.h>
 
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
@@ -204,11 +203,6 @@ static void rcar_du_escr_divider(struct clk *clk, unsigned long target,
 	}
 }
 
-static const struct soc_device_attribute rcar_du_r8a7795_es1[] = {
-	{ .soc_id = "r8a7795", .revision = "ES1.*" },
-	{ /* sentinel */ }
-};
-
 static void rcar_du_crtc_set_display_timing(struct rcar_du_crtc *rcrtc)
 {
 	const struct drm_display_mode *mode = &rcrtc->crtc.state->adjusted_mode;
@@ -238,7 +232,7 @@ static void rcar_du_crtc_set_display_timing(struct rcar_du_crtc *rcrtc)
 		 * no post-divider when a display PLL is present (as shown by
 		 * the workaround breaking HDMI output on M3-W during testing).
 		 */
-		if (soc_device_match(rcar_du_r8a7795_es1)) {
+		if (rcdu->info->quirks & RCAR_DU_QUIRK_H3_ES1_PCLK_STABILITY) {
 			target *= 2;
 			div = 1;
 		}
drivers/gpu/drm/rcar-du/rcar_du_drv.c +48 -0

80 lines not on this page · raw message

drivers/gpu/drm/rcar-du/rcar_du_drv.h +1 -0
index 5cfa2bb7ad93d..df87ccab146f4 100644
@@ -34,6 +34,7 @@ struct rcar_du_device;
 #define RCAR_DU_FEATURE_NO_BLENDING	BIT(5)	/* PnMR.SPIM does not have ALP nor EOR bits */
 
 #define RCAR_DU_QUIRK_ALIGN_128B	BIT(0)	/* Align pitches to 128 bytes */
+#define RCAR_DU_QUIRK_H3_ES1_PCLK_STABILITY BIT(1)	/* H3 ES1 has pclk stability issue */
 
 enum rcar_du_output {
 	RCAR_DU_OUTPUT_DPAD0,
Signature · 1 line
2.39.2
author maintainer
#616 lore raw

PATCH 615/885 drm: rcar-du: Fix setting a reserved bit in DPLLCR

In reply to Greg Kroah-Hartman

From: Tomi Valkeinen <[email protected]>

[ Upstream commit 5fbc2f3b91d27e12b614947048764099570cbb55 ]

On H3 ES1.x two bits in DPLLCR are used to select the DU input dot clock source. These are bits 20 and 21 for DU2, and bits 22 and 23 for DU1. On non-ES1.x, only the higher bits are used (bits 21 and 23), and the lower bits are reserved and should be set to 0.

The current code always sets the lower bits, even on non-ES1.x.

For both DU1 and DU2, on all SoC versions, when writing zeroes to those bits the input clock is DCLKIN, and thus there's no difference between ES1.x and non-ES1.x.

For DU1, writing 0b10 to the bits (or only writing the higher bit) results in using PLL0 as the input clock, so in this case there's also no difference between ES1.x and non-ES1.x.

However, for DU2, writing 0b10 to the bits results in using PLL0 as the input clock on ES1.x, whereas on non-ES1.x it results in using PLL1. On ES1.x you need to write 0b11 to select PLL1.

The current code always writes 0b11 to PLCS0 field to select PLL1 on all SoC versions, which works but causes an illegal (in the sense of not allowed by the documentation) write to a reserved bit field.

To remove the illegal bit write on PLSC0 we need to handle the input dot clock selection differently for ES1.x and non-ES1.x.

Add a new quirk, RCAR_DU_QUIRK_H3_ES1_PLL, for this. This way we can always set the bit 21 on PLSC0 when choosing the PLL as the source clock, and additionally set the bit 20 when on ES1.x.

---
 drivers/gpu/drm/rcar-du/rcar_du_crtc.c | 23 ++++++++++++++++++++---
 drivers/gpu/drm/rcar-du/rcar_du_drv.c  |  3 ++-
 drivers/gpu/drm/rcar-du/rcar_du_drv.h  |  1 +
 drivers/gpu/drm/rcar-du/rcar_du_regs.h |  8 ++------
 4 files changed, 25 insertions(+), 10 deletions(-)
drivers/gpu/drm/rcar-du/rcar_du_crtc.c +20 -3
index f2d3266509cc1..b7dd59fe119e6 100644
@@ -245,13 +245,30 @@ static void rcar_du_crtc_set_display_timing(struct rcar_du_crtc *rcrtc)
 		       | DPLLCR_N(dpll.n) | DPLLCR_M(dpll.m)
 		       | DPLLCR_STBY;
 
-		if (rcrtc->index == 1)
+		if (rcrtc->index == 1) {
 			dpllcr |= DPLLCR_PLCS1
 			       |  DPLLCR_INCS_DOTCLKIN1;
-		else
-			dpllcr |= DPLLCR_PLCS0
+		} else {
+			dpllcr |= DPLLCR_PLCS0_PLL
 			       |  DPLLCR_INCS_DOTCLKIN0;
 
+			/*
+			 * On ES2.x we have a single mux controlled via bit 21,
+			 * which selects between DCLKIN source (bit 21 = 0) and
+			 * a PLL source (bit 21 = 1), where the PLL is always
+			 * PLL1.
+			 *
+			 * On ES1.x we have an additional mux, controlled
+			 * via bit 20, for choosing between PLL0 (bit 20 = 0)
+			 * and PLL1 (bit 20 = 1). We always want to use PLL1,
+			 * so on ES1.x, in addition to setting bit 21, we need
+			 * to set the bit 20.
+			 */
+
+			if (rcdu->info->quirks & RCAR_DU_QUIRK_H3_ES1_PLL)
+				dpllcr |= DPLLCR_PLCS0_H3ES1X_PLL1;
+		}
+
 		rcar_du_group_write(rcrtc->group, DPLLCR, dpllcr);
 
 		escr = ESCR_DCLKSEL_DCLKIN | div;
drivers/gpu/drm/rcar-du/rcar_du_drv.c +2 -1
index 0dada0646b2eb..6381578c4db58 100644
@@ -394,7 +394,8 @@ static const struct rcar_du_device_info rcar_du_r8a7795_es1_info = {
 		  | RCAR_DU_FEATURE_VSP1_SOURCE
 		  | RCAR_DU_FEATURE_INTERLACED
 		  | RCAR_DU_FEATURE_TVM_SYNC,
-	.quirks = RCAR_DU_QUIRK_H3_ES1_PCLK_STABILITY,
+	.quirks = RCAR_DU_QUIRK_H3_ES1_PCLK_STABILITY
+		| RCAR_DU_QUIRK_H3_ES1_PLL,
 	.channels_mask = BIT(3) | BIT(2) | BIT(1) | BIT(0),
 	.routes = {
 		/*
drivers/gpu/drm/rcar-du/rcar_du_drv.h +1 -0
index df87ccab146f4..acc3673fefe18 100644
@@ -35,6 +35,7 @@ struct rcar_du_device;
 
 #define RCAR_DU_QUIRK_ALIGN_128B	BIT(0)	/* Align pitches to 128 bytes */
 #define RCAR_DU_QUIRK_H3_ES1_PCLK_STABILITY BIT(1)	/* H3 ES1 has pclk stability issue */
+#define RCAR_DU_QUIRK_H3_ES1_PLL	BIT(2)	/* H3 ES1 PLL setup differs from non-ES1 */
 
 enum rcar_du_output {
 	RCAR_DU_OUTPUT_DPAD0,
drivers/gpu/drm/rcar-du/rcar_du_regs.h +2 -6
index c1bcb0e8b5b4e..789ae9285108e 100644
@@ -283,12 +283,8 @@
 #define DPLLCR			0x20044
 #define DPLLCR_CODE		(0x95 << 24)
 #define DPLLCR_PLCS1		(1 << 23)
-/*
- * PLCS0 is bit 21, but H3 ES1.x requires bit 20 to be set as well. As bit 20
- * isn't implemented by other SoC in the Gen3 family it can safely be set
- * unconditionally.
- */
-#define DPLLCR_PLCS0		(3 << 20)
+#define DPLLCR_PLCS0_PLL	(1 << 21)
+#define DPLLCR_PLCS0_H3ES1X_PLL1	(1 << 20)
 #define DPLLCR_CLKE		(1 << 18)
 #define DPLLCR_FDPLL(n)		((n) << 12)
 #define DPLLCR_N(n)		((n) << 5)
Signature · 1 line
2.39.2
author maintainer
#617 lore raw

PATCH 616/885 drm/drm_print: correct format problem

In reply to Greg Kroah-Hartman

From: Wayne Lin <[email protected]>

[ Upstream commit d987150b539271b0394f24c1c648d2846662adb4 ]

[why & how]
__drm_dbg() parameter set format is wrong and not aligned with the format under CONFIG_DRM_USE_DYNAMIC_DEBUG is on. Fix it.

---
 include/drm/drm_print.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/drm/drm_print.h +1 -1
index a44fb7ef257f6..094ded23534c7 100644
@@ -521,7 +521,7 @@ __printf(1, 2)
 void __drm_err(const char *format, ...);
 
 #if !defined(CONFIG_DRM_USE_DYNAMIC_DEBUG)
-#define __drm_dbg(fmt, ...)		___drm_dbg(NULL, fmt, ##__VA_ARGS__)
+#define __drm_dbg(cat, fmt, ...)		___drm_dbg(NULL, cat, fmt, ##__VA_ARGS__)
 #else
 #define __drm_dbg(cat, fmt, ...)					\
 	_dynamic_func_call_cls(cat, fmt, ___drm_dbg,			\
Signature · 1 line
2.39.2
author maintainer
#618 lore raw

PATCH 617/885 drm/amd/display: Set hvm_enabled flag for S/G mode

In reply to Greg Kroah-Hartman

From: Roman Li <[email protected]>

[ Upstream commit 40e9f3f067bc6fb47b878f8ba0a9cc7b93abbf49 ]

[Why]
After enabling S/G on dcn314 a screen corruption may be observed. HostVM flag should be set in S/G mode to be included in DML calculations.

[How]
In S/G mode gpu_vm_support flag is set.
Use its value to init is_hvm_enabled.

---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c +1 -1
index 25ffc51ae6fa7..050e7a52c8f62 100644
@@ -1240,7 +1240,7 @@ static void mmhub_read_system_context(struct amdgpu_device *adev, struct dc_phy_
 	pa_config->gart_config.page_table_end_addr = page_table_end.quad_part << 12;
 	pa_config->gart_config.page_table_base_addr = page_table_base.quad_part;
 
-	pa_config->is_hvm_enabled = 0;
+	pa_config->is_hvm_enabled = adev->mode_info.gpu_vm_support;
 
 }
 
Signature · 1 line
2.39.2
author maintainer
#619 lore raw

PATCH 618/885 habanalabs: extend fatal messages to contain PCI info

In reply to Greg Kroah-Hartman

From: Moti Haimovski <[email protected]>

[ Upstream commit 2a0a839b6a28f7c4c528bb75b740c7f38ef79a37 ]

This commit attaches the PCI device address to driver fatal messages in order to ease debugging in multi-device setups.

---
 drivers/misc/habanalabs/common/device.c | 38 ++++++++++++++++---------
 1 file changed, 25 insertions(+), 13 deletions(-)
drivers/misc/habanalabs/common/device.c +25 -13
index 233d8b46c831f..e0dca445abf14 100644
@@ -1458,7 +1458,8 @@ int hl_device_reset(struct hl_device *hdev, u32 flags)
 		if (rc == -EBUSY) {
 			if (hdev->device_fini_pending) {
 				dev_crit(hdev->dev,
-					"Failed to kill all open processes, stopping hard reset\n");
+					"%s Failed to kill all open processes, stopping hard reset\n",
+					dev_name(&(hdev)->pdev->dev));
 				goto out_err;
 			}
 
@@ -1468,7 +1469,8 @@ int hl_device_reset(struct hl_device *hdev, u32 flags)
 
 		if (rc) {
 			dev_crit(hdev->dev,
-				"Failed to kill all open processes, stopping hard reset\n");
+				"%s Failed to kill all open processes, stopping hard reset\n",
+				dev_name(&(hdev)->pdev->dev));
 			goto out_err;
 		}
 
@@ -1519,14 +1521,16 @@ int hl_device_reset(struct hl_device *hdev, u32 flags)
 			 * ensure driver puts the driver in a unusable state
 			 */
 			dev_crit(hdev->dev,
-				"Consecutive FW fatal errors received, stopping hard reset\n");
+				"%s Consecutive FW fatal errors received, stopping hard reset\n",
+				dev_name(&(hdev)->pdev->dev));
 			rc = -EIO;
 			goto out_err;
 		}
 
 		if (hdev->kernel_ctx) {
 			dev_crit(hdev->dev,
-				"kernel ctx was alive during hard reset, something is terribly wrong\n");
+				"%s kernel ctx was alive during hard reset, something is terribly wrong\n",
+				dev_name(&(hdev)->pdev->dev));
 			rc = -EBUSY;
 			goto out_err;
 		}
@@ -1645,9 +1649,13 @@ int hl_device_reset(struct hl_device *hdev, u32 flags)
 	hdev->reset_info.needs_reset = false;
 
 	if (hard_reset)
-		dev_info(hdev->dev, "Successfully finished resetting the device\n");
+		dev_info(hdev->dev,
+			 "Successfully finished resetting the %s device\n",
+			 dev_name(&(hdev)->pdev->dev));
 	else
-		dev_dbg(hdev->dev, "Successfully finished resetting the device\n");
+		dev_dbg(hdev->dev,
+			"Successfully finished resetting the %s device\n",
+			dev_name(&(hdev)->pdev->dev));
 
 	if (hard_reset) {
 		hdev->reset_info.hard_reset_cnt++;
@@ -1681,7 +1689,9 @@ int hl_device_reset(struct hl_device *hdev, u32 flags)
 	hdev->reset_info.in_compute_reset = 0;
 
 	if (hard_reset) {
-		dev_err(hdev->dev, "Failed to reset! Device is NOT usable\n");
+		dev_err(hdev->dev,
+			"%s Failed to reset! Device is NOT usable\n",
+			dev_name(&(hdev)->pdev->dev));
 		hdev->reset_info.hard_reset_cnt++;
 	} else if (reset_upon_device_release) {
 		spin_unlock(&hdev->reset_info.lock);
@@ -2004,7 +2014,8 @@ int hl_device_init(struct hl_device *hdev, struct class *hclass)
 	}
 
 	dev_notice(hdev->dev,
-		"Successfully added device to habanalabs driver\n");
+		"Successfully added device %s to habanalabs driver\n",
+		dev_name(&(hdev)->pdev->dev));
 
 	hdev->init_done = true;
 
@@ -2053,11 +2064,11 @@ int hl_device_init(struct hl_device *hdev, struct class *hclass)
 		device_cdev_sysfs_add(hdev);
 	if (hdev->pdev)
 		dev_err(&hdev->pdev->dev,
-			"Failed to initialize hl%d. Device is NOT usable !\n",
-			hdev->cdev_idx);
+			"Failed to initialize hl%d. Device %s is NOT usable !\n",
+			hdev->cdev_idx, dev_name(&(hdev)->pdev->dev));
 	else
-		pr_err("Failed to initialize hl%d. Device is NOT usable !\n",
-			hdev->cdev_idx);
+		pr_err("Failed to initialize hl%d. Device %s is NOT usable !\n",
+			hdev->cdev_idx, dev_name(&(hdev)->pdev->dev));
 
 	return rc;
 }
@@ -2113,7 +2124,8 @@ void hl_device_fini(struct hl_device *hdev)
 
 		if (ktime_compare(ktime_get(), timeout) > 0) {
 			dev_crit(hdev->dev,
-				"Failed to remove device because reset function did not finish\n");
+				"%s Failed to remove device because reset function did not finish\n",
+				dev_name(&(hdev)->pdev->dev));
 			return;
 		}
 	}
Signature · 1 line
2.39.2
author maintainer
#620 lore raw

PATCH 619/885 habanalabs: fix bug in timestamps registration code

In reply to Greg Kroah-Hartman

From: farah kassabri <[email protected]>

[ Upstream commit ac5af9900f82b7034de7c9eb1d70d030ba325607 ]

Protect re-using the same timestamp buffer record before actually adding it to the to interrupt wait list.
Mark ts buff offset as in use in the spinlock protection area of the interrupt wait list to avoid getting in the re-use section in ts_buff_get_kernel_ts_record before adding the node to the list. this scenario might happen when multiple threads are racing on same offset and one thread could set data in the ts buff in ts_buff_get_kernel_ts_record then the other thread takes over and get to ts_buff_get_kernel_ts_record and we will try to re-use the same ts buff offset then we will try to delete a non existing node from the list.

---
 .../habanalabs/common/command_submission.c    | 33 ++++++++++++-------
 1 file changed, 22 insertions(+), 11 deletions(-)
drivers/misc/habanalabs/common/command_submission.c +22 -11
index fa05770865c65..1071bf492e423 100644
@@ -3091,19 +3091,18 @@ static int ts_buff_get_kernel_ts_record(struct hl_mmap_mem_buf *buf,
 			goto start_over;
 		}
 	} else {
+		/* Fill up the new registration node info */
+		requested_offset_record->ts_reg_info.buf = buf;
+		requested_offset_record->ts_reg_info.cq_cb = cq_cb;
+		requested_offset_record->ts_reg_info.timestamp_kernel_addr =
+				(u64 *) ts_buff->user_buff_address + ts_offset;
+		requested_offset_record->cq_kernel_addr =
+				(u64 *) cq_cb->kernel_address + cq_offset;
+		requested_offset_record->cq_target_value = target_value;
+
 		spin_unlock_irqrestore(wait_list_lock, flags);
 	}
 
-	/* Fill up the new registration node info */
-	requested_offset_record->ts_reg_info.in_use = 1;
-	requested_offset_record->ts_reg_info.buf = buf;
-	requested_offset_record->ts_reg_info.cq_cb = cq_cb;
-	requested_offset_record->ts_reg_info.timestamp_kernel_addr =
-			(u64 *) ts_buff->user_buff_address + ts_offset;
-	requested_offset_record->cq_kernel_addr =
-			(u64 *) cq_cb->kernel_address + cq_offset;
-	requested_offset_record->cq_target_value = target_value;
-
 	*pend = requested_offset_record;
 
 	dev_dbg(buf->mmg->dev, "Found available node in TS kernel CB %p\n",
@@ -3151,7 +3150,7 @@ static int _hl_interrupt_wait_ioctl(struct hl_device *hdev, struct hl_ctx *ctx,
 			goto put_cq_cb;
 		}
 
-		/* Find first available record */
+		/* get ts buffer record */
 		rc = ts_buff_get_kernel_ts_record(buf, cq_cb, ts_offset,
 						cq_counters_offset, target_value,
 						&interrupt->wait_list_lock, &pend);
@@ -3199,7 +3198,19 @@ static int _hl_interrupt_wait_ioctl(struct hl_device *hdev, struct hl_ctx *ctx,
 	 * Note that we cannot have sorted list by target value,
 	 * in order to shorten the list pass loop, since
 	 * same list could have nodes for different cq counter handle.
+	 * Note:
+	 * Mark ts buff offset as in use here in the spinlock protection area
+	 * to avoid getting in the re-use section in ts_buff_get_kernel_ts_record
+	 * before adding the node to the list. this scenario might happen when
+	 * multiple threads are racing on same offset and one thread could
+	 * set the ts buff in ts_buff_get_kernel_ts_record then the other thread
+	 * takes over and get to ts_buff_get_kernel_ts_record and then we will try
+	 * to re-use the same ts buff offset, and will try to delete a non existing
+	 * node from the list.
 	 */
+	if (register_ts_record)
+		pend->ts_reg_info.in_use = 1;
+
 	list_add_tail(&pend->wait_list_node, &interrupt->wait_list_head);
 	spin_unlock_irqrestore(&interrupt->wait_list_lock, flags);
 
Signature · 1 line
2.39.2
author maintainer
#621 lore raw

PATCH 620/885 docs/scripts/gdb: add necessary make scripts_gdb step

In reply to Greg Kroah-Hartman

From: Jakob Koschel <[email protected]>

[ Upstream commit 6b219431037bf98c9efd49716aea9b68440477a3 ]

In order to debug the kernel successfully with gdb you need to run 'make scripts_gdb' nowadays.

This was changed with the following commit:

Commit 67274c083438340ad16c ("scripts/gdb: delay generation of gdb constants.py")

In order to have a complete guide for beginners this remark should be added to the offial documentation.

---
 Documentation/dev-tools/gdb-kernel-debugging.rst | 4 ++++
 1 file changed, 4 insertions(+)
Documentation/dev-tools/gdb-kernel-debugging.rst +4 -0
index 8e0f1fe8d17ad..895285c037c72 100644
@@ -39,6 +39,10 @@ Setup
   this mode. In this case, you should build the kernel with
   CONFIG_RANDOMIZE_BASE disabled if the architecture supports KASLR.
 
+- Build the gdb scripts (required on kernels v5.1 and above)::
+
+    make scripts_gdb
+
 - Enable the gdb stub of QEMU/KVM, either
 
     - at VM startup time by appending "-s" to the QEMU command line
Signature · 1 line
2.39.2
author maintainer
#622 lore raw

PATCH 621/885 drm/msm/dpu: Add DSC hardware blocks to register snapshot

In reply to Greg Kroah-Hartman

From: Marijn Suijten <[email protected]>

[ Upstream commit a7efe60e36b9c0e966d7f82ac90a89b591d984e9 ]

Add missing DSC hardware block register ranges to the snapshot utility to include them in dmesg (on MSM_DISP_SNAPSHOT_DUMP_IN_CONSOLE) and the kms debugfs file.

Patchwork: https://patchwork.freedesktop.org/patch/520175/

---
 drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c | 5 +++++
 1 file changed, 5 insertions(+)
drivers/gpu/drm/msm/disp/dpu1/dpu_kms.c +5 -0
index 5e6e2626151e8..b7901b666612a 100644
@@ -942,6 +942,11 @@ static void dpu_kms_mdp_snapshot(struct msm_disp_state *disp_state, struct msm_k
 	msm_disp_snapshot_add_block(disp_state, cat->mdp[0].len,
 			dpu_kms->mmio + cat->mdp[0].base, "top");
 
+	/* dump DSC sub-blocks HW regs info */
+	for (i = 0; i < cat->dsc_count; i++)
+		msm_disp_snapshot_add_block(disp_state, cat->dsc[i].len,
+				dpu_kms->mmio + cat->dsc[i].base, "dsc_%d", i);
+
 	pm_runtime_put_sync(&dpu_kms->pdev->dev);
 }
 
Signature · 1 line
2.39.2
author maintainer
#623 lore raw

PATCH 622/885 ASoC: soc-compress: Reposition and add pcm_mutex

In reply to Greg Kroah-Hartman

From: 강신형 <[email protected]>

[ Upstream commit aa9ff6a4955fdba02b54fbc4386db876603703b7 ]

If panic_on_warn is set and compress stream(DPCM) is started, then kernel panic occurred because card->pcm_mutex isn't held appropriately. In the following functions, warning were issued at this line "snd_soc_dpcm_mutex_assert_held".

static int dpcm_be_connect(struct snd_soc_pcm_runtime *fe,
struct snd_soc_pcm_runtime *be, int stream)

{
	...
	snd_soc_dpcm_mutex_assert_held(fe);
	...
}

void dpcm_be_disconnect(struct snd_soc_pcm_runtime *fe, int stream)

{
	...
	snd_soc_dpcm_mutex_assert_held(fe);
	...
}

void snd_soc_runtime_action(struct snd_soc_pcm_runtime *rtd,
int stream, int action)

{
	...
	snd_soc_dpcm_mutex_assert_held(rtd);
	...
}

int dpcm_dapm_stream_event(struct snd_soc_pcm_runtime *fe, int dir,
int event)

{
	...
	snd_soc_dpcm_mutex_assert_held(fe);
	...
}

These functions are called by soc_compr_set_params_fe, soc_compr_open_fe and soc_compr_free_fe
without pcm_mutex locking. And this is call stack.

[  414.527841][ T2179] pc : dpcm_process_paths+0x5a4/0x750
[  414.527848][ T2179] lr : dpcm_process_paths+0x37c/0x750
[  414.527945][ T2179] Call trace:
[  414.527949][ T2179]  dpcm_process_paths+0x5a4/0x750
[  414.527955][ T2179]  soc_compr_open_fe+0xb0/0x2cc
[  414.527972][ T2179]  snd_compr_open+0x180/0x248
[  414.527981][ T2179]  snd_open+0x15c/0x194
[  414.528003][ T2179]  chrdev_open+0x1b0/0x220
[  414.528023][ T2179]  do_dentry_open+0x30c/0x594
[  414.528045][ T2179]  vfs_open+0x34/0x44
[  414.528053][ T2179]  path_openat+0x914/0xb08
[  414.528062][ T2179]  do_filp_open+0xc0/0x170
[  414.528068][ T2179]  do_sys_openat2+0x94/0x18c
[  414.528076][ T2179]  __arm64_sys_openat+0x78/0xa4
[  414.528084][ T2179]  invoke_syscall+0x48/0x10c
[  414.528094][ T2179]  el0_svc_common+0xbc/0x104
[  414.528099][ T2179]  do_el0_svc+0x34/0xd8
[  414.528103][ T2179]  el0_svc+0x34/0xc4
[  414.528125][ T2179]  el0t_64_sync_handler+0x8c/0xfc
[  414.528133][ T2179]  el0t_64_sync+0x1a0/0x1a4
[  414.528142][ T2179] Kernel panic - not syncing: panic_on_warn set ...

So, I reposition and add pcm_mutex to resolve lockdep error.

---
 sound/soc/soc-compress.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)
sound/soc/soc-compress.c +6 -3
index cb0ed2fea893a..e7aa6f360cabe 100644
@@ -149,6 +149,8 @@ static int soc_compr_open_fe(struct snd_compr_stream *cstream)
 	if (ret < 0)
 		goto be_err;
 
+	mutex_lock_nested(&fe->card->pcm_mutex, fe->card->pcm_subclass);
+
 	/* calculate valid and active FE <-> BE dpcms */
 	dpcm_process_paths(fe, stream, &list, 1);
 	fe->dpcm[stream].runtime = fe_substream->runtime;
@@ -184,7 +186,6 @@ static int soc_compr_open_fe(struct snd_compr_stream *cstream)
 	fe->dpcm[stream].state = SND_SOC_DPCM_STATE_OPEN;
 	fe->dpcm[stream].runtime_update = SND_SOC_DPCM_UPDATE_NO;
 
-	mutex_lock_nested(&fe->card->pcm_mutex, fe->card->pcm_subclass);
 	snd_soc_runtime_activate(fe, stream);
 	mutex_unlock(&fe->card->pcm_mutex);
 
@@ -215,7 +216,6 @@ static int soc_compr_free_fe(struct snd_compr_stream *cstream)
 
 	mutex_lock_nested(&fe->card->pcm_mutex, fe->card->pcm_subclass);
 	snd_soc_runtime_deactivate(fe, stream);
-	mutex_unlock(&fe->card->pcm_mutex);
 
 	fe->dpcm[stream].runtime_update = SND_SOC_DPCM_UPDATE_FE;
 
@@ -234,6 +234,8 @@ static int soc_compr_free_fe(struct snd_compr_stream *cstream)
 
 	dpcm_be_disconnect(fe, stream);
 
+	mutex_unlock(&fe->card->pcm_mutex);
+
 	fe->dpcm[stream].runtime = NULL;
 
 	snd_soc_link_compr_shutdown(cstream, 0);
@@ -409,8 +411,9 @@ static int soc_compr_set_params_fe(struct snd_compr_stream *cstream,
 	ret = snd_soc_link_compr_set_params(cstream);
 	if (ret < 0)
 		goto out;
-
+	mutex_lock_nested(&fe->card->pcm_mutex, fe->card->pcm_subclass);
 	dpcm_dapm_stream_event(fe, stream, SND_SOC_DAPM_STREAM_START);
+	mutex_unlock(&fe->card->pcm_mutex);
 	fe->dpcm[stream].state = SND_SOC_DPCM_STATE_PREPARE;
 
 out:
Signature · 1 line
2.39.2
author maintainer
#624 lore raw

PATCH 623/885 ASoC: kirkwood: Iterate over array indexes instead of using pointer math

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

[ Upstream commit b3bcedc0402fcdc5c8624c433562d9d1882749d8 ]

Walking the dram->cs array was seen as accesses beyond the first array item by the compiler. Instead, use the array index directly. This allows for run-time bounds checking under CONFIG_UBSAN_BOUNDS as well. Seen with GCC 13 with -fstrict-flex-arrays:

../sound/soc/kirkwood/kirkwood-dma.c: In function
'kirkwood_dma_conf_mbus_windows.constprop':
../sound/soc/kirkwood/kirkwood-dma.c:90:24: warning: array subscript 0 is outside array bounds of 'const struct mbus_dram_window[0]' [-Warray-bounds=]

   90 |                 if ((cs->base & 0xffff0000) < (dma & 0xffff0000)) {
      |                      ~~^~~~~~
---
 sound/soc/kirkwood/kirkwood-dma.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
sound/soc/kirkwood/kirkwood-dma.c +1 -1
index 700a18561a940..640cebd2983e2 100644
@@ -86,7 +86,7 @@ kirkwood_dma_conf_mbus_windows(void __iomem *base, int win,
 
 	/* try to find matching cs for current dma address */
 	for (i = 0; i < dram->num_cs; i++) {
-		const struct mbus_dram_window *cs = dram->cs + i;
+		const struct mbus_dram_window *cs = &dram->cs[i];
 		if ((cs->base & 0xffff0000) < (dma & 0xffff0000)) {
 			writel(cs->base & 0xffff0000,
 				base + KIRKWOOD_AUDIO_WIN_BASE_REG(win));
Signature · 1 line
2.39.2
author maintainer
#625 lore raw

PATCH 624/885 regulator: max77802: Bounds check regulator id against opmode

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

[ Upstream commit 4fd8bcec5fd7c0d586206fa2f42bd67b06cdaa7e ]

Explicitly bounds-check the id before accessing the opmode array. Seen with GCC 13:

../drivers/regulator/max77802-regulator.c: In function 'max77802_enable':
../drivers/regulator/max77802-regulator.c:217:29: warning: array subscript [0, 41] is outside array bounds of 'unsigned int[42]' [-Warray-bounds=]
217 | if (max77802->opmode[id] == MAX77802_OFF_PWRREQ)

      |             ~~~~~~~~~~~~~~~~^~~~

../drivers/regulator/max77802-regulator.c:62:22: note: while referencing 'opmode'
62 | unsigned int opmode[MAX77802_REG_MAX];
| ^~~~~~

---
 drivers/regulator/max77802-regulator.c | 34 ++++++++++++++++++--------
 1 file changed, 24 insertions(+), 10 deletions(-)
drivers/regulator/max77802-regulator.c +24 -10
index 21e0eb0f43f94..befe5f319819b 100644
@@ -94,9 +94,11 @@ static int max77802_set_suspend_disable(struct regulator_dev *rdev)
 {
 	unsigned int val = MAX77802_OFF_PWRREQ;
 	struct max77802_regulator_prv *max77802 = rdev_get_drvdata(rdev);
-	int id = rdev_get_id(rdev);
+	unsigned int id = rdev_get_id(rdev);
 	int shift = max77802_get_opmode_shift(id);
 
+	if (WARN_ON_ONCE(id >= ARRAY_SIZE(max77802->opmode)))
+		return -EINVAL;
 	max77802->opmode[id] = val;
 	return regmap_update_bits(rdev->regmap, rdev->desc->enable_reg,
 				  rdev->desc->enable_mask, val << shift);
@@ -110,7 +112,7 @@ static int max77802_set_suspend_disable(struct regulator_dev *rdev)
 static int max77802_set_mode(struct regulator_dev *rdev, unsigned int mode)
 {
 	struct max77802_regulator_prv *max77802 = rdev_get_drvdata(rdev);
-	int id = rdev_get_id(rdev);
+	unsigned int id = rdev_get_id(rdev);
 	unsigned int val;
 	int shift = max77802_get_opmode_shift(id);
 
@@ -127,6 +129,9 @@ static int max77802_set_mode(struct regulator_dev *rdev, unsigned int mode)
 		return -EINVAL;
 	}
 
+	if (WARN_ON_ONCE(id >= ARRAY_SIZE(max77802->opmode)))
+		return -EINVAL;
+
 	max77802->opmode[id] = val;
 	return regmap_update_bits(rdev->regmap, rdev->desc->enable_reg,
 				  rdev->desc->enable_mask, val << shift);
@@ -135,8 +140,10 @@ static int max77802_set_mode(struct regulator_dev *rdev, unsigned int mode)
 static unsigned max77802_get_mode(struct regulator_dev *rdev)
 {
 	struct max77802_regulator_prv *max77802 = rdev_get_drvdata(rdev);
-	int id = rdev_get_id(rdev);
+	unsigned int id = rdev_get_id(rdev);
 
+	if (WARN_ON_ONCE(id >= ARRAY_SIZE(max77802->opmode)))
+		return -EINVAL;
 	return max77802_map_mode(max77802->opmode[id]);
 }
 
@@ -160,10 +167,13 @@ static int max77802_set_suspend_mode(struct regulator_dev *rdev,
 				     unsigned int mode)
 {
 	struct max77802_regulator_prv *max77802 = rdev_get_drvdata(rdev);
-	int id = rdev_get_id(rdev);
+	unsigned int id = rdev_get_id(rdev);
 	unsigned int val;
 	int shift = max77802_get_opmode_shift(id);
 
+	if (WARN_ON_ONCE(id >= ARRAY_SIZE(max77802->opmode)))
+		return -EINVAL;
+
 	/*
 	 * If the regulator has been disabled for suspend
 	 * then is invalid to try setting a suspend mode.
@@ -209,9 +219,11 @@ static int max77802_set_suspend_mode(struct regulator_dev *rdev,
 static int max77802_enable(struct regulator_dev *rdev)
 {
 	struct max77802_regulator_prv *max77802 = rdev_get_drvdata(rdev);
-	int id = rdev_get_id(rdev);
+	unsigned int id = rdev_get_id(rdev);
 	int shift = max77802_get_opmode_shift(id);
 
+	if (WARN_ON_ONCE(id >= ARRAY_SIZE(max77802->opmode)))
+		return -EINVAL;
 	if (max77802->opmode[id] == MAX77802_OFF_PWRREQ)
 		max77802->opmode[id] = MAX77802_OPMODE_NORMAL;
 
@@ -495,7 +507,7 @@ static int max77802_pmic_probe(struct platform_device *pdev)
 
 	for (i = 0; i < MAX77802_REG_MAX; i++) {
 		struct regulator_dev *rdev;
-		int id = regulators[i].id;
+		unsigned int id = regulators[i].id;
 		int shift = max77802_get_opmode_shift(id);
 		int ret;
 
@@ -513,10 +525,12 @@ static int max77802_pmic_probe(struct platform_device *pdev)
 		 * the hardware reports OFF as the regulator operating mode.
 		 * Default to operating mode NORMAL in that case.
 		 */
-		if (val == MAX77802_STATUS_OFF)
-			max77802->opmode[id] = MAX77802_OPMODE_NORMAL;
-		else
-			max77802->opmode[id] = val;
+		if (id < ARRAY_SIZE(max77802->opmode)) {
+			if (val == MAX77802_STATUS_OFF)
+				max77802->opmode[id] = MAX77802_OPMODE_NORMAL;
+			else
+				max77802->opmode[id] = val;
+		}
 
 		rdev = devm_regulator_register(&pdev->dev,
 					       &regulators[i], &config);
Signature · 1 line
2.39.2
author maintainer
#626 lore raw

PATCH 625/885 regulator: s5m8767: Bounds check id indexing into arrays

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

[ Upstream commit e314e15a0b58f9d051c00b25951073bcdae61953 ]

The compiler has no way to know if "id" is within the array bounds of the regulators array. Add a check for this and a build-time check that the regulators and reg_voltage_map arrays are sized the same. Seen with GCC 13:

../drivers/regulator/s5m8767.c: In function 's5m8767_pmic_probe':
../drivers/regulator/s5m8767.c:936:35: warning: array subscript [0, 36] is outside array bounds of 'struct regulator_desc[37]' [-Warray-bounds=]

  936 |                         regulators[id].vsel_reg =
      |                         ~~~~~~~~~~^~~~
---
 drivers/regulator/s5m8767.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/regulator/s5m8767.c +5 -1
index 35269f9982105..754c6fcc6e642 100644
@@ -923,10 +923,14 @@ static int s5m8767_pmic_probe(struct platform_device *pdev)
 
 	for (i = 0; i < pdata->num_regulators; i++) {
 		const struct sec_voltage_desc *desc;
-		int id = pdata->regulators[i].id;
+		unsigned int id = pdata->regulators[i].id;
 		int enable_reg, enable_val;
 		struct regulator_dev *rdev;
 
+		BUILD_BUG_ON(ARRAY_SIZE(regulators) != ARRAY_SIZE(reg_voltage_map));
+		if (WARN_ON_ONCE(id >= ARRAY_SIZE(regulators)))
+			continue;
+
 		desc = reg_voltage_map[id];
 		if (desc) {
 			regulators[id].n_voltages =
Signature · 1 line
2.39.2
author maintainer
#627 lore raw

PATCH 626/885 Revert "drm/amdgpu: TA unload messages are not actually sent to psp when amdgpu is uninstalled"

In reply to Greg Kroah-Hartman

From: Vitaly Prosyak <[email protected]>

[ Upstream commit 39934d3ed5725c5e3570ed1b67f612f1ea60ce03 ]

This reverts commit fac53471d0ea9693d314aa2df08d62b2e7e3a0f8. The following change: move the drm_dev_unplug call after amdgpu_driver_unload_kms in amdgpu_pci_remove. The reason is the following: amdgpu_pci_remove calls drm_dev_unregister and it should be called first to ensure userspace can't access the device instance anymore. If we call drm_dev_unplug after amdgpu_driver_unload_kms then we observe IGT PCI software unplug test failure (kernel hung) for all ASICs. This is how this regression was found.

