Skip to contentSkip to conversation

patches for FPGA

PATCH New Alan Tull 1 participant 5 messages May 9, 2019

Subsystems (2)
lore
Share & tools
lore

Alan Tull

author
#0 lore raw

cover patches for FPGA

Hi Greg,

Please take these four fpga fixes patches. They have been reviewed on the mailing list and apply cleanly on current linux-next and char-misc-testing.

Thanks,
Alan

Chengguang Xu (1):
fpga: dfl: expand minor range when registering chrdev region

Scott Wood (2):
fpga: dfl: afu: Pass the correct device to dma_mapping_error()
fpga: dfl: Add lockdep classes for pdata->lock

Wen Yang (1):
fpga: stratix10-soc: fix use-after-free on s10_init()

 drivers/fpga/dfl-afu-dma-region.c |  2 +-
 drivers/fpga/dfl.c                | 22 ++++++++++++++++++----
 drivers/fpga/stratix10-soc.c      |  6 +++++-
 3 files changed, 24 insertions(+), 6 deletions(-)
Signature · 1 line
2.21.0
author
#1 lore raw

PATCH 1/4 fpga: stratix10-soc: fix use-after-free on s10_init()

From: Wen Yang <[email protected]>

The refcount of fw_np has already been decreased by of_find_matching_node() so it shouldn't be used anymore.
This patch adds an of_node_get() before of_find_matching_node() to avoid the use-after-free problem.

Fixes e7eef1d7633a ("fpga: add intel stratix10 soc fpga manager driver")
Cc Alan Tull <[email protected]>
Cc Moritz Fischer <[email protected]>
Cc Nicolas Saenz Julienne <[email protected]>
---
 drivers/fpga/stratix10-soc.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)
drivers/fpga/stratix10-soc.c +5 -1
index 13851b3d1c56..215d33789c74 100644
@@ -507,12 +507,16 @@ static int __init s10_init(void)
 	if (!fw_np)
 		return -ENODEV;
 
+	of_node_get(fw_np);
 	np = of_find_matching_node(fw_np, s10_of_match);
-	if (!np)
+	if (!np) {
+		of_node_put(fw_np);
 		return -ENODEV;
+	}
 
 	of_node_put(np);
 	ret = of_platform_populate(fw_np, s10_of_match, NULL, NULL);
+	of_node_put(fw_np);
 	if (ret)
 		return ret;
 
Signature · 1 line
2.21.0
author
#2 lore raw

PATCH 2/4 fpga: dfl: afu: Pass the correct device to dma_mapping_error()

In reply to Alan Tull

From: Scott Wood <[email protected]>

dma_mapping_error() was being called on a different device struct than what was passed to map/unmap. Besides rendering the error checking ineffective, it caused a debug splat with CONFIG_DMA_API_DEBUG.

---
 drivers/fpga/dfl-afu-dma-region.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/fpga/dfl-afu-dma-region.c +1 -1
index 0bbc7142f1dc..f7d268f45df0 100644
@@ -391,7 +391,7 @@ int afu_dma_map_region(struct dfl_feature_platform_data *pdata,
 				    region->pages[0], 0,
 				    region->length,
 				    DMA_BIDIRECTIONAL);
-	if (dma_mapping_error(&pdata->dev->dev, region->iova)) {
+	if (dma_mapping_error(dfl_fpga_pdata_to_parent(pdata), region->iova)) {
 		dev_err(&pdata->dev->dev, "failed to map for dma\n");
 		ret = -EFAULT;
 		goto unpin_pages;
Signature · 1 line
2.21.0
author
#3 lore raw

PATCH 3/4 fpga: dfl: Add lockdep classes for pdata->lock

In reply to Alan Tull

From: Scott Wood <[email protected]>

struct dfl_feature_platform_data (and it's mutex) is used by both fme and port devices, and when lockdep is enabled it complains about nesting between these locks. Tell lockdep about the difference so it can track each class separately.

