Merge branch 'for-linus-4.21-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml

Pull UML updates from Richard Weinberger:

 - DISCARD support for our block device driver

 - Many TLB flush optimizations

 - Various smaller fixes

 - And most important, Anton agreed to help me maintaining UML

* 'for-linus-4.21-rc1' of git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml:
  um: Remove obsolete reenable_XX calls
  um: writev needs <sys/uio.h>
  Add Anton Ivanov to UML maintainers
  um: remove redundant generic-y
  um: Optimize Flush TLB for force/fork case
  um: Avoid marking pages with "changed protection"
  um: Skip TLB flushing where not needed
  um: Optimize TLB operations v2
  um: Remove unnecessary faulted check in uaccess.c
  um: Add support for DISCARD in the UBD Driver
  um: Remove unsafe printks from the io thread
  um: Clean-up command processing in UML UBD driver
  um: Switch to block-mq constants in the UML UBD driver
  um: Make GCOV depend on !KCOV
  um: Include sys/uio.h to have writev()
  um: Add HAVE_DEBUG_BUGVERBOSE
  um: Update maintainers file entry

MAINTAINERS

@@ -15951,15 +15951,16 @@ F:	drivers/media/usb/zr364xx/
 USER-MODE LINUX (UML)
 M:	Jeff Dike <jdike@addtoit.com>
 M:	Richard Weinberger <richard@nod.at>
+M:	Anton Ivanov <anton.ivanov@cambridgegreys.com>
 L:	linux-um@lists.infradead.org
 W:	http://user-mode-linux.sourceforge.net
+Q:	https://patchwork.ozlabs.org/project/linux-um/list/
 T:	git git://git.kernel.org/pub/scm/linux/kernel/git/rw/uml.git
 S:	Maintained
 F:	Documentation/virtual/uml/
 F:	arch/um/
 F:	arch/x86/um/
 F:	fs/hostfs/
-F:	fs/hppfs/
 
 USERSPACE COPYIN/COPYOUT (UIOVEC)
 M:	Alexander Viro <viro@zeniv.linux.org.uk>

arch/um/Kconfig

@@ -12,6 +12,7 @@ config UML
 	select HAVE_UID16
 	select HAVE_FUTEX_CMPXCHG if FUTEX
 	select HAVE_DEBUG_KMEMLEAK
+	select HAVE_DEBUG_BUGVERBOSE
 	select GENERIC_IRQ_SHOW
 	select GENERIC_CPU_DEVICES
 	select GENERIC_CLOCKEVENTS

arch/um/Kconfig.debug

@@ -16,6 +16,7 @@ config GPROF
 config GCOV
 	bool "Enable gcov support"
 	depends on DEBUG_INFO
+	depends on !KCOV
 	help
 	  This option allows developers to retrieve coverage data from a UML
 	  session.

arch/um/drivers/chan_kern.c

@@ -211,12 +211,6 @@ void deactivate_chan(struct chan *chan, int irq)
 		deactivate_fd(chan->fd, irq);
 }
 
-void reactivate_chan(struct chan *chan, int irq)
-{
-	if (chan && chan->enabled)
-		reactivate_fd(chan->fd, irq);
-}
-
 int write_chan(struct chan *chan, const char *buf, int len,
 	       int write_irq)
 {
@@ -228,8 +222,6 @@ int write_chan(struct chan *chan, const char *buf, int len,
 	n = chan->ops->write(chan->fd, buf, len, chan->data);
 	if (chan->primary) {
 		ret = n;
-		if ((ret == -EAGAIN) || ((ret >= 0) && (ret < len)))
-			reactivate_fd(chan->fd, write_irq);
 	}
 	return ret;
 }
@@ -527,8 +519,6 @@ void chan_interrupt(struct line *line, int irq)
 			tty_insert_flip_char(port, c, TTY_NORMAL);
 	} while (err > 0);
 
-	if (err == 0)
-		reactivate_fd(chan->fd, irq);
 	if (err == -EIO) {
 		if (chan->primary) {
 			tty_port_tty_hangup(&line->port, false);

arch/um/drivers/line.c

@@ -235,14 +235,6 @@ void line_unthrottle(struct tty_struct *tty)
 
 	line->throttled = 0;
 	chan_interrupt(line, line->driver->read_irq);
-
-	/*
-	 * Maybe there is enough stuff pending that calling the interrupt
-	 * throttles us again.  In this case, line->throttled will be 1
-	 * again and we shouldn't turn the interrupt back on.
-	 */
-	if (!line->throttled)
-		reactivate_chan(line->chan_in, line->driver->read_irq);
 }
 
 static irqreturn_t line_write_interrupt(int irq, void *data)
@@ -667,8 +659,6 @@ static irqreturn_t winch_interrupt(int irq, void *data)
 		tty_kref_put(tty);
 	}
  out:
-	if (winch->fd != -1)
-		reactivate_fd(winch->fd, WINCH_IRQ);
 	return IRQ_HANDLED;
 }
 

arch/um/drivers/mconsole_kern.c

@@ -96,7 +96,6 @@ static irqreturn_t mconsole_interrupt(int irq, void *dev_id)
 	}
 	if (!list_empty(&mc_requests))
 		schedule_work(&mconsole_work);
-	reactivate_fd(fd, MCONSOLE_IRQ);
 	return IRQ_HANDLED;
 }
 