After this revert, the following commands do work not, but it would be fixed in the next commit:
- sudo modprobe -r amdgpu
- sudo modprobe amdgpu

Reviewed-by Alex Deucher <[email protected]>

---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c | 3 ++-
 drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c    | 4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)
drivers/gpu/drm/amd/amdgpu/amdgpu_device.c +2 -1
index a21b3f66fd708..824b0b356b3ce 100644
@@ -4012,7 +4012,8 @@ void amdgpu_device_fini_hw(struct amdgpu_device *adev)
 
 	amdgpu_gart_dummy_page_fini(adev);
 
-	amdgpu_device_unmap_mmio(adev);
+	if (drm_dev_is_unplugged(adev_to_drm(adev)))
+		amdgpu_device_unmap_mmio(adev);
 
 }
 
drivers/gpu/drm/amd/amdgpu/amdgpu_drv.c +2 -2
index 2e5d78b6635c4..dfbeef2c4a9e2 100644
@@ -2226,6 +2226,8 @@ amdgpu_pci_remove(struct pci_dev *pdev)
 	struct drm_device *dev = pci_get_drvdata(pdev);
 	struct amdgpu_device *adev = drm_to_adev(dev);
 
+	drm_dev_unplug(dev);
+
 	if (adev->pm.rpm_mode != AMDGPU_RUNPM_NONE) {
 		pm_runtime_get_sync(dev->dev);
 		pm_runtime_forbid(dev->dev);
@@ -2265,8 +2267,6 @@ amdgpu_pci_remove(struct pci_dev *pdev)
 
 	amdgpu_driver_unload_kms(dev);
 
-	drm_dev_unplug(dev);
-
 	/*
 	 * Flush any in flight DMA operations from device.
 	 * Clear the Bus Master Enable bit and then wait on the PCIe Device
Signature · 1 line
2.39.2
author maintainer
#628 lore raw

PATCH 627/885 drm/amd/display: fix FCLK pstate change underflow

In reply to Greg Kroah-Hartman

From: Vladimir Stempen <[email protected]>

[ Upstream commit 972243f973eb0821084e5833d5f7f4ed025f42da ]

[Why]
Currently we set FCLK p-state change
watermark calculated based on dummy
p-state latency when UCLK p-state is
not supported

[How]
Calculate FCLK p-state change watermark
based on on FCLK pstate change latency
in case UCLK p-state is not supported

---
 drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/gpu/drm/amd/display/dc/dml/dcn32/dcn32_fpu.c +4 -0
index d90216d2fe3a8..04cc96e700981 100644
@@ -1963,6 +1963,10 @@ void dcn32_calculate_wm_and_dlg_fpu(struct dc *dc, struct dc_state *context,
 		 */
 		context->bw_ctx.bw.dcn.watermarks.a = context->bw_ctx.bw.dcn.watermarks.c;
 		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.pstate_change_ns = 0;
+		/* Calculate FCLK p-state change watermark based on FCLK pstate change latency in case
+		 * UCLK p-state is not supported, to avoid underflow in case FCLK pstate is supported
+		 */
+		context->bw_ctx.bw.dcn.watermarks.a.cstate_pstate.fclk_pstate_change_ns = get_fclk_watermark(&context->bw_ctx.dml, pipes, pipe_cnt) * 1000;
 	} else {
 		/* Set A:
 		 * All clocks min.
Signature · 1 line
2.39.2
author maintainer
#629 lore raw

PATCH 628/885 gfs2: Improve gfs2_make_fs_rw error handling

In reply to Greg Kroah-Hartman

From: Andreas Gruenbacher <[email protected]>

[ Upstream commit b66f723bb552ad59c2acb5d45ea45c890f84498b ]

In gfs2_make_fs_rw(), make sure to call gfs2_consist() to report an inconsistency and mark the filesystem as withdrawn when gfs2_find_jhead() fails.

At the end of gfs2_make_fs_rw(), when we discover that the filesystem has been withdrawn, make sure we report an error. This also replaces the gfs2_withdrawn() check after gfs2_find_jhead().

---
 fs/gfs2/super.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
fs/gfs2/super.c +6 -2
index 011f9e7660ef8..2015bd05cba10 100644
@@ -138,8 +138,10 @@ int gfs2_make_fs_rw(struct gfs2_sbd *sdp)
 		return -EIO;
 
 	error = gfs2_find_jhead(sdp->sd_jdesc, &head, false);
-	if (error || gfs2_withdrawn(sdp))
+	if (error) {
+		gfs2_consist(sdp);
 		return error;
+	}
 
 	if (!(head.lh_flags & GFS2_LOG_HEAD_UNMOUNT)) {
 		gfs2_consist(sdp);
@@ -151,7 +153,9 @@ int gfs2_make_fs_rw(struct gfs2_sbd *sdp)
 	gfs2_log_pointers_init(sdp, head.lh_blkno);
 
 	error = gfs2_quota_init(sdp);
-	if (!error && !gfs2_withdrawn(sdp))
+	if (!error && gfs2_withdrawn(sdp))
+		error = -EIO;
+	if (!error)
 		set_bit(SDF_JOURNAL_LIVE, &sdp->sd_flags);
 	return error;
 }
Signature · 1 line
2.39.2
author maintainer
#630 lore raw

PATCH 629/885 hwmon: (coretemp) Simplify platform device handling

In reply to Greg Kroah-Hartman

From: Robin Murphy <[email protected]>

[ Upstream commit 6d03bbff456befeccdd4d663177c4d6c75d0c4ff ]

Coretemp's platform driver is unconventional. All the real work is done globally by the initcall and CPU hotplug notifiers, while the "driver" effectively just wraps an allocation and the registration of the hwmon interface in a long-winded round-trip through the driver core. The whole logic of dynamically creating and destroying platform devices to bring the interfaces up and down is error prone, since it assumes platform_device_add() will synchronously bind the driver and set drvdata before it returns, thus results in a NULL dereference if drivers_autoprobe is turned off for the platform bus. Furthermore, the unusual approach of doing that from within a CPU hotplug notifier, already commented in the code that it deadlocks suspend, also causes lockdep issues for other drivers or subsystems which may want to legitimately register a CPU hotplug notifier from a platform bus notifier.

All of these issues can be solved by ripping this unusual behaviour out completely, simply tying the platform devices to the lifetime of the module itself, and directly managing the hwmon interfaces from the hotplug notifiers. There is a slight user-visible change in that /sys/bus/platform/drivers/coretemp will no longer appear, and /sys/devices/platform/coretemp.n will remain present if package n is hotplugged off, but hwmon users should really only be looking for the presence of the hwmon interfaces, whose behaviour remains unchanged.

---
 drivers/hwmon/coretemp.c | 128 ++++++++++++++++++---------------------
 1 file changed, 58 insertions(+), 70 deletions(-)
drivers/hwmon/coretemp.c +58 -70
index 9bee4d33fbdf0..baaf8af4cb443 100644
@@ -550,66 +550,49 @@ static void coretemp_remove_core(struct platform_data *pdata, int indx)
 		ida_free(&pdata->ida, indx - BASE_SYSFS_ATTR_NO);
 }
 
-static int coretemp_probe(struct platform_device *pdev)
+static int coretemp_device_add(int zoneid)
 {
-	struct device *dev = &pdev->dev;
+	struct platform_device *pdev;
 	struct platform_data *pdata;
+	int err;
 
 	/* Initialize the per-zone data structures */
-	pdata = devm_kzalloc(dev, sizeof(struct platform_data), GFP_KERNEL);
+	pdata = kzalloc(sizeof(*pdata), GFP_KERNEL);
 	if (!pdata)
 		return -ENOMEM;
 
-	pdata->pkg_id = pdev->id;
+	pdata->pkg_id = zoneid;
 	ida_init(&pdata->ida);
-	platform_set_drvdata(pdev, pdata);
 
-	pdata->hwmon_dev = devm_hwmon_device_register_with_groups(dev, DRVNAME,
-								  pdata, NULL);
-	return PTR_ERR_OR_ZERO(pdata->hwmon_dev);
-}
-
-static int coretemp_remove(struct platform_device *pdev)
-{
-	struct platform_data *pdata = platform_get_drvdata(pdev);
-	int i;
+	pdev = platform_device_alloc(DRVNAME, zoneid);
+	if (!pdev) {
+		err = -ENOMEM;
+		goto err_free_pdata;
+	}
 
-	for (i = MAX_CORE_DATA - 1; i >= 0; --i)
-		if (pdata->core_data[i])
-			coretemp_remove_core(pdata, i);
+	err = platform_device_add(pdev);
+	if (err)
+		goto err_put_dev;
 
-	ida_destroy(&pdata->ida);
+	platform_set_drvdata(pdev, pdata);
+	zone_devices[zoneid] = pdev;
 	return 0;
-}
 
-static struct platform_driver coretemp_driver = {
-	.driver = {
-		.name = DRVNAME,
-	},
-	.probe = coretemp_probe,
-	.remove = coretemp_remove,
-};
+err_put_dev:
+	platform_device_put(pdev);
+err_free_pdata:
+	kfree(pdata);
+	return err;
+}
 
-static struct platform_device *coretemp_device_add(unsigned int cpu)
+static void coretemp_device_remove(int zoneid)
 {
-	int err, zoneid = topology_logical_die_id(cpu);
-	struct platform_device *pdev;
-
-	if (zoneid < 0)
-		return ERR_PTR(-ENOMEM);
-
-	pdev = platform_device_alloc(DRVNAME, zoneid);
-	if (!pdev)
-		return ERR_PTR(-ENOMEM);
-
-	err = platform_device_add(pdev);
-	if (err) {
-		platform_device_put(pdev);
-		return ERR_PTR(err);
-	}
+	struct platform_device *pdev = zone_devices[zoneid];
+	struct platform_data *pdata = platform_get_drvdata(pdev);
 
-	zone_devices[zoneid] = pdev;
-	return pdev;
+	ida_destroy(&pdata->ida);
+	kfree(pdata);
+	platform_device_unregister(pdev);
 }
 
 static int coretemp_cpu_online(unsigned int cpu)
@@ -633,7 +616,10 @@ static int coretemp_cpu_online(unsigned int cpu)
 	if (!cpu_has(c, X86_FEATURE_DTHERM))
 		return -ENODEV;
 
-	if (!pdev) {
+	pdata = platform_get_drvdata(pdev);
+	if (!pdata->hwmon_dev) {
+		struct device *hwmon;
+
 		/* Check the microcode version of the CPU */
 		if (chk_ucode_version(cpu))
 			return -EINVAL;
@@ -644,9 +630,11 @@ static int coretemp_cpu_online(unsigned int cpu)
 		 * online. So, initialize per-pkg data structures and
 		 * then bring this core online.
 		 */
-		pdev = coretemp_device_add(cpu);
-		if (IS_ERR(pdev))
-			return PTR_ERR(pdev);
+		hwmon = hwmon_device_register_with_groups(&pdev->dev, DRVNAME,
+							  pdata, NULL);
+		if (IS_ERR(hwmon))
+			return PTR_ERR(hwmon);
+		pdata->hwmon_dev = hwmon;
 
 		/*
 		 * Check whether pkgtemp support is available.
@@ -656,7 +644,6 @@ static int coretemp_cpu_online(unsigned int cpu)
 			coretemp_add_core(pdev, cpu, 1);
 	}
 
-	pdata = platform_get_drvdata(pdev);
 	/*
 	 * Check whether a thread sibling is already online. If not add the
 	 * interface for this CPU core.
@@ -675,18 +662,14 @@ static int coretemp_cpu_offline(unsigned int cpu)
 	struct temp_data *tdata;
 	int i, indx = -1, target;
 
-	/*
-	 * Don't execute this on suspend as the device remove locks
-	 * up the machine.
-	 */
+	/* No need to tear down any interfaces for suspend */
 	if (cpuhp_tasks_frozen)
 		return 0;
 
 	/* If the physical CPU device does not exist, just return */
-	if (!pdev)
-		return 0;
-
 	pd = platform_get_drvdata(pdev);
+	if (!pd->hwmon_dev)
+		return 0;
 
 	for (i = 0; i < NUM_REAL_CORES; i++) {
 		if (pd->cpu_map[i] == topology_core_id(cpu)) {
@@ -718,13 +701,14 @@ static int coretemp_cpu_offline(unsigned int cpu)
 	}
 
 	/*
-	 * If all cores in this pkg are offline, remove the device. This
-	 * will invoke the platform driver remove function, which cleans up
-	 * the rest.
+	 * If all cores in this pkg are offline, remove the interface.
 	 */
+	tdata = pd->core_data[PKG_SYSFS_ATTR_NO];
 	if (cpumask_empty(&pd->cpumask)) {
-		zone_devices[topology_logical_die_id(cpu)] = NULL;
-		platform_device_unregister(pdev);
+		if (tdata)
+			coretemp_remove_core(pd, PKG_SYSFS_ATTR_NO);
+		hwmon_device_unregister(pd->hwmon_dev);
+		pd->hwmon_dev = NULL;
 		return 0;
 	}
 
@@ -732,7 +716,6 @@ static int coretemp_cpu_offline(unsigned int cpu)
 	 * Check whether this core is the target for the package
 	 * interface. We need to assign it to some other cpu.
 	 */
-	tdata = pd->core_data[PKG_SYSFS_ATTR_NO];
 	if (tdata && tdata->cpu == cpu) {
 		target = cpumask_first(&pd->cpumask);
 		mutex_lock(&tdata->update_lock);
@@ -751,7 +734,7 @@ static enum cpuhp_state coretemp_hp_online;
 
 static int __init coretemp_init(void)
 {
-	int err;
+	int i, err;
 
 	/*
 	 * CPUID.06H.EAX[0] indicates whether the CPU has thermal
@@ -767,20 +750,22 @@ static int __init coretemp_init(void)
 	if (!zone_devices)
 		return -ENOMEM;
 
-	err = platform_driver_register(&coretemp_driver);
-	if (err)
-		goto outzone;
+	for (i = 0; i < max_zones; i++) {
+		err = coretemp_device_add(i);
+		if (err)
+			goto outzone;
+	}
 
 	err = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "hwmon/coretemp:online",
 				coretemp_cpu_online, coretemp_cpu_offline);
 	if (err < 0)
-		goto outdrv;
+		goto outzone;
 	coretemp_hp_online = err;
 	return 0;
 
-outdrv:
-	platform_driver_unregister(&coretemp_driver);
 outzone:
+	while (i--)
+		coretemp_device_remove(i);
 	kfree(zone_devices);
 	return err;
 }
@@ -788,8 +773,11 @@ module_init(coretemp_init)
 
 static void __exit coretemp_exit(void)
 {
+	int i;
+
 	cpuhp_remove_state(coretemp_hp_online);
-	platform_driver_unregister(&coretemp_driver);
+	for (i = 0; i < max_zones; i++)
+		coretemp_device_remove(i);
 	kfree(zone_devices);
 }
 module_exit(coretemp_exit)
Signature · 1 line
2.39.2
author maintainer
#631 lore raw

PATCH 630/885 hwmon: (nct6775) Directly call ASUS ACPI WMI method

In reply to Greg Kroah-Hartman

From: Denis Pauk <[email protected]>

[ Upstream commit c3b3747d02f571da2543e719066a50dd966989d8 ]

New ASUS B650/B660/X670 boards firmware have not exposed WMI monitoring GUID and entrypoint method WMBD could be implemented for different device UID.

Implement the direct call to entrypoint method for monitoring the device UID of B550/X570 boards.

---
 drivers/hwmon/Kconfig            |  2 +-
 drivers/hwmon/nct6775-platform.c | 98 ++++++++++++++++++++++----------
 2 files changed, 70 insertions(+), 30 deletions(-)
drivers/hwmon/Kconfig +1 -1
index d3bccc8176c51..a5143d01b95f8 100644
@@ -1508,7 +1508,7 @@ config SENSORS_NCT6775_CORE
 config SENSORS_NCT6775
 	tristate "Platform driver for Nuvoton NCT6775F and compatibles"
 	depends on !PPC
-	depends on ACPI_WMI || ACPI_WMI=n
+	depends on ACPI || ACPI=n
 	select HWMON_VID
 	select SENSORS_NCT6775_CORE
 	help
drivers/hwmon/nct6775-platform.c +69 -29

144 lines not on this page · raw message

Signature · 1 line
2.39.2
author maintainer
#632 lore raw

PATCH 631/885 hwmon: (nct6775) B650/B660/X670 ASUS boards support

In reply to Greg Kroah-Hartman

From: Denis Pauk <[email protected]>

[ Upstream commit e2e09989ccc21ad428d6393450add78584b143bd ]

Boards such as:
"EX-B660M-V5 PRO D4",
"PRIME B650-PLUS",
"PRIME B650M-A",
"PRIME B650M-A AX",
"PRIME B650M-A II",
"PRIME B650M-A WIFI",
"PRIME B650M-A WIFI II",
"PRIME B660M-A D4",
"PRIME B660M-A WIFI D4",
"PRIME X670-P",
"PRIME X670-P WIFI",
"PRIME X670E-PRO WIFI",
"Pro B660M-C-D4",
"ProArt B660-CREATOR D4",
"ProArt X670E-CREATOR WIFI",
"ROG CROSSHAIR X670E EXTREME",
"ROG CROSSHAIR X670E GENE",
"ROG CROSSHAIR X670E HERO",
"ROG MAXIMUS XIII EXTREME GLACIAL",
"ROG MAXIMUS Z690 EXTREME",
"ROG MAXIMUS Z690 EXTREME GLACIAL",
"ROG STRIX B650-A GAMING WIFI",
"ROG STRIX B650E-E GAMING WIFI",
"ROG STRIX B650E-F GAMING WIFI",
"ROG STRIX B650E-I GAMING WIFI",
"ROG STRIX B660-A GAMING WIFI D4",
"ROG STRIX B660-F GAMING WIFI",
"ROG STRIX B660-G GAMING WIFI",
"ROG STRIX B660-I GAMING WIFI",
"ROG STRIX X670E-A GAMING WIFI",
"ROG STRIX X670E-E GAMING WIFI",
"ROG STRIX X670E-F GAMING WIFI",
"ROG STRIX X670E-I GAMING WIFI",
"ROG STRIX Z590-A GAMING WIFI II",
"ROG STRIX Z690-A GAMING WIFI D4",
"TUF GAMING B650-PLUS",
"TUF GAMING B650-PLUS WIFI",
"TUF GAMING B650M-PLUS",
"TUF GAMING B650M-PLUS WIFI",
"TUF GAMING B660M-PLUS WIFI",
"TUF GAMING X670E-PLUS",
"TUF GAMING X670E-PLUS WIFI",
"TUF GAMING Z590-PLUS WIFI",
have got a NCT6799D chip, but by default there's no use of it because of resource conflict with WMI method.

This commit adds such boards to the monitoring list with new ACPI device UID.

---
 drivers/hwmon/nct6775-platform.c | 52 ++++++++++++++++++++++++++++++++
 1 file changed, 52 insertions(+)
drivers/hwmon/nct6775-platform.c +52 -0
index e5d4a79cd5f7d..76c6b564d7fc4 100644
@@ -114,6 +114,7 @@ struct nct6775_sio_data {
 #define ASUSWMI_UNSUPPORTED_METHOD	0xFFFFFFFE
 #define ASUSWMI_DEVICE_HID		"PNP0C14"
 #define ASUSWMI_DEVICE_UID		"ASUSWMI"
+#define ASUSMSI_DEVICE_UID		"AsusMbSwInterface"
 
 #if IS_ENABLED(CONFIG_ACPI)
 /*
@@ -1108,6 +1109,52 @@ static const char * const asus_wmi_boards[] = {
 	"TUF GAMING Z490-PLUS (WI-FI)",
 };
 
+static const char * const asus_msi_boards[] = {
+	"EX-B660M-V5 PRO D4",
+	"PRIME B650-PLUS",
+	"PRIME B650M-A",
+	"PRIME B650M-A AX",
+	"PRIME B650M-A II",
+	"PRIME B650M-A WIFI",
+	"PRIME B650M-A WIFI II",
+	"PRIME B660M-A D4",
+	"PRIME B660M-A WIFI D4",
+	"PRIME X670-P",
+	"PRIME X670-P WIFI",
+	"PRIME X670E-PRO WIFI",
+	"Pro B660M-C-D4",
+	"ProArt B660-CREATOR D4",
+	"ProArt X670E-CREATOR WIFI",
+	"ROG CROSSHAIR X670E EXTREME",
+	"ROG CROSSHAIR X670E GENE",
+	"ROG CROSSHAIR X670E HERO",
+	"ROG MAXIMUS XIII EXTREME GLACIAL",
+	"ROG MAXIMUS Z690 EXTREME",
+	"ROG MAXIMUS Z690 EXTREME GLACIAL",
+	"ROG STRIX B650-A GAMING WIFI",
+	"ROG STRIX B650E-E GAMING WIFI",
+	"ROG STRIX B650E-F GAMING WIFI",
+	"ROG STRIX B650E-I GAMING WIFI",
+	"ROG STRIX B660-A GAMING WIFI D4",
+	"ROG STRIX B660-F GAMING WIFI",
+	"ROG STRIX B660-G GAMING WIFI",
+	"ROG STRIX B660-I GAMING WIFI",
+	"ROG STRIX X670E-A GAMING WIFI",
+	"ROG STRIX X670E-E GAMING WIFI",
+	"ROG STRIX X670E-F GAMING WIFI",
+	"ROG STRIX X670E-I GAMING WIFI",
+	"ROG STRIX Z590-A GAMING WIFI II",
+	"ROG STRIX Z690-A GAMING WIFI D4",
+	"TUF GAMING B650-PLUS",
+	"TUF GAMING B650-PLUS WIFI",
+	"TUF GAMING B650M-PLUS",
+	"TUF GAMING B650M-PLUS WIFI",
+	"TUF GAMING B660M-PLUS WIFI",
+	"TUF GAMING X670E-PLUS",
+	"TUF GAMING X670E-PLUS WIFI",
+	"TUF GAMING Z590-PLUS WIFI",
+};
+
 #if IS_ENABLED(CONFIG_ACPI)
 /*
  * Callback for acpi_bus_for_each_dev() to find the right device
@@ -1171,6 +1218,11 @@ static int __init sensors_nct6775_platform_init(void)
 				   board_name);
 		if (err >= 0)
 			access = nct6775_determine_access(ASUSWMI_DEVICE_UID);
+
+		err = match_string(asus_msi_boards, ARRAY_SIZE(asus_msi_boards),
+				   board_name);
+		if (err >= 0)
+			access = nct6775_determine_access(ASUSMSI_DEVICE_UID);
 	}
 
 	/*
Signature · 1 line
2.39.2
author maintainer
#633 lore raw

PATCH 632/885 pinctrl: at91: use devm_kasprintf() to avoid potential leaks

In reply to Greg Kroah-Hartman

From: Claudiu Beznea <[email protected]>

[ Upstream commit 1c4e5c470a56f7f7c649c0c70e603abc1eab15c4 ]

Use devm_kasprintf() instead of kasprintf() to avoid any potential leaks. At the moment drivers have no remove functionality thus there is no need for fixes tag.

---
 drivers/pinctrl/pinctrl-at91-pio4.c | 4 ++--
 drivers/pinctrl/pinctrl-at91.c      | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)
drivers/pinctrl/pinctrl-at91-pio4.c +2 -2
index 82b921fd630d5..7f193f2b1566a 100644
@@ -1120,8 +1120,8 @@ static int atmel_pinctrl_probe(struct platform_device *pdev)
 
 		pin_desc[i].number = i;
 		/* Pin naming convention: P(bank_name)(bank_pin_number). */
-		pin_desc[i].name = kasprintf(GFP_KERNEL, "P%c%d",
-					     bank + 'A', line);
+		pin_desc[i].name = devm_kasprintf(&pdev->dev, GFP_KERNEL, "P%c%d",
+						  bank + 'A', line);
 
 		group->name = group_names[i] = pin_desc[i].name;
 		group->pin = pin_desc[i].number;
drivers/pinctrl/pinctrl-at91.c +1 -1
index 81dbffab621fb..ff3b6a8a0b170 100644
@@ -1883,7 +1883,7 @@ static int at91_gpio_probe(struct platform_device *pdev)
 	}
 
 	for (i = 0; i < chip->ngpio; i++)
-		names[i] = kasprintf(GFP_KERNEL, "pio%c%d", alias_idx + 'A', i);
+		names[i] = devm_kasprintf(&pdev->dev, GFP_KERNEL, "pio%c%d", alias_idx + 'A', i);
 
 	chip->names = (const char *const *)names;
 
Signature · 1 line
2.39.2
author maintainer
#634 lore raw

PATCH 633/885 drm/amd/display: Do not commit pipe when updating DRR

In reply to Greg Kroah-Hartman

From: Wesley Chalmers <[email protected]>

[ Upstream commit 8f0d304d21b351d65e8c434c5399a40231876ba1 ]

[WHY]
DRR and Pipe cannot be updated on
the same frame, or else underflow will
occur.

---
 drivers/gpu/drm/amd/display/dc/core/dc.c          | 15 +++++++++++++++
 drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h |  3 ++-
 drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c |  9 +++++++++
 drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.h |  2 ++
 .../drm/amd/display/dc/inc/hw/timing_generator.h  |  1 +
 5 files changed, 29 insertions(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/core/dc.c +15 -0
index 24015f8cac75a..af7aefe285ffd 100644
@@ -3222,6 +3222,21 @@ static void commit_planes_for_stream(struct dc *dc,
 
 	dc_z10_restore(dc);
 
+	if (update_type == UPDATE_TYPE_FULL) {
+		/* wait for all double-buffer activity to clear on all pipes */
+		int pipe_idx;
+
+		for (pipe_idx = 0; pipe_idx < dc->res_pool->pipe_count; pipe_idx++) {
+			struct pipe_ctx *pipe_ctx = &context->res_ctx.pipe_ctx[pipe_idx];
+
+			if (!pipe_ctx->stream)
+				continue;
+
+			if (pipe_ctx->stream_res.tg->funcs->wait_drr_doublebuffer_pending_clear)
+				pipe_ctx->stream_res.tg->funcs->wait_drr_doublebuffer_pending_clear(pipe_ctx->stream_res.tg);
+		}
+	}
+
 	if (get_seamless_boot_stream_count(context) > 0 && surface_count > 0) {
 		/* Optimize seamless boot flag keeps clocks and watermarks high until
 		 * first flip. After first flip, optimization is required to lower
drivers/gpu/drm/amd/display/dc/dcn10/dcn10_optc.h +2 -1
index 88ac5f6f4c96c..0b37bb0e184b2 100644
@@ -519,7 +519,8 @@ struct dcn_optc_registers {
 	type OTG_CRC_DATA_STREAM_COMBINE_MODE;\
 	type OTG_CRC_DATA_STREAM_SPLIT_MODE;\
 	type OTG_CRC_DATA_FORMAT;\
-	type OTG_V_TOTAL_LAST_USED_BY_DRR;
+	type OTG_V_TOTAL_LAST_USED_BY_DRR;\
+	type OTG_DRR_TIMING_DBUF_UPDATE_PENDING;
 
 #define TG_REG_FIELD_LIST_DCN3_2(type) \
 	type OTG_H_TIMING_DIV_MODE_MANUAL;
drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.c +9 -0
index 892d3c4d01a1e..25749f7d88366 100644
@@ -282,6 +282,14 @@ static void optc3_set_timing_double_buffer(struct timing_generator *optc, bool e
 		   OTG_DRR_TIMING_DBUF_UPDATE_MODE, mode);
 }
 
+void optc3_wait_drr_doublebuffer_pending_clear(struct timing_generator *optc)
+{
+	struct optc *optc1 = DCN10TG_FROM_TG(optc);
+
+	REG_WAIT(OTG_DOUBLE_BUFFER_CONTROL, OTG_DRR_TIMING_DBUF_UPDATE_PENDING, 0, 2, 100000); /* 1 vupdate at 5hz */
+
+}
+
 void optc3_set_vtotal_min_max(struct timing_generator *optc, int vtotal_min, int vtotal_max)
 {
 	optc1_set_vtotal_min_max(optc, vtotal_min, vtotal_max);
@@ -351,6 +359,7 @@ static struct timing_generator_funcs dcn30_tg_funcs = {
 		.program_manual_trigger = optc2_program_manual_trigger,
 		.setup_manual_trigger = optc2_setup_manual_trigger,
 		.get_hw_timing = optc1_get_hw_timing,
+		.wait_drr_doublebuffer_pending_clear = optc3_wait_drr_doublebuffer_pending_clear,
 };
 
 void dcn30_timing_generator_init(struct optc *optc1)
drivers/gpu/drm/amd/display/dc/dcn30/dcn30_optc.h +2 -0
index dd45a5499b078..fb06dc9a48937 100644
@@ -279,6 +279,7 @@
 	SF(OTG0_OTG_DRR_TRIGGER_WINDOW, OTG_DRR_TRIGGER_WINDOW_END_X, mask_sh),\
 	SF(OTG0_OTG_DRR_V_TOTAL_CHANGE, OTG_DRR_V_TOTAL_CHANGE_LIMIT, mask_sh),\
 	SF(OTG0_OTG_H_TIMING_CNTL, OTG_H_TIMING_DIV_BY2, mask_sh),\
+	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_DRR_TIMING_DBUF_UPDATE_PENDING, mask_sh),\
 	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_DRR_TIMING_DBUF_UPDATE_MODE, mask_sh),\
 	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_BLANK_DATA_DOUBLE_BUFFER_EN, mask_sh)
 
@@ -317,6 +318,7 @@
 	SF(OTG0_OTG_DRR_TRIGGER_WINDOW, OTG_DRR_TRIGGER_WINDOW_END_X, mask_sh),\
 	SF(OTG0_OTG_DRR_V_TOTAL_CHANGE, OTG_DRR_V_TOTAL_CHANGE_LIMIT, mask_sh),\
 	SF(OTG0_OTG_H_TIMING_CNTL, OTG_H_TIMING_DIV_MODE, mask_sh),\
+	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_DRR_TIMING_DBUF_UPDATE_PENDING, mask_sh),\
 	SF(OTG0_OTG_DOUBLE_BUFFER_CONTROL, OTG_DRR_TIMING_DBUF_UPDATE_MODE, mask_sh)
 
 void dcn30_timing_generator_init(struct optc *optc1);
drivers/gpu/drm/amd/display/dc/inc/hw/timing_generator.h +1 -0
index 25a1df45b2641..f96fb425345e4 100644
@@ -325,6 +325,7 @@ struct timing_generator_funcs {
 			uint32_t vtotal_change_limit);
 
 	void (*init_odm)(struct timing_generator *tg);
+	void (*wait_drr_doublebuffer_pending_clear)(struct timing_generator *tg);
 };
 
 #endif
Signature · 1 line
2.39.2
author maintainer
#635 lore raw

PATCH 634/885 scsi: snic: Fix memory leak with using debugfs_lookup()

In reply to Greg Kroah-Hartman

From: Greg Kroah-Hartman <[email protected]>

[ Upstream commit ad0e4e2fab928477f74d742e6e77d79245d3d3e7 ]

When calling debugfs_lookup() the result must have dput() called on it, otherwise the memory will leak over time. To make things simpler, just call debugfs_lookup_and_remove() instead which handles all of the logic at once.

---
 drivers/scsi/snic/snic_debugfs.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
drivers/scsi/snic/snic_debugfs.c +2 -2
index 57bdc3ba49d9c..9dd975b36b5bd 100644
@@ -437,6 +437,6 @@ void snic_trc_debugfs_init(void)
 void
 snic_trc_debugfs_term(void)
 {
-	debugfs_remove(debugfs_lookup(TRC_FILE, snic_glob->trc_root));
-	debugfs_remove(debugfs_lookup(TRC_ENABLE_FILE, snic_glob->trc_root));
+	debugfs_lookup_and_remove(TRC_FILE, snic_glob->trc_root);
+	debugfs_lookup_and_remove(TRC_ENABLE_FILE, snic_glob->trc_root);
 }
Signature · 1 line
2.39.2
author maintainer
#636 lore raw

PATCH 635/885 scsi: ufs: core: Fix device management cmd timeout flow

In reply to Greg Kroah-Hartman

From: Mason Zhang <[email protected]>

[ Upstream commit 36822124f9de200cedc2f42516301b50d386a6cd ]

In the UFS error handling flow, the host will send a device management cmd (NOP OUT) to the device for link recovery. If this cmd times out and clearing the doorbell fails, ufshcd_wait_for_dev_cmd() will do nothing and return. hba->dev_cmd.complete struct is not set to NULL.