Here's the lockdep complaint:

[  409.680668] WARNING: possible recursive locking detected
[  409.685983] 5.1.0-rc3.fpga+ #1 Tainted: G            E
[  409.691469] --------------------------------------------
[  409.696779] fpgaconf/9348 is trying to acquire lock:
[  409.701746] 00000000a443fe2e (&pdata->lock){+.+.}, at: port_enable_set+0x24/0x60 [dfl_afu]
[  409.710006]
[  409.710006] but task is already holding lock:
[  409.715837] 0000000063b78782 (&pdata->lock){+.+.}, at: fme_pr_ioctl+0x21d/0x330 [dfl_fme]
[  409.724012]
[  409.724012] other info that might help us debug this:
[  409.730535]  Possible unsafe locking scenario:
[  409.730535]
[  409.736457]        CPU0
[  409.738910]        ----
[  409.741360]   lock(&pdata->lock);
[  409.744679]   lock(&pdata->lock);
[  409.747999]
[  409.747999]  *** DEADLOCK ***
[  409.747999]
[  409.753920]  May be due to missing lock nesting notation
[  409.753920]
[  409.760704] 4 locks held by fpgaconf/9348:
[  409.764805]  #0: 0000000063b78782 (&pdata->lock){+.+.}, at: fme_pr_ioctl+0x21d/0x330 [dfl_fme]
[  409.773408]  #1: 00000000213c8a66 (&region->mutex){+.+.}, at: fpga_region_program_fpga+0x24/0x200 [fpga_region]
[  409.783489]  #2: 00000000fe63afb9 (&mgr->ref_mutex){+.+.}, at: fpga_mgr_lock+0x15/0x40 [fpga_mgr]
[  409.792354]  #3: 000000000b2285c5 (&bridge->mutex){+.+.}, at: __fpga_bridge_get+0x26/0xa0 [fpga_bridge]
[  409.801740]
[  409.801740] stack backtrace:
[  409.806102] CPU: 45 PID: 9348 Comm: fpgaconf Kdump: loaded Tainted: G            E     5.1.0-rc3.fpga+ #1
[  409.815658] Hardware name: Intel Corporation S2600BT/S2600BT, BIOS SE5C620.86B.01.00.0763.022420181017 02/24/2018
[  409.825911] Call Trace:
[  409.828369]  dump_stack+0x5e/0x8b
[  409.831686]  __lock_acquire+0xf3d/0x10e0
[  409.835612]  ? find_held_lock+0x3c/0xa0
[  409.839451]  lock_acquire+0xbc/0x1d0
[  409.843030]  ? port_enable_set+0x24/0x60 [dfl_afu]
[  409.847823]  ? port_enable_set+0x24/0x60 [dfl_afu]
[  409.852616]  __mutex_lock+0x86/0x970
[  409.856195]  ? port_enable_set+0x24/0x60 [dfl_afu]
[  409.860989]  ? port_enable_set+0x24/0x60 [dfl_afu]
[  409.865777]  ? __mutex_unlock_slowpath+0x4b/0x290
[  409.870486]  port_enable_set+0x24/0x60 [dfl_afu]
[  409.875106]  fpga_bridges_disable+0x36/0x50 [fpga_bridge]
[  409.880502]  fpga_region_program_fpga+0xea/0x200 [fpga_region]
[  409.886338]  fme_pr_ioctl+0x13e/0x330 [dfl_fme]
[  409.890870]  fme_ioctl+0x66/0xe0 [dfl_fme]
[  409.894973]  do_vfs_ioctl+0xa9/0x720
[  409.898548]  ? lockdep_hardirqs_on+0xf0/0x1a0
[  409.902907]  ksys_ioctl+0x60/0x90
[  409.906225]  __x64_sys_ioctl+0x16/0x20
[  409.909981]  do_syscall_64+0x5a/0x220
[  409.913644]  entry_SYSCALL_64_after_hwframe+0x49/0xbe
[  409.918698] RIP: 0033:0x7f9d31b9b8d7
[  409.922276] Code: 44 00 00 48 8b 05 b9 15 2d 00 64 c7 00 26 00 00 00 48 c7 c0 ff ff ff ff c3 66 2e 0f 1f 84 00 00 00 00 00 b8 10 00 00 00 0f 05 <48> 3d 01 f0 ff ff 73 01 c3 48 8b 0d 89 15 2d 00 f7 d8 64 89 01 48
[  409.941020] RSP: 002b:00007ffe4cae0d68 EFLAGS: 00000202 ORIG_RAX: 0000000000000010
[  409.948588] RAX: ffffffffffffffda RBX: 00007f9d32ade6a0 RCX: 00007f9d31b9b8d7
[  409.955719] RDX: 00007ffe4cae0df0 RSI: 000000000000b680 RDI: 0000000000000003
[  409.962852] RBP: 0000000000000003 R08: 00007f9d2b70a177 R09: 00007ffe4cae0e40
[  409.969984] R10: 00007ffe4cae0160 R11: 0000000000000202 R12: 00007ffe4cae0df0
[  409.977115] R13: 000000000000b680 R14: 0000000000000000 R15: 00007ffe4cae0f60
---
 drivers/fpga/dfl.c | 16 +++++++++++++++-
 1 file changed, 15 insertions(+), 1 deletion(-)
drivers/fpga/dfl.c +15 -1
index 2c09e502e721..c25217cde5ca 100644
@@ -40,6 +40,13 @@ enum dfl_fpga_devt_type {
 	DFL_FPGA_DEVT_MAX,
 };
 
