Merge tag 'for-linus-5.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml

Pull UML updates from Richard Weinberger:

 - Support for VMAP_STACK

 - Support for splice_write in hostfs

 - Fixes for virt-pci

 - Fixes for virtio_uml

 - Various fixes

* tag 'for-linus-5.15-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml:
  um: fix stub location calculation
  um: virt-pci: fix uapi documentation
  um: enable VMAP_STACK
  um: virt-pci: don't do DMA from stack
  hostfs: support splice_write
  um: virtio_uml: fix memory leak on init failures
  um: virtio_uml: include linux/virtio-uml.h
  lib/logic_iomem: fix sparse warnings
  um: make PCI emulation driver init/exit static

Author: torvalds

arch/um/Kconfig

@@ -24,6 +24,7 @@ config UML
 	select SET_FS
 	select TRACE_IRQFLAGS_SUPPORT
 	select TTY # Needed for line.c
+	select HAVE_ARCH_VMAP_STACK
 
 config MMU
 	bool

arch/um/drivers/virt-pci.c

@@ -56,6 +56,13 @@ static unsigned long um_pci_msi_used[BITS_TO_LONGS(MAX_MSI_VECTORS)];
 
 #define UM_VIRT_PCI_MAXDELAY 40000
 
+struct um_pci_message_buffer {
+	struct virtio_pcidev_msg hdr;
+	u8 data[8];
+};
+
+static struct um_pci_message_buffer __percpu *um_pci_msg_bufs;
+
 static int um_pci_send_cmd(struct um_pci_device *dev,
 			   struct virtio_pcidev_msg *cmd,
 			   unsigned int cmd_size,
@@ -68,11 +75,12 @@ static int um_pci_send_cmd(struct um_pci_device *dev,
 		[1] = extra ? &extra_sg : &in_sg,
 		[2] = extra ? &in_sg : NULL,
 	};
+	struct um_pci_message_buffer *buf;
 	int delay_count = 0;
 	int ret, len;
 	bool posted;
 
-	if (WARN_ON(cmd_size < sizeof(*cmd)))
+	if (WARN_ON(cmd_size < sizeof(*cmd) || cmd_size > sizeof(*buf)))
 		return -EINVAL;
 
 	switch (cmd->op) {
@@ -88,6 +96,9 @@ static int um_pci_send_cmd(struct um_pci_device *dev,
 		break;
 	}
 
+	buf = get_cpu_var(um_pci_msg_bufs);
+	memcpy(buf, cmd, cmd_size);
+
 	if (posted) {
 		u8 *ncmd = kmalloc(cmd_size + extra_size, GFP_ATOMIC);
 
@@ -102,7 +113,10 @@ static int um_pci_send_cmd(struct um_pci_device *dev,
 		} else {
 			/* try without allocating memory */
 			posted = false;
+			cmd = (void *)buf;
 		}
+	} else {
+		cmd = (void *)buf;
 	}
 
 	sg_init_one(&out_sg, cmd, cmd_size);
@@ -118,11 +132,12 @@ static int um_pci_send_cmd(struct um_pci_device *dev,
 				posted ? cmd : HANDLE_NO_FREE(cmd),
 				GFP_ATOMIC);
 	if (ret)
-		return ret;
+		goto out;
 
 	if (posted) {
 		virtqueue_kick(dev->cmd_vq);
-		return 0;
+		ret = 0;
+		goto out;
 	}
 
 	/* kick and poll for getting a response on the queue */
@@ -148,6 +163,8 @@ static int um_pci_send_cmd(struct um_pci_device *dev,
 	}
 	clear_bit(UM_PCI_STAT_WAITING, &dev->status);
 
+out:
+	put_cpu_var(um_pci_msg_bufs);
 	return ret;
 }
 
@@ -161,12 +178,17 @@ static unsigned long um_pci_cfgspace_read(void *priv, unsigned int offset,
 		.size = size,
 		.addr = offset,
 	};
-	/* maximum size - we may only use parts of it */
-	u8 data[8];
+	/* buf->data is maximum size - we may only use parts of it */
+	struct um_pci_message_buffer *buf;
+	u8 *data;
+	unsigned long ret = ~0ULL;
 
 	if (!dev)
 		return ~0ULL;
 
+	buf = get_cpu_var(um_pci_msg_bufs);
+	data = buf->data;
+
 	memset(data, 0xff, sizeof(data));
 
 	switch (size) {
@@ -179,27 +201,34 @@ static unsigned long um_pci_cfgspace_read(void *priv, unsigned int offset,
 		break;
 	default:
 		WARN(1, "invalid config space read size %d\n", size);
-		return ~0ULL;
+		goto out;
 	}
 