When this happens, if cmd has been completed by device, then we will call complete() in __ufshcd_transfer_req_compl(). Because the complete struct is allocated on the stack, the following crash will occur:

  ipanic_die+0x24/0x38 [mrdump]
  die+0x344/0x748
  arm64_notify_die+0x44/0x104
  do_debug_exception+0x104/0x1e0
  el1_dbg+0x38/0x54
  el1_sync_handler+0x40/0x88
  el1_sync+0x8c/0x140
  queued_spin_lock_slowpath+0x2e4/0x3c0
  __ufshcd_transfer_req_compl+0x3b0/0x1164
  ufshcd_trc_handler+0x15c/0x308
  ufshcd_host_reset_and_restore+0x54/0x260
  ufshcd_reset_and_restore+0x28c/0x57c
  ufshcd_err_handler+0xeb8/0x1b6c
  process_one_work+0x288/0x964
  worker_thread+0x4bc/0xc7c
  kthread+0x15c/0x264
  ret_from_fork+0x10/0x30
---
 drivers/ufs/core/ufshcd.c | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)
drivers/ufs/core/ufshcd.c +16 -0
index e08ce7f2ff3af..edd34dac91b1d 100644
@@ -3006,6 +3006,22 @@ static int ufshcd_wait_for_dev_cmd(struct ufs_hba *hba,
 		} else {
 			dev_err(hba->dev, "%s: failed to clear tag %d\n",
 				__func__, lrbp->task_tag);
+
+			spin_lock_irqsave(&hba->outstanding_lock, flags);
+			pending = test_bit(lrbp->task_tag,
+					   &hba->outstanding_reqs);
+			if (pending)
+				hba->dev_cmd.complete = NULL;
+			spin_unlock_irqrestore(&hba->outstanding_lock, flags);
+
+			if (!pending) {
+				/*
+				 * The completion handler ran while we tried to
+				 * clear the command.
+				 */
+				time_left = 1;
+				goto retry;
+			}
 		}
 	}
 
Signature · 1 line
2.39.2
author maintainer
#637 lore raw

PATCH 636/885 HID: logitech-hidpp: Dont restart communication if not necessary

In reply to Greg Kroah-Hartman

From: Bastien Nocera <[email protected]>

[ Upstream commit 498ba20690357691103de0f766960355247c78be ]

Don't stop and restart communication with the device unless we need to modify the connect flags used because of a device quirk.

---
 drivers/hid/hid-logitech-hidpp.c | 32 ++++++++++++++++++++------------
 1 file changed, 20 insertions(+), 12 deletions(-)
drivers/hid/hid-logitech-hidpp.c +20 -12
index 19fce82d03f02..fdb66dc065822 100644
@@ -4102,6 +4102,7 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id)
 	bool connected;
 	unsigned int connect_mask = HID_CONNECT_DEFAULT;
 	struct hidpp_ff_private_data data;
+	bool will_restart = false;
 
 	/* report_fixup needs drvdata to be set before we call hid_parse */
 	hidpp = devm_kzalloc(&hdev->dev, sizeof(*hidpp), GFP_KERNEL);
@@ -4157,6 +4158,10 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id)
 			return ret;
 	}
 
+	if (hidpp->quirks & HIDPP_QUIRK_DELAYED_INIT ||
+	    hidpp->quirks & HIDPP_QUIRK_UNIFYING)
+		will_restart = true;
+
 	INIT_WORK(&hidpp->work, delayed_work_cb);
 	mutex_init(&hidpp->send_mutex);
 	init_waitqueue_head(&hidpp->wait);
@@ -4171,7 +4176,7 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id)
 	 * Plain USB connections need to actually call start and open
 	 * on the transport driver to allow incoming data.
 	 */
-	ret = hid_hw_start(hdev, 0);
+	ret = hid_hw_start(hdev, will_restart ? 0 : connect_mask);
 	if (ret) {
 		hid_err(hdev, "hw start failed\n");
 		goto hid_hw_start_fail;
@@ -4208,6 +4213,7 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id)
 			hidpp->wireless_feature_index = 0;
 		else if (ret)
 			goto hid_hw_init_fail;
+		ret = 0;
 	}
 
 	if (connected && (hidpp->quirks & HIDPP_QUIRK_CLASS_WTP)) {
@@ -4222,19 +4228,21 @@ static int hidpp_probe(struct hid_device *hdev, const struct hid_device_id *id)
 
 	hidpp_connect_event(hidpp);
 
-	/* Reset the HID node state */
-	hid_device_io_stop(hdev);
-	hid_hw_close(hdev);
-	hid_hw_stop(hdev);
+	if (will_restart) {
+		/* Reset the HID node state */
+		hid_device_io_stop(hdev);
+		hid_hw_close(hdev);
+		hid_hw_stop(hdev);
 
-	if (hidpp->quirks & HIDPP_QUIRK_NO_HIDINPUT)
-		connect_mask &= ~HID_CONNECT_HIDINPUT;
+		if (hidpp->quirks & HIDPP_QUIRK_NO_HIDINPUT)
+			connect_mask &= ~HID_CONNECT_HIDINPUT;
 
-	/* Now export the actual inputs and hidraw nodes to the world */
-	ret = hid_hw_start(hdev, connect_mask);
-	if (ret) {
-		hid_err(hdev, "%s:hid_hw_start returned error\n", __func__);
-		goto hid_hw_start_fail;
+		/* Now export the actual inputs and hidraw nodes to the world */
+		ret = hid_hw_start(hdev, connect_mask);
+		if (ret) {
+			hid_err(hdev, "%s:hid_hw_start returned error\n", __func__);
+			goto hid_hw_start_fail;
+		}
 	}
 
 	if (hidpp->quirks & HIDPP_QUIRK_CLASS_G920) {
Signature · 1 line
2.39.2
author maintainer
#638 lore raw

PATCH 637/885 drm/amd/display: Enable P-state validation checks for DCN314

In reply to Greg Kroah-Hartman

From: Nicholas Kazlauskas <[email protected]>

[ Upstream commit 37d184b548db0f64d4a878960b2c6988b38a3e7e ]

[Why]
To align with DCN31 behavior. This helps avoid p-state hangs in the case where underflow does occur.

[How]
Flip the bit to true.

---
 drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c +1 -1
index c80c8c8f51e97..5985ce8df4e08 100644
@@ -897,7 +897,7 @@ static const struct dc_debug_options debug_defaults_drv = {
 	.max_downscale_src_width = 4096,/*upto true 4k*/
 	.disable_pplib_wm_range = false,
 	.scl_reset_length10 = true,
-	.sanity_checks = false,
+	.sanity_checks = true,
 	.underflow_assert_delay_us = 0xFFFFFFFF,
 	.dwb_fi_phase = -1, // -1 = disable,
 	.dmub_command_table = true,
Signature · 1 line
2.39.2
author maintainer
#639 lore raw

PATCH 638/885 drm: panel-orientation-quirks: Add quirk for Lenovo IdeaPad Duet 3 10IGL5

In reply to Greg Kroah-Hartman

From: Darrell Kavanagh <[email protected]>

[ Upstream commit 38b2d8efd03d2e56431b611e3523f0158306451d ]

Another Lenovo convertable where the panel is installed landscape but is reported to the kernel as portrait.

---
 drivers/gpu/drm/drm_panel_orientation_quirks.c | 6 ++++++
 1 file changed, 6 insertions(+)
drivers/gpu/drm/drm_panel_orientation_quirks.c +6 -0
index b409fe256fd0a..5522d610c5cfd 100644
@@ -322,6 +322,12 @@ static const struct dmi_system_id orientation_data[] = {
 		  DMI_EXACT_MATCH(DMI_PRODUCT_VERSION, "Lenovo ideapad D330-10IGL"),
 		},
 		.driver_data = (void *)&lcd800x1280_rightside_up,
+	}, {	/* Lenovo IdeaPad Duet 3 10IGL5 */
+		.matches = {
+		  DMI_EXACT_MATCH(DMI_SYS_VENDOR, "LENOVO"),
+		  DMI_EXACT_MATCH(DMI_PRODUCT_VERSION, "IdeaPad Duet 3 10IGL5"),
+		},
+		.driver_data = (void *)&lcd1200x1920_rightside_up,
 	}, {	/* Lenovo Yoga Book X90F / X91F / X91L */
 		.matches = {
 		  /* Non exact match to match all versions */
Signature · 1 line
2.39.2
author maintainer
#640 lore raw

PATCH 639/885 drm/amd/display: Disable HUBP/DPP PG on DCN314 for now

In reply to Greg Kroah-Hartman

From: Nicholas Kazlauskas <[email protected]>

[ Upstream commit b7c67f72408b11b922f23f06c7df0f6743a2e89d ]

[Why]
The DMCUB implementation required to workaround corruption is not currently stable and may cause intermittent corruption or hangs.

[How]
Disable PG until the sequence is stable.

---
 drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/gpu/drm/amd/display/dc/dcn314/dcn314_resource.c +2 -0
index 5985ce8df4e08..9918bccd6defb 100644
@@ -888,6 +888,8 @@ static const struct dc_debug_options debug_defaults_drv = {
 	.force_abm_enable = false,
 	.timing_trace = false,
 	.clock_trace = true,
+	.disable_dpp_power_gate = true,
+	.disable_hubp_power_gate = true,
 	.disable_pplib_clock_request = false,
 	.pipe_split_policy = MPC_SPLIT_DYNAMIC,
 	.force_single_disp_pipe_split = false,
Signature · 1 line
2.39.2
author maintainer
#641 lore raw

PATCH 640/885 dm thin: add cond_resched() to various workqueue loops

In reply to Greg Kroah-Hartman

From: Mike Snitzer <[email protected]>

[ Upstream commit e4f80303c2353952e6e980b23914e4214487f2a6 ]

Otherwise on resource constrained systems these workqueues may be too greedy.

---
 drivers/md/dm-thin.c | 2 ++
 1 file changed, 2 insertions(+)
drivers/md/dm-thin.c +2 -0
index 196f82559ad6b..d28c9077d6ed2 100644
@@ -2207,6 +2207,7 @@ static void process_thin_deferred_bios(struct thin_c *tc)
 			throttle_work_update(&pool->throttle);
 			dm_pool_issue_prefetches(pool->pmd);
 		}
+		cond_resched();
 	}
 	blk_finish_plug(&plug);
 }
@@ -2289,6 +2290,7 @@ static void process_thin_deferred_cells(struct thin_c *tc)
 			else
 				pool->process_cell(tc, cell);
 		}
+		cond_resched();
 	} while (!list_empty(&cells));
 }
 
Signature · 1 line
2.39.2
author maintainer
#642 lore raw

PATCH 641/885 dm cache: add cond_resched() to various workqueue loops

In reply to Greg Kroah-Hartman

From: Mike Snitzer <[email protected]>

[ Upstream commit 76227f6dc805e9e960128bcc6276647361e0827c ]

Otherwise on resource constrained systems these workqueues may be too greedy.

---
 drivers/md/dm-cache-target.c | 4 ++++
 1 file changed, 4 insertions(+)
drivers/md/dm-cache-target.c +4 -0
index 5e92fac90b675..17fde3e5a1f7b 100644
@@ -1805,6 +1805,7 @@ static void process_deferred_bios(struct work_struct *ws)
 
 		else
 			commit_needed = process_bio(cache, bio) || commit_needed;
+		cond_resched();
 	}
 
 	if (commit_needed)
@@ -1827,6 +1828,7 @@ static void requeue_deferred_bios(struct cache *cache)
 	while ((bio = bio_list_pop(&bios))) {
 		bio->bi_status = BLK_STS_DM_REQUEUE;
 		bio_endio(bio);
+		cond_resched();
 	}
 }
 
@@ -1867,6 +1869,8 @@ static void check_migrations(struct work_struct *ws)
 		r = mg_start(cache, op, NULL);
 		if (r)
 			break;
+
+		cond_resched();
 	}
 }
 
Signature · 1 line
2.39.2
author maintainer
#643 lore raw

PATCH 642/885 nfsd: zero out pointers after putting nfsd_files on COPY setup error

In reply to Greg Kroah-Hartman

From: Jeff Layton <[email protected]>

[ Upstream commit 1f0001d43d0c0ac2a19a34a914f6595ad97cbc1d ]

At first, I thought this might be a source of nfsd_file overputs, but the current callers seem to avoid an extra put when nfsd4_verify_copy returns an error.

Still, it's "bad form" to leave the pointers filled out when we don't have a reference to them anymore, and that might lead to bugs later. Zero them out as a defensive coding measure.

---
 fs/nfsd/nfs4proc.c | 2 ++
 1 file changed, 2 insertions(+)
fs/nfsd/nfs4proc.c +2 -0
index 53113976e6424..a90e792a94d77 100644
@@ -1227,8 +1227,10 @@ nfsd4_verify_copy(struct svc_rqst *rqstp, struct nfsd4_compound_state *cstate,
 	return status;
 out_put_dst:
 	nfsd_file_put(*dst);
+	*dst = NULL;
 out_put_src:
 	nfsd_file_put(*src);
+	*src = NULL;
 	goto out;
 }
 
Signature · 1 line
2.39.2
author maintainer
#644 lore raw

PATCH 643/885 nfsd: dont hand out delegation on setuid files being opened for write

In reply to Greg Kroah-Hartman

From: Jeff Layton <[email protected]>

[ Upstream commit 826b67e6376c2a788e3a62c4860dcd79500a27d5 ]

We had a bug report that xfstest generic/355 was failing on NFSv4.0. This test sets various combinations of setuid/setgid modes and tests whether DIO writes will cause them to be stripped.

What I found was that the server did properly strip those bits, but the client didn't notice because it held a delegation that was not recalled. The recall didn't occur because the client itself was the one generating the activity and we avoid recalls in that case.

Clearing setuid bits is an "implicit" activity. The client didn't specifically request that we do that, so we need the server to issue a CB_RECALL, or avoid the situation entirely by not issuing a delegation.

The easiest fix here is to simply not give out a delegation if the file is being opened for write, and the mode has the setuid and/or setgid bit set. Note that there is a potential race between the mode and lease being set, so we test for this condition both before and after setting the lease.

This patch fixes generic/355, generic/683 and generic/684 for me. (Note that 355 fails only on v4.0, and 683 and 684 require NFSv4.2 to run and fail).

---
 fs/nfsd/nfs4state.c | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)
fs/nfsd/nfs4state.c +27 -0
index a768b8701b44e..34561764e5c97 100644
@@ -5399,6 +5399,23 @@ nfsd4_verify_deleg_dentry(struct nfsd4_open *open, struct nfs4_file *fp,
 	return 0;
 }
 
+/*
+ * We avoid breaking delegations held by a client due to its own activity, but
+ * clearing setuid/setgid bits on a write is an implicit activity and the client
+ * may not notice and continue using the old mode. Avoid giving out a delegation
+ * on setuid/setgid files when the client is requesting an open for write.
+ */
+static int
+nfsd4_verify_setuid_write(struct nfsd4_open *open, struct nfsd_file *nf)
+{
+	struct inode *inode = file_inode(nf->nf_file);
+
+	if ((open->op_share_access & NFS4_SHARE_ACCESS_WRITE) &&
+	    (inode->i_mode & (S_ISUID|S_ISGID)))
+		return -EAGAIN;
+	return 0;
+}
+
 static struct nfs4_delegation *
 nfs4_set_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 		    struct svc_fh *parent)
@@ -5432,6 +5449,8 @@ nfs4_set_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 	spin_lock(&fp->fi_lock);
 	if (nfs4_delegation_exists(clp, fp))
 		status = -EAGAIN;
+	else if (nfsd4_verify_setuid_write(open, nf))
+		status = -EAGAIN;
 	else if (!fp->fi_deleg_file) {
 		fp->fi_deleg_file = nf;
 		/* increment early to prevent fi_deleg_file from being
@@ -5472,6 +5491,14 @@ nfs4_set_delegation(struct nfsd4_open *open, struct nfs4_ol_stateid *stp,
 	if (status)
 		goto out_unlock;
 
+	/*
+	 * Now that the deleg is set, check again to ensure that nothing
+	 * raced in and changed the mode while we weren't lookng.
+	 */
+	status = nfsd4_verify_setuid_write(open, fp->fi_deleg_file);
+	if (status)
+		goto out_unlock;
+
 	spin_lock(&state_lock);
 	spin_lock(&fp->fi_lock);
 	if (fp->fi_had_conflict)
Signature · 1 line
2.39.2
author maintainer
#645 lore raw

PATCH 644/885 cifs: prevent data race in smb2_reconnect()

In reply to Greg Kroah-Hartman

From: Paulo Alcantara <[email protected]>

[ Upstream commit 3c0070f54b3128de498c2dd9934a21f0dd867111 ]

Make sure to get an up-to-date TCP_Server_Info::nr_targets value prior to waiting the server to be reconnected in smb2_reconnect(). It is set in cifs_tcp_ses_needs_reconnect() and protected by TCP_Server_Info::srv_lock.

---
 fs/cifs/smb2pdu.c | 119 +++++++++++++++++++++++++---------------------
 1 file changed, 64 insertions(+), 55 deletions(-)
fs/cifs/smb2pdu.c +64 -55
index 3b93680a319e4..5a063af8a094e 100644
@@ -139,6 +139,66 @@ smb2_hdr_assemble(struct smb2_hdr *shdr, __le16 smb2_cmd,
 	return;
 }
 
+static int wait_for_server_reconnect(struct TCP_Server_Info *server,
+				     __le16 smb2_command, bool retry)
+{
+	int timeout = 10;
+	int rc;
+
+	spin_lock(&server->srv_lock);
+	if (server->tcpStatus != CifsNeedReconnect) {
+		spin_unlock(&server->srv_lock);
+		return 0;
+	}
+	timeout *= server->nr_targets;
+	spin_unlock(&server->srv_lock);
+
+	/*
+	 * Return to caller for TREE_DISCONNECT and LOGOFF and CLOSE
+	 * here since they are implicitly done when session drops.
+	 */
+	switch (smb2_command) {
+	/*
+	 * BB Should we keep oplock break and add flush to exceptions?
+	 */
+	case SMB2_TREE_DISCONNECT:
+	case SMB2_CANCEL:
+	case SMB2_CLOSE:
+	case SMB2_OPLOCK_BREAK:
+		return -EAGAIN;
+	}
+
+	/*
+	 * Give demultiplex thread up to 10 seconds to each target available for
+	 * reconnect -- should be greater than cifs socket timeout which is 7
+	 * seconds.
+	 *
+	 * On "soft" mounts we wait once. Hard mounts keep retrying until
+	 * process is killed or server comes back on-line.
+	 */
+	do {
+		rc = wait_event_interruptible_timeout(server->response_q,
+						      (server->tcpStatus != CifsNeedReconnect),
+						      timeout * HZ);
+		if (rc < 0) {
+			cifs_dbg(FYI, "%s: aborting reconnect due to received signal\n",
+				 __func__);
+			return -ERESTARTSYS;
+		}
+
+		/* are we still trying to reconnect? */
+		spin_lock(&server->srv_lock);
+		if (server->tcpStatus != CifsNeedReconnect) {
+			spin_unlock(&server->srv_lock);
+			return 0;
+		}
+		spin_unlock(&server->srv_lock);
+	} while (retry);
+
+	cifs_dbg(FYI, "%s: gave up waiting on reconnect\n", __func__);
+	return -EHOSTDOWN;
+}
+
 static int
 smb2_reconnect(__le16 smb2_command, struct cifs_tcon *tcon,
 	       struct TCP_Server_Info *server)
@@ -146,7 +206,6 @@ smb2_reconnect(__le16 smb2_command, struct cifs_tcon *tcon,
 	int rc = 0;
 	struct nls_table *nls_codepage;
 	struct cifs_ses *ses;
-	int retries;
 
 	/*
 	 * SMB2s NegProt, SessSetup, Logoff do not have tcon yet so
@@ -184,61 +243,11 @@ smb2_reconnect(__le16 smb2_command, struct cifs_tcon *tcon,
 	    (!tcon->ses->server) || !server)
 		return -EIO;
 
-	ses = tcon->ses;
-	retries = server->nr_targets;
-
-	/*
-	 * Give demultiplex thread up to 10 seconds to each target available for
-	 * reconnect -- should be greater than cifs socket timeout which is 7
-	 * seconds.
-	 */
-	while (server->tcpStatus == CifsNeedReconnect) {
-		/*
-		 * Return to caller for TREE_DISCONNECT and LOGOFF and CLOSE
-		 * here since they are implicitly done when session drops.
-		 */
-		switch (smb2_command) {
-		/*
-		 * BB Should we keep oplock break and add flush to exceptions?
-		 */
-		case SMB2_TREE_DISCONNECT:
-		case SMB2_CANCEL:
-		case SMB2_CLOSE:
-		case SMB2_OPLOCK_BREAK:
-			return -EAGAIN;
-		}
-
-		rc = wait_event_interruptible_timeout(server->response_q,
-						      (server->tcpStatus != CifsNeedReconnect),
-						      10 * HZ);
-		if (rc < 0) {
-			cifs_dbg(FYI, "%s: aborting reconnect due to a received signal by the process\n",
-				 __func__);
-			return -ERESTARTSYS;
-		}
-
-		/* are we still trying to reconnect? */
-		spin_lock(&server->srv_lock);
-		if (server->tcpStatus != CifsNeedReconnect) {
-			spin_unlock(&server->srv_lock);
-			break;
-		}
-		spin_unlock(&server->srv_lock);
-
-		if (retries && --retries)
-			continue;
+	rc = wait_for_server_reconnect(server, smb2_command, tcon->retry);
+	if (rc)
+		return rc;
 
-		/*
-		 * on "soft" mounts we wait once. Hard mounts keep
-		 * retrying until process is killed or server comes
-		 * back on-line
-		 */
-		if (!tcon->retry) {
-			cifs_dbg(FYI, "gave up waiting on reconnect in smb_init\n");
-			return -EHOSTDOWN;
-		}
-		retries = server->nr_targets;
-	}
+	ses = tcon->ses;
 
 	spin_lock(&ses->chan_lock);
 	if (!cifs_chan_needs_reconnect(ses, server) && !tcon->need_reconnect) {
Signature · 1 line
2.39.2
author maintainer
#646 lore raw

PATCH 645/885 drm/shmem-helper: Revert accidental non-GPL export

In reply to Greg Kroah-Hartman

From: Asahi Lina <[email protected]>

[ Upstream commit 047a754558d640eaa080fce3b22ca9f3d4e04626 ]

The referenced commit added a wrapper for drm_gem_shmem_get_pages_sgt(), but in the process it accidentally changed the export type from GPL to non-GPL. Switch it back to GPL.

Fixes ddddedaa0db9 ("drm/shmem-helper: Fix locking for drm_gem_shmem_get_pages_sgt()")
---
 drivers/gpu/drm/drm_gem_shmem_helper.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/gpu/drm/drm_gem_shmem_helper.c +1 -1
index 2c559b310cad3..7af9da886d4e5 100644
@@ -747,7 +747,7 @@ struct sg_table *drm_gem_shmem_get_pages_sgt(struct drm_gem_shmem_object *shmem)
 
 	return sgt;
 }
-EXPORT_SYMBOL(drm_gem_shmem_get_pages_sgt);
+EXPORT_SYMBOL_GPL(drm_gem_shmem_get_pages_sgt);
 
 /**
  * drm_gem_shmem_prime_import_sg_table - Produce a shmem GEM object from
Signature · 1 line
2.39.2
author maintainer
#647 lore raw

PATCH 646/885 driver core: fw_devlink: Avoid spurious error message

In reply to Greg Kroah-Hartman

From: Saravana Kannan <[email protected]>

[ Upstream commit 6309872413f14f3d58c13ae4dc85b1a7004b4193 ]

fw_devlink can sometimes try to create a device link with the consumer and supplier as the same device. These attempts will fail (correctly), but are harmless. So, avoid printing an error for these cases. Also, add more detail to the error message.

---
 drivers/base/core.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/base/core.c +3 -3
index ac08d475e2828..e30223c2672fc 100644
@@ -2087,9 +2087,9 @@ static int fw_devlink_create_devlink(struct device *con,
 			goto out;
 		}
 
-		if (!device_link_add(con, sup_dev, flags)) {
-			dev_err(con, "Failed to create device link with %s\n",
-				dev_name(sup_dev));
+		if (con != sup_dev && !device_link_add(con, sup_dev, flags)) {
+			dev_err(con, "Failed to create device link (0x%x) with %s\n",
+				flags, dev_name(sup_dev));
 			ret = -EINVAL;
 		}
 
Signature · 1 line
2.39.2
author maintainer
#648 lore raw

PATCH 647/885 wifi: rtl8xxxu: fixing transmisison failure for rtl8192eu

In reply to Greg Kroah-Hartman

From: Jun ASAKA <[email protected]>

commit c6015bf3ff1ffb3caa27eb913797438a0fc634a0 upstream.

Fixing transmission failure which results in "authentication with ... timed out". This can be fixed by disable the REG_TXPAUSE.

---
 drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_8192e.c |    5 +++++
 1 file changed, 5 insertions(+)
drivers/net/wireless/realtek/rtl8xxxu/rtl8xxxu_8192e.c +5 -0
@@ -1669,6 +1669,11 @@ static void rtl8192e_enable_rf(struct rt
 	val8 = rtl8xxxu_read8(priv, REG_PAD_CTRL1);
 	val8 &= ~BIT(0);
 	rtl8xxxu_write8(priv, REG_PAD_CTRL1, val8);
+
+	/*
+	 * Fix transmission failure of rtl8192e.
+	 */
+	rtl8xxxu_write8(priv, REG_TXPAUSE, 0x00);
 }
 
 struct rtl8xxxu_fileops rtl8192eu_fops = {
author maintainer
#649 lore raw

PATCH 648/885 scsi: mpt3sas: Remove usage of dma_get_required_mask() API

In reply to Greg Kroah-Hartman

From: Sreekanth Reddy <[email protected]>

commit 06e472acf964649a58b7de35fc9cdc3151acb970 upstream.

Remove the usage of dma_get_required_mask() API. Directly set the DMA mask to 63/64 if the system is a 64bit machine.

---
 drivers/scsi/mpt3sas/mpt3sas_base.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
drivers/scsi/mpt3sas/mpt3sas_base.c +1 -2
@@ -2992,8 +2992,7 @@ _base_config_dma_addressing(struct MPT3S
 	struct sysinfo s;
 	u64 coherent_dma_mask, dma_mask;
 
-	if (ioc->is_mcpu_endpoint || sizeof(dma_addr_t) == 4 ||
-	    dma_get_required_mask(&pdev->dev) <= DMA_BIT_MASK(32)) {
+	if (ioc->is_mcpu_endpoint || sizeof(dma_addr_t) == 4) {
 		ioc->dma_mask = 32;
 		coherent_dma_mask = dma_mask = DMA_BIT_MASK(32);
 	/* Set 63 bit DMA mask for all SAS3 and SAS35 controllers */
author maintainer
#650 lore raw

PATCH 649/885 firmware: coreboot: framebuffer: Ignore reserved pixel color bits

In reply to Greg Kroah-Hartman

From: Alper Nebi Yasak <[email protected]>

commit e6acaf25cba14661211bb72181c35dd13b24f5b3 upstream.

The coreboot framebuffer doesn't support transparency, its 'reserved' bit field is merely padding for byte/word alignment of pixel colors [1]. When trying to match the framebuffer to a simplefb format, the kernel driver unnecessarily requires the format's transparency bit field to exactly match this padding, even if the former is zero-width.

Due to a coreboot bug [2] (fixed upstream), some boards misreport the reserved field's size as equal to its position (0x18 for both on a 'Lick' Chromebook), and the driver fails to probe where it would have otherwise worked fine with e.g. the a8r8g8b8 or x8r8g8b8 formats.

Remove the transparency comparison with reserved bits. When the bits-per-pixel and other color components match, transparency will already be in a subset of the reserved field. Not forcing it to match reserved bits allows the driver to work on the boards which misreport the reserved field. It also enables using simplefb formats that don't have transparency bits, although this doesn't currently happen due to format support and ordering in linux/platform_data/simplefb.h.

[1] https://review.coreboot.org/plugins/gitiles/coreboot/+/4.19/src/commonlib/include/commonlib/coreboot_tables.h#255
[2] https://review.coreboot.org/plugins/gitiles/coreboot/+/4.13/src/drivers/intel/fsp2_0/graphics.c#82

---
 drivers/firmware/google/framebuffer-coreboot.c |    4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)
drivers/firmware/google/framebuffer-coreboot.c +1 -3
@@ -43,9 +43,7 @@ static int framebuffer_probe(struct core
 		    fb->green_mask_pos     == formats[i].green.offset &&
 		    fb->green_mask_size    == formats[i].green.length &&
 		    fb->blue_mask_pos      == formats[i].blue.offset &&
-		    fb->blue_mask_size     == formats[i].blue.length &&
-		    fb->reserved_mask_pos  == formats[i].transp.offset &&
-		    fb->reserved_mask_size == formats[i].transp.length)
+		    fb->blue_mask_size     == formats[i].blue.length)
 			pdata.format = formats[i].name;
 	}
 	if (!pdata.format)
author maintainer
#651 lore raw

PATCH 650/885 block: dont allow multiple bios for IOCB_NOWAIT issue

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit 67d59247d4b52c917e373f05a807027756ab216f upstream.

If we're doing a large IO request which needs to be split into multiple bios for issue, then we can run into the same situation as the below marked commit fixes - parts will complete just fine, one or more parts will fail to allocate a request. This will result in a partially completed read or write request, where the caller gets EAGAIN even though parts of the IO completed just fine.

Do the same for large bios as we do for splits - fail a NOWAIT request with EAGAIN. This isn't technically fixing an issue in the below marked patch, but for stable purposes, we should have either none of them or both.

This depends on: 613b14884b85 ("block: handle bio_split_to_limits() NULL return")

Fixes 9cea62b2cbab ("block: don't allow splitting of a REQ_NOWAIT bio")
---
 block/fops.c |   21 ++++++++++++++++++---
 1 file changed, 18 insertions(+), 3 deletions(-)
block/fops.c +18 -3
@@ -221,6 +221,24 @@ static ssize_t __blkdev_direct_IO(struct
 			bio_endio(bio);
 			break;
 		}
+		if (iocb->ki_flags & IOCB_NOWAIT) {
+			/*
+			 * This is nonblocking IO, and we need to allocate
+			 * another bio if we have data left to map. As we
+			 * cannot guarantee that one of the sub bios will not
+			 * fail getting issued FOR NOWAIT and as error results
+			 * are coalesced across all of them, be safe and ask for
+			 * a retry of this from blocking context.
+			 */
+			if (unlikely(iov_iter_count(iter))) {
+				bio_release_pages(bio, false);
+				bio_clear_flag(bio, BIO_REFFED);
+				bio_put(bio);
+				blk_finish_plug(&plug);
+				return -EAGAIN;
+			}
+			bio->bi_opf |= REQ_NOWAIT;
+		}
 
 		if (is_read) {
 			if (dio->flags & DIO_SHOULD_DIRTY)
@@ -228,9 +246,6 @@ static ssize_t __blkdev_direct_IO(struct
 		} else {
 			task_io_account_write(bio->bi_iter.bi_size);
 		}
-		if (iocb->ki_flags & IOCB_NOWAIT)
-			bio->bi_opf |= REQ_NOWAIT;
-
 		dio->size += bio->bi_iter.bi_size;
 		pos += bio->bi_iter.bi_size;
 
author maintainer
#652 lore raw

PATCH 651/885 block: clear bio->bi_bdev when putting a bio back in the cache

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit 11eb695feb636fa5211067189cad25ac073e7fe5 upstream.

This isn't strictly needed in terms of correctness, but it does allow polling to know if the bio has been put already by a different task and hence avoid polling something that we don't need to.

Fixes be4d234d7aeb ("bio: add allocation cache abstraction")
---
 block/bio.c |    1 +
 1 file changed, 1 insertion(+)
block/bio.c +1 -0
@@ -747,6 +747,7 @@ void bio_put(struct bio *bio)
 		bio_uninit(bio);
 		cache = per_cpu_ptr(bio->bi_pool->cache, get_cpu());
 		bio->bi_next = cache->free_list;
+		bio->bi_bdev = NULL;
 		cache->free_list = bio;
 		if (++cache->nr > ALLOC_CACHE_MAX + ALLOC_CACHE_SLACK)
 			bio_alloc_cache_prune(cache, ALLOC_CACHE_SLACK);
author maintainer
#653 lore raw

PATCH 652/885 block: be a bit more careful in checking for NULL bdev while polling

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit 310726c33ad76cebdee312dbfafc12c1b44bf977 upstream.

Wei reports a crash with an application using polled IO:

PGD 14265e067 P4D 14265e067 PUD 47ec50067 PMD 0

Oops: 0000 [#1] SMP
CPU: 0 PID: 21915 Comm: iocore_0 Kdump: loaded Tainted: G S                5.12.0-0_fbk12_clang_7346_g1bb6f2e7058f #1
Hardware name: Wiwynn Delta Lake MP T8/Delta Lake-Class2, BIOS Y3DLM08 04/10/2022
RIP: 0010:bio_poll+0x25/0x200
Code: 0f 1f 44 00 00 0f 1f 44 00 00 55 41 57 41 56 41 55 41 54 53 48 83 ec 28 65 48 8b 04 25 28 00 00 00 48 89 44 24 20 48 8b 47 08 <48> 8b 80 70 02 00 00 4c 8b 70 50 8b 6f 34 31 db 83 fd ff 75 25 65
RSP: 0018:ffffc90005fafdf8 EFLAGS: 00010292
RAX: 0000000000000000 RBX: 0000000000000000 RCX: 74b43cd65dd66600
RDX: 0000000000000003 RSI: ffffc90005fafe78 RDI: ffff8884b614e140
RBP: ffff88849964df78 R08: 0000000000000000 R09: 0000000000000008
R10: 0000000000000000 R11: 0000000000000000 R12: ffff88849964df00
R13: ffffc90005fafe78 R14: ffff888137d3c378 R15: 0000000000000001
FS:  00007fd195000640(0000) GS:ffff88903f400000(0000) knlGS:0000000000000000
CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
CR2: 0000000000000270 CR3: 0000000466121001 CR4: 00000000007706f0
DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
PKRU: 55555554
Call Trace:
 iocb_bio_iopoll+0x1d/0x30
 io_do_iopoll+0xac/0x250
 __se_sys_io_uring_enter+0x3c5/0x5a0
 ? __x64_sys_write+0x89/0xd0
 do_syscall_64+0x2d/0x40
 entry_SYSCALL_64_after_hwframe+0x44/0xae
RIP: 0033:0x94f225d
Code: 24 cc 00 00 00 41 8b 84 24 d0 00 00 00 c1 e0 04 83 e0 10 41 09 c2 8b 33 8b 53 04 4c 8b 43 18 4c 63 4b 0c b8 aa 01 00 00 0f 05 <85> c0 0f 88 85 00 00 00 29 03 45 84 f6 0f 84 88 00 00 00 41 f6 c7
RSP: 002b:00007fd194ffcd88 EFLAGS: 00000202 ORIG_RAX: 00000000000001aa
RAX: ffffffffffffffda RBX: 00007fd194ffcdc0 RCX: 00000000094f225d
RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000007
RBP: 00007fd194ffcdb0 R08: 0000000000000000 R09: 0000000000000008
R10: 0000000000000001 R11: 0000000000000202 R12: 00007fd269d68030
R13: 0000000000000000 R14: 0000000000000001 R15: 0000000000000000

which is due to bio->bi_bdev being NULL. This can happen if we have two tasks doing polled IO, and task B ends up completing IO from task A if they are sharing a poll queue. If task B completes the IO and puts the bio into our cache, then it can allocate that bio again before task A is done polling for it. As that would necessitate a preempt between the two tasks, it's enough to just be a bit more careful in checking for whether or not bio->bi_bdev is NULL.