+static struct lock_class_key dfl_pdata_keys[DFL_ID_MAX];
+
+static const char *dfl_pdata_key_strings[DFL_ID_MAX] = {
+	"dfl-fme-pdata",
+	"dfl-port-pdata",
+};
+
 /**
  * dfl_dev_info - dfl feature device information.
  * @name: name string of the feature platform device.
@@ -443,11 +450,16 @@ static int build_info_commit_dev(struct build_feature_devs_info *binfo)
 	struct platform_device *fdev = binfo->feature_dev;
 	struct dfl_feature_platform_data *pdata;
 	struct dfl_feature_info *finfo, *p;
+	enum dfl_id_type type;
 	int ret, index = 0;
 
 	if (!fdev)
 		return 0;
 
+	type = feature_dev_id_type(fdev);
+	if (WARN_ON_ONCE(type >= DFL_ID_MAX))
+		return -EINVAL;
+
 	/*
 	 * we do not need to care for the memory which is associated with
 	 * the platform device. After calling platform_device_unregister(),
@@ -463,6 +475,8 @@ static int build_info_commit_dev(struct build_feature_devs_info *binfo)
 	pdata->num = binfo->feature_num;
 	pdata->dfl_cdev = binfo->cdev;
 	mutex_init(&pdata->lock);
+	lockdep_set_class_and_name(&pdata->lock, &dfl_pdata_keys[type],
+				   dfl_pdata_key_strings[type]);
 
 	/*
 	 * the count should be initialized to 0 to make sure
@@ -497,7 +511,7 @@ static int build_info_commit_dev(struct build_feature_devs_info *binfo)
 
 	ret = platform_device_add(binfo->feature_dev);
 	if (!ret) {
-		if (feature_dev_id_type(binfo->feature_dev) == PORT_ID)
+		if (type == PORT_ID)
 			dfl_fpga_cdev_add_port_dev(binfo->cdev,
 						   binfo->feature_dev);
 		else
Signature · 1 line
2.21.0
author
#4 lore raw

PATCH 4/4 fpga: dfl: expand minor range when registering chrdev region

In reply to Alan Tull

From: Chengguang Xu <[email protected]>

Actually, total amount of available minor number for a single major is MINORMASK + 1. So expand minor range when registering chrdev region.

---
 drivers/fpga/dfl.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
drivers/fpga/dfl.c +3 -3
index c25217cde5ca..4b66aaa32b5a 100644
@@ -322,7 +322,7 @@ static void dfl_chardev_uinit(void)
 	for (i = 0; i < DFL_FPGA_DEVT_MAX; i++)
 		if (MAJOR(dfl_chrdevs[i].devt)) {
 			unregister_chrdev_region(dfl_chrdevs[i].devt,
-						 MINORMASK);
+						 MINORMASK + 1);
 			dfl_chrdevs[i].devt = MKDEV(0, 0);
 		}
 }
@@ -332,8 +332,8 @@ static int dfl_chardev_init(void)
 	int i, ret;
 
 	for (i = 0; i < DFL_FPGA_DEVT_MAX; i++) {
-		ret = alloc_chrdev_region(&dfl_chrdevs[i].devt, 0, MINORMASK,
-					  dfl_chrdevs[i].name);
+		ret = alloc_chrdev_region(&dfl_chrdevs[i].devt, 0,
+					  MINORMASK + 1, dfl_chrdevs[i].name);
 		if (ret)
 			goto exit;
 	}
Signature · 1 line
2.21.0