-	if (um_pci_send_cmd(dev, &hdr, sizeof(hdr), NULL, 0,
-			    data, sizeof(data)))
-		return ~0ULL;
+	if (um_pci_send_cmd(dev, &hdr, sizeof(hdr), NULL, 0, data, 8))
+		goto out;
 
 	switch (size) {
 	case 1:
-		return data[0];
+		ret = data[0];
+		break;
 	case 2:
-		return le16_to_cpup((void *)data);
+		ret = le16_to_cpup((void *)data);
+		break;
 	case 4:
-		return le32_to_cpup((void *)data);
+		ret = le32_to_cpup((void *)data);
+		break;
 #ifdef CONFIG_64BIT
 	case 8:
-		return le64_to_cpup((void *)data);
+		ret = le64_to_cpup((void *)data);
+		break;
 #endif
 	default:
-		return ~0ULL;
+		break;
 	}
+
+out:
+	put_cpu_var(um_pci_msg_bufs);
+	return ret;
 }
 
 static void um_pci_cfgspace_write(void *priv, unsigned int offset, int size,
@@ -272,8 +301,13 @@ static void um_pci_bar_copy_from(void *priv, void *buffer,
 static unsigned long um_pci_bar_read(void *priv, unsigned int offset,
 				     int size)
 {
-	/* maximum size - we may only use parts of it */
-	u8 data[8];
+	/* buf->data is maximum size - we may only use parts of it */
+	struct um_pci_message_buffer *buf;
+	u8 *data;
+	unsigned long ret = ~0ULL;
+
+	buf = get_cpu_var(um_pci_msg_bufs);
+	data = buf->data;
 
 	switch (size) {
 	case 1:
@@ -285,25 +319,33 @@ static unsigned long um_pci_bar_read(void *priv, unsigned int offset,
 		break;
 	default:
 		WARN(1, "invalid config space read size %d\n", size);
-		return ~0ULL;
+		goto out;
 	}
 
 	um_pci_bar_copy_from(priv, data, offset, size);
 
 	switch (size) {
 	case 1:
-		return data[0];
+		ret = data[0];
+		break;
 	case 2:
-		return le16_to_cpup((void *)data);
+		ret = le16_to_cpup((void *)data);
+		break;
 	case 4:
-		return le32_to_cpup((void *)data);
+		ret = le32_to_cpup((void *)data);
+		break;
 #ifdef CONFIG_64BIT
 	case 8:
-		return le64_to_cpup((void *)data);
+		ret = le64_to_cpup((void *)data);
+		break;
 #endif
 	default:
-		return ~0ULL;
+		break;
 	}
+
+out:
+	put_cpu_var(um_pci_msg_bufs);
+	return ret;
 }
 
 static void um_pci_bar_copy_to(void *priv, unsigned int offset,
@@ -810,7 +852,7 @@ void *pci_root_bus_fwnode(struct pci_bus *bus)
 	return um_pci_fwnode;
 }
 
-int um_pci_init(void)
+static int um_pci_init(void)
 {
 	int err, i;
 
@@ -823,10 +865,16 @@ int um_pci_init(void)
 		 "No virtio device ID configured for PCI - no PCI support\n"))
 		return 0;
 
-	bridge = pci_alloc_host_bridge(0);
-	if (!bridge)
+	um_pci_msg_bufs = alloc_percpu(struct um_pci_message_buffer);
+	if (!um_pci_msg_bufs)
 		return -ENOMEM;
 
+	bridge = pci_alloc_host_bridge(0);
+	if (!bridge) {
+		err = -ENOMEM;
+		goto free;
+	}
+
 	um_pci_fwnode = irq_domain_alloc_named_fwnode("um-pci");
 	if (!um_pci_fwnode) {
 		err = -ENOMEM;
@@ -878,18 +926,22 @@ int um_pci_init(void)
 		irq_domain_remove(um_pci_inner_domain);
 	if (um_pci_fwnode)
 		irq_domain_free_fwnode(um_pci_fwnode);
-	pci_free_resource_list(&bridge->windows);
-	pci_free_host_bridge(bridge);
+	if (bridge) {
+		pci_free_resource_list(&bridge->windows);
+		pci_free_host_bridge(bridge);
+	}
+	free_percpu(um_pci_msg_bufs);
 	return err;
 }
 module_init(um_pci_init);
 