Reported-and-tested-by Wei Zhang <[email protected]>
Fixes be4d234d7aeb ("bio: add allocation cache abstraction")
---
 block/blk-core.c |   10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)
block/blk-core.c +8 -2
@@ -842,10 +842,16 @@ EXPORT_SYMBOL(submit_bio);
  */
 int bio_poll(struct bio *bio, struct io_comp_batch *iob, unsigned int flags)
 {
-	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
 	blk_qc_t cookie = READ_ONCE(bio->bi_cookie);
+	struct block_device *bdev;
+	struct request_queue *q;
 	int ret = 0;
 
+	bdev = READ_ONCE(bio->bi_bdev);
+	if (!bdev)
+		return 0;
+
+	q = bdev_get_queue(bdev);
 	if (cookie == BLK_QC_T_NONE ||
 	    !test_bit(QUEUE_FLAG_POLL, &q->queue_flags))
 		return 0;
@@ -905,7 +911,7 @@ int iocb_bio_iopoll(struct kiocb *kiocb,
 	 */
 	rcu_read_lock();
 	bio = READ_ONCE(kiocb->private);
-	if (bio && bio->bi_bdev)
+	if (bio)
 		ret = bio_poll(bio, iob, flags);
 	rcu_read_unlock();
 
author maintainer
#654 lore raw

PATCH 653/885 rtc: pm8xxx: fix set-alarm race

In reply to Greg Kroah-Hartman

From: Johan Hovold <[email protected]>

commit c88db0eff9722fc2b6c4d172a50471d20e08ecc6 upstream.

Make sure to disable the alarm before updating the four alarm time registers to avoid spurious alarms during the update.

Note that the disable needs to be done outside of the ctrl_reg_lock section to prevent a racing alarm interrupt from disabling the newly set alarm when the lock is released.

---
 drivers/rtc/rtc-pm8xxx.c |   24 ++++++++++--------------
 1 file changed, 10 insertions(+), 14 deletions(-)
drivers/rtc/rtc-pm8xxx.c +10 -14
@@ -221,7 +221,6 @@ static int pm8xxx_rtc_set_alarm(struct d
 {
 	int rc, i;
 	u8 value[NUM_8_BIT_RTC_REGS];
-	unsigned int ctrl_reg;
 	unsigned long secs, irq_flags;
 	struct pm8xxx_rtc *rtc_dd = dev_get_drvdata(dev);
 	const struct pm8xxx_rtc_regs *regs = rtc_dd->regs;
@@ -233,6 +232,11 @@ static int pm8xxx_rtc_set_alarm(struct d
 		secs >>= 8;
 	}
 
+	rc = regmap_update_bits(rtc_dd->regmap, regs->alarm_ctrl,
+				regs->alarm_en, 0);
+	if (rc)
+		return rc;
+
 	spin_lock_irqsave(&rtc_dd->ctrl_reg_lock, irq_flags);
 
 	rc = regmap_bulk_write(rtc_dd->regmap, regs->alarm_rw, value,
@@ -242,19 +246,11 @@ static int pm8xxx_rtc_set_alarm(struct d
 		goto rtc_rw_fail;
 	}
 
-	rc = regmap_read(rtc_dd->regmap, regs->alarm_ctrl, &ctrl_reg);
-	if (rc)
-		goto rtc_rw_fail;
-
-	if (alarm->enabled)
-		ctrl_reg |= regs->alarm_en;
-	else
-		ctrl_reg &= ~regs->alarm_en;
-
-	rc = regmap_write(rtc_dd->regmap, regs->alarm_ctrl, ctrl_reg);
-	if (rc) {
-		dev_err(dev, "Write to RTC alarm control register failed\n");
-		goto rtc_rw_fail;
+	if (alarm->enabled) {
+		rc = regmap_update_bits(rtc_dd->regmap, regs->alarm_ctrl,
+					regs->alarm_en, regs->alarm_en);
+		if (rc)
+			goto rtc_rw_fail;
 	}
 
 	dev_dbg(dev, "Alarm Set for h:m:s=%ptRt, y-m-d=%ptRdr\n",
author maintainer
#655 lore raw

PATCH 654/885 ipmi: ipmb: Fix the MODULE_PARM_DESC associated to retry_time_ms

In reply to Greg Kroah-Hartman

From: Christophe JAILLET <[email protected]>

commit befb28f2676a65a5a4cc4626ae224461d8785af6 upstream.

'This should be 'retry_time_ms' instead of 'max_retries'.

Fixes 63c4eb347164 ("ipmi:ipmb: Add initial support for IPMI over IPMB")

Message-Id: <0d8670cff2c656e99a832a249e77dc90578f67de.1675591429.git.christophe.jaillet@wanadoo.fr>

---
 drivers/char/ipmi/ipmi_ipmb.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/char/ipmi/ipmi_ipmb.c +1 -1
@@ -27,7 +27,7 @@ MODULE_PARM_DESC(bmcaddr, "Address to us
 
 static unsigned int retry_time_ms = 250;
 module_param(retry_time_ms, uint, 0644);
-MODULE_PARM_DESC(max_retries, "Timeout time between retries, in milliseconds.");
+MODULE_PARM_DESC(retry_time_ms, "Timeout time between retries, in milliseconds.");
 
 static unsigned int max_retries = 1;
 module_param(max_retries, uint, 0644);
author maintainer
#656 lore raw

PATCH 655/885 ipmi:ssif: resend_msg() cannot fail

In reply to Greg Kroah-Hartman

From: Corey Minyard <[email protected]>

commit 95767ed78a181d5404202627499f9cde56053b96 upstream.

The resend_msg() function cannot fail, but there was error handling around using it. Rework the handling of the error, and fix the out of retries debug reporting that was wrong around this, too.

---
 drivers/char/ipmi/ipmi_ssif.c |   28 +++++++---------------------
 1 file changed, 7 insertions(+), 21 deletions(-)
drivers/char/ipmi/ipmi_ssif.c +7 -21
@@ -602,7 +602,7 @@ static void ssif_alert(struct i2c_client
 		start_get(ssif_info);
 }
 
-static int start_resend(struct ssif_info *ssif_info);
+static void start_resend(struct ssif_info *ssif_info);
 
 static void msg_done_handler(struct ssif_info *ssif_info, int result,
 			     unsigned char *data, unsigned int len)
@@ -909,31 +909,17 @@ static void msg_written_handler(struct s
 	if (result < 0) {
 		ssif_info->retries_left--;
 		if (ssif_info->retries_left > 0) {
-			if (!start_resend(ssif_info)) {
-				ssif_inc_stat(ssif_info, send_retries);
-				return;
-			}
-			/* request failed, just return the error. */
-			ssif_inc_stat(ssif_info, send_errors);
-
-			if (ssif_info->ssif_debug & SSIF_DEBUG_MSG)
-				dev_dbg(&ssif_info->client->dev,
-					"%s: Out of retries\n", __func__);
-			msg_done_handler(ssif_info, -EIO, NULL, 0);
+			start_resend(ssif_info);
 			return;
 		}
 
 		ssif_inc_stat(ssif_info, send_errors);
 
-		/*
-		 * Got an error on transmit, let the done routine
-		 * handle it.
-		 */
 		if (ssif_info->ssif_debug & SSIF_DEBUG_MSG)
 			dev_dbg(&ssif_info->client->dev,
-				"%s: Error  %d\n", __func__, result);
+				"%s: Out of retries\n", __func__);
 
-		msg_done_handler(ssif_info, result, NULL, 0);
+		msg_done_handler(ssif_info, -EIO, NULL, 0);
 		return;
 	}
 
@@ -996,7 +982,7 @@ static void msg_written_handler(struct s
 	}
 }
 
-static int start_resend(struct ssif_info *ssif_info)
+static void start_resend(struct ssif_info *ssif_info)
 {
 	int command;
 
@@ -1021,7 +1007,6 @@ static int start_resend(struct ssif_info
 
 	ssif_i2c_send(ssif_info, msg_written_handler, I2C_SMBUS_WRITE,
 		   command, ssif_info->data, I2C_SMBUS_BLOCK_DATA);
-	return 0;
 }
 
 static int start_send(struct ssif_info *ssif_info,
@@ -1036,7 +1021,8 @@ static int start_send(struct ssif_info *
 	ssif_info->retries_left = SSIF_SEND_RETRIES;
 	memcpy(ssif_info->data + 1, data, len);
 	ssif_info->data_len = len;
-	return start_resend(ssif_info);
+	start_resend(ssif_info);
+	return 0;
 }
 
 /* Must be called with the message lock held. */
author maintainer
#657 lore raw

PATCH 656/885 ipmi_ssif: Rename idle state and check

In reply to Greg Kroah-Hartman

From: Corey Minyard <[email protected]>

commit 8230831c43a328c2be6d28c65d3f77e14c59986b upstream.

Rename the SSIF_IDLE() to IS_SSIF_IDLE(), since that is more clear, and rename SSIF_NORMAL to SSIF_IDLE, since that's more accurate.

---
 drivers/char/ipmi/ipmi_ssif.c |   46 +++++++++++++++++++++---------------------
 1 file changed, 23 insertions(+), 23 deletions(-)
drivers/char/ipmi/ipmi_ssif.c +23 -23
@@ -92,7 +92,7 @@
 #define SSIF_WATCH_WATCHDOG_TIMEOUT	msecs_to_jiffies(250)
 
 enum ssif_intf_state {
-	SSIF_NORMAL,
+	SSIF_IDLE,
 	SSIF_GETTING_FLAGS,
 	SSIF_GETTING_EVENTS,
 	SSIF_CLEARING_FLAGS,
@@ -100,8 +100,8 @@ enum ssif_intf_state {
 	/* FIXME - add watchdog stuff. */
 };
 
-#define SSIF_IDLE(ssif)	 ((ssif)->ssif_state == SSIF_NORMAL \
-			  && (ssif)->curr_msg == NULL)
+#define IS_SSIF_IDLE(ssif) ((ssif)->ssif_state == SSIF_IDLE \
+			    && (ssif)->curr_msg == NULL)
 
 /*
  * Indexes into stats[] in ssif_info below.
@@ -348,9 +348,9 @@ static void return_hosed_msg(struct ssif
 
 /*
  * Must be called with the message lock held.  This will release the
- * message lock.  Note that the caller will check SSIF_IDLE and start a
- * new operation, so there is no need to check for new messages to
- * start in here.
+ * message lock.  Note that the caller will check IS_SSIF_IDLE and
+ * start a new operation, so there is no need to check for new
+ * messages to start in here.
  */
 static void start_clear_flags(struct ssif_info *ssif_info, unsigned long *flags)
 {
@@ -367,7 +367,7 @@ static void start_clear_flags(struct ssi
 
 	if (start_send(ssif_info, msg, 3) != 0) {
 		/* Error, just go to normal state. */
-		ssif_info->ssif_state = SSIF_NORMAL;
+		ssif_info->ssif_state = SSIF_IDLE;
 	}
 }
 
@@ -382,7 +382,7 @@ static void start_flag_fetch(struct ssif
 	mb[0] = (IPMI_NETFN_APP_REQUEST << 2);
 	mb[1] = IPMI_GET_MSG_FLAGS_CMD;
 	if (start_send(ssif_info, mb, 2) != 0)
-		ssif_info->ssif_state = SSIF_NORMAL;
+		ssif_info->ssif_state = SSIF_IDLE;
 }
 
 static void check_start_send(struct ssif_info *ssif_info, unsigned long *flags,
@@ -393,7 +393,7 @@ static void check_start_send(struct ssif
 
 		flags = ipmi_ssif_lock_cond(ssif_info, &oflags);
 		ssif_info->curr_msg = NULL;
-		ssif_info->ssif_state = SSIF_NORMAL;
+		ssif_info->ssif_state = SSIF_IDLE;
 		ipmi_ssif_unlock_cond(ssif_info, flags);
 		ipmi_free_smi_msg(msg);
 	}
@@ -407,7 +407,7 @@ static void start_event_fetch(struct ssi
 
 	msg = ipmi_alloc_smi_msg();
 	if (!msg) {
-		ssif_info->ssif_state = SSIF_NORMAL;
+		ssif_info->ssif_state = SSIF_IDLE;
 		ipmi_ssif_unlock_cond(ssif_info, flags);
 		return;
 	}
@@ -430,7 +430,7 @@ static void start_recv_msg_fetch(struct
 
 	msg = ipmi_alloc_smi_msg();
 	if (!msg) {
-		ssif_info->ssif_state = SSIF_NORMAL;
+		ssif_info->ssif_state = SSIF_IDLE;
 		ipmi_ssif_unlock_cond(ssif_info, flags);
 		return;
 	}
@@ -448,9 +448,9 @@ static void start_recv_msg_fetch(struct
 
 /*
  * Must be called with the message lock held.  This will release the
- * message lock.  Note that the caller will check SSIF_IDLE and start a
- * new operation, so there is no need to check for new messages to
- * start in here.
+ * message lock.  Note that the caller will check IS_SSIF_IDLE and
+ * start a new operation, so there is no need to check for new
+ * messages to start in here.
  */
 static void handle_flags(struct ssif_info *ssif_info, unsigned long *flags)
 {
@@ -466,7 +466,7 @@ static void handle_flags(struct ssif_inf
 		/* Events available. */
 		start_event_fetch(ssif_info, flags);
 	else {
-		ssif_info->ssif_state = SSIF_NORMAL;
+		ssif_info->ssif_state = SSIF_IDLE;
 		ipmi_ssif_unlock_cond(ssif_info, flags);
 	}
 }
@@ -568,7 +568,7 @@ static void watch_timeout(struct timer_l
 	if (ssif_info->watch_timeout) {
 		mod_timer(&ssif_info->watch_timer,
 			  jiffies + ssif_info->watch_timeout);
-		if (SSIF_IDLE(ssif_info)) {
+		if (IS_SSIF_IDLE(ssif_info)) {
 			start_flag_fetch(ssif_info, flags); /* Releases lock */
 			return;
 		}
@@ -756,7 +756,7 @@ static void msg_done_handler(struct ssif
 	}
 
 	switch (ssif_info->ssif_state) {
-	case SSIF_NORMAL:
+	case SSIF_IDLE:
 		ipmi_ssif_unlock_cond(ssif_info, flags);
 		if (!msg)
 			break;
@@ -774,7 +774,7 @@ static void msg_done_handler(struct ssif
 			 * Error fetching flags, or invalid length,
 			 * just give up for now.
 			 */
-			ssif_info->ssif_state = SSIF_NORMAL;
+			ssif_info->ssif_state = SSIF_IDLE;
 			ipmi_ssif_unlock_cond(ssif_info, flags);
 			dev_warn(&ssif_info->client->dev,
 				 "Error getting flags: %d %d, %x\n",
@@ -809,7 +809,7 @@ static void msg_done_handler(struct ssif
 				 "Invalid response clearing flags: %x %x\n",
 				 data[0], data[1]);
 		}
-		ssif_info->ssif_state = SSIF_NORMAL;
+		ssif_info->ssif_state = SSIF_IDLE;
 		ipmi_ssif_unlock_cond(ssif_info, flags);
 		break;
 
@@ -887,7 +887,7 @@ static void msg_done_handler(struct ssif
 	}
 
 	flags = ipmi_ssif_lock_cond(ssif_info, &oflags);
-	if (SSIF_IDLE(ssif_info) && !ssif_info->stopping) {
+	if (IS_SSIF_IDLE(ssif_info) && !ssif_info->stopping) {
 		if (ssif_info->req_events)
 			start_event_fetch(ssif_info, flags);
 		else if (ssif_info->req_flags)
@@ -1032,7 +1032,7 @@ static void start_next_msg(struct ssif_i
 	unsigned long oflags;
 
  restart:
-	if (!SSIF_IDLE(ssif_info)) {
+	if (!IS_SSIF_IDLE(ssif_info)) {
 		ipmi_ssif_unlock_cond(ssif_info, flags);
 		return;
 	}
@@ -1255,7 +1255,7 @@ static void shutdown_ssif(void *send_inf
 	dev_set_drvdata(&ssif_info->client->dev, NULL);
 
 	/* make sure the driver is not looking for flags any more. */
-	while (ssif_info->ssif_state != SSIF_NORMAL)
+	while (ssif_info->ssif_state != SSIF_IDLE)
 		schedule_timeout(1);
 
 	ssif_info->stopping = true;
@@ -1825,7 +1825,7 @@ static int ssif_probe(struct i2c_client
 	}
 
 	spin_lock_init(&ssif_info->lock);
-	ssif_info->ssif_state = SSIF_NORMAL;
+	ssif_info->ssif_state = SSIF_IDLE;
 	timer_setup(&ssif_info->retry_timer, retry_timeout, 0);
 	timer_setup(&ssif_info->watch_timer, watch_timeout, 0);
 
author maintainer
#658 lore raw

PATCH 657/885 io_uring: Replace 0-length array with flexible array

In reply to Greg Kroah-Hartman

From: Kees Cook <[email protected]>

commit 36632d062975a9ff4410c90dd6d37922b68d0920 upstream.

Zero-length arrays are deprecated[1]. Replace struct io_uring_buf_ring's "bufs" with a flexible array member. (How is the size of this array verified?) Detected with GCC 13, using -fstrict-flex-arrays=3:

In function 'io_ring_buffer_select',

    inlined from 'io_buffer_select' at io_uring/kbuf.c:183:10:
io_uring/kbuf.c:141:23: warning: array subscript 255 is outside the bounds of an interior zero-length array 'struct io_uring_buf[0]' [-Wzero-length-bounds]
  141 |                 buf = &br->bufs[head];
      |                       ^~~~~~~~~~~~~~~
In file included from include/linux/io_uring.h:7,
                 from io_uring/kbuf.c:10:
include/uapi/linux/io_uring.h: In function 'io_buffer_select':
include/uapi/linux/io_uring.h:628:41: note: while referencing 'bufs'
  628 |                 struct io_uring_buf     bufs[0];
      |                                         ^~~~

[1] https://www.kernel.org/doc/html/latest/process/deprecated.html#zero-length-and-one-element-arrays

Fixes c7fb19428d67 ("io_uring: add support for ring mapped supplied buffers")
Cc Jens Axboe <[email protected]>
Cc Pavel Begunkov <[email protected]>
Cc "Gustavo A. R. Silva" <[email protected]>
---
 include/uapi/linux/io_uring.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
include/uapi/linux/io_uring.h +1 -1
@@ -617,7 +617,7 @@ struct io_uring_buf_ring {
 			__u16	resv3;
 			__u16	tail;
 		};
-		struct io_uring_buf	bufs[0];
+		__DECLARE_FLEX_ARRAY(struct io_uring_buf, bufs);
 	};
 };
 
author maintainer
#659 lore raw

PATCH 658/885 io_uring: use user visible tail in io_uring_poll()

In reply to Greg Kroah-Hartman

From: Pavel Begunkov <[email protected]>

commit c10bb64684813a326174c3eebcafb3ee5af52ca3 upstream.

We return POLLIN from io_uring_poll() depending on whether there are CQEs for the userspace, and so we should use the user visible tail pointer instead of a transient cached value.

---
 io_uring/io_uring.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
io_uring/io_uring.c +1 -1
@@ -2653,7 +2653,7 @@ static __poll_t io_uring_poll(struct fil
 	 * pushs them to do the flush.
 	 */
 
-	if (io_cqring_events(ctx) || io_has_work(ctx))
+	if (__io_cqring_events_user(ctx) || io_has_work(ctx))
 		mask |= EPOLLIN | EPOLLRDNORM;
 
 	return mask;
author maintainer
#660 lore raw

PATCH 659/885 io_uring: handle TIF_NOTIFY_RESUME when checking for task_work

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit b5d3ae202fbfe055aa2a8ae8524531ee1dcab717 upstream.

If TIF_NOTIFY_RESUME is set, then we need to call resume_user_mode_work() for PF_IO_WORKER threads. They never return to usermode, hence never get a chance to process any items that are marked by this flag. Most notably this includes the final put of files, but also any throttling markers set by block cgroups.

---
 io_uring/io_uring.h |    8 ++++++++
 1 file changed, 8 insertions(+)
io_uring/io_uring.h +8 -0
@@ -3,6 +3,7 @@
 
 #include <linux/errno.h>
 #include <linux/lockdep.h>
+#include <linux/resume_user_mode.h>
 #include <linux/io_uring_types.h>
 #include <uapi/linux/eventpoll.h>
 #include "io-wq.h"
@@ -255,6 +256,13 @@ static inline int io_run_task_work(void)
 	 */
 	if (test_thread_flag(TIF_NOTIFY_SIGNAL))
 		clear_notify_signal();
+	/*
+	 * PF_IO_WORKER never returns to userspace, so check here if we have
+	 * notify work that needs processing.
+	 */
+	if (current->flags & PF_IO_WORKER &&
+	    test_thread_flag(TIF_NOTIFY_RESUME))
+		resume_user_mode_work(NULL);
 	if (task_work_pending(current)) {
 		__set_current_state(TASK_RUNNING);
 		task_work_run();
author maintainer
#661 lore raw

PATCH 660/885 io_uring: add a conditional reschedule to the IOPOLL cancelation loop

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit fcc926bb857949dbfa51a7d95f3f5ebc657f198c upstream.

If the kernel is configured with CONFIG_PREEMPT_NONE, we could be sitting in a tight loop reaping events but not giving them a chance to finish. This results in a trace ala:

rcu: INFO: rcu_sched self-detected stall on CPU
rcu: 	2-...!: (5249 ticks this GP) idle=935c/1/0x4000000000000000 softirq=4265/4274 fqs=1
	(t=5251 jiffies g=465 q=4135 ncpus=4)

rcu: rcu_sched kthread starved for 5249 jiffies! g465 f0x0 RCU_GP_WAIT_FQS(5) ->state=0x0 ->cpu=0
rcu: Unless rcu_sched kthread gets sufficient CPU time, OOM is now expected behavior.
rcu: RCU grace-period kthread stack dump:
task:rcu_sched state:R running task stack:0 pid:12 ppid:2 flags:0x00000008

Call trace:
 __switch_to+0xb0/0xc8
 __schedule+0x43c/0x520
 schedule+0x4c/0x98
 schedule_timeout+0xbc/0xdc
 rcu_gp_fqs_loop+0x308/0x344
 rcu_gp_kthread+0xd8/0xf0
 kthread+0xb8/0xc8
 ret_from_fork+0x10/0x20
rcu: Stack dump where RCU GP kthread last ran:
Task dump for CPU 0:
task:kworker/u8:10   state:R  running task     stack:0     pid:89    ppid:2      flags:0x0000000a
Workqueue: events_unbound io_ring_exit_work
Call trace:
 __switch_to+0xb0/0xc8
 0xffff0000c8fefd28
CPU: 2 PID: 95 Comm: kworker/u8:13 Not tainted 6.2.0-rc5-00042-g40316e337c80-dirty #2759
Hardware name: linux,dummy-virt (DT)
Workqueue: events_unbound io_ring_exit_work
pstate: 61400005 (nZCv daif +PAN -UAO -TCO +DIT -SSBS BTYPE=--)
pc : io_do_iopoll+0x344/0x360
lr : io_do_iopoll+0xb8/0x360

sp : ffff800009bebc60
x29: ffff800009bebc60 x28: 0000000000000000 x27: 0000000000000000
x26: ffff0000c0f67d48 x25: ffff0000c0f67840 x24: ffff800008950024
x23: 0000000000000001 x22: 0000000000000000 x21: ffff0000c27d3200
x20: ffff0000c0f67840 x19: ffff0000c0f67800 x18: 0000000000000000
x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000000
x14: 0000000000000001 x13: 0000000000000001 x12: 0000000000000000
x11: 0000000000000179 x10: 0000000000000870 x9 : ffff800009bebd60 x8 : ffff0000c27d3ad0 x7 : fefefefefefefeff x6 : 0000646e756f626e x5 : ffff0000c0f67840 x4 : 0000000000000000 x3 : ffff0000c2398000 x2 : 0000000000000000 x1 : 0000000000000000 x0 : 0000000000000000

Call trace:
 io_do_iopoll+0x344/0x360
 io_uring_try_cancel_requests+0x21c/0x334
 io_ring_exit_work+0x90/0x40c
 process_one_work+0x1a4/0x254
 worker_thread+0x1ec/0x258
 kthread+0xb8/0xc8
 ret_from_fork+0x10/0x20

Add a cond_resched() in the cancelation IOPOLL loop to fix this.

---
 io_uring/io_uring.c |    1 +
 1 file changed, 1 insertion(+)
io_uring/io_uring.c +1 -0
@@ -2912,6 +2912,7 @@ static __cold bool io_uring_try_cancel_r
 		while (!wq_list_empty(&ctx->iopoll_list)) {
 			io_iopoll_try_reap_events(ctx);
 			ret = true;
+			cond_resched();
 		}
 	}
 
author maintainer
#662 lore raw

PATCH 661/885 io_uring: add reschedule point to handle_tw_list()

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit f58680085478dd292435727210122960d38e8014 upstream.

If CONFIG_PREEMPT_NONE is set and the task_work chains are long, we could be running into issues blocking others for too long. Add a reschedule check in handle_tw_list(), and flush the ctx if we need to reschedule.

---
 io_uring/io_uring.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
io_uring/io_uring.c +7 -1
@@ -1030,10 +1030,16 @@ static unsigned int handle_tw_list(struc
 			/* if not contended, grab and improve batching */
 			*locked = mutex_trylock(&(*ctx)->uring_lock);
 			percpu_ref_get(&(*ctx)->refs);
-		}
+		} else if (!*locked)
+			*locked = mutex_trylock(&(*ctx)->uring_lock);
 		req->io_task_work.func(req, locked);
 		node = next;
 		count++;
+		if (unlikely(need_resched())) {
+			ctx_flush_and_put(*ctx, locked);
+			*ctx = NULL;
+			cond_resched();
+		}
 	}
 
 	return count;
author maintainer
#663 lore raw

PATCH 662/885 io_uring/rsrc: disallow multi-source reg buffers

In reply to Greg Kroah-Hartman

From: Pavel Begunkov <[email protected]>

commit edd478269640b360c6f301f2baa04abdda563ef3 upstream.

If two or more mappings go back to back to each other they can be passed into io_uring to be registered as a single registered buffer. That would even work if mappings came from different sources, e.g. it's possible to mix in this way anon pages and pages from shmem or hugetlb. That is not a problem but it'd rather be less prone if we forbid such mixing.

---
 io_uring/rsrc.c |   13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
io_uring/rsrc.c +8 -5
@@ -1147,14 +1147,17 @@ struct page **io_pin_pages(unsigned long
 	pret = pin_user_pages(ubuf, nr_pages, FOLL_WRITE | FOLL_LONGTERM,
 			      pages, vmas);
 	if (pret == nr_pages) {
+		struct file *file = vmas[0]->vm_file;
+
 		/* don't support file backed memory */
 		for (i = 0; i < nr_pages; i++) {
-			struct vm_area_struct *vma = vmas[i];
-
-			if (vma_is_shmem(vma))
+			if (vmas[i]->vm_file != file) {
+				ret = -EINVAL;
+				break;
+			}
+			if (!file)
 				continue;
-			if (vma->vm_file &&
-			    !is_file_hugepages(vma->vm_file)) {
+			if (!vma_is_shmem(vmas[i]) && !is_file_hugepages(file)) {
 				ret = -EOPNOTSUPP;
 				break;
 			}
author maintainer
#664 lore raw

PATCH 663/885 io_uring: remove MSG_NOSIGNAL from recvmsg

In reply to Greg Kroah-Hartman

From: David Lamparter <[email protected]>

commit 7605c43d67face310b4b87dee1a28bc0c8cd8c0f upstream.

MSG_NOSIGNAL is not applicable for the receiving side, SIGPIPE is generated when trying to write to a "broken pipe". AF_PACKET's packet_recvmsg() does enforce this, giving back EINVAL when MSG_NOSIGNAL is set - making it unuseable in io_uring's recvmsg.

Remove MSG_NOSIGNAL from io_recvmsg_prep().

---
 io_uring/net.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
io_uring/net.c +1 -1
@@ -553,7 +553,7 @@ int io_recvmsg_prep(struct io_kiocb *req
 	sr->flags = READ_ONCE(sqe->ioprio);
 	if (sr->flags & ~(RECVMSG_FLAGS))
 		return -EINVAL;
-	sr->msg_flags = READ_ONCE(sqe->msg_flags) | MSG_NOSIGNAL;
+	sr->msg_flags = READ_ONCE(sqe->msg_flags);
 	if (sr->msg_flags & MSG_DONTWAIT)
 		req->flags |= REQ_F_NOWAIT;
 	if (sr->msg_flags & MSG_ERRQUEUE)
author maintainer
#665 lore raw

PATCH 664/885 io_uring: fix fget leak when fs dont support nowait buffered read

In reply to Greg Kroah-Hartman

From: Joseph Qi <[email protected]>

commit 54aa7f2330b82884f4a1afce0220add6e8312f8b upstream.

Heming reported a BUG when using io_uring doing link-cp on ocfs2. [1]

Do the following steps can reproduce this BUG:
mount -t ocfs2 /dev/vdc /mnt/ocfs2
cp testfile /mnt/ocfs2/
./link-cp /mnt/ocfs2/testfile /mnt/ocfs2/testfile.1 umount /mnt/ocfs2

Then umount will fail, and it outputs:
umount: /mnt/ocfs2: target is busy.

While tracing umount, it blames mnt_get_count() not return as expected. Do a deep investigation for fget()/fput() on related code flow, I've finally found that fget() leaks since ocfs2 doesn't support nowait buffered read.

io_issue_sqe
|-io_assign_file // do fget() first
|-io_read
|-io_iter_do_read
|-ocfs2_file_read_iter // return -EOPNOTSUPP
|-kiocb_done

    |-io_rw_done
      |-__io_complete_rw_common  // set REQ_F_REISSUE
    |-io_resubmit_prep
      |-io_req_prep_async  // override req->file, leak happens

This was introduced by commit a196c78b5443 in v5.18. Fix it by don't re-assign req->file if it has already been assigned.

[1] https://lore.kernel.org/ocfs2-devel/[email protected]/T/#t

Fixes a196c78b5443 ("io_uring: assign non-fixed early for async work")
Cc Xiaoguang Wang <[email protected]>
---
 io_uring/io_uring.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
io_uring/io_uring.c +1 -1
@@ -1597,7 +1597,7 @@ int io_req_prep_async(struct io_kiocb *r
 	const struct io_op_def *def = &io_op_defs[req->opcode];
 
 	/* assign early for deferred execution for non-fixed file */
-	if (def->needs_file && !(req->flags & REQ_F_FIXED_FILE))
+	if (def->needs_file && !(req->flags & REQ_F_FIXED_FILE) && !req->file)
 		req->file = io_file_get_normal(req, req->cqe.fd);
 	if (!def->prep_async)
 		return 0;
author maintainer
#666 lore raw

PATCH 665/885 s390/extmem: return correct segment type in __segment_load()

In reply to Greg Kroah-Hartman

From: Gerald Schaefer <[email protected]>

commit 8c42dd78df148c90e48efff204cce38743906a79 upstream.

Commit f05f62d04271f ("s390/vmem: get rid of memory segment list") reshuffled the call to vmem_add_mapping() in __segment_load(), which now overwrites rc after it was set to contain the segment type code.

As result, __segment_load() will now always return 0 on success, which corresponds to the segment type code SEG_TYPE_SW, i.e. a writeable segment. This results in a kernel crash when loading a read-only segment as dcssblk block device, and trying to write to it.

Instead of reshuffling code again, make sure to return the segment type on success, and also describe this rather delicate and unexpected logic in the function comment. Also initialize new segtype variable with invalid value, to prevent possible future confusion.

Fixes f05f62d04271 ("s390/vmem: get rid of memory segment list")
Cc <[email protected]> # 5.9+
---
 arch/s390/mm/extmem.c |   12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)
arch/s390/mm/extmem.c +7 -5
@@ -289,15 +289,17 @@ segment_overlaps_others (struct dcss_seg
 
 /*
  * real segment loading function, called from segment_load
+ * Must return either an error code < 0, or the segment type code >= 0
  */
 static int
 __segment_load (char *name, int do_nonshared, unsigned long *addr, unsigned long *end)
 {
 	unsigned long start_addr, end_addr, dummy;
 	struct dcss_segment *seg;
-	int rc, diag_cc;
+	int rc, diag_cc, segtype;
 
 	start_addr = end_addr = 0;
+	segtype = -1;
 	seg = kmalloc(sizeof(*seg), GFP_KERNEL | GFP_DMA);
 	if (seg == NULL) {
 		rc = -ENOMEM;
@@ -326,9 +328,9 @@ __segment_load (char *name, int do_nonsh
 	seg->res_name[8] = '\0';
 	strlcat(seg->res_name, " (DCSS)", sizeof(seg->res_name));
 	seg->res->name = seg->res_name;
-	rc = seg->vm_segtype;
-	if (rc == SEG_TYPE_SC ||
-	    ((rc == SEG_TYPE_SR || rc == SEG_TYPE_ER) && !do_nonshared))
+	segtype = seg->vm_segtype;
+	if (segtype == SEG_TYPE_SC ||
+	    ((segtype == SEG_TYPE_SR || segtype == SEG_TYPE_ER) && !do_nonshared))
 		seg->res->flags |= IORESOURCE_READONLY;
 
 	/* Check for overlapping resources before adding the mapping. */
@@ -386,7 +388,7 @@ __segment_load (char *name, int do_nonsh
  out_free:
 	kfree(seg);
  out:
-	return rc;
+	return rc < 0 ? rc : segtype;
 }
 
 /*
author maintainer
#667 lore raw

PATCH 666/885 s390: discard .interp section

In reply to Greg Kroah-Hartman

From: Ilya Leoshkevich <[email protected]>

commit e9c9cb90e76ffaabcc7ca8f275d9e82195fd6367 upstream.

When debugging vmlinux with QEMU + GDB, the following GDB error may occur:

    (gdb) c
    Continuing.
    Warning:
    Cannot insert breakpoint -1.
    Cannot access memory at address 0xffffffffffff95c0

    Command aborted.
    (gdb)

The reason is that, when .interp section is present, GDB tries to locate the file specified in it in memory and put a number of breakpoints there (see enable_break() function in gdb/solib-svr4.c). Sometimes GDB finds a bogus location that matches its heuristics, fails to set a breakpoint and stops. This makes further debugging impossible.

The .interp section contains misleading information anyway (vmlinux does not need ld.so), so fix by discarding it.

---
 arch/s390/kernel/vmlinux.lds.S |    1 +
 1 file changed, 1 insertion(+)
arch/s390/kernel/vmlinux.lds.S +1 -0
@@ -228,5 +228,6 @@ SECTIONS
 	DISCARDS
 	/DISCARD/ : {
 		*(.eh_frame)
+		*(.interp)
 	}
 }
author maintainer
#668 lore raw

PATCH 667/885 s390/kprobes: fix irq mask clobbering on kprobe reenter from post_handler

In reply to Greg Kroah-Hartman

From: Vasily Gorbik <[email protected]>

commit 42e19e6f04984088b6f9f0507c4c89a8152d9730 upstream.

Recent test_kprobe_missed kprobes kunit test uncovers the following error (reported when CONFIG_DEBUG_ATOMIC_SLEEP is enabled):

BUG: sleeping function called from invalid context at kernel/locking/mutex.c:580

in_atomic(): 0, irqs_disabled(): 1, non_block: 0, pid: 662, name: kunit_try_catch
preempt_count: 0, expected: 0
RCU nest depth: 0, expected: 0
no locks held by kunit_try_catch/662.

irq event stamp: 280
hardirqs last  enabled at (279): [<00000003e60a3d42>] __do_pgm_check+0x17a/0x1c0
hardirqs last disabled at (280): [<00000003e3bd774a>] kprobe_exceptions_notify+0x27a/0x318
softirqs last  enabled at (0): [<00000003e3c5c890>] copy_process+0x14a8/0x4c80
softirqs last disabled at (0): [<0000000000000000>] 0x0
CPU: 46 PID: 662 Comm: kunit_try_catch Tainted: G                 N 6.2.0-173644-g44c18d77f0c0 #2
Hardware name: IBM 3931 A01 704 (LPAR)
Call Trace:
 [<00000003e60a3a00>] dump_stack_lvl+0x120/0x198
 [<00000003e3d02e82>] __might_resched+0x60a/0x668
 [<00000003e60b9908>] __mutex_lock+0xc0/0x14e0
 [<00000003e60bad5a>] mutex_lock_nested+0x32/0x40
 [<00000003e3f7b460>] unregister_kprobe+0x30/0xd8
 [<00000003e51b2602>] test_kprobe_missed+0xf2/0x268
 [<00000003e51b5406>] kunit_try_run_case+0x10e/0x290
 [<00000003e51b7dfa>] kunit_generic_run_threadfn_adapter+0x62/0xb8
 [<00000003e3ce30f8>] kthread+0x2d0/0x398
 [<00000003e3b96afa>] __ret_from_fork+0x8a/0xe8
 [<00000003e60ccada>] ret_from_fork+0xa/0x40

The reason for this error report is that kprobes handling code failed to restore irqs.

The problem is that when kprobe is triggered from another kprobe post_handler current sequence of enable_singlestep / disable_singlestep is the following:
enable_singlestep <- original kprobe (saves kprobe_saved_imask) enable_singlestep <- kprobe triggered from post_handler (clobbers kprobe_saved_imask) disable_singlestep <- kprobe triggered from post_handler (restores kprobe_saved_imask) disable_singlestep <- original kprobe (restores wrong clobbered kprobe_saved_imask)

There is just one kprobe_ctlblk per cpu and both calls saves and loads irq mask to kprobe_saved_imask. To fix the problem simply move resume_execution (which calls disable_singlestep) before calling post_handler. This also fixes the problem that post_handler is called with pt_regs which were not yet adjusted after single-stepping.

---
 arch/s390/kernel/kprobes.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
arch/s390/kernel/kprobes.c +1 -2
@@ -433,12 +433,11 @@ static int post_kprobe_handler(struct pt
 	if (!p)
 		return 0;
 
+	resume_execution(p, regs);
 	if (kcb->kprobe_status != KPROBE_REENTER && p->post_handler) {
 		kcb->kprobe_status = KPROBE_HIT_SSDONE;
 		p->post_handler(p, regs, 0);
 	}
-
-	resume_execution(p, regs);
 	pop_kprobe(kcb);
 	preempt_enable_no_resched();
 
author maintainer
#669 lore raw

PATCH 668/885 s390/kprobes: fix current_kprobe never cleared after kprobes reenter

In reply to Greg Kroah-Hartman

From: Vasily Gorbik <[email protected]>

commit cd57953936f2213dfaccce10d20f396956222c7d upstream.

Recent test_kprobe_missed kprobes kunit test uncovers the following problem. Once kprobe is triggered from another kprobe (kprobe reenter), all future kprobes on this cpu are considered as kprobe reenter, thus pre_handler and post_handler are not being called and kprobes are counted as "missed".

Commit b9599798f953 ("[S390] kprobes: activation and deactivation") introduced a simpler scheme for kprobes (de)activation and status tracking by using push_kprobe/pop_kprobe, which supposed to work for both initial kprobe entry as well as kprobe reentry and helps to avoid handling those two cases differently. The problem is that a sequence of calls in case of kprobes reenter:
push_kprobe() <- NULL (current_kprobe)
push_kprobe() <- kprobe1 (current_kprobe)
pop_kprobe() -> kprobe1 (current_kprobe)
pop_kprobe() -> kprobe1 (current_kprobe)
leaves "kprobe1" as "current_kprobe" on this cpu, instead of setting it to NULL. In fact push_kprobe/pop_kprobe can only store a single state (there is just one prev_kprobe in kprobe_ctlblk). Which is a hack but sufficient, there is no need to have another prev_kprobe just to store NULL. To make a simple and backportable fix simply reset "prev_kprobe" when kprobe is poped from this "stack". No need to worry about "kprobe_status" in this case, because its value is only checked when current_kprobe != NULL.

Fixes b9599798f953 ("[S390] kprobes: activation and deactivation")
---
 arch/s390/kernel/kprobes.c |    1 +
 1 file changed, 1 insertion(+)
arch/s390/kernel/kprobes.c +1 -0
@@ -279,6 +279,7 @@ static void pop_kprobe(struct kprobe_ctl
 {
 	__this_cpu_write(current_kprobe, kcb->prev_kprobe.kp);
 	kcb->kprobe_status = kcb->prev_kprobe.status;
+	kcb->prev_kprobe.kp = NULL;
 }
 NOKPROBE_SYMBOL(pop_kprobe);
 
author maintainer
#670 lore raw

PATCH 669/885 KVM: s390: disable migration mode when dirty tracking is disabled

In reply to Greg Kroah-Hartman

From: Nico Boehr <[email protected]>

commit f2d3155e2a6bac44d16f04415a321e8707d895c6 upstream.

Migration mode is a VM attribute which enables tracking of changes in storage attributes (PGSTE). It assumes dirty tracking is enabled on all memslots to keep a dirty bitmap of pages with changed storage attributes.

When enabling migration mode, we currently check that dirty tracking is enabled for all memslots. However, userspace can disable dirty tracking without disabling migration mode.

Since migration mode is pointless with dirty tracking disabled, disable migration mode whenever userspace disables dirty tracking on any slot.

Also update the documentation to clarify that dirty tracking must be enabled when enabling migration mode, which is already enforced by the code in kvm_s390_vm_start_migration().

Also highlight in the documentation for KVM_S390_GET_CMMA_BITS that it can now fail with -EINVAL when dirty tracking is disabled while migration mode is on. Move all the error codes to a table so this stays readable.

To disable migration mode, slots_lock should be held, which is taken in kvm_set_memory_region() and thus held in kvm_arch_prepare_memory_region().

Restructure the prepare code a bit so all the sanity checking is done before disabling migration mode. This ensures migration mode isn't disabled when some sanity check fails.

Fixes 190df4a212a7 ("KVM: s390: CMMA tracking, ESSA emulation, migration mode")

Message-Id: <[email protected]>

[[email protected]: fixed commit message typo, moved api.rst error table upwards]
---
 Documentation/virt/kvm/api.rst        |   18 ++++++++-----
 Documentation/virt/kvm/devices/vm.rst |    4 ++
 arch/s390/kvm/kvm-s390.c              |   47 +++++++++++++++++++++++-----------
 3 files changed, 48 insertions(+), 21 deletions(-)
Documentation/virt/kvm/api.rst +12 -6
@@ -4483,6 +4483,18 @@ not holding a previously reported uncorr
 :Parameters: struct kvm_s390_cmma_log (in, out)
 :Returns: 0 on success, a negative value on error
 
+Errors:
+
+  ======     =============================================================
+  ENOMEM     not enough memory can be allocated to complete the task
+  ENXIO      if CMMA is not enabled
+  EINVAL     if KVM_S390_CMMA_PEEK is not set but migration mode was not enabled
+  EINVAL     if KVM_S390_CMMA_PEEK is not set but dirty tracking has been
+             disabled (and thus migration mode was automatically disabled)
+  EFAULT     if the userspace address is invalid or if no page table is
+             present for the addresses (e.g. when using hugepages).
+  ======     =============================================================
+
 This ioctl is used to get the values of the CMMA bits on the s390
 architecture. It is meant to be used in two scenarios:
 
@@ -4563,12 +4575,6 @@ mask is unused.
 
 values points to the userspace buffer where the result will be stored.
 
-This ioctl can fail with -ENOMEM if not enough memory can be allocated to
-complete the task, with -ENXIO if CMMA is not enabled, with -EINVAL if
-KVM_S390_CMMA_PEEK is not set but migration mode was not enabled, with
--EFAULT if the userspace address is invalid or if no page table is
-present for the addresses (e.g. when using hugepages).
-
 4.108 KVM_S390_SET_CMMA_BITS
 ----------------------------
 
Documentation/virt/kvm/devices/vm.rst +4 -0
@@ -302,6 +302,10 @@ Allows userspace to start migration mode
 Setting this attribute when migration mode is already active will have
 no effects.
 
+Dirty tracking must be enabled on all memslots, else -EINVAL is returned. When
+dirty tracking is disabled on any memslot, migration mode is automatically
+stopped.
+
 :Parameters: none
 :Returns:   -ENOMEM if there is not enough free memory to start migration mode;
 	    -EINVAL if the state of the VM is invalid (e.g. no memory defined);
arch/s390/kvm/kvm-s390.c +32 -15

55 lines not on this page · raw message

author maintainer
#671 lore raw

PATCH 670/885 cifs: Fix uninitialized memory read in smb3_qfs_tcon()

In reply to Greg Kroah-Hartman

From: Volker Lendecke <[email protected]>

commit d447e794a37288ec7a080aa1b044a8d9deebbab7 upstream.

oparms was not fully initialized

---
 fs/cifs/smb2ops.c |   13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
fs/cifs/smb2ops.c +7 -6
@@ -729,12 +729,13 @@ smb3_qfs_tcon(const unsigned int xid, st
 	struct cifs_fid fid;
 	struct cached_fid *cfid = NULL;
 
-	oparms.tcon = tcon;
-	oparms.desired_access = FILE_READ_ATTRIBUTES;
-	oparms.disposition = FILE_OPEN;
-	oparms.create_options = cifs_create_options(cifs_sb, 0);
-	oparms.fid = &fid;
-	oparms.reconnect = false;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.desired_access = FILE_READ_ATTRIBUTES,
+		.disposition = FILE_OPEN,
+		.create_options = cifs_create_options(cifs_sb, 0),
+		.fid = &fid,
+	};
 
 	rc = open_cached_dir(xid, tcon, "", cifs_sb, false, &cfid);
 	if (rc == 0)
author maintainer
#672 lore raw

PATCH 671/885 cifs: Fix uninitialized memory reads for oparms.mode

In reply to Greg Kroah-Hartman

From: Volker Lendecke <[email protected]>

commit de036dcaca65cf94bf7ff09c571c077f02bc92b4 upstream.

Use a struct assignment with implicit member initialization

---
 fs/cifs/cached_dir.c |   13 +--
 fs/cifs/cifsacl.c    |   34 ++++-----
 fs/cifs/cifssmb.c    |   17 ++--
 fs/cifs/dir.c        |   19 ++---
 fs/cifs/file.c       |   35 +++++----
 fs/cifs/inode.c      |   53 +++++++-------
 fs/cifs/link.c       |   66 +++++++++--------
 fs/cifs/smb1ops.c    |   72 ++++++++++---------
 fs/cifs/smb2inode.c  |   17 ++--
 fs/cifs/smb2ops.c    |  191 ++++++++++++++++++++++++++-------------------------
 10 files changed, 274 insertions(+), 243 deletions(-)
fs/cifs/cached_dir.c +7 -6
@@ -181,12 +181,13 @@ int open_cached_dir(unsigned int xid, st
 	rqst[0].rq_iov = open_iov;
 	rqst[0].rq_nvec = SMB2_CREATE_IOV_SIZE;
 
-	oparms.tcon = tcon;
-	oparms.create_options = cifs_create_options(cifs_sb, CREATE_NOT_FILE);
-	oparms.desired_access = FILE_READ_ATTRIBUTES;
-	oparms.disposition = FILE_OPEN;
-	oparms.fid = pfid;
-	oparms.reconnect = false;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.create_options = cifs_create_options(cifs_sb, CREATE_NOT_FILE),
+		.desired_access = FILE_READ_ATTRIBUTES,
+		.disposition = FILE_OPEN,
+		.fid = pfid,
+	};
 
 	rc = SMB2_open_init(tcon, server,
 			    &rqst[0], &oplock, &oparms, utf16_path);
fs/cifs/cifsacl.c +18 -16
@@ -1423,14 +1423,15 @@ static struct cifs_ntsd *get_cifs_acl_by
 	tcon = tlink_tcon(tlink);
 	xid = get_xid();
 
-	oparms.tcon = tcon;
-	oparms.cifs_sb = cifs_sb;
-	oparms.desired_access = READ_CONTROL;
-	oparms.create_options = cifs_create_options(cifs_sb, 0);
-	oparms.disposition = FILE_OPEN;
-	oparms.path = path;
-	oparms.fid = &fid;
-	oparms.reconnect = false;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.cifs_sb = cifs_sb,
+		.desired_access = READ_CONTROL,
+		.create_options = cifs_create_options(cifs_sb, 0),
+		.disposition = FILE_OPEN,
+		.path = path,
+		.fid = &fid,
+	};
 
 	rc = CIFS_open(xid, &oparms, &oplock, NULL);
 	if (!rc) {
@@ -1489,14 +1490,15 @@ int set_cifs_acl(struct cifs_ntsd *pnnts
 	else
 		access_flags = WRITE_DAC;
 
-	oparms.tcon = tcon;
-	oparms.cifs_sb = cifs_sb;
-	oparms.desired_access = access_flags;
-	oparms.create_options = cifs_create_options(cifs_sb, 0);
-	oparms.disposition = FILE_OPEN;
-	oparms.path = path;
-	oparms.fid = &fid;
-	oparms.reconnect = false;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.cifs_sb = cifs_sb,
+		.desired_access = access_flags,
+		.create_options = cifs_create_options(cifs_sb, 0),
+		.disposition = FILE_OPEN,
+		.path = path,
+		.fid = &fid,
+	};
 
 	rc = CIFS_open(xid, &oparms, &oplock, NULL);
 	if (rc) {
fs/cifs/cifssmb.c +9 -8
@@ -5314,14 +5314,15 @@ CIFSSMBSetPathInfoFB(const unsigned int
 	struct cifs_fid fid;
 	int rc;
 
-	oparms.tcon = tcon;
-	oparms.cifs_sb = cifs_sb;
-	oparms.desired_access = GENERIC_WRITE;
-	oparms.create_options = cifs_create_options(cifs_sb, 0);
-	oparms.disposition = FILE_OPEN;
-	oparms.path = fileName;
-	oparms.fid = &fid;
-	oparms.reconnect = false;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.cifs_sb = cifs_sb,
+		.desired_access = GENERIC_WRITE,
+		.create_options = cifs_create_options(cifs_sb, 0),
+		.disposition = FILE_OPEN,
+		.path = fileName,
+		.fid = &fid,
+	};
 
 	rc = CIFS_open(xid, &oparms, &oplock, NULL);
 	if (rc)
fs/cifs/dir.c +10 -9
@@ -295,15 +295,16 @@ static int cifs_do_create(struct inode *
 	if (!tcon->unix_ext && (mode & S_IWUGO) == 0)
 		create_options |= CREATE_OPTION_READONLY;
 
-	oparms.tcon = tcon;
-	oparms.cifs_sb = cifs_sb;
-	oparms.desired_access = desired_access;
-	oparms.create_options = cifs_create_options(cifs_sb, create_options);
-	oparms.disposition = disposition;
-	oparms.path = full_path;
-	oparms.fid = fid;
-	oparms.reconnect = false;
-	oparms.mode = mode;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.cifs_sb = cifs_sb,
+		.desired_access = desired_access,
+		.create_options = cifs_create_options(cifs_sb, create_options),
+		.disposition = disposition,
+		.path = full_path,
+		.fid = fid,
+		.mode = mode,
+	};
 	rc = server->ops->open(xid, &oparms, oplock, buf);
 	if (rc) {
 		cifs_dbg(FYI, "cifs_create returned 0x%x\n", rc);
fs/cifs/file.c +19 -16
@@ -260,14 +260,15 @@ static int cifs_nt_open(const char *full
 	if (f_flags & O_DIRECT)
 		create_options |= CREATE_NO_BUFFER;
 
-	oparms.tcon = tcon;
-	oparms.cifs_sb = cifs_sb;
-	oparms.desired_access = desired_access;
-	oparms.create_options = cifs_create_options(cifs_sb, create_options);
-	oparms.disposition = disposition;
-	oparms.path = full_path;
-	oparms.fid = fid;
-	oparms.reconnect = false;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.cifs_sb = cifs_sb,
+		.desired_access = desired_access,
+		.create_options = cifs_create_options(cifs_sb, create_options),
+		.disposition = disposition,
+		.path = full_path,
+		.fid = fid,
+	};
 
 	rc = server->ops->open(xid, &oparms, oplock, buf);
 	if (rc)
@@ -848,14 +849,16 @@ cifs_reopen_file(struct cifsFileInfo *cf
 	if (server->ops->get_lease_key)
 		server->ops->get_lease_key(inode, &cfile->fid);
 
-	oparms.tcon = tcon;
-	oparms.cifs_sb = cifs_sb;
-	oparms.desired_access = desired_access;
-	oparms.create_options = cifs_create_options(cifs_sb, create_options);
-	oparms.disposition = disposition;
-	oparms.path = full_path;
-	oparms.fid = &cfile->fid;
-	oparms.reconnect = true;
+	oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.cifs_sb = cifs_sb,
+		.desired_access = desired_access,
+		.create_options = cifs_create_options(cifs_sb, create_options),
+		.disposition = disposition,
+		.path = full_path,
+		.fid = &cfile->fid,
+		.reconnect = true,
+	};
 
 	/*
 	 * Can not refresh inode by passing in file_info buf to be returned by
fs/cifs/inode.c +28 -25

71 lines not on this page · raw message

fs/cifs/link.c +35 -31

90 lines not on this page · raw message

fs/cifs/smb1ops.c +38 -34

96 lines not on this page · raw message

fs/cifs/smb2inode.c +9 -8
@@ -104,14 +104,15 @@ static int smb2_compound_op(const unsign
 		goto finished;
 	}
 
-	vars->oparms.tcon = tcon;
-	vars->oparms.desired_access = desired_access;
-	vars->oparms.disposition = create_disposition;
-	vars->oparms.create_options = cifs_create_options(cifs_sb, create_options);
-	vars->oparms.fid = &fid;
-	vars->oparms.reconnect = false;
-	vars->oparms.mode = mode;
-	vars->oparms.cifs_sb = cifs_sb;
+	vars->oparms = (struct cifs_open_parms) {
+		.tcon = tcon,
+		.desired_access = desired_access,
+		.disposition = create_disposition,
+		.create_options = cifs_create_options(cifs_sb, create_options),
+		.fid = &fid,
+		.mode = mode,
+		.cifs_sb = cifs_sb,
+	};
 
 	rqst[num_rqst].rq_iov = &vars->open_iov[0];
 	rqst[num_rqst].rq_nvec = SMB2_CREATE_IOV_SIZE;
fs/cifs/smb2ops.c +101 -90

269 lines not on this page · raw message

author maintainer
#673 lore raw

PATCH 672/885 cifs: fix mount on old smb servers

In reply to Greg Kroah-Hartman

From: Paulo Alcantara <[email protected]>

commit d99e86ebde2d7b3a04190f8d14de5bf6814bf10f upstream.

The client was sending rfc1002 session request packet with a wrong length field set, therefore failing to mount shares against old SMB servers over port 139.

Fix this by calculating the correct length as specified in rfc1002.

Fixes d7173623bf0b ("cifs: use ALIGN() and round_up() macros")
---
 fs/cifs/connect.c | 94 ++++++++++++++++++-----------------------------
 1 file changed, 35 insertions(+), 59 deletions(-)
fs/cifs/connect.c +35 -59
index b2a04b4e89a5..af49ae53aaf4 100644
@@ -2843,72 +2843,48 @@ ip_rfc1001_connect(struct TCP_Server_Info *server)
 	 * negprot - BB check reconnection in case where second
 	 * sessinit is sent but no second negprot
 	 */
-	struct rfc1002_session_packet *ses_init_buf;
-	unsigned int req_noscope_len;
-	struct smb_hdr *smb_buf;
+	struct rfc1002_session_packet req = {};
+	struct smb_hdr *smb_buf = (struct smb_hdr *)&req;
+	unsigned int len;
 
-	ses_init_buf = kzalloc(sizeof(struct rfc1002_session_packet),
-			       GFP_KERNEL);
+	req.trailer.session_req.called_len = sizeof(req.trailer.session_req.called_name);
 
-	if (ses_init_buf) {
-		ses_init_buf->trailer.session_req.called_len = 32;
+	if (server->server_RFC1001_name[0] != 0)
+		rfc1002mangle(req.trailer.session_req.called_name,
+			      server->server_RFC1001_name,
+			      RFC1001_NAME_LEN_WITH_NULL);
+	else
+		rfc1002mangle(req.trailer.session_req.called_name,
+			      DEFAULT_CIFS_CALLED_NAME,
+			      RFC1001_NAME_LEN_WITH_NULL);
 
-		if (server->server_RFC1001_name[0] != 0)
-			rfc1002mangle(ses_init_buf->trailer.
-				      session_req.called_name,
-				      server->server_RFC1001_name,
-				      RFC1001_NAME_LEN_WITH_NULL);
-		else
-			rfc1002mangle(ses_init_buf->trailer.
-				      session_req.called_name,
-				      DEFAULT_CIFS_CALLED_NAME,
-				      RFC1001_NAME_LEN_WITH_NULL);
+	req.trailer.session_req.calling_len = sizeof(req.trailer.session_req.calling_name);
 
-		ses_init_buf->trailer.session_req.calling_len = 32;
+	/* calling name ends in null (byte 16) from old smb convention */
+	if (server->workstation_RFC1001_name[0] != 0)
+		rfc1002mangle(req.trailer.session_req.calling_name,
+			      server->workstation_RFC1001_name,
+			      RFC1001_NAME_LEN_WITH_NULL);
+	else
+		rfc1002mangle(req.trailer.session_req.calling_name,
+			      "LINUX_CIFS_CLNT",
+			      RFC1001_NAME_LEN_WITH_NULL);
 
-		/*
-		 * calling name ends in null (byte 16) from old smb
-		 * convention.
-		 */
-		if (server->workstation_RFC1001_name[0] != 0)
-			rfc1002mangle(ses_init_buf->trailer.
-				      session_req.calling_name,
-				      server->workstation_RFC1001_name,
-				      RFC1001_NAME_LEN_WITH_NULL);
-		else
-			rfc1002mangle(ses_init_buf->trailer.
-				      session_req.calling_name,
-				      "LINUX_CIFS_CLNT",
-				      RFC1001_NAME_LEN_WITH_NULL);
-
-		ses_init_buf->trailer.session_req.scope1 = 0;
-		ses_init_buf->trailer.session_req.scope2 = 0;
-		smb_buf = (struct smb_hdr *)ses_init_buf;
-
-		/* sizeof RFC1002_SESSION_REQUEST with no scopes */
-		req_noscope_len = sizeof(struct rfc1002_session_packet) - 2;
-
-		/* == cpu_to_be32(0x81000044) */
-		smb_buf->smb_buf_length =
-			cpu_to_be32((RFC1002_SESSION_REQUEST << 24) | req_noscope_len);
-		rc = smb_send(server, smb_buf, 0x44);
-		kfree(ses_init_buf);
-		/*
-		 * RFC1001 layer in at least one server
-		 * requires very short break before negprot
-		 * presumably because not expecting negprot
-		 * to follow so fast.  This is a simple
-		 * solution that works without
-		 * complicating the code and causes no
-		 * significant slowing down on mount
-		 * for everyone else
-		 */
-		usleep_range(1000, 2000);
-	}
 	/*
-	 * else the negprot may still work without this
-	 * even though malloc failed
+	 * As per rfc1002, @len must be the number of bytes that follows the
+	 * length field of a rfc1002 session request payload.
+	 */
+	len = sizeof(req) - offsetof(struct rfc1002_session_packet, trailer.session_req);
+
+	smb_buf->smb_buf_length = cpu_to_be32((RFC1002_SESSION_REQUEST << 24) | len);
+	rc = smb_send(server, smb_buf, len);
+	/*
+	 * RFC1001 layer in at least one server requires very short break before
+	 * negprot presumably because not expecting negprot to follow so fast.
+	 * This is a simple solution that works without complicating the code
+	 * and causes no significant slowing down on mount for everyone else
 	 */
+	usleep_range(1000, 2000);
 
 	return rc;
 }
Signature · 1 line
2.39.2
author maintainer
#674 lore raw

PATCH 673/885 cifs: introduce cifs_io_parms in smb2_async_writev()

In reply to Greg Kroah-Hartman

From: Stefan Metzmacher <[email protected]>

commit d643a8a446fc46c06837d08a056f69da2ff16025 upstream.

This will simplify the following changes and makes it easy to get in passed in from the caller in future.

---
 fs/cifs/smb2pdu.c |   53 +++++++++++++++++++++++++++++++++++++++--------------
 1 file changed, 39 insertions(+), 14 deletions(-)
fs/cifs/smb2pdu.c +39 -14
@@ -4504,10 +4504,27 @@ smb2_async_writev(struct cifs_writedata
 	struct kvec iov[1];
 	struct smb_rqst rqst = { };
 	unsigned int total_len;
+	struct cifs_io_parms _io_parms;
+	struct cifs_io_parms *io_parms = NULL;
 
 	if (!wdata->server)
 		server = wdata->server = cifs_pick_channel(tcon->ses);
 
+	/*
+	 * in future we may get cifs_io_parms passed in from the caller,
+	 * but for now we construct it here...
+	 */
+	_io_parms = (struct cifs_io_parms) {
+		.tcon = tcon,
+		.server = server,
+		.offset = wdata->offset,
+		.length = wdata->bytes,
+		.persistent_fid = wdata->cfile->fid.persistent_fid,
+		.volatile_fid = wdata->cfile->fid.volatile_fid,
+		.pid = wdata->pid,
+	};
+	io_parms = &_io_parms;
+
 	rc = smb2_plain_req_init(SMB2_WRITE, tcon, server,
 				 (void **) &req, &total_len);
 	if (rc)
@@ -4517,26 +4534,31 @@ smb2_async_writev(struct cifs_writedata
 		flags |= CIFS_TRANSFORM_REQ;
 
 	shdr = (struct smb2_hdr *)req;
-	shdr->Id.SyncId.ProcessId = cpu_to_le32(wdata->cfile->pid);
+	shdr->Id.SyncId.ProcessId = cpu_to_le32(io_parms->pid);
 
-	req->PersistentFileId = wdata->cfile->fid.persistent_fid;
-	req->VolatileFileId = wdata->cfile->fid.volatile_fid;
+	req->PersistentFileId = io_parms->persistent_fid;
+	req->VolatileFileId = io_parms->volatile_fid;
 	req->WriteChannelInfoOffset = 0;
 	req->WriteChannelInfoLength = 0;
 	req->Channel = 0;
-	req->Offset = cpu_to_le64(wdata->offset);
+	req->Offset = cpu_to_le64(io_parms->offset);
 	req->DataOffset = cpu_to_le16(
 				offsetof(struct smb2_write_req, Buffer));
 	req->RemainingBytes = 0;
 
-	trace_smb3_write_enter(0 /* xid */, wdata->cfile->fid.persistent_fid,
-		tcon->tid, tcon->ses->Suid, wdata->offset, wdata->bytes);
+	trace_smb3_write_enter(0 /* xid */,
+			       io_parms->persistent_fid,
+			       io_parms->tcon->tid,
+			       io_parms->tcon->ses->Suid,
+			       io_parms->offset,
+			       io_parms->length);
+
 #ifdef CONFIG_CIFS_SMB_DIRECT
 	/*
 	 * If we want to do a server RDMA read, fill in and append
 	 * smbd_buffer_descriptor_v1 to the end of write request
 	 */
-	if (server->rdma && !server->sign && wdata->bytes >=
+	if (server->rdma && !server->sign && io_parms->length >=
 		server->smbd_conn->rdma_readwrite_threshold) {
 
 		struct smbd_buffer_descriptor_v1 *v1;
@@ -4590,14 +4612,14 @@ smb2_async_writev(struct cifs_writedata
 	}
 #endif
 	cifs_dbg(FYI, "async write at %llu %u bytes\n",
-		 wdata->offset, wdata->bytes);
+		 io_parms->offset, io_parms->length);
 
 #ifdef CONFIG_CIFS_SMB_DIRECT
 	/* For RDMA read, I/O size is in RemainingBytes not in Length */
 	if (!wdata->mr)
-		req->Length = cpu_to_le32(wdata->bytes);
+		req->Length = cpu_to_le32(io_parms->length);
 #else
-	req->Length = cpu_to_le32(wdata->bytes);
+	req->Length = cpu_to_le32(io_parms->length);
 #endif
 
 	if (wdata->credits.value > 0) {
@@ -4605,7 +4627,7 @@ smb2_async_writev(struct cifs_writedata
 						    SMB2_MAX_BUFFER_SIZE));
 		shdr->CreditRequest = cpu_to_le16(le16_to_cpu(shdr->CreditCharge) + 8);
 
-		rc = adjust_credits(server, &wdata->credits, wdata->bytes);
+		rc = adjust_credits(server, &wdata->credits, io_parms->length);
 		if (rc)
 			goto async_writev_out;
 
@@ -4618,9 +4640,12 @@ smb2_async_writev(struct cifs_writedata
 
 	if (rc) {
 		trace_smb3_write_err(0 /* no xid */,
-				     req->PersistentFileId,
-				     tcon->tid, tcon->ses->Suid, wdata->offset,
-				     wdata->bytes, rc);
+				     io_parms->persistent_fid,
+				     io_parms->tcon->tid,
+				     io_parms->tcon->ses->Suid,
+				     io_parms->offset,
+				     io_parms->length,
+				     rc);
 		kref_put(&wdata->refcount, release);
 		cifs_stats_fail_inc(tcon, SMB2_WRITE_HE);
 	}
author maintainer
#675 lore raw

PATCH 674/885 cifs: split out smb3_use_rdma_offload() helper

In reply to Greg Kroah-Hartman

From: Stefan Metzmacher <[email protected]>

commit a6559cc1d35d3eeafb0296aca347b2f745a28a74 upstream.

We should have the logic to decide if we want rdma offload in a single spot in order to advance it in future.

---
 fs/cifs/smb2pdu.c |   34 ++++++++++++++++++++++++++++------
 1 file changed, 28 insertions(+), 6 deletions(-)
fs/cifs/smb2pdu.c +28 -6
@@ -4063,6 +4063,32 @@ SMB2_flush(const unsigned int xid, struc
 	return rc;
 }
 
+#ifdef CONFIG_CIFS_SMB_DIRECT
+static inline bool smb3_use_rdma_offload(struct cifs_io_parms *io_parms)
+{
+	struct TCP_Server_Info *server = io_parms->server;
+	struct cifs_tcon *tcon = io_parms->tcon;
+
+	/* we can only offload if we're connected */
+	if (!server || !tcon)
+		return false;
+
+	/* we can only offload on an rdma connection */
+	if (!server->rdma || !server->smbd_conn)
+		return false;
+
+	/* we don't support signed offload yet */
+	if (server->sign)
+		return false;
+
+	/* offload also has its overhead, so only do it if desired */
+	if (io_parms->length < server->smbd_conn->rdma_readwrite_threshold)
+		return false;
+
+	return true;
+}
+#endif /* CONFIG_CIFS_SMB_DIRECT */
+
 /*
  * To form a chain of read requests, any read requests after the first should
  * have the end_of_chain boolean set to true.
@@ -4106,9 +4132,7 @@ smb2_new_read_req(void **buf, unsigned i
 	 * If we want to do a RDMA write, fill in and append
 	 * smbd_buffer_descriptor_v1 to the end of read request
 	 */
-	if (server->rdma && rdata && !server->sign &&
-		rdata->bytes >= server->smbd_conn->rdma_readwrite_threshold) {
-
+	if (smb3_use_rdma_offload(io_parms)) {
 		struct smbd_buffer_descriptor_v1 *v1;
 		bool need_invalidate = server->dialect == SMB30_PROT_ID;
 
@@ -4558,9 +4582,7 @@ smb2_async_writev(struct cifs_writedata
 	 * If we want to do a server RDMA read, fill in and append
 	 * smbd_buffer_descriptor_v1 to the end of write request
 	 */
-	if (server->rdma && !server->sign && io_parms->length >=
-		server->smbd_conn->rdma_readwrite_threshold) {
-
+	if (smb3_use_rdma_offload(io_parms)) {
 		struct smbd_buffer_descriptor_v1 *v1;
 		bool need_invalidate = server->dialect == SMB30_PROT_ID;
 
author maintainer
#676 lore raw

PATCH 675/885 cifs: dont try to use rdma offload on encrypted connections

In reply to Greg Kroah-Hartman

From: Stefan Metzmacher <[email protected]>

commit 3891f6c7655a39065e44980f51ba46bb32be3133 upstream.

The aim of using encryption on a connection is to keep the data confidential, so we must not use plaintext rdma offload for that data!

It seems that current windows servers and ksmbd would allow this, but that's no reason to expose the users data in plaintext! And servers hopefully reject this in future.

Note modern windows servers support signed or encrypted offload, see MS-SMB2 2.2.3.1.6 SMB2_RDMA_TRANSFORM_CAPABILITIES, but we don't support that yet.