-void um_pci_exit(void)
+static void um_pci_exit(void)
 {
 	unregister_virtio_driver(&um_pci_virtio_driver);
 	irq_domain_remove(um_pci_msi_domain);
 	irq_domain_remove(um_pci_inner_domain);
 	pci_free_resource_list(&bridge->windows);
 	pci_free_host_bridge(bridge);
+	free_percpu(um_pci_msg_bufs);
 }
 module_exit(um_pci_exit);

arch/um/drivers/virtio_uml.c

@@ -27,6 +27,7 @@
 #include <linux/virtio_config.h>
 #include <linux/virtio_ring.h>
 #include <linux/time-internal.h>
+#include <linux/virtio-uml.h>
 #include <shared/as-layout.h>
 #include <irq_kern.h>
 #include <init.h>
@@ -1139,7 +1140,7 @@ static int virtio_uml_probe(struct platform_device *pdev)
 		rc = os_connect_socket(pdata->socket_path);
 	} while (rc == -EINTR);
 	if (rc < 0)
-		return rc;
+		goto error_free;
 	vu_dev->sock = rc;
 
 	spin_lock_init(&vu_dev->sock_lock);
@@ -1160,6 +1161,8 @@ static int virtio_uml_probe(struct platform_device *pdev)
 
 error_init:
 	os_close_file(vu_dev->sock);
+error_free:
+	kfree(vu_dev);
 	return rc;
 }
 

arch/um/kernel/skas/clone.c

@@ -24,8 +24,7 @@
 void __attribute__ ((__section__ (".__syscall_stub")))
 stub_clone_handler(void)
 {
-	int stack;
-	struct stub_data *data = (void *) ((unsigned long)&stack & ~(UM_KERN_PAGE_SIZE - 1));
+	struct stub_data *data = get_stub_page();
 	long err;
 
 	err = stub_syscall2(__NR_clone, CLONE_PARENT | CLONE_FILES | SIGCHLD,

arch/x86/um/shared/sysdep/stub_32.h

@@ -101,4 +101,16 @@ static inline void remap_stack_and_trap(void)
 		"memory");
 }
 
+static __always_inline void *get_stub_page(void)
+{
+	unsigned long ret;
+
+	asm volatile (
+		"movl %%esp,%0 ;"
+		"andl %1,%0"
+		: "=a" (ret)
+		: "g" (~(UM_KERN_PAGE_SIZE - 1)));
+
+	return (void *)ret;
+}
 #endif

arch/x86/um/shared/sysdep/stub_64.h

@@ -108,4 +108,16 @@ static inline void remap_stack_and_trap(void)
 		__syscall_clobber, "r10", "r8", "r9");
 }
 
+static __always_inline void *get_stub_page(void)
+{
+	unsigned long ret;
+
+	asm volatile (
+		"movq %%rsp,%0 ;"
+		"andq %1,%0"
+		: "=a" (ret)
+		: "g" (~(UM_KERN_PAGE_SIZE - 1)));
+
+	return (void *)ret;
+}
 #endif

arch/x86/um/stub_segv.c

@@ -11,9 +11,8 @@
 void __attribute__ ((__section__ (".__syscall_stub")))
 stub_segv_handler(int sig, siginfo_t *info, void *p)
 {
-	int stack;
+	struct faultinfo *f = get_stub_page();
 	ucontext_t *uc = p;
-	struct faultinfo *f = (void *)(((unsigned long)&stack) & ~(UM_KERN_PAGE_SIZE - 1));
 
 	GET_FAULTINFO_FROM_MC(*f, &uc->uc_mcontext);
 	trap_myself();

fs/hostfs/hostfs_kern.c

@@ -381,6 +381,7 @@ static int hostfs_fsync(struct file *file, loff_t start, loff_t end,
 static const struct file_operations hostfs_file_fops = {
 	.llseek		= generic_file_llseek,
 	.splice_read	= generic_file_splice_read,
+	.splice_write	= iter_file_splice_write,
 	.read_iter	= generic_file_read_iter,
 	.write_iter	= generic_file_write_iter,
 	.mmap		= generic_file_mmap,

include/uapi/linux/virtio_pcidev.h

@@ -9,13 +9,14 @@
 
 /**
  * enum virtio_pcidev_ops - virtual PCI device operations
+ * @VIRTIO_PCIDEV_OP_RESERVED: reserved to catch errors
  * @VIRTIO_PCIDEV_OP_CFG_READ: read config space, size is 1, 2, 4 or 8;
  *	the @data field should be filled in by the device (in little endian).
  * @VIRTIO_PCIDEV_OP_CFG_WRITE: write config space, size is 1, 2, 4 or 8;
  *	the @data field contains the data to write (in little endian).
- * @VIRTIO_PCIDEV_OP_BAR_READ: read BAR mem/pio, size can be variable;
+ * @VIRTIO_PCIDEV_OP_MMIO_READ: read BAR mem/pio, size can be variable;
  *	the @data field should be filled in by the device (in little endian).
- * @VIRTIO_PCIDEV_OP_BAR_WRITE: write BAR mem/pio, size can be variable;
+ * @VIRTIO_PCIDEV_OP_MMIO_WRITE: write BAR mem/pio, size can be variable;
  *	the @data field contains the data to write (in little endian).
  * @VIRTIO_PCIDEV_OP_MMIO_MEMSET: memset MMIO, size is variable but
  *	the @data field only has one byte (unlike @VIRTIO_PCIDEV_OP_MMIO_WRITE)