---
 fs/cifs/smb2pdu.c |    4 ++++
 1 file changed, 4 insertions(+)
fs/cifs/smb2pdu.c +4 -0
@@ -4081,6 +4081,10 @@ static inline bool smb3_use_rdma_offload
 	if (server->sign)
 		return false;
 
+	/* we don't support encrypted offload yet */
+	if (smb3_encryption_required(tcon))
+		return false;
+
 	/* offload also has its overhead, so only do it if desired */
 	if (io_parms->length < server->smbd_conn->rdma_readwrite_threshold)
 		return false;
author maintainer
#677 lore raw

PATCH 676/885 cifs: Check the lease context if we actually got a lease

In reply to Greg Kroah-Hartman

From: Ronnie Sahlberg <[email protected]>

commit 66d45ca1350a3bb8d5f4db8879ccad3ed492337a upstream.

Some servers may return that we got a lease in rsp->OplockLevel but then in the lease context contradict this and say we got no lease at all. Thus we need to check the context if we have a lease. Additionally, If we do not get a lease we need to make sure we close the handle before we return an error to the caller.

---
 fs/cifs/cached_dir.c |   14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)
fs/cifs/cached_dir.c +9 -5
@@ -221,8 +221,7 @@ int open_cached_dir(unsigned int xid, st
 		}
 		goto oshr_free;
 	}
-
-	atomic_inc(&tcon->num_remote_opens);
+	cfid->is_open = true;
 
 	o_rsp = (struct smb2_create_rsp *)rsp_iov[0].iov_base;
 	oparms.fid->persistent_fid = o_rsp->PersistentFileId;
@@ -239,7 +238,8 @@ int open_cached_dir(unsigned int xid, st
 			    &oparms.fid->epoch,
 			    oparms.fid->lease_key, &oplock,
 			    NULL, NULL);
-
+	if (!(oplock & SMB2_LEASE_READ_CACHING_HE))
+		goto oshr_free;
 	qi_rsp = (struct smb2_query_info_rsp *)rsp_iov[1].iov_base;
 	if (le32_to_cpu(qi_rsp->OutputBufferLength) < sizeof(struct smb2_file_all_info))
 		goto oshr_free;
@@ -262,7 +262,6 @@ int open_cached_dir(unsigned int xid, st
 	cfid->dentry = dentry;
 	cfid->tcon = tcon;
 	cfid->time = jiffies;
-	cfid->is_open = true;
 	cfid->has_lease = true;
 
 oshr_free:
@@ -282,12 +281,17 @@ oshr_free:
 	}
 	spin_unlock(&cfids->cfid_list_lock);
 	if (rc) {
+		if (cfid->is_open)
+			SMB2_close(0, cfid->tcon, cfid->fid.persistent_fid,
+				   cfid->fid.volatile_fid);
 		free_cached_dir(cfid);
 		cfid = NULL;
 	}
 
-	if (rc == 0)
+	if (rc == 0) {
 		*ret_cfid = cfid;
+		atomic_inc(&tcon->num_remote_opens);
+	}
 
 	return rc;
 }
author maintainer
#678 lore raw

PATCH 677/885 cifs: return a single-use cfid if we did not get a lease

In reply to Greg Kroah-Hartman

From: Ronnie Sahlberg <[email protected]>

commit 8e843bf38f7be0766642a91523cfa65f2b021a8a upstream.

If we did not get a lease we can still return a single use cfid to the caller. The cfid will not have has_lease set and will thus not be shared with any other concurrent users and will be freed immediately when the caller drops the handle.

This avoids extra roundtrips for servers that do not support directory leases where they would first fail to get a cfid with a lease and then fallback to try a normal SMB2_open()

---
 fs/cifs/cached_dir.c |   16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)
fs/cifs/cached_dir.c +13 -3
@@ -14,6 +14,7 @@
 
 static struct cached_fid *init_cached_dir(const char *path);
 static void free_cached_dir(struct cached_fid *cfid);
+static void smb2_close_cached_fid(struct kref *ref);
 
 static struct cached_fid *find_or_create_cached_dir(struct cached_fids *cfids,
 						    const char *path,
@@ -221,6 +222,7 @@ int open_cached_dir(unsigned int xid, st
 		}
 		goto oshr_free;
 	}
+	cfid->tcon = tcon;
 	cfid->is_open = true;
 
 	o_rsp = (struct smb2_create_rsp *)rsp_iov[0].iov_base;
@@ -233,7 +235,6 @@ int open_cached_dir(unsigned int xid, st
 	if (o_rsp->OplockLevel != SMB2_OPLOCK_LEVEL_LEASE)
 		goto oshr_free;
 
-
 	smb2_parse_contexts(server, o_rsp,
 			    &oparms.fid->epoch,
 			    oparms.fid->lease_key, &oplock,
@@ -260,7 +261,6 @@ int open_cached_dir(unsigned int xid, st
 		}
 	}
 	cfid->dentry = dentry;
-	cfid->tcon = tcon;
 	cfid->time = jiffies;
 	cfid->has_lease = true;
 
@@ -271,7 +271,7 @@ oshr_free:
 	free_rsp_buf(resp_buftype[0], rsp_iov[0].iov_base);
 	free_rsp_buf(resp_buftype[1], rsp_iov[1].iov_base);
 	spin_lock(&cfids->cfid_list_lock);
-	if (!cfid->has_lease) {
+	if (rc && !cfid->has_lease) {
 		if (cfid->on_list) {
 			list_del(&cfid->entry);
 			cfid->on_list = false;
@@ -280,6 +280,15 @@ oshr_free:
 		rc = -ENOENT;
 	}
 	spin_unlock(&cfids->cfid_list_lock);
+	if (!rc && !cfid->has_lease) {
+		/*
+		 * We are guaranteed to have two references at this point.
+		 * One for the caller and one for a potential lease.
+		 * Release the Lease-ref so that the directory will be closed
+		 * when the caller closes the cached handle.
+		 */
+		kref_put(&cfid->refcount, smb2_close_cached_fid);
+	}
 	if (rc) {
 		if (cfid->is_open)
 			SMB2_close(0, cfid->tcon, cfid->fid.persistent_fid,
@@ -340,6 +349,7 @@ smb2_close_cached_fid(struct kref *ref)
 	if (cfid->is_open) {
 		SMB2_close(0, cfid->tcon, cfid->fid.persistent_fid,
 			   cfid->fid.volatile_fid);
+		atomic_dec(&cfid->tcon->num_remote_opens);
 	}
 
 	free_cached_dir(cfid);
author maintainer
#679 lore raw

PATCH 678/885 scsi: mpi3mr: Fix missing mrioc->evtack_cmds initialization

In reply to Greg Kroah-Hartman

From: Shin'ichiro Kawasaki <[email protected]>

commit e39ea831ebad4ab15c4748cb62a397a8abcca36e upstream.

Commit c1af985d27da ("scsi: mpi3mr: Add Event acknowledgment logic") introduced an array mrioc->evtack_cmds but initialization of the array elements was missed. They are just zero cleared. The function mpi3mr_complete_evt_ack() refers host_tag field of the elements. Due to the zero value of the host_tag field, the function calls clear_bit() for mrico->evtack_cmds_bitmap with wrong bit index. This results in memory access to invalid address and "BUG: KASAN: use-after-free". This BUG was observed at eHBA-9600 firmware update to version 8.3.1.0. To fix it, add the missing initialization of mrioc->evtack_cmds.

---
 drivers/scsi/mpi3mr/mpi3mr_os.c |    4 ++++
 1 file changed, 4 insertions(+)
drivers/scsi/mpi3mr/mpi3mr_os.c +4 -0
@@ -4952,6 +4952,10 @@ mpi3mr_probe(struct pci_dev *pdev, const
 		mpi3mr_init_drv_cmd(&mrioc->dev_rmhs_cmds[i],
 		    MPI3MR_HOSTTAG_DEVRMCMD_MIN + i);
 
+	for (i = 0; i < MPI3MR_NUM_EVTACKCMD; i++)
+		mpi3mr_init_drv_cmd(&mrioc->evtack_cmds[i],
+				    MPI3MR_HOSTTAG_EVTACKCMD_MIN + i);
+
 	if (pdev->revision)
 		mrioc->enable_segqueue = true;
 
author maintainer
#680 lore raw

PATCH 679/885 scsi: mpi3mr: Fix issues in mpi3mr_get_all_tgt_info()

In reply to Greg Kroah-Hartman

From: Shin'ichiro Kawasaki <[email protected]>

commit fb428a2005fc1260d18b989cc5199f281617f44d upstream.

The function mpi3mr_get_all_tgt_info() has four issues:

1) It calculates valid entry length in alltgt_info assuming the header part of the struct mpi3mr_device_map_info would equal to sizeof(u32). The correct size is sizeof(u64).

2) When it calculates the valid entry length kern_entrylen, it excludes one entry by subtracting 1 from num_devices.

3) It copies num_device by calling memcpy(). Substitution is enough.

4) It does not specify the calculated length to sg_copy_from_buffer(). Instead, it specifies the payload length which is larger than the alltgt_info size. It causes "BUG: KASAN: slab-out-of-bounds".

Fix the issues by using the correct header size, removing the subtraction from num_devices, replacing the memcpy() with substitution and specifying the correct length to sg_copy_from_buffer().

---
 drivers/scsi/mpi3mr/mpi3mr_app.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)
drivers/scsi/mpi3mr/mpi3mr_app.c +8 -7
index 9baac224b213..72054e3a26cb 100644
@@ -312,7 +312,7 @@ static long mpi3mr_get_all_tgt_info(struct mpi3mr_ioc *mrioc,
 		num_devices++;
 	spin_unlock_irqrestore(&mrioc->tgtdev_lock, flags);
 
-	if ((job->request_payload.payload_len == sizeof(u32)) ||
+	if ((job->request_payload.payload_len <= sizeof(u64)) ||
 		list_empty(&mrioc->tgtdev_list)) {
 		sg_copy_from_buffer(job->request_payload.sg_list,
 				    job->request_payload.sg_cnt,
@@ -320,14 +320,14 @@ static long mpi3mr_get_all_tgt_info(struct mpi3mr_ioc *mrioc,
 		return 0;
 	}
 
-	kern_entrylen = (num_devices - 1) * sizeof(*devmap_info);
-	size = sizeof(*alltgt_info) + kern_entrylen;
+	kern_entrylen = num_devices * sizeof(*devmap_info);
+	size = sizeof(u64) + kern_entrylen;
 	alltgt_info = kzalloc(size, GFP_KERNEL);
 	if (!alltgt_info)
 		return -ENOMEM;
 
 	devmap_info = alltgt_info->dmi;
-	memset((u8 *)devmap_info, 0xFF, (kern_entrylen + sizeof(*devmap_info)));
+	memset((u8 *)devmap_info, 0xFF, kern_entrylen);
 	spin_lock_irqsave(&mrioc->tgtdev_lock, flags);
 	list_for_each_entry(tgtdev, &mrioc->tgtdev_list, list) {
 		if (i < num_devices) {
@@ -344,9 +344,10 @@ static long mpi3mr_get_all_tgt_info(struct mpi3mr_ioc *mrioc,
 	num_devices = i;
 	spin_unlock_irqrestore(&mrioc->tgtdev_lock, flags);
 
-	memcpy(&alltgt_info->num_devices, &num_devices, sizeof(num_devices));
+	alltgt_info->num_devices = num_devices;
 
-	usr_entrylen = (job->request_payload.payload_len - sizeof(u32)) / sizeof(*devmap_info);
+	usr_entrylen = (job->request_payload.payload_len - sizeof(u64)) /
+		sizeof(*devmap_info);
 	usr_entrylen *= sizeof(*devmap_info);
 	min_entrylen = min(usr_entrylen, kern_entrylen);
 	if (min_entrylen && (!memcpy(&alltgt_info->dmi, devmap_info, min_entrylen))) {
@@ -358,7 +359,7 @@ static long mpi3mr_get_all_tgt_info(struct mpi3mr_ioc *mrioc,
 
 	sg_copy_from_buffer(job->request_payload.sg_list,
 			    job->request_payload.sg_cnt,
-			    alltgt_info, job->request_payload.payload_len);
+			    alltgt_info, (min_entrylen + sizeof(u64)));
 	rval = 0;
 out:
 	kfree(alltgt_info);
Signature · 1 line
2.39.2
author maintainer
#681 lore raw

PATCH 680/885 scsi: mpi3mr: Remove unnecessary memcpy() to alltgt_info->dmi

In reply to Greg Kroah-Hartman

From: Shin'ichiro Kawasaki <[email protected]>

commit eeb270aee3e085411399f129fc14fa04bd6d83cf upstream.

In the function mpi3mr_get_all_tgt_info(), devmap_info points to alltgt_info->dmi then there is no need to memcpy() data from devmap_info to alltgt_info->dmi. Remove the unnecessary memcpy(). This also allows to remove the local variable 'rval' and the goto label 'out'.

---
 drivers/scsi/mpi3mr/mpi3mr_app.c | 13 ++-----------
 1 file changed, 2 insertions(+), 11 deletions(-)
drivers/scsi/mpi3mr/mpi3mr_app.c +2 -11
index 72054e3a26cb..bff637702397 100644
@@ -293,7 +293,6 @@ static long mpi3mr_bsg_pel_enable(struct mpi3mr_ioc *mrioc,
 static long mpi3mr_get_all_tgt_info(struct mpi3mr_ioc *mrioc,
 	struct bsg_job *job)
 {
-	long rval = -EINVAL;
 	u16 num_devices = 0, i = 0, size;
 	unsigned long flags;
 	struct mpi3mr_tgt_dev *tgtdev;
@@ -304,7 +303,7 @@ static long mpi3mr_get_all_tgt_info(struct mpi3mr_ioc *mrioc,
 	if (job->request_payload.payload_len < sizeof(u32)) {
 		dprint_bsg_err(mrioc, "%s: invalid size argument\n",
 		    __func__);
-		return rval;
+		return -EINVAL;
 	}
 
 	spin_lock_irqsave(&mrioc->tgtdev_lock, flags);
@@ -350,20 +349,12 @@ static long mpi3mr_get_all_tgt_info(struct mpi3mr_ioc *mrioc,
 		sizeof(*devmap_info);
 	usr_entrylen *= sizeof(*devmap_info);
 	min_entrylen = min(usr_entrylen, kern_entrylen);
-	if (min_entrylen && (!memcpy(&alltgt_info->dmi, devmap_info, min_entrylen))) {
-		dprint_bsg_err(mrioc, "%s:%d: device map info copy failed\n",
-		    __func__, __LINE__);
-		rval = -EFAULT;
-		goto out;
-	}
 
 	sg_copy_from_buffer(job->request_payload.sg_list,
 			    job->request_payload.sg_cnt,
 			    alltgt_info, (min_entrylen + sizeof(u64)));
-	rval = 0;
-out:
 	kfree(alltgt_info);
-	return rval;
+	return 0;
 }
 /**
  * mpi3mr_get_change_count - Get topology change count
Signature · 1 line
2.39.2
author maintainer
#682 lore raw

PATCH 681/885 btrfs: hold block group refcount during async discard

In reply to Greg Kroah-Hartman

From: Boris Burkov <[email protected]>

commit 2b5463fcbdfb24e898916bcae2b1359042d26963 upstream.

Async discard does not acquire the block group reference count while it holds a reference on the discard list. This is generally OK, as the paths which destroy block groups tend to try to synchronize on cancelling async discard work. However, relying on cancelling work requires careful analysis to be sure it is safe from races with unpinning scheduling more work.

While I am unable to find a race with unpinning in the current code for either the unused bgs or relocation paths, I believe we have one in an older version of auto relocation in a Meta internal build. This suggests that this is in fact an error prone model, and could be fragile to future changes to these bg deletion paths.

To make this ownership more clear, add a refcount for async discard. If work is queued for a block group, its refcount should be incremented, and when work is completed or canceled, it should be decremented.

---
 fs/btrfs/discard.c |   41 ++++++++++++++++++++++++++++++++++++++---
 1 file changed, 38 insertions(+), 3 deletions(-)
fs/btrfs/discard.c +38 -3
@@ -77,6 +77,7 @@ static struct list_head *get_discard_lis
 static void __add_to_discard_list(struct btrfs_discard_ctl *discard_ctl,
 				  struct btrfs_block_group *block_group)
 {
+	lockdep_assert_held(&discard_ctl->lock);
 	if (!btrfs_run_discard_work(discard_ctl))
 		return;
 
@@ -88,6 +89,8 @@ static void __add_to_discard_list(struct
 						      BTRFS_DISCARD_DELAY);
 		block_group->discard_state = BTRFS_DISCARD_RESET_CURSOR;
 	}
+	if (list_empty(&block_group->discard_list))
+		btrfs_get_block_group(block_group);
 
 	list_move_tail(&block_group->discard_list,
 		       get_discard_list(discard_ctl, block_group));
@@ -107,8 +110,12 @@ static void add_to_discard_list(struct b
 static void add_to_discard_unused_list(struct btrfs_discard_ctl *discard_ctl,
 				       struct btrfs_block_group *block_group)
 {
+	bool queued;
+
 	spin_lock(&discard_ctl->lock);
 
+	queued = !list_empty(&block_group->discard_list);
+
 	if (!btrfs_run_discard_work(discard_ctl)) {
 		spin_unlock(&discard_ctl->lock);
 		return;
@@ -120,6 +127,8 @@ static void add_to_discard_unused_list(s
 	block_group->discard_eligible_time = (ktime_get_ns() +
 					      BTRFS_DISCARD_UNUSED_DELAY);
 	block_group->discard_state = BTRFS_DISCARD_RESET_CURSOR;
+	if (!queued)
+		btrfs_get_block_group(block_group);
 	list_add_tail(&block_group->discard_list,
 		      &discard_ctl->discard_list[BTRFS_DISCARD_INDEX_UNUSED]);
 
@@ -130,6 +139,7 @@ static bool remove_from_discard_list(str
 				     struct btrfs_block_group *block_group)
 {
 	bool running = false;
+	bool queued = false;
 
 	spin_lock(&discard_ctl->lock);
 
@@ -139,7 +149,16 @@ static bool remove_from_discard_list(str
 	}
 
 	block_group->discard_eligible_time = 0;
+	queued = !list_empty(&block_group->discard_list);
 	list_del_init(&block_group->discard_list);
+	/*
+	 * If the block group is currently running in the discard workfn, we
+	 * don't want to deref it, since it's still being used by the workfn.
+	 * The workfn will notice this case and deref the block group when it is
+	 * finished.
+	 */
+	if (queued && !running)
+		btrfs_put_block_group(block_group);
 
 	spin_unlock(&discard_ctl->lock);
 
@@ -212,10 +231,12 @@ again:
 	if (block_group && now >= block_group->discard_eligible_time) {
 		if (block_group->discard_index == BTRFS_DISCARD_INDEX_UNUSED &&
 		    block_group->used != 0) {
-			if (btrfs_is_block_group_data_only(block_group))
+			if (btrfs_is_block_group_data_only(block_group)) {
 				__add_to_discard_list(discard_ctl, block_group);
-			else
+			} else {
 				list_del_init(&block_group->discard_list);
+				btrfs_put_block_group(block_group);
+			}
 			goto again;
 		}
 		if (block_group->discard_state == BTRFS_DISCARD_RESET_CURSOR) {
@@ -502,6 +523,15 @@ static void btrfs_discard_workfn(struct
 	spin_lock(&discard_ctl->lock);
 	discard_ctl->prev_discard = trimmed;
 	discard_ctl->prev_discard_time = now;
+	/*
+	 * If the block group was removed from the discard list while it was
+	 * running in this workfn, then we didn't deref it, since this function
+	 * still owned that reference. But we set the discard_ctl->block_group
+	 * back to NULL, so we can use that condition to know that now we need
+	 * to deref the block_group.
+	 */
+	if (discard_ctl->block_group == NULL)
+		btrfs_put_block_group(block_group);
 	discard_ctl->block_group = NULL;
 	__btrfs_discard_schedule_work(discard_ctl, now, false);
 	spin_unlock(&discard_ctl->lock);
@@ -638,8 +668,12 @@ void btrfs_discard_punt_unused_bgs_list(
 	list_for_each_entry_safe(block_group, next, &fs_info->unused_bgs,
 				 bg_list) {
 		list_del_init(&block_group->bg_list);
-		btrfs_put_block_group(block_group);
 		btrfs_discard_queue_work(&fs_info->discard_ctl, block_group);
+		/*
+		 * This put is for the get done by btrfs_mark_bg_unused.
+		 * Queueing discard incremented it for discard's reference.
+		 */
+		btrfs_put_block_group(block_group);
 	}
 	spin_unlock(&fs_info->unused_bgs_lock);
 }
@@ -669,6 +703,7 @@ static void btrfs_discard_purge_list(str
 			if (block_group->used == 0)
 				btrfs_mark_bg_unused(block_group);
 			spin_lock(&discard_ctl->lock);
+			btrfs_put_block_group(block_group);
 		}
 	}
 	spin_unlock(&discard_ctl->lock);
author maintainer
#683 lore raw

PATCH 682/885 locking/rwsem: Prevent non-first waiter from spinning in down_write() slowpath

In reply to Greg Kroah-Hartman

From: Waiman Long <[email protected]>

commit b613c7f31476c44316bfac1af7cac714b7d6bef9 upstream.

A non-first waiter can potentially spin in the for loop of rwsem_down_write_slowpath() without sleeping but fail to acquire the lock even if the rwsem is free if the following sequence happens:

  Non-first RT waiter    First waiter      Lock holder
  -------------------    ------------      -----------
  Acquire wait_lock
  rwsem_try_write_lock():
    Set handoff bit if RT or
      wait too long
    Set waiter->handoff_set
  Release wait_lock
                         Acquire wait_lock
                         Inherit waiter->handoff_set
                         Release wait_lock
					   Clear owner
                                           Release lock
  if (waiter.handoff_set) {
    rwsem_spin_on_owner(();
    if (OWNER_NULL)
      goto trylock_again;
  }
  trylock_again:
  Acquire wait_lock
  rwsem_try_write_lock():
     if (first->handoff_set && (waiter != first))
	return false;

Release wait_lock

A non-first waiter cannot really acquire the rwsem even if it mistakenly believes that it can spin on OWNER_NULL value. If that waiter happens to be an RT task running on the same CPU as the first waiter, it can block the first waiter from acquiring the rwsem leading to live lock. Fix this problem by making sure that a non-first waiter cannot spin in the slowpath loop without sleeping.

---
 kernel/locking/rwsem.c |   19 +++++++++----------
 1 file changed, 9 insertions(+), 10 deletions(-)
kernel/locking/rwsem.c +9 -10
@@ -624,18 +624,16 @@ static inline bool rwsem_try_write_lock(
 			 */
 			if (first->handoff_set && (waiter != first))
 				return false;
-
-			/*
-			 * First waiter can inherit a previously set handoff
-			 * bit and spin on rwsem if lock acquisition fails.
-			 */
-			if (waiter == first)
-				waiter->handoff_set = true;
 		}
 
 		new = count;
 
 		if (count & RWSEM_LOCK_MASK) {
+			/*
+			 * A waiter (first or not) can set the handoff bit
+			 * if it is an RT task or wait in the wait queue
+			 * for too long.
+			 */
 			if (has_handoff || (!rt_task(waiter->task) &&
 					    !time_after(jiffies, waiter->timeout)))
 				return false;
@@ -651,11 +649,12 @@ static inline bool rwsem_try_write_lock(
 	} while (!atomic_long_try_cmpxchg_acquire(&sem->count, &count, new));
 
 	/*
-	 * We have either acquired the lock with handoff bit cleared or
-	 * set the handoff bit.
+	 * We have either acquired the lock with handoff bit cleared or set
+	 * the handoff bit. Only the first waiter can have its handoff_set
+	 * set here to enable optimistic spinning in slowpath loop.
 	 */
 	if (new & RWSEM_FLAG_HANDOFF) {
-		waiter->handoff_set = true;
+		first->handoff_set = true;
 		lockevent_inc(rwsem_wlock_handoff);
 		return false;
 	}
author maintainer
#684 lore raw

PATCH 683/885 ksmbd: fix wrong data area length for smb2 lock request

In reply to Greg Kroah-Hartman

From: Namjae Jeon <[email protected]>

commit 8f8c43b125882ac14372f8dca0c8e50a59e78d79 upstream.

When turning debug mode on, The following error message from ksmbd_smb2_check_message() is coming.

ksmbd: cli req padded more than expected. Length 112 not 88 for cmd:10 mid:14

data area length calculation for smb2 lock request in smb2_get_data_area_len() is incorrect.

Fixes e2f34481b24d ("cifsd: add server-side procedures for SMB3")
---
 fs/ksmbd/smb2misc.c |   10 +++-------
 1 file changed, 3 insertions(+), 7 deletions(-)
fs/ksmbd/smb2misc.c +3 -7
@@ -149,15 +149,11 @@ static int smb2_get_data_area_len(unsign
 		break;
 	case SMB2_LOCK:
 	{
-		int lock_count;
+		unsigned short lock_count;
 
-		/*
-		 * smb2_lock request size is 48 included single
-		 * smb2_lock_element structure size.
-		 */
-		lock_count = le16_to_cpu(((struct smb2_lock_req *)hdr)->LockCount) - 1;
+		lock_count = le16_to_cpu(((struct smb2_lock_req *)hdr)->LockCount);
 		if (lock_count > 0) {
-			*off = __SMB2_HEADER_STRUCTURE_SIZE + 48;
+			*off = offsetof(struct smb2_lock_req, locks);
 			*len = sizeof(struct smb2_lock_element) * lock_count;
 		}
 		break;
author maintainer
#685 lore raw

PATCH 684/885 ksmbd: do not allow the actual frame length to be smaller than the rfc1002 length

In reply to Greg Kroah-Hartman

From: Namjae Jeon <[email protected]>

commit fb533473d1595fe79ecb528fda1de33552b07178 upstream.

ksmbd allowed the actual frame length to be smaller than the rfc1002 length. If allowed, it is possible to allocates a large amount of memory that can be limited by credit management and can eventually cause memory exhaustion problem. This patch do not allow it except SMB2 Negotiate request which will be validated when message handling proceeds. Also, Allow a message that padded to 8byte boundary.

Fixes e2f34481b24d ("cifsd: add server-side procedures for SMB3")
---
 fs/ksmbd/smb2misc.c |   21 ++++++++++-----------
 1 file changed, 10 insertions(+), 11 deletions(-)
fs/ksmbd/smb2misc.c +10 -11
@@ -408,20 +408,19 @@ int ksmbd_smb2_check_message(struct ksmb
 			goto validate_credit;
 
 		/*
-		 * windows client also pad up to 8 bytes when compounding.
-		 * If pad is longer than eight bytes, log the server behavior
-		 * (once), since may indicate a problem but allow it and
-		 * continue since the frame is parseable.
+		 * SMB2 NEGOTIATE request will be validated when message
+		 * handling proceeds.
 		 */
-		if (clc_len < len) {
-			ksmbd_debug(SMB,
-				    "cli req padded more than expected. Length %d not %d for cmd:%d mid:%llu\n",
-				    len, clc_len, command,
-				    le64_to_cpu(hdr->MessageId));
+		if (command == SMB2_NEGOTIATE_HE)
 			goto validate_credit;
-		}
 
-		ksmbd_debug(SMB,
+		/*
+		 * Allow a message that padded to 8byte boundary.
+		 */
+		if (clc_len < len && (len - clc_len) < 8)
+			goto validate_credit;
+
+		pr_err_ratelimited(
 			    "cli req too short, len %d not %d. cmd:%d mid:%llu\n",
 			    len, clc_len, command,
 			    le64_to_cpu(hdr->MessageId));
author maintainer
#686 lore raw

PATCH 685/885 ksmbd: fix possible memory leak in smb2_lock()

In reply to Greg Kroah-Hartman

From: Hangyu Hua <[email protected]>

commit d3ca9f7aeba793d74361d88a8800b2f205c9236b upstream.

argv needs to be free when setup_async_work fails or when the current process is woken up.

Fixes e2f34481b24d ("cifsd: add server-side procedures for SMB3")
---
 fs/ksmbd/smb2pdu.c   |   28 +++++++++++++---------------
 fs/ksmbd/vfs_cache.c |    5 ++---
 2 files changed, 15 insertions(+), 18 deletions(-)
fs/ksmbd/smb2pdu.c +13 -15
@@ -6642,7 +6642,7 @@ int smb2_cancel(struct ksmbd_work *work)
 	struct ksmbd_conn *conn = work->conn;
 	struct smb2_hdr *hdr = smb2_get_msg(work->request_buf);
 	struct smb2_hdr *chdr;
-	struct ksmbd_work *cancel_work = NULL, *iter;
+	struct ksmbd_work *iter;
 	struct list_head *command_list;
 
 	ksmbd_debug(SMB, "smb2 cancel called on mid %llu, async flags 0x%x\n",
@@ -6664,7 +6664,9 @@ int smb2_cancel(struct ksmbd_work *work)
 				    "smb2 with AsyncId %llu cancelled command = 0x%x\n",
 				    le64_to_cpu(hdr->Id.AsyncId),
 				    le16_to_cpu(chdr->Command));
-			cancel_work = iter;
+			iter->state = KSMBD_WORK_CANCELLED;
+			if (iter->cancel_fn)
+				iter->cancel_fn(iter->cancel_argv);
 			break;
 		}
 		spin_unlock(&conn->request_lock);
@@ -6683,18 +6685,12 @@ int smb2_cancel(struct ksmbd_work *work)
 				    "smb2 with mid %llu cancelled command = 0x%x\n",
 				    le64_to_cpu(hdr->MessageId),
 				    le16_to_cpu(chdr->Command));
-			cancel_work = iter;
+			iter->state = KSMBD_WORK_CANCELLED;
 			break;
 		}
 		spin_unlock(&conn->request_lock);
 	}
 
-	if (cancel_work) {
-		cancel_work->state = KSMBD_WORK_CANCELLED;
-		if (cancel_work->cancel_fn)
-			cancel_work->cancel_fn(cancel_work->cancel_argv);
-	}
-
 	/* For SMB2_CANCEL command itself send no response*/
 	work->send_no_response = 1;
 	return 0;
@@ -7055,6 +7051,14 @@ skip:
 
 				ksmbd_vfs_posix_lock_wait(flock);
 
+				spin_lock(&work->conn->request_lock);
+				spin_lock(&fp->f_lock);
+				list_del(&work->fp_entry);
+				work->cancel_fn = NULL;
+				kfree(argv);
+				spin_unlock(&fp->f_lock);
+				spin_unlock(&work->conn->request_lock);
+
 				if (work->state != KSMBD_WORK_ACTIVE) {
 					list_del(&smb_lock->llist);
 					spin_lock(&work->conn->llist_lock);
@@ -7063,9 +7067,6 @@ skip:
 					locks_free_lock(flock);
 
 					if (work->state == KSMBD_WORK_CANCELLED) {
-						spin_lock(&fp->f_lock);
-						list_del(&work->fp_entry);
-						spin_unlock(&fp->f_lock);
 						rsp->hdr.Status =
 							STATUS_CANCELLED;
 						kfree(smb_lock);
@@ -7087,9 +7088,6 @@ skip:
 				list_del(&smb_lock->clist);
 				spin_unlock(&work->conn->llist_lock);
 
-				spin_lock(&fp->f_lock);
-				list_del(&work->fp_entry);
-				spin_unlock(&fp->f_lock);
 				goto retry;
 			} else if (!rc) {
 				spin_lock(&work->conn->llist_lock);
fs/ksmbd/vfs_cache.c +2 -3
@@ -364,12 +364,11 @@ static void __put_fd_final(struct ksmbd_
 
 static void set_close_state_blocked_works(struct ksmbd_file *fp)
 {
-	struct ksmbd_work *cancel_work, *ctmp;
+	struct ksmbd_work *cancel_work;
 
 	spin_lock(&fp->f_lock);
-	list_for_each_entry_safe(cancel_work, ctmp, &fp->blocked_works,
+	list_for_each_entry(cancel_work, &fp->blocked_works,
 				 fp_entry) {
-		list_del(&cancel_work->fp_entry);
 		cancel_work->state = KSMBD_WORK_CLOSED;
 		cancel_work->cancel_fn(cancel_work->cancel_argv);
 	}
author maintainer
#687 lore raw

PATCH 686/885 torture: Fix hang during kthread shutdown phase

In reply to Greg Kroah-Hartman

From: Joel Fernandes (Google) <[email protected]>

commit d52d3a2bf408ff86f3a79560b5cce80efb340239 upstream.

During rcutorture shutdown, the rcu_torture_cleanup() function calls torture_cleanup_begin(), which sets the fullstop global variable to FULLSTOP_RMMOD. This causes the rcutorture threads for readers and fakewriters to exit all of their "while" loops and start shutting down.

They then call torture_kthread_stopping(), which in turn waits for kthread_stop() to be called. However, rcu_torture_cleanup() has not yet called kthread_stop() on those threads, and before it gets a chance to do so, multiple instances of torture_kthread_stopping() invoke schedule_timeout_interruptible(1) in a tight loop. Tracing confirms that TIMER_SOFTIRQ can then continuously execute timer callbacks. If that TIMER_SOFTIRQ preempts the task executing rcu_torture_cleanup(), that task might never invoke kthread_stop().

This commit improves this situation by increasing the timeout passed to schedule_timeout_interruptible() from one jiffy to 1/20th of a second. This change prevents TIMER_SOFTIRQ from monopolizing its CPU, thus allowing rcu_torture_cleanup() to carry out the needed kthread_stop() invocations. Testing has shown 100 runs of TREE07 passing reliably, as oppose to the tens-of-percent failure rates seen beforehand.

---
 kernel/torture.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
kernel/torture.c +1 -1
index 29afc62f2bfe..1a0519b836ac 100644
@@ -915,7 +915,7 @@ void torture_kthread_stopping(char *title)
 	VERBOSE_TOROUT_STRING(buf);
 	while (!kthread_should_stop()) {
 		torture_shutdown_absorb(title);
-		schedule_timeout_uninterruptible(1);
+		schedule_timeout_uninterruptible(HZ / 20);
 	}
 }
 EXPORT_SYMBOL_GPL(torture_kthread_stopping);
Signature · 1 line
2.39.2
author maintainer
#688 lore raw

PATCH 687/885 ARM: dts: exynos: correct HDMI phy compatible in Exynos4

In reply to Greg Kroah-Hartman

From: Krzysztof Kozlowski <[email protected]>

commit af1c89ddb74f170eccd5a57001d7317560b638ea upstream.

The HDMI phy compatible was missing vendor prefix.

Fixes ed80d4cab772 ("ARM: dts: add hdmi related nodes for exynos4 SoCs")
---
 arch/arm/boot/dts/exynos4.dtsi |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/arm/boot/dts/exynos4.dtsi +1 -1
@@ -605,7 +605,7 @@
 			status = "disabled";
 
 			hdmi_i2c_phy: hdmiphy@38 {
-				compatible = "exynos4210-hdmiphy";
+				compatible = "samsung,exynos4210-hdmiphy";
 				reg = <0x38>;
 			};
 		};
author maintainer
#689 lore raw

PATCH 688/885 io_uring: mark task TASK_RUNNING before handling resume/task work

In reply to Greg Kroah-Hartman

From: Jens Axboe <[email protected]>

commit 2f2bb1ffc9983e227424d0787289da5483b0c74f upstream.

Just like for task_work, set the task mode to TASK_RUNNING before doing any potential resume work. We're not holding any locks at this point, but we may have already set the task state to TASK_INTERRUPTIBLE in preparation for going to sleep waiting for events. Ensure that we set it back to TASK_RUNNING if we have work to process, to avoid warnings on calling blocking operations with !TASK_RUNNING.

Fixes b5d3ae202fbf ("io_uring: handle TIF_NOTIFY_RESUME when checking for task_work")
---
 io_uring/io_uring.h |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
io_uring/io_uring.h +3 -1
@@ -261,8 +261,10 @@ static inline int io_run_task_work(void)
 	 * notify work that needs processing.
 	 */
 	if (current->flags & PF_IO_WORKER &&
-	    test_thread_flag(TIF_NOTIFY_RESUME))
+	    test_thread_flag(TIF_NOTIFY_RESUME)) {
+		__set_current_state(TASK_RUNNING);
 		resume_user_mode_work(NULL);
+	}
 	if (task_work_pending(current)) {
 		__set_current_state(TASK_RUNNING);
 		task_work_run();
author maintainer
#690 lore raw

PATCH 689/885 hfs: fix missing hfs_bnode_get() in __hfs_bnode_create

In reply to Greg Kroah-Hartman

From: Liu Shixin <[email protected]>

commit a9dc087fd3c484fd1ed18c5efb290efaaf44ce03 upstream.

Syzbot found a kernel BUG in hfs_bnode_put():

kernel BUG at fs/hfs/bnode.c:466!
invalid opcode: 0000 [#1] PREEMPT SMP KASAN

 CPU: 0 PID: 3634 Comm: kworker/u4:5 Not tainted 6.1.0-rc7-syzkaller-00190-g97ee9d1c1696 #0
 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 10/26/2022
 Workqueue: writeback wb_workfn (flush-7:0)
 RIP: 0010:hfs_bnode_put+0x46f/0x480 fs/hfs/bnode.c:466
 Code: 8a 80 ff e9 73 fe ff ff 89 d9 80 e1 07 80 c1 03 38 c1 0f 8c a0 fe ff ff 48 89 df e8 db 8a 80 ff e9 93 fe ff ff e8 a1 68 2c ff <0f> 0b e8 9a 68 2c ff 0f 0b 0f 1f 84 00 00 00 00 00 55 41 57 41 56
 RSP: 0018:ffffc90003b4f258 EFLAGS: 00010293
 RAX: ffffffff825e318f RBX: 0000000000000000 RCX: ffff8880739dd7c0
 RDX: 0000000000000000 RSI: 0000000000000000 RDI: 0000000000000000
 RBP: ffffc90003b4f430 R08: ffffffff825e2d9b R09: ffffed10045157d1
 R10: ffffed10045157d1 R11: 1ffff110045157d0 R12: ffff8880228abe80
 R13: ffff88807016c000 R14: dffffc0000000000 R15: ffff8880228abe00
 FS:  0000000000000000(0000) GS:ffff8880b9800000(0000) knlGS:0000000000000000
 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
 CR2: 00007fa6ebe88718 CR3: 000000001e93d000 CR4: 00000000003506f0
 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000
 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400
 Call Trace:
  <TASK>
  hfs_write_inode+0x1bc/0xb40
  write_inode fs/fs-writeback.c:1440 [inline]
  __writeback_single_inode+0x4d6/0x670 fs/fs-writeback.c:1652
  writeback_sb_inodes+0xb3b/0x18f0 fs/fs-writeback.c:1878
  __writeback_inodes_wb+0x125/0x420 fs/fs-writeback.c:1949
  wb_writeback+0x440/0x7b0 fs/fs-writeback.c:2054
  wb_check_start_all fs/fs-writeback.c:2176 [inline]
  wb_do_writeback fs/fs-writeback.c:2202 [inline]
  wb_workfn+0x827/0xef0 fs/fs-writeback.c:2235
  process_one_work+0x877/0xdb0 kernel/workqueue.c:2289
  worker_thread+0xb14/0x1330 kernel/workqueue.c:2436
  kthread+0x266/0x300 kernel/kthread.c:376
  ret_from_fork+0x1f/0x30 arch/x86/entry/entry_64.S:306
  </TASK>

The BUG_ON() is triggered at here:

/* Dispose of resources used by a node */
void hfs_bnode_put(struct hfs_bnode *node)

{
	if (node) {

<skipped>
BUG_ON(!atomic_read(&node->refcnt)); <- we have issue here!!!! <skipped>
}
}

By tracing the refcnt, I found the node is created by hfs_bmap_alloc() with refcnt 1. Then the node is used by hfs_btree_write(). There is a missing of hfs_bnode_get() after find the node. The issue happened in following path:

<alloc>
hfs_bmap_alloc
hfs_bnode_find
__hfs_bnode_create <- allocate a new node with refcnt 1.
hfs_bnode_put <- decrease the refcnt

<write>
hfs_btree_write
hfs_bnode_find
__hfs_bnode_create
hfs_bnode_findhash <- find the node without refcnt increased.
hfs_bnode_put <- trigger the BUG_ON() since refcnt is 0.

---
 fs/hfs/bnode.c |    1 +
 1 file changed, 1 insertion(+)
fs/hfs/bnode.c +1 -0
@@ -274,6 +274,7 @@ static struct hfs_bnode *__hfs_bnode_cre
 		tree->node_hash[hash] = node;
 		tree->node_hash_cnt++;
 	} else {
+		hfs_bnode_get(node2);
 		spin_unlock(&tree->hash_lock);
 		kfree(node);
 		wait_event(node2->lock_wq, !test_bit(HFS_BNODE_NEW, &node2->flags));
author maintainer
#691 lore raw

PATCH 690/885 fs: hfsplus: fix UAF issue in hfsplus_put_super

In reply to Greg Kroah-Hartman

From: Dongliang Mu <[email protected]>

commit 07db5e247ab5858439b14dd7cc1fe538b9efcf32 upstream.

The current hfsplus_put_super first calls hfs_btree_close on sbi->ext_tree, then invokes iput on sbi->hidden_dir, resulting in an use-after-free issue in hfsplus_release_folio.

As shown in hfsplus_fill_super, the error handling code also calls iput before hfs_btree_close.

To fix this error, we move all iput calls before hfsplus_btree_close.

Note that this patch is tested on Syzbot.

---
 fs/hfsplus/super.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
fs/hfsplus/super.c +2 -2
@@ -295,11 +295,11 @@ static void hfsplus_put_super(struct sup
 		hfsplus_sync_fs(sb, 1);
 	}
 
+	iput(sbi->alloc_file);
+	iput(sbi->hidden_dir);
 	hfs_btree_close(sbi->attr_tree);
 	hfs_btree_close(sbi->cat_tree);
 	hfs_btree_close(sbi->ext_tree);
-	iput(sbi->alloc_file);
-	iput(sbi->hidden_dir);
 	kfree(sbi->s_vhdr_buf);
 	kfree(sbi->s_backup_vhdr_buf);
 	unload_nls(sbi->nls);
author maintainer
#692 lore raw

PATCH 691/885 exfat: fix reporting fs error when reading dir beyond EOF

In reply to Greg Kroah-Hartman

From: Yuezhang Mo <[email protected]>

commit 706fdcac002316893434d753be8cfb549fe1d40d upstream.

Since seekdir() does not check whether the position is valid, the position may exceed the size of the directory. We found that for a directory with discontinuous clusters, if the position exceeds the size of the directory and the excess size is greater than or equal to the cluster size, exfat_readdir() will return -EIO, causing a file system error and making the file system unavailable.

Reproduce this bug by:

seekdir(dir, dir_size + cluster_size);
dirent = readdir(dir);

The following log will be printed if mount with 'errors=remount-ro'.

[11166.712896] exFAT-fs (sdb1): error, invalid access to FAT (entry 0xffffffff)
[11166.712905] exFAT-fs (sdb1): Filesystem has been set read-only
Fixes 1e5654de0f51 ("exfat: handle wrong stream entry size in exfat_readdir()")
---
 fs/exfat/dir.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/exfat/dir.c +1 -1
@@ -102,7 +102,7 @@ static int exfat_readdir(struct inode *i
 			clu.dir = ei->hint_bmap.clu;
 		}
 
-		while (clu_offset > 0) {
+		while (clu_offset > 0 && clu.dir != EXFAT_EOF_CLUSTER) {
 			if (exfat_get_next_cluster(sb, &(clu.dir)))
 				return -EIO;
 
author maintainer
#693 lore raw

PATCH 692/885 exfat: fix unexpected EOF while reading dir

In reply to Greg Kroah-Hartman

From: Yuezhang Mo <[email protected]>

commit 6cb5d1a16a51d080fbc1649a5144cbc5ca7d6f88 upstream.

If the position is not aligned with the dentry size, the return value of readdir() will be NULL and errno is 0, which means the end of the directory stream is reached.

If the position is aligned with dentry size, but there is no file or directory at the position, exfat_readdir() will continue to get dentry from the next dentry. So the dentry gotten by readdir() may not be at the position.

After this commit, if the position is not aligned with the dentry size, round the position up to the dentry size and continue to get the dentry.

---
 fs/exfat/dir.c |    5 +----
 1 file changed, 1 insertion(+), 4 deletions(-)
fs/exfat/dir.c +1 -4
@@ -236,10 +236,7 @@ static int exfat_iterate(struct file *fi
 		fake_offset = 1;
 	}
 
-	if (cpos & (DENTRY_SIZE - 1)) {
-		err = -ENOENT;
-		goto unlock;
-	}
+	cpos = round_up(cpos, DENTRY_SIZE);
 
 	/* name buffer should be allocated before use */
 	err = exfat_alloc_namebuf(nb);
author maintainer
#694 lore raw

PATCH 693/885 exfat: redefine DIR_DELETED as the bad cluster number

In reply to Greg Kroah-Hartman

From: Sungjong Seo <[email protected]>

commit bdaadfd343e3cba49ad0b009ff4b148dad0fa404 upstream.

When a file or a directory is deleted, the hint for the cluster of its parent directory in its in-memory inode is set as DIR_DELETED. Therefore, DIR_DELETED must be one of invalid cluster numbers. According to the exFAT specification, a volume can have at most 2^32-11 clusters. However, DIR_DELETED is wrongly defined as 0xFFFF0321, which could be a valid cluster number. To fix it, let's redefine DIR_DELETED as 0xFFFFFFF7, the bad cluster number.

Fixes 1acf1a564b60 ("exfat: add in-memory and on-disk structures and headers")
---
 fs/exfat/exfat_fs.h |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/exfat/exfat_fs.h +1 -1
@@ -41,7 +41,7 @@ enum {
 #define ES_2_ENTRIES		2
 #define ES_ALL_ENTRIES		0
 
-#define DIR_DELETED		0xFFFF0321
+#define DIR_DELETED		0xFFFFFFF7
 
 /* type values */
 #define TYPE_UNUSED		0x0000
author maintainer
#695 lore raw

PATCH 694/885 exfat: fix inode->i_blocks for non-512 byte sector size device

In reply to Greg Kroah-Hartman

From: Yuezhang Mo <[email protected]>

commit 39c1ce8eafc0ff64fb9e28536ccc7df6a8e2999d upstream.

inode->i_blocks is not real number of blocks, but 512 byte ones.

---
 fs/exfat/file.c  |    3 +--
 fs/exfat/inode.c |    6 ++----
 fs/exfat/namei.c |    2 +-
 fs/exfat/super.c |    3 +--
 4 files changed, 5 insertions(+), 9 deletions(-)
fs/exfat/file.c +1 -2
@@ -211,8 +211,7 @@ void exfat_truncate(struct inode *inode,
 	if (err)
 		goto write_size;
 
-	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >>
-				inode->i_blkbits;
+	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9;
 write_size:
 	aligned_size = i_size_read(inode);
 	if (aligned_size & (blocksize - 1)) {
fs/exfat/inode.c +2 -4
@@ -221,8 +221,7 @@ static int exfat_map_cluster(struct inod
 		num_clusters += num_to_be_allocated;
 		*clu = new_clu.dir;
 
-		inode->i_blocks +=
-			num_to_be_allocated << sbi->sect_per_clus_bits;
+		inode->i_blocks += EXFAT_CLU_TO_B(num_to_be_allocated, sbi) >> 9;
 
 		/*
 		 * Move *clu pointer along FAT chains (hole care) because the
@@ -582,8 +581,7 @@ static int exfat_fill_inode(struct inode
 
 	exfat_save_attr(inode, info->attr);
 
-	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >>
-				inode->i_blkbits;
+	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9;
 	inode->i_mtime = info->mtime;
 	inode->i_ctime = info->mtime;
 	ei->i_crtime = info->crtime;
fs/exfat/namei.c +1 -1
@@ -387,7 +387,7 @@ static int exfat_find_empty_entry(struct
 		ei->i_size_ondisk += sbi->cluster_size;
 		ei->i_size_aligned += sbi->cluster_size;
 		ei->flags = p_dir->flags;
-		inode->i_blocks += 1 << sbi->sect_per_clus_bits;
+		inode->i_blocks += sbi->cluster_size >> 9;
 	}
 
 	return dentry;
fs/exfat/super.c +1 -2
@@ -373,8 +373,7 @@ static int exfat_read_root(struct inode
 	inode->i_op = &exfat_dir_inode_operations;
 	inode->i_fop = &exfat_dir_operations;
 
-	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >>
-				inode->i_blkbits;
+	inode->i_blocks = round_up(i_size_read(inode), sbi->cluster_size) >> 9;
 	ei->i_pos = ((loff_t)sbi->root_dir << 32) | 0xffffffff;
 	ei->i_size_aligned = i_size_read(inode);
 	ei->i_size_ondisk = i_size_read(inode);
author maintainer
#696 lore raw

PATCH 695/885 fs: dlm: dont set stop rx flag after node reset

In reply to Greg Kroah-Hartman

From: Alexander Aring <[email protected]>

commit 15c63db8e86a72e0d5cfb9bf0cd1870e39a3e5fe upstream.

Similar to the stop tx flag, the rx flag should warn about a dlm message being received at DLM_FIN state change, when we are assuming no other dlm application messages. If we receive a FIN message and we are in the state DLM_FIN_WAIT2 we call midcomms_node_reset() which puts the midcomms node into DLM_CLOSED state. Afterwards we should not set the DLM_NODE_FLAG_STOP_RX flag any more. This patch changes the setting DLM_NODE_FLAG_STOP_RX in those state changes when we receive a FIN message and we assume there will be no other dlm application messages received until we hit DLM_CLOSED state.

Fixes 489d8e559c65 ("fs: dlm: add reliable connection if reconnect")
---
 fs/dlm/midcomms.c |    3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)
fs/dlm/midcomms.c +1 -2
@@ -509,6 +509,7 @@ static void dlm_midcomms_receive_buffer(
 				break;
 			case DLM_FIN_WAIT1:
 				node->state = DLM_CLOSING;
+				set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
 				pr_debug("switch node %d to state %s\n",
 					 node->nodeid, dlm_state_str(node->state));
 				break;
@@ -529,8 +530,6 @@ static void dlm_midcomms_receive_buffer(
 				return;
 			}
 			spin_unlock(&node->state_lock);
-
-			set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
 			break;
 		default:
 			WARN_ON(test_bit(DLM_NODE_FLAG_STOP_RX, &node->flags));
author maintainer
#697 lore raw

PATCH 696/885 fs: dlm: move sending fin message into state change handling

In reply to Greg Kroah-Hartman

From: Alexander Aring <[email protected]>

commit a58496361802070996f9bd76e941d109c4a85ebd upstream.

This patch moves the send fin handling, which should appear in a specific state change, into the state change handling while the per node state_lock is held. I experienced issues with other messages because we changed the state and a fin message was sent out in a different state.

Fixes 489d8e559c65 ("fs: dlm: add reliable connection if reconnect")
---
 fs/dlm/midcomms.c |   33 +++++++++------------------------
 1 file changed, 9 insertions(+), 24 deletions(-)
fs/dlm/midcomms.c +9 -24
@@ -401,7 +401,7 @@ static int dlm_send_fin(struct midcomms_
 	struct dlm_mhandle *mh;
 	char *ppc;
 
-	mh = dlm_midcomms_get_mhandle(node->nodeid, mb_len, GFP_NOFS, &ppc);
+	mh = dlm_midcomms_get_mhandle(node->nodeid, mb_len, GFP_ATOMIC, &ppc);
 	if (!mh)
 		return -ENOMEM;
 
@@ -503,8 +503,8 @@ static void dlm_midcomms_receive_buffer(
 					node->state = DLM_LAST_ACK;
 					pr_debug("switch node %d to state %s case 1\n",
 						 node->nodeid, dlm_state_str(node->state));
-					spin_unlock(&node->state_lock);
-					goto send_fin;
+					set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
+					dlm_send_fin(node, dlm_pas_fin_ack_rcv);
 				}
 				break;
 			case DLM_FIN_WAIT1:
@@ -547,12 +547,6 @@ static void dlm_midcomms_receive_buffer(
 		log_print_ratelimited("ignore dlm msg because seq mismatch, seq: %u, expected: %u, nodeid: %d",
 				      seq, node->seq_next, node->nodeid);
 	}
-
-	return;
-
-send_fin:
-	set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
-	dlm_send_fin(node, dlm_pas_fin_ack_rcv);
 }
 
 static struct midcomms_node *
@@ -1286,11 +1280,11 @@ void dlm_midcomms_remove_member(int node
 		case DLM_CLOSE_WAIT:
 			/* passive shutdown DLM_LAST_ACK case 2 */
 			node->state = DLM_LAST_ACK;
-			spin_unlock(&node->state_lock);
-
 			pr_debug("switch node %d to state %s case 2\n",
 				 node->nodeid, dlm_state_str(node->state));
-			goto send_fin;
+			set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
+			dlm_send_fin(node, dlm_pas_fin_ack_rcv);
+			break;
 		case DLM_LAST_ACK:
 			/* probably receive fin caught it, do nothing */
 			break;
@@ -1306,12 +1300,6 @@ void dlm_midcomms_remove_member(int node
 	spin_unlock(&node->state_lock);
 
 	srcu_read_unlock(&nodes_srcu, idx);
-	return;
-
-send_fin:
-	set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
-	dlm_send_fin(node, dlm_pas_fin_ack_rcv);
-	srcu_read_unlock(&nodes_srcu, idx);
 }
 
 static void midcomms_node_release(struct rcu_head *rcu)
@@ -1342,6 +1330,7 @@ static void midcomms_shutdown(struct mid
 		node->state = DLM_FIN_WAIT1;
 		pr_debug("switch node %d to state %s case 2\n",
 			 node->nodeid, dlm_state_str(node->state));
+		dlm_send_fin(node, dlm_act_fin_ack_rcv);
 		break;
 	case DLM_CLOSED:
 		/* we have what we want */
@@ -1355,12 +1344,8 @@ static void midcomms_shutdown(struct mid
 	}
 	spin_unlock(&node->state_lock);
 
-	if (node->state == DLM_FIN_WAIT1) {
-		dlm_send_fin(node, dlm_act_fin_ack_rcv);
-
-		if (DLM_DEBUG_FENCE_TERMINATION)
-			msleep(5000);
-	}
+	if (DLM_DEBUG_FENCE_TERMINATION)
+		msleep(5000);
 
 	/* wait for other side dlm + fin */
 	ret = wait_event_timeout(node->shutdown_wait,
author maintainer
#698 lore raw

PATCH 697/885 fs: dlm: send FIN ack back in right cases

In reply to Greg Kroah-Hartman

From: Alexander Aring <[email protected]>

commit 00908b3388255fc1d3782b744d07f327712f401f upstream.

This patch moves to send a ack back for receiving a FIN message only when we are in valid states. In other cases and there might be a sender waiting for a ack we just let it timeout at the senders time and hopefully all other cleanups will remove the FIN message on their sending queue. As an example we should never send out an ACK being in LAST_ACK state or we cannot assume a working socket communication when we are in CLOSED state.

Fixes 489d8e559c65 ("fs: dlm: add reliable connection if reconnect")
---
 fs/dlm/midcomms.c |    9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)
fs/dlm/midcomms.c +5 -4
@@ -374,7 +374,7 @@ static int dlm_send_ack(int nodeid, uint
 	struct dlm_msg *msg;
 	char *ppc;
 
-	msg = dlm_lowcomms_new_msg(nodeid, mb_len, GFP_NOFS, &ppc,
+	msg = dlm_lowcomms_new_msg(nodeid, mb_len, GFP_ATOMIC, &ppc,
 				   NULL, NULL);
 	if (!msg)
 		return -ENOMEM;
@@ -483,15 +483,14 @@ static void dlm_midcomms_receive_buffer(
 
 		switch (p->header.h_cmd) {
 		case DLM_FIN:
-			/* send ack before fin */
-			dlm_send_ack(node->nodeid, node->seq_next);
-
 			spin_lock(&node->state_lock);
 			pr_debug("receive fin msg from node %d with state %s\n",
 				 node->nodeid, dlm_state_str(node->state));
 
 			switch (node->state) {
 			case DLM_ESTABLISHED:
+				dlm_send_ack(node->nodeid, node->seq_next);
+
 				node->state = DLM_CLOSE_WAIT;
 				pr_debug("switch node %d to state %s\n",
 					 node->nodeid, dlm_state_str(node->state));
@@ -508,12 +507,14 @@ static void dlm_midcomms_receive_buffer(
 				}
 				break;
 			case DLM_FIN_WAIT1:
+				dlm_send_ack(node->nodeid, node->seq_next);
 				node->state = DLM_CLOSING;
 				set_bit(DLM_NODE_FLAG_STOP_RX, &node->flags);
 				pr_debug("switch node %d to state %s\n",
 					 node->nodeid, dlm_state_str(node->state));
 				break;
 			case DLM_FIN_WAIT2:
+				dlm_send_ack(node->nodeid, node->seq_next);
 				midcomms_node_reset(node);
 				pr_debug("switch node %d to state %s\n",
 					 node->nodeid, dlm_state_str(node->state));
author maintainer
#699 lore raw

PATCH 698/885 f2fs: fix information leak in f2fs_move_inline_dirents()

In reply to Greg Kroah-Hartman

From: Eric Biggers <[email protected]>

commit 9a5571cff4ffcfc24847df9fd545cc5799ac0ee5 upstream.

When converting an inline directory to a regular one, f2fs is leaking uninitialized memory to disk because it doesn't initialize the entire directory block. Fix this by zero-initializing the block.

This bug was introduced by commit 4ec17d688d74 ("f2fs: avoid unneeded initializing when converting inline dentry"), which didn't consider the security implications of leaking uninitialized memory to disk.

This was found by running xfstest generic/435 on a KMSAN-enabled kernel.

Fixes 4ec17d688d74 ("f2fs: avoid unneeded initializing when converting inline dentry")
Cc <[email protected]> # v4.3+
---
 fs/f2fs/inline.c |   13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)
fs/f2fs/inline.c +6 -7
@@ -422,18 +422,17 @@ static int f2fs_move_inline_dirents(stru
 
 	dentry_blk = page_address(page);
 
+	/*
+	 * Start by zeroing the full block, to ensure that all unused space is
+	 * zeroed and no uninitialized memory is leaked to disk.
+	 */
+	memset(dentry_blk, 0, F2FS_BLKSIZE);
+
 	make_dentry_ptr_inline(dir, &src, inline_dentry);
 	make_dentry_ptr_block(dir, &dst, dentry_blk);
 
 	/* copy data from inline dentry block to new dentry block */
 	memcpy(dst.bitmap, src.bitmap, src.nr_bitmap);
-	memset(dst.bitmap + src.nr_bitmap, 0, dst.nr_bitmap - src.nr_bitmap);
-	/*
-	 * we do not need to zero out remainder part of dentry and filename
-	 * field, since we have used bitmap for marking the usage status of
-	 * them, besides, we can also ignore copying/zeroing reserved space
-	 * of dentry block, because them haven't been used so far.
-	 */
 	memcpy(dst.dentry, src.dentry, SIZE_OF_DIR_ENTRY * src.max);
 	memcpy(dst.filename, src.filename, src.max * F2FS_SLOT_LEN);
 
author maintainer
#700 lore raw

PATCH 699/885 f2fs: retry to update the inode page given data corruption

In reply to Greg Kroah-Hartman

From: Jaegeuk Kim <[email protected]>

commit 3aa51c61cb4a4dcb40df51ac61171e9ac5a35321 upstream.

If the storage gives a corrupted node block due to short power failure and reset, f2fs stops the entire operations by setting the checkpoint failure flag.

Let's give more chances to live by re-issuing IOs for a while in such critical path.

---
 fs/f2fs/inode.c |   13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)
fs/f2fs/inode.c +7 -6
@@ -708,18 +708,19 @@ void f2fs_update_inode_page(struct inode
 {
 	struct f2fs_sb_info *sbi = F2FS_I_SB(inode);
 	struct page *node_page;
+	int count = 0;
 retry:
 	node_page = f2fs_get_node_page(sbi, inode->i_ino);
 	if (IS_ERR(node_page)) {
 		int err = PTR_ERR(node_page);
 
-		if (err == -ENOMEM) {
-			cond_resched();
+		/* The node block was truncated. */
+		if (err == -ENOENT)
+			return;
+
+		if (err == -ENOMEM || ++count <= DEFAULT_RETRY_IO_COUNT)
 			goto retry;
-		} else if (err != -ENOENT) {
-			f2fs_stop_checkpoint(sbi, false,
-					STOP_CP_REASON_UPDATE_INODE);
-		}
+		f2fs_stop_checkpoint(sbi, false, STOP_CP_REASON_UPDATE_INODE);
 		return;
 	}
 	f2fs_update_inode(inode, node_page);
author maintainer
#701 lore raw

PATCH 700/885 f2fs: fix cgroup writeback accounting with fs-layer encryption

In reply to Greg Kroah-Hartman

From: Eric Biggers <[email protected]>

commit 844545c51a5b2a524b22a2fe9d0b353b827d24b4 upstream.

When writing a page from an encrypted file that is using filesystem-layer encryption (not inline encryption), f2fs encrypts the pagecache page into a bounce page, then writes the bounce page.

It also passes the bounce page to wbc_account_cgroup_owner(). That's incorrect, because the bounce page is a newly allocated temporary page that doesn't have the memory cgroup of the original pagecache page. This makes wbc_account_cgroup_owner() not account the I/O to the owner of the pagecache page as it should.

Fix this by always passing the pagecache page to
wbc_account_cgroup_owner().

---
 fs/f2fs/data.c |    6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
fs/f2fs/data.c +3 -3
@@ -726,7 +726,7 @@ int f2fs_submit_page_bio(struct f2fs_io_
 	}
 
 	if (fio->io_wbc && !is_read_io(fio->op))
-		wbc_account_cgroup_owner(fio->io_wbc, page, PAGE_SIZE);
+		wbc_account_cgroup_owner(fio->io_wbc, fio->page, PAGE_SIZE);
 
 	inc_page_count(fio->sbi, is_read_io(fio->op) ?
 			__read_io_type(page) : WB_DATA_TYPE(fio->page));
@@ -933,7 +933,7 @@ alloc_new:
 	}
 
 	if (fio->io_wbc)
-		wbc_account_cgroup_owner(fio->io_wbc, page, PAGE_SIZE);
+		wbc_account_cgroup_owner(fio->io_wbc, fio->page, PAGE_SIZE);
 
 	inc_page_count(fio->sbi, WB_DATA_TYPE(page));
 
@@ -1007,7 +1007,7 @@ alloc_new:
 	}
 
 	if (fio->io_wbc)
-		wbc_account_cgroup_owner(fio->io_wbc, bio_page, PAGE_SIZE);
+		wbc_account_cgroup_owner(fio->io_wbc, fio->page, PAGE_SIZE);
 
 	io->last_block_in_bio = fio->new_blkaddr;
 
author maintainer
#702 lore raw

PATCH 701/885 f2fs: fix kernel crash due to null io->bio

In reply to Greg Kroah-Hartman

From: Jaegeuk Kim <[email protected]>

commit 267c159f9c7bcb7009dae16889b880c5ed8759a8 upstream.

We should return when io->bio is null before doing anything. Otherwise, panic.

BUG: kernel NULL pointer dereference, address: 0000000000000010
RIP: 0010:__submit_merged_write_cond+0x164/0x240 [f2fs]
Call Trace:
 <TASK>
 f2fs_submit_merged_write+0x1d/0x30 [f2fs]
 commit_checkpoint+0x110/0x1e0 [f2fs]
 f2fs_write_checkpoint+0x9f7/0xf00 [f2fs]
 ? __pfx_issue_checkpoint_thread+0x10/0x10 [f2fs]
 __checkpoint_and_complete_reqs+0x84/0x190 [f2fs]
 ? preempt_count_add+0x82/0xc0
 ? __pfx_issue_checkpoint_thread+0x10/0x10 [f2fs]
 issue_checkpoint_thread+0x4c/0xf0 [f2fs]
 ? __pfx_autoremove_wake_function+0x10/0x10
 kthread+0xff/0x130
 ? __pfx_kthread+0x10/0x10
 ret_from_fork+0x2c/0x50
 </TASK>
Cc [email protected] # v5.18+
Fixes 64bf0eef0171 ("f2fs: pass the bio operation to bio_alloc_bioset")
---
 fs/f2fs/data.c |    4 ++++
 1 file changed, 4 insertions(+)
fs/f2fs/data.c +4 -0
@@ -640,6 +640,9 @@ static void __f2fs_submit_merged_write(s
 
 	f2fs_down_write(&io->io_rwsem);
 
+	if (!io->bio)
+		goto unlock_out;
+
 	/* change META to META_FLUSH in the checkpoint procedure */
 	if (type >= META_FLUSH) {
 		io->fio.type = META_FLUSH;
@@ -648,6 +651,7 @@ static void __f2fs_submit_merged_write(s
 			io->bio->bi_opf |= REQ_PREFLUSH | REQ_FUA;
 	}
 	__submit_merged_bio(io);
+unlock_out:
 	f2fs_up_write(&io->io_rwsem);
 }
 
author maintainer
#703 lore raw

PATCH 702/885 ocfs2: fix defrag path triggering jbd2 ASSERT

In reply to Greg Kroah-Hartman

From: Heming Zhao via Ocfs2-devel <[email protected]>

commit 60eed1e3d45045623e46944ebc7c42c30a4350f0 upstream.

code path:

ocfs2_ioctl_move_extents
ocfs2_move_extents
ocfs2_defrag_extent
__ocfs2_move_extent
+ ocfs2_journal_access_di
+ ocfs2_split_extent //sub-paths call jbd2_journal_restart
+ ocfs2_journal_dirty //crash by jbs2 ASSERT

crash stacks:

PID: 11297 TASK: ffff974a676dcd00 CPU: 67 COMMAND: "defragfs.ocfs2"
#0 [ffffb25d8dad3900] machine_kexec at ffffffff8386fe01
#1 [ffffb25d8dad3958] __crash_kexec at ffffffff8395959d
#2 [ffffb25d8dad3a20] crash_kexec at ffffffff8395a45d
#3 [ffffb25d8dad3a38] oops_end at ffffffff83836d3f
#4 [ffffb25d8dad3a58] do_trap at ffffffff83833205
#5 [ffffb25d8dad3aa0] do_invalid_op at ffffffff83833aa6
#6 [ffffb25d8dad3ac0] invalid_op at ffffffff84200d18

    [exception RIP: jbd2_journal_dirty_metadata+0x2ba]
    RIP: ffffffffc09ca54a  RSP: ffffb25d8dad3b70  RFLAGS: 00010207
    RAX: 0000000000000000  RBX: ffff9706eedc5248  RCX: 0000000000000000
    RDX: 0000000000000001  RSI: ffff97337029ea28  RDI: ffff9706eedc5250
    RBP: ffff9703c3520200   R8: 000000000f46b0b2   R9: 0000000000000000
    R10: 0000000000000001  R11: 00000001000000fe  R12: ffff97337029ea28
    R13: 0000000000000000  R14: ffff9703de59bf60  R15: ffff9706eedc5250
    ORIG_RAX: ffffffffffffffff  CS: 0010  SS: 0018

#7 [ffffb25d8dad3ba8] ocfs2_journal_dirty at ffffffffc137fb95 [ocfs2] #8 [ffffb25d8dad3be8] __ocfs2_move_extent at ffffffffc139a950 [ocfs2] #9 [ffffb25d8dad3c80] ocfs2_defrag_extent at ffffffffc139b2d2 [ocfs2]

Analysis

This bug has the same root cause of 'commit 7f27ec978b0e ("ocfs2: call ocfs2_journal_access_di() before ocfs2_journal_dirty() in ocfs2_write_end_nolock()")'. For this bug, jbd2_journal_restart() is called by ocfs2_split_extent() during defragmenting.

How to fix

For ocfs2_split_extent() can handle journal operations totally by itself. Caller doesn't need to call journal access/dirty pair, and caller only needs to call journal start/stop pair. The fix method is to remove journal access/dirty from __ocfs2_move_extent().

The discussion for this patch:
https://oss.oracle.com/pipermail/ocfs2-devel/2023-February/000647.html

---
 fs/ocfs2/move_extents.c |   10 ----------
 1 file changed, 10 deletions(-)
fs/ocfs2/move_extents.c +0 -10
@@ -105,14 +105,6 @@ static int __ocfs2_move_extent(handle_t
 	 */
 	replace_rec.e_flags = ext_flags & ~OCFS2_EXT_REFCOUNTED;
 
-	ret = ocfs2_journal_access_di(handle, INODE_CACHE(inode),
-				      context->et.et_root_bh,
-				      OCFS2_JOURNAL_ACCESS_WRITE);
-	if (ret) {
-		mlog_errno(ret);
-		goto out;
-	}
-
 	ret = ocfs2_split_extent(handle, &context->et, path, index,
 				 &replace_rec, context->meta_ac,
 				 &context->dealloc);
@@ -121,8 +113,6 @@ static int __ocfs2_move_extent(handle_t
 		goto out;
 	}
 
-	ocfs2_journal_dirty(handle, context->et.et_root_bh);
-
 	context->new_phys_cpos = new_p_cpos;
 
 	/*
author maintainer
#704 lore raw

PATCH 703/885 ocfs2: fix non-auto defrag path not working issue

In reply to Greg Kroah-Hartman

From: Heming Zhao via Ocfs2-devel <[email protected]>

commit 236b9254f8d1edc273ad88b420aa85fbd84f492d upstream.

This fixes three issues on move extents ioctl without auto defrag:

a) In ocfs2_find_victim_alloc_group(), we have to convert bits to block first in case of global bitmap.

b) In ocfs2_probe_alloc_group(), when finding enough bits in block group bitmap, we have to back off move_len to start pos as well, otherwise it may corrupt filesystem.

c) In ocfs2_ioctl_move_extents(), set me_threshold both for non-auto and auto defrag paths. Otherwise it will set move_max_hop to 0 and finally cause unexpectedly ENOSPC error.

Currently there are no tools triggering the above issues since defragfs.ocfs2 enables auto defrag by default. Tested with manually changing defragfs.ocfs2 to run non auto defrag path.

---
 fs/ocfs2/move_extents.c |   24 +++++++++++++-----------
 1 file changed, 13 insertions(+), 11 deletions(-)
fs/ocfs2/move_extents.c +13 -11
@@ -434,7 +434,7 @@ static int ocfs2_find_victim_alloc_group
 			bg = (struct ocfs2_group_desc *)gd_bh->b_data;
 
 			if (vict_blkno < (le64_to_cpu(bg->bg_blkno) +
-						le16_to_cpu(bg->bg_bits))) {
+						(le16_to_cpu(bg->bg_bits) << bits_per_unit))) {
 
 				*ret_bh = gd_bh;
 				*vict_bit = (vict_blkno - blkno) >>
@@ -549,6 +549,7 @@ static void ocfs2_probe_alloc_group(stru
 			last_free_bits++;
 
 		if (last_free_bits == move_len) {
+			i -= move_len;
 			*goal_bit = i;
 			*phys_cpos = base_cpos + i;
 			break;
@@ -1020,18 +1021,19 @@ int ocfs2_ioctl_move_extents(struct file
 
 	context->range = &range;
 
+	/*
+	 * ok, the default theshold for the defragmentation
+	 * is 1M, since our maximum clustersize was 1M also.
+	 * any thought?
+	 */
+	if (!range.me_threshold)
+		range.me_threshold = 1024 * 1024;
+
+	if (range.me_threshold > i_size_read(inode))
+		range.me_threshold = i_size_read(inode);
+
 	if (range.me_flags & OCFS2_MOVE_EXT_FL_AUTO_DEFRAG) {
 		context->auto_defrag = 1;
-		/*
-		 * ok, the default theshold for the defragmentation
-		 * is 1M, since our maximum clustersize was 1M also.
-		 * any thought?
-		 */
-		if (!range.me_threshold)
-			range.me_threshold = 1024 * 1024;
-
-		if (range.me_threshold > i_size_read(inode))
-			range.me_threshold = i_size_read(inode);
 
 		if (range.me_flags & OCFS2_MOVE_EXT_FL_PART_DEFRAG)
 			context->partial = 1;
author maintainer
#705 lore raw

PATCH 704/885 fs/cramfs/inode.c: initialize file_ra_state

In reply to Greg Kroah-Hartman

From: Andrew Morton <[email protected]>

commit 3e35102666f873a135d31a726ac1ec8af4905206 upstream.

file_ra_state_init() assumes that the file_ra_state has been zeroed out. Fixes a KMSAN used-unintialized issue (at least).

Fixes cf948cbc35e80 ("cramfs: read_mapping_page() is synchronous")
Cc Matthew Wilcox <[email protected]>
Cc Nicolas Pitre <[email protected]>
---
 fs/cramfs/inode.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
fs/cramfs/inode.c +1 -1
index e3d168911dbe..006ef68d7ff6 100644
@@ -183,7 +183,7 @@ static void *cramfs_blkdev_read(struct super_block *sb, unsigned int offset,
 				unsigned int len)
 {
 	struct address_space *mapping = sb->s_bdev->bd_inode->i_mapping;
-	struct file_ra_state ra;
+	struct file_ra_state ra = {};
 	struct page *pages[BLKS_PER_BUF];
 	unsigned i, blocknr, buffer;
 	unsigned long devsize;
Signature · 1 line
2.39.2
author maintainer
#706 lore raw

PATCH 705/885 selftests/landlock: Skip overlayfs tests when not supported

In reply to Greg Kroah-Hartman

From: Jeff Xu <[email protected]>

commit 366617a69e60610912836570546f118006ebc7cb upstream.

overlayfs may be disabled in the kernel configuration, causing related tests to fail. Check that overlayfs is supported at runtime, so we can skip layout2_overlay.* accordingly.

---
 tools/testing/selftests/landlock/fs_test.c |   47 +++++++++++++++++++++++++++++
 1 file changed, 47 insertions(+)
tools/testing/selftests/landlock/fs_test.c +47 -0
@@ -11,6 +11,7 @@
 #include <fcntl.h>
 #include <linux/landlock.h>
 #include <sched.h>
+#include <stdio.h>
 #include <string.h>
 #include <sys/capability.h>
 #include <sys/mount.h>
@@ -87,6 +88,40 @@ static const char dir_s3d3[] = TMP_DIR "
  *         └── s3d3
  */
 
+static bool fgrep(FILE *const inf, const char *const str)
+{
+	char line[32];
+	const int slen = strlen(str);
+
+	while (!feof(inf)) {
+		if (!fgets(line, sizeof(line), inf))
+			break;
+		if (strncmp(line, str, slen))
+			continue;
+
+		return true;
+	}
+
+	return false;
+}
+
+static bool supports_overlayfs(void)
+{
+	bool res;
+	FILE *const inf = fopen("/proc/filesystems", "r");
+
+	/*
+	 * Consider that the filesystem is supported if we cannot get the
+	 * supported ones.
+	 */
+	if (!inf)
+		return true;
+
+	res = fgrep(inf, "nodev\toverlay\n");
+	fclose(inf);
+	return res;
+}
+
 static void mkdir_parents(struct __test_metadata *const _metadata,
 			  const char *const path)
 {
@@ -3539,6 +3574,9 @@ FIXTURE(layout2_overlay) {};
 
 FIXTURE_SETUP(layout2_overlay)
 {
+	if (!supports_overlayfs())
+		SKIP(return, "overlayfs is not supported");
+
 	prepare_layout(_metadata);
 
 	create_directory(_metadata, LOWER_BASE);
@@ -3575,6 +3613,9 @@ FIXTURE_SETUP(layout2_overlay)
 
 FIXTURE_TEARDOWN(layout2_overlay)
 {
+	if (!supports_overlayfs())
+		SKIP(return, "overlayfs is not supported");
+
 	EXPECT_EQ(0, remove_path(lower_do1_fl3));
 	EXPECT_EQ(0, remove_path(lower_dl1_fl2));
 	EXPECT_EQ(0, remove_path(lower_fl1));
@@ -3606,6 +3647,9 @@ FIXTURE_TEARDOWN(layout2_overlay)
 
 TEST_F_FORK(layout2_overlay, no_restriction)
 {
+	if (!supports_overlayfs())
+		SKIP(return, "overlayfs is not supported");
+
 	ASSERT_EQ(0, test_open(lower_fl1, O_RDONLY));
 	ASSERT_EQ(0, test_open(lower_dl1, O_RDONLY));
 	ASSERT_EQ(0, test_open(lower_dl1_fl2, O_RDONLY));
@@ -3769,6 +3813,9 @@ TEST_F_FORK(layout2_overlay, same_conten
 	size_t i;
 	const char *path_entry;
 
+	if (!supports_overlayfs())
+		SKIP(return, "overlayfs is not supported");
+
 	/* Sets rules on base directories (i.e. outside overlay scope). */
 	ruleset_fd = create_ruleset(_metadata, ACCESS_RW, layer1_base);
 	ASSERT_LE(0, ruleset_fd);
author maintainer
#707 lore raw

PATCH 706/885 selftests/landlock: Test ptrace as much as possible with Yama

In reply to Greg Kroah-Hartman

From: Jeff Xu <[email protected]>

commit 8677e555f17f51321d0730b945aeb7d4b95f998f upstream.

Update ptrace tests according to all potential Yama security policies. This is required to make such tests pass even if Yama is enabled.

Tests are not skipped but they now check both Landlock and Yama boundary restrictions at run time to keep a maximum test coverage (i.e. positive and negative testing).

[mic: Add curly braces around EXPECT_EQ() to make it build, and improve commit message]

---
 tools/testing/selftests/landlock/ptrace_test.c |  113 +++++++++++++++++++++----
 1 file changed, 96 insertions(+), 17 deletions(-)
tools/testing/selftests/landlock/ptrace_test.c +96 -17
@@ -19,6 +19,12 @@
 
 #include "common.h"
 
+/* Copied from security/yama/yama_lsm.c */
+#define YAMA_SCOPE_DISABLED 0
+#define YAMA_SCOPE_RELATIONAL 1
+#define YAMA_SCOPE_CAPABILITY 2
+#define YAMA_SCOPE_NO_ATTACH 3
+
 static void create_domain(struct __test_metadata *const _metadata)
 {
 	int ruleset_fd;
@@ -60,6 +66,25 @@ static int test_ptrace_read(const pid_t
 	return 0;
 }
 
+static int get_yama_ptrace_scope(void)
+{
+	int ret;
+	char buf[2] = {};
+	const int fd = open("/proc/sys/kernel/yama/ptrace_scope", O_RDONLY);
+
+	if (fd < 0)
+		return 0;
+
+	if (read(fd, buf, 1) < 0) {
+		close(fd);
+		return -1;
+	}
+
+	ret = atoi(buf);
+	close(fd);
+	return ret;
+}
+
 /* clang-format off */
 FIXTURE(hierarchy) {};
 /* clang-format on */
@@ -232,8 +257,51 @@ TEST_F(hierarchy, trace)
 	pid_t child, parent;
 	int status, err_proc_read;
 	int pipe_child[2], pipe_parent[2];
+	int yama_ptrace_scope;
 	char buf_parent;
 	long ret;
+	bool can_read_child, can_trace_child, can_read_parent, can_trace_parent;
+
+	yama_ptrace_scope = get_yama_ptrace_scope();
+	ASSERT_LE(0, yama_ptrace_scope);
+
+	if (yama_ptrace_scope > YAMA_SCOPE_DISABLED)
+		TH_LOG("Incomplete tests due to Yama restrictions (scope %d)",
+		       yama_ptrace_scope);
+
+	/*
+	 * can_read_child is true if a parent process can read its child
+	 * process, which is only the case when the parent process is not
+	 * isolated from the child with a dedicated Landlock domain.
+	 */
+	can_read_child = !variant->domain_parent;
+
+	/*
+	 * can_trace_child is true if a parent process can trace its child
+	 * process.  This depends on two conditions:
+	 * - The parent process is not isolated from the child with a dedicated
+	 *   Landlock domain.
+	 * - Yama allows tracing children (up to YAMA_SCOPE_RELATIONAL).
+	 */
+	can_trace_child = can_read_child &&
+			  yama_ptrace_scope <= YAMA_SCOPE_RELATIONAL;
+
+	/*
+	 * can_read_parent is true if a child process can read its parent
+	 * process, which is only the case when the child process is not
+	 * isolated from the parent with a dedicated Landlock domain.
+	 */
+	can_read_parent = !variant->domain_child;
+
+	/*
+	 * can_trace_parent is true if a child process can trace its parent
+	 * process.  This depends on two conditions:
+	 * - The child process is not isolated from the parent with a dedicated
+	 *   Landlock domain.
+	 * - Yama is disabled (YAMA_SCOPE_DISABLED).
+	 */
+	can_trace_parent = can_read_parent &&
+			   yama_ptrace_scope <= YAMA_SCOPE_DISABLED;
 
 	/*
 	 * Removes all effective and permitted capabilities to not interfere
@@ -264,16 +332,21 @@ TEST_F(hierarchy, trace)
 		/* Waits for the parent to be in a domain, if any. */
 		ASSERT_EQ(1, read(pipe_parent[0], &buf_child, 1));
 
-		/* Tests PTRACE_ATTACH and PTRACE_MODE_READ on the parent. */
+		/* Tests PTRACE_MODE_READ on the parent. */
 		err_proc_read = test_ptrace_read(parent);
+		if (can_read_parent) {
+			EXPECT_EQ(0, err_proc_read);
+		} else {
+			EXPECT_EQ(EACCES, err_proc_read);
+		}
+
+		/* Tests PTRACE_ATTACH on the parent. */
 		ret = ptrace(PTRACE_ATTACH, parent, NULL, 0);
-		if (variant->domain_child) {
+		if (can_trace_parent) {
+			EXPECT_EQ(0, ret);
+		} else {
 			EXPECT_EQ(-1, ret);
 			EXPECT_EQ(EPERM, errno);
-			EXPECT_EQ(EACCES, err_proc_read);
-		} else {
-			EXPECT_EQ(0, ret);
-			EXPECT_EQ(0, err_proc_read);
 		}
 		if (ret == 0) {
 			ASSERT_EQ(parent, waitpid(parent, &status, 0));
@@ -283,11 +356,11 @@ TEST_F(hierarchy, trace)
 
 		/* Tests child PTRACE_TRACEME. */
 		ret = ptrace(PTRACE_TRACEME);
-		if (variant->domain_parent) {
+		if (can_trace_child) {
+			EXPECT_EQ(0, ret);
+		} else {
 			EXPECT_EQ(-1, ret);
 			EXPECT_EQ(EPERM, errno);
-		} else {
-			EXPECT_EQ(0, ret);
 		}
 
 		/*
@@ -296,7 +369,7 @@ TEST_F(hierarchy, trace)
 		 */
 		ASSERT_EQ(1, write(pipe_child[1], ".", 1));
 
-		if (!variant->domain_parent) {
+		if (can_trace_child) {
 			ASSERT_EQ(0, raise(SIGSTOP));
 		}
 
@@ -321,7 +394,7 @@ TEST_F(hierarchy, trace)
 	ASSERT_EQ(1, read(pipe_child[0], &buf_parent, 1));
 
 	/* Tests child PTRACE_TRACEME. */
-	if (!variant->domain_parent) {
+	if (can_trace_child) {
 		ASSERT_EQ(child, waitpid(child, &status, 0));
 		ASSERT_EQ(1, WIFSTOPPED(status));
 		ASSERT_EQ(0, ptrace(PTRACE_DETACH, child, NULL, 0));
@@ -331,17 +404,23 @@ TEST_F(hierarchy, trace)
 		EXPECT_EQ(ESRCH, errno);
 	}
 
-	/* Tests PTRACE_ATTACH and PTRACE_MODE_READ on the child. */
+	/* Tests PTRACE_MODE_READ on the child. */
 	err_proc_read = test_ptrace_read(child);
+	if (can_read_child) {
+		EXPECT_EQ(0, err_proc_read);
+	} else {
+		EXPECT_EQ(EACCES, err_proc_read);
+	}
+
+	/* Tests PTRACE_ATTACH on the child. */
 	ret = ptrace(PTRACE_ATTACH, child, NULL, 0);
-	if (variant->domain_parent) {
+	if (can_trace_child) {
+		EXPECT_EQ(0, ret);
+	} else {
 		EXPECT_EQ(-1, ret);
 		EXPECT_EQ(EPERM, errno);
-		EXPECT_EQ(EACCES, err_proc_read);
-	} else {
-		EXPECT_EQ(0, ret);
-		EXPECT_EQ(0, err_proc_read);
 	}
+
 	if (ret == 0) {
 		ASSERT_EQ(child, waitpid(child, &status, 0));
 		ASSERT_EQ(1, WIFSTOPPED(status));
author maintainer
#708 lore raw

PATCH 707/885 udf: Truncate added extents on failed expansion

In reply to Greg Kroah-Hartman

From: Jan Kara <[email protected]>

commit 70bfb3a8d661d4fdc742afc061b88a7f3fc9f500 upstream.

When a file expansion failed because we didn't have enough space for indirect extents make sure we truncate extents created so far so that we don't leave extents beyond EOF.

---
 fs/udf/inode.c |   15 +++++++++++----
 1 file changed, 11 insertions(+), 4 deletions(-)
fs/udf/inode.c +11 -4
@@ -526,8 +526,10 @@ static int udf_do_extend_file(struct ino
 	}
 
 	if (fake) {
-		udf_add_aext(inode, last_pos, &last_ext->extLocation,
-			     last_ext->extLength, 1);
+		err = udf_add_aext(inode, last_pos, &last_ext->extLocation,
+				   last_ext->extLength, 1);
+		if (err < 0)
+			goto out_err;
 		count++;
 	} else {
 		struct kernel_lb_addr tmploc;
@@ -561,7 +563,7 @@ static int udf_do_extend_file(struct ino
 		err = udf_add_aext(inode, last_pos, &last_ext->extLocation,
 				   last_ext->extLength, 1);
 		if (err)
-			return err;
+			goto out_err;
 		count++;
 	}
 	if (new_block_bytes) {
@@ -570,7 +572,7 @@ static int udf_do_extend_file(struct ino
 		err = udf_add_aext(inode, last_pos, &last_ext->extLocation,
 				   last_ext->extLength, 1);
 		if (err)
-			return err;
+			goto out_err;
 		count++;
 	}
 
@@ -584,6 +586,11 @@ out:
 		return -EIO;
 
 	return count;
+out_err:
+	/* Remove extents we've created so far */
+	udf_clear_extent_cache(inode);
+	udf_truncate_extents(inode);
+	return err;
 }
 
 /* Extend the final block of the file to final_block_len bytes */
author maintainer
#709 lore raw

PATCH 708/885 udf: Do not bother merging very long extents

In reply to Greg Kroah-Hartman

From: Jan Kara <[email protected]>

commit 53cafe1d6d8ef9f93318e5bfccc0d24f27d41ced upstream.

When merging very long extents we try to push as much length as possible to the first extent. However this is unnecessarily complicated and not really worth the trouble. Furthermore there was a bug in the logic resulting in corrupting extents in the file as syzbot reproducer shows. So just don't bother with the merging of extents that are too long together.

---
 fs/udf/inode.c |   19 ++-----------------
 1 file changed, 2 insertions(+), 17 deletions(-)
fs/udf/inode.c +2 -17
@@ -1094,23 +1094,8 @@ static void udf_merge_extents(struct ino
 			blocksize - 1) >> blocksize_bits)))) {
 
 			if (((li->extLength & UDF_EXTENT_LENGTH_MASK) +
-				(lip1->extLength & UDF_EXTENT_LENGTH_MASK) +
-				blocksize - 1) & ~UDF_EXTENT_LENGTH_MASK) {
-				lip1->extLength = (lip1->extLength -
-						  (li->extLength &
-						   UDF_EXTENT_LENGTH_MASK) +
-						   UDF_EXTENT_LENGTH_MASK) &
-							~(blocksize - 1);
-				li->extLength = (li->extLength &
-						 UDF_EXTENT_FLAG_MASK) +
-						(UDF_EXTENT_LENGTH_MASK + 1) -
-						blocksize;
-				lip1->extLocation.logicalBlockNum =
-					li->extLocation.logicalBlockNum +
-					((li->extLength &
-						UDF_EXTENT_LENGTH_MASK) >>
-						blocksize_bits);
-			} else {
+			     (lip1->extLength & UDF_EXTENT_LENGTH_MASK) +
+			     blocksize - 1) <= UDF_EXTENT_LENGTH_MASK) {
 				li->extLength = lip1->extLength +
 					(((li->extLength &
 						UDF_EXTENT_LENGTH_MASK) +
author maintainer
#710 lore raw

PATCH 709/885 udf: Do not update file length for failed writes to inline files

In reply to Greg Kroah-Hartman

From: Jan Kara <[email protected]>

commit 256fe4162f8b5a1625b8603ca5f7ff79725bfb47 upstream.

When write to inline file fails (or happens only partly), we still updated length of inline data as if the whole write succeeded. Fix the update of length of inline data to happen only if the write succeeds.

---
 fs/udf/file.c |   26 ++++++++++++--------------
 1 file changed, 12 insertions(+), 14 deletions(-)
fs/udf/file.c +12 -14
@@ -149,26 +149,24 @@ static ssize_t udf_file_write_iter(struc
 		goto out;
 
 	down_write(&iinfo->i_data_sem);
-	if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB) {
-		loff_t end = iocb->ki_pos + iov_iter_count(from);
-
-		if (inode->i_sb->s_blocksize <
-				(udf_file_entry_alloc_offset(inode) + end)) {
-			err = udf_expand_file_adinicb(inode);
-			if (err) {
-				inode_unlock(inode);
-				udf_debug("udf_expand_adinicb: err=%d\n", err);
-				return err;
-			}
-		} else {
-			iinfo->i_lenAlloc = max(end, inode->i_size);
-			up_write(&iinfo->i_data_sem);
+	if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB &&
+	    inode->i_sb->s_blocksize < (udf_file_entry_alloc_offset(inode) +
+				 iocb->ki_pos + iov_iter_count(from))) {
+		err = udf_expand_file_adinicb(inode);
+		if (err) {
+			inode_unlock(inode);
+			udf_debug("udf_expand_adinicb: err=%d\n", err);
+			return err;
 		}
 	} else
 		up_write(&iinfo->i_data_sem);
 
 	retval = __generic_file_write_iter(iocb, from);
 out:
+	down_write(&iinfo->i_data_sem);
+	if (iinfo->i_alloc_type == ICBTAG_FLAG_AD_IN_ICB && retval > 0)
+		iinfo->i_lenAlloc = inode->i_size;
+	up_write(&iinfo->i_data_sem);
 	inode_unlock(inode);
 
 	if (retval > 0) {
author maintainer
#711 lore raw

PATCH 710/885 udf: Preserve link count of system files

In reply to Greg Kroah-Hartman

From: Jan Kara <[email protected]>

commit fc8033a34a3ca7d23353e645e6dde5d364ac5f12 upstream.

System files in UDF filesystem have link count 0. To not confuse VFS we fudge the link count to be 1 when reading such inodes however we forget to restore the link count of 0 when writing such inodes. Fix that.

---
 fs/udf/inode.c |    9 +++++++--
 fs/udf/super.c |    1 +
 fs/udf/udf_i.h |    3 ++-
 3 files changed, 10 insertions(+), 3 deletions(-)
fs/udf/inode.c +7 -2
@@ -1380,6 +1380,7 @@ reread:
 		ret = -EIO;
 		goto out;
 	}
+	iinfo->i_hidden = hidden_inode;
 	iinfo->i_unique = 0;
 	iinfo->i_lenEAttr = 0;
 	iinfo->i_lenExtents = 0;
@@ -1715,8 +1716,12 @@ static int udf_update_inode(struct inode
 
 	if (S_ISDIR(inode->i_mode) && inode->i_nlink > 0)
 		fe->fileLinkCount = cpu_to_le16(inode->i_nlink - 1);
-	else
-		fe->fileLinkCount = cpu_to_le16(inode->i_nlink);
+	else {
+		if (iinfo->i_hidden)
+			fe->fileLinkCount = cpu_to_le16(0);
+		else
+			fe->fileLinkCount = cpu_to_le16(inode->i_nlink);
+	}
 
 	fe->informationLength = cpu_to_le64(inode->i_size);
 
fs/udf/super.c +1 -0
@@ -147,6 +147,7 @@ static struct inode *udf_alloc_inode(str
 	ei->i_next_alloc_goal = 0;
 	ei->i_strat4096 = 0;
 	ei->i_streamdir = 0;
+	ei->i_hidden = 0;
 	init_rwsem(&ei->i_data_sem);
 	ei->cached_extent.lstart = -1;
 	spin_lock_init(&ei->i_extent_cache_lock);
fs/udf/udf_i.h +2 -1
@@ -44,7 +44,8 @@ struct udf_inode_info {
 	unsigned		i_use : 1;	/* unallocSpaceEntry */
 	unsigned		i_strat4096 : 1;
 	unsigned		i_streamdir : 1;
-	unsigned		reserved : 25;
+	unsigned		i_hidden : 1;	/* hidden system inode */
+	unsigned		reserved : 24;
 	__u8			*i_data;
 	struct kernel_lb_addr	i_locStreamdir;
 	__u64			i_lenStreams;
author maintainer
#712 lore raw

PATCH 711/885 udf: Detect system inodes linked into directory hierarchy

In reply to Greg Kroah-Hartman

From: Jan Kara <[email protected]>

commit 85a37983ec69cc9fcd188bc37c4de15ee326355a upstream.

When UDF filesystem is corrupted, hidden system inodes can be linked into directory hierarchy which is an avenue for further serious corruption of the filesystem and kernel confusion as noticed by syzbot fuzzed images. Refuse to access system inodes linked into directory hierarchy and vice versa.

---
 fs/udf/inode.c |    7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
fs/udf/inode.c +6 -1
@@ -1892,8 +1892,13 @@ struct inode *__udf_iget(struct super_bl
 	if (!inode)
 		return ERR_PTR(-ENOMEM);
 
-	if (!(inode->i_state & I_NEW))
+	if (!(inode->i_state & I_NEW)) {
+		if (UDF_I(inode)->i_hidden != hidden_inode) {
+			iput(inode);
+			return ERR_PTR(-EFSCORRUPTED);
+		}
 		return inode;
+	}
 
 	memcpy(&UDF_I(inode)->i_location, ino, sizeof(struct kernel_lb_addr));
 	err = udf_read_inode(inode, hidden_inode);
author maintainer
#713 lore raw

PATCH 712/885 udf: Fix file corruption when appending just after end of preallocated extent

In reply to Greg Kroah-Hartman

From: Jan Kara <[email protected]>

commit 36ec52ea038b18a53e198116ef7d7e70c87db046 upstream.

When we append new block just after the end of preallocated extent, the code in inode_getblk() wrongly determined we're going to use the preallocated extent which resulted in adding block into a wrong logical offset in the file. Sequence like this manifests it:

xfs_io -f -c "pwrite 0x2cacf 0xd122" -c "truncate 0x2dd6f" \
-c "pwrite 0x27fd9 0x69a9" -c "pwrite 0x32981 0x7244" <file>

The code that determined the use of preallocated extent is actually stale because udf_do_extend_file() does not create preallocation anymore so after calling that function we are sure there's no usable preallocation. Just remove the faulty condition.

Fixes 16d055656814 ("udf: Discard preallocation before extending file with a hole")
---
 fs/udf/inode.c |   24 +++++++++++-------------
 1 file changed, 11 insertions(+), 13 deletions(-)
fs/udf/inode.c +11 -13
@@ -805,19 +805,17 @@ static sector_t inode_getblk(struct inod
 		c = 0;
 		offset = 0;
 		count += ret;
-		/* We are not covered by a preallocated extent? */
-		if ((laarr[0].extLength & UDF_EXTENT_FLAG_MASK) !=
-						EXT_NOT_RECORDED_ALLOCATED) {
-			/* Is there any real extent? - otherwise we overwrite
-			 * the fake one... */
-			if (count)
-				c = !c;
-			laarr[c].extLength = EXT_NOT_RECORDED_NOT_ALLOCATED |
-				inode->i_sb->s_blocksize;
-			memset(&laarr[c].extLocation, 0x00,
-				sizeof(struct kernel_lb_addr));
-			count++;
-		}
+		/*
+		 * Is there any real extent? - otherwise we overwrite the fake
+		 * one...
+		 */
+		if (count)
+			c = !c;
+		laarr[c].extLength = EXT_NOT_RECORDED_NOT_ALLOCATED |
+			inode->i_sb->s_blocksize;
+		memset(&laarr[c].extLocation, 0x00,
+			sizeof(struct kernel_lb_addr));
+		count++;
 		endnum = c + 1;
 		lastblock = 1;
 	} else {
author maintainer
#714 lore raw

PATCH 713/885 md: dont update recovery_cp when curr_resync is ACTIVE

In reply to Greg Kroah-Hartman

From: Hou Tao <[email protected]>

commit 1d1f25bfda432a6b61bd0205d426226bbbd73504 upstream.

Don't update recovery_cp when curr_resync is MD_RESYNC_ACTIVE, otherwise md may skip the resync of the first 3 sectors if the resync procedure is interrupted before the first calling of ->sync_request() as shown below:

md_do_sync thread control thread
// setup resync
mddev->recovery_cp = 0
j = 0
mddev->curr_resync = MD_RESYNC_ACTIVE

                             // e.g., set array as idle
                             set_bit(MD_RECOVERY_INTR, &&mddev_recovery)

// resync loop
// check INTR before calling sync_request
!test_bit(MD_RECOVERY_INTR, &mddev->recovery

// resync interrupted
// update recovery_cp from 0 to 3
// the resync of three 3 sectors will be skipped mddev->recovery_cp = 3

Fixes eac58d08d493 ("md: Use enum for overloaded magic numbers used by mddev->curr_resync")
---
 drivers/md/md.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/md/md.c +1 -1
@@ -9039,7 +9039,7 @@ void md_do_sync(struct md_thread *thread
 	mddev->pers->sync_request(mddev, max_sectors, &skipped);
 
 	if (!test_bit(MD_RECOVERY_CHECK, &mddev->recovery) &&
-	    mddev->curr_resync >= MD_RESYNC_ACTIVE) {
+	    mddev->curr_resync > MD_RESYNC_ACTIVE) {
 		if (test_bit(MD_RECOVERY_SYNC, &mddev->recovery)) {
 			if (test_bit(MD_RECOVERY_INTR, &mddev->recovery)) {
 				if (mddev->curr_resync >= mddev->recovery_cp) {
author maintainer
#715 lore raw

PATCH 714/885 RDMA/siw: Fix user page pinning accounting

In reply to Greg Kroah-Hartman

From: Bernard Metzler <[email protected]>

[ Upstream commit 65a8fc30fb6722fc25adec6d7dd5b53b0bb85820 ]

To avoid racing with other user memory reservations, immediately account full amount of pages to be pinned.

---
 drivers/infiniband/sw/siw/siw_mem.c | 23 ++++++++++++-----------
 1 file changed, 12 insertions(+), 11 deletions(-)
drivers/infiniband/sw/siw/siw_mem.c +12 -11
index 61c17db70d658..bf69566e2eb63 100644
@@ -398,7 +398,7 @@ struct siw_umem *siw_umem_get(u64 start, u64 len, bool writable)
 
 	mlock_limit = rlimit(RLIMIT_MEMLOCK) >> PAGE_SHIFT;
 
-	if (num_pages + atomic64_read(&mm_s->pinned_vm) > mlock_limit) {
+	if (atomic64_add_return(num_pages, &mm_s->pinned_vm) > mlock_limit) {
 		rv = -ENOMEM;
 		goto out_sem_up;
 	}
@@ -411,18 +411,16 @@ struct siw_umem *siw_umem_get(u64 start, u64 len, bool writable)
 		goto out_sem_up;
 	}
 	for (i = 0; num_pages; i++) {
-		int got, nents = min_t(int, num_pages, PAGES_PER_CHUNK);
-
-		umem->page_chunk[i].plist =
+		int nents = min_t(int, num_pages, PAGES_PER_CHUNK);
+		struct page **plist =
 			kcalloc(nents, sizeof(struct page *), GFP_KERNEL);
-		if (!umem->page_chunk[i].plist) {
+
+		if (!plist) {
 			rv = -ENOMEM;
 			goto out_sem_up;
 		}
-		got = 0;
+		umem->page_chunk[i].plist = plist;
 		while (nents) {
-			struct page **plist = &umem->page_chunk[i].plist[got];
-
 			rv = pin_user_pages(first_page_va, nents,
 					    foll_flags | FOLL_LONGTERM,
 					    plist, NULL);
@@ -430,12 +428,11 @@ struct siw_umem *siw_umem_get(u64 start, u64 len, bool writable)
 				goto out_sem_up;
 
 			umem->num_pages += rv;
-			atomic64_add(rv, &mm_s->pinned_vm);
 			first_page_va += rv * PAGE_SIZE;
+			plist += rv;
 			nents -= rv;
-			got += rv;
+			num_pages -= rv;
 		}
-		num_pages -= got;
 	}
 out_sem_up:
 	mmap_read_unlock(mm_s);
@@ -443,6 +440,10 @@ struct siw_umem *siw_umem_get(u64 start, u64 len, bool writable)
 	if (rv > 0)
 		return umem;
 
+	/* Adjust accounting for pages not pinned */
+	if (num_pages)
+		atomic64_sub(num_pages, &mm_s->pinned_vm);
+
 	siw_umem_release(umem, false);
 
 	return ERR_PTR(rv);
Signature · 1 line
2.39.2
author maintainer
#716 lore raw

PATCH 715/885 KVM: Destroy target device if coalesced MMIO unregistration fails

In reply to Greg Kroah-Hartman

From: Sean Christopherson <[email protected]>

commit b1cb1fac22abf102ffeb29dd3eeca208a3869d54 upstream.

Destroy and free the target coalesced MMIO device if unregistering said device fails. As clearly noted in the code, kvm_io_bus_unregister_dev() does not destroy the target device.

  BUG: memory leak
  unreferenced object 0xffff888112a54880 (size 64):
    comm "syz-executor.2", pid 5258, jiffies 4297861402 (age 14.129s)
    hex dump (first 32 bytes):
      38 c7 67 15 00 c9 ff ff 38 c7 67 15 00 c9 ff ff  8.g.....8.g.....
      e0 c7 e1 83 ff ff ff ff 00 30 67 15 00 c9 ff ff  .........0g.....
    backtrace:
      [<0000000006995a8a>] kmalloc include/linux/slab.h:556 [inline]
      [<0000000006995a8a>] kzalloc include/linux/slab.h:690 [inline]
      [<0000000006995a8a>] kvm_vm_ioctl_register_coalesced_mmio+0x8e/0x3d0 arch/x86/kvm/../../../virt/kvm/coalesced_mmio.c:150
      [<00000000022550c2>] kvm_vm_ioctl+0x47d/0x1600 arch/x86/kvm/../../../virt/kvm/kvm_main.c:3323
      [<000000008a75102f>] vfs_ioctl fs/ioctl.c:46 [inline]
      [<000000008a75102f>] file_ioctl fs/ioctl.c:509 [inline]
      [<000000008a75102f>] do_vfs_ioctl+0xbab/0x1160 fs/ioctl.c:696
      [<0000000080e3f669>] ksys_ioctl+0x76/0xa0 fs/ioctl.c:713
      [<0000000059ef4888>] __do_sys_ioctl fs/ioctl.c:720 [inline]
      [<0000000059ef4888>] __se_sys_ioctl fs/ioctl.c:718 [inline]
      [<0000000059ef4888>] __x64_sys_ioctl+0x6f/0xb0 fs/ioctl.c:718
      [<000000006444fa05>] do_syscall_64+0x9f/0x4e0 arch/x86/entry/common.c:290
      [<000000009a4ed50b>] entry_SYSCALL_64_after_hwframe+0x49/0xbe

  BUG: leak checking failed
---
 virt/kvm/coalesced_mmio.c |    8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)
virt/kvm/coalesced_mmio.c +5 -3
@@ -187,15 +187,17 @@ int kvm_vm_ioctl_unregister_coalesced_mm
 			r = kvm_io_bus_unregister_dev(kvm,
 				zone->pio ? KVM_PIO_BUS : KVM_MMIO_BUS, &dev->dev);
 
+			kvm_iodevice_destructor(&dev->dev);
+
 			/*
 			 * On failure, unregister destroys all devices on the
 			 * bus _except_ the target device, i.e. coalesced_zones
-			 * has been modified.  No need to restart the walk as
-			 * there aren't any zones left.
+			 * has been modified.  Bail after destroying the target
+			 * device, there's no need to restart the walk as there
+			 * aren't any zones left.
 			 */
 			if (r)
 				break;
-			kvm_iodevice_destructor(&dev->dev);
 		}
 	}