@@ -240,7 +239,6 @@ void mconsole_stop(struct mc_request *req)
 		(*req->cmd->handler)(req);
 	}
 	os_set_fd_block(req->originating_fd, 0);
-	reactivate_fd(req->originating_fd, MCONSOLE_IRQ);
 	mconsole_reply(req, "", 0, 0);
 }
 

arch/um/drivers/net_kern.c

@@ -137,8 +137,6 @@ static irqreturn_t uml_net_interrupt(int irq, void *dev_id)
 		schedule_work(&lp->work);
 		goto out;
 	}
-	reactivate_fd(lp->fd, UM_ETH_IRQ);
-
 out:
 	spin_unlock(&lp->lock);
 	return IRQ_HANDLED;

arch/um/drivers/port_kern.c

@@ -137,7 +137,6 @@ static void port_work_proc(struct work_struct *unused)
 		if (!port->has_connection)
 			continue;
 
-		reactivate_fd(port->fd, ACCEPT_IRQ);
 		while (port_accept(port))
 			;
 		port->has_connection = 0;

arch/um/drivers/random.c

@@ -73,7 +73,6 @@ static ssize_t rng_dev_read (struct file *filp, char __user *buf, size_t size,
 				return ret ? : -EAGAIN;
 
 			atomic_inc(&host_sleep_count);
-			reactivate_fd(random_fd, RANDOM_IRQ);
 			add_sigio_fd(random_fd);
 
 			add_wait_queue(&host_read_wait, &wait);

arch/um/drivers/ubd_kern.c

 /*
+ * Copyright (C) 2018 Cambridge Greys Ltd
  * Copyright (C) 2015-2016 Anton Ivanov (aivanov@brocade.com)
  * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
  * Licensed under the GPL
@@ -43,11 +44,11 @@
 #include <os.h>
 #include "cow.h"
 
-enum ubd_req { UBD_READ, UBD_WRITE, UBD_FLUSH };
+/* Max request size is determined by sector mask - 32K */
+#define UBD_MAX_REQUEST (8 * sizeof(long))
 
 struct io_thread_req {
 	struct request *req;
-	enum ubd_req op;
 	int fds[2];
 	unsigned long offsets[2];
 	unsigned long long offset;
@@ -153,6 +154,7 @@ struct ubd {
 	struct openflags openflags;
 	unsigned shared:1;
 	unsigned no_cow:1;
+	unsigned no_trim:1;
 	struct cow cow;
 	struct platform_device pdev;
 	struct request_queue *queue;
@@ -176,6 +178,7 @@ struct ubd {
 	.boot_openflags =	OPEN_FLAGS, \
 	.openflags =		OPEN_FLAGS, \
 	.no_cow =               0, \
+	.no_trim =		0, \
 	.shared =		0, \
 	.cow =			DEFAULT_COW, \
 	.lock =			__SPIN_LOCK_UNLOCKED(ubd_devs.lock), \
@@ -322,7 +325,7 @@ static int ubd_setup_common(char *str, int *index_out, char **error_out)
 		*index_out = n;
 
 	err = -EINVAL;
-	for (i = 0; i < sizeof("rscd="); i++) {
+	for (i = 0; i < sizeof("rscdt="); i++) {
 		switch (*str) {
 		case 'r':
 			flags.w = 0;
@@ -336,12 +339,15 @@ static int ubd_setup_common(char *str, int *index_out, char **error_out)
 		case 'c':
 			ubd_dev->shared = 1;
 			break;
+		case 't':
+			ubd_dev->no_trim = 1;
+			break;
 		case '=':
 			str++;
 			goto break_loop;
 		default:
 			*error_out = "Expected '=' or flag letter "
-				"(r, s, c, or d)";
+				"(r, s, c, t or d)";
 			goto out;
 		}
 		str++;
@@ -414,6 +420,7 @@ __uml_help(ubd_setup,
 "    'c' will cause the device to be treated as being shared between multiple\n"
 "    UMLs and file locking will be turned off - this is appropriate for a\n"
 "    cluster filesystem and inappropriate at almost all other times.\n\n"
+"    't' will disable trim/discard support on the device (enabled by default).\n\n"
 );
 
 static int udb_setup(char *str)
@@ -511,16 +518,21 @@ static void ubd_handler(void)
 		}
 		for (count = 0; count < n/sizeof(struct io_thread_req *); count++) {
 			struct io_thread_req *io_req = (*irq_req_buffer)[count];
-			int err = io_req->error ? BLK_STS_IOERR : BLK_STS_OK;
-
-			if (!blk_update_request(io_req->req, err, io_req->length))
-				__blk_mq_end_request(io_req->req, err);
 
+			if ((io_req->error == BLK_STS_NOTSUPP) && (req_op(io_req->req) == REQ_OP_DISCARD)) {
+				blk_queue_max_discard_sectors(io_req->req->q, 0);
+				blk_queue_max_write_zeroes_sectors(io_req->req->q, 0);
+				blk_queue_flag_clear(QUEUE_FLAG_DISCARD, io_req->req->q);
+			}
+			if ((io_req->error) || (io_req->buffer == NULL))
+				blk_mq_end_request(io_req->req, io_req->error);
+			else {
+				if (!blk_update_request(io_req->req, io_req->error, io_req->length))
+					__blk_mq_end_request(io_req->req, io_req->error);
+			}
 			kfree(io_req);
 		}
 	}
-
-	reactivate_fd(thread_fd, UBD_IRQ);
 }
 
 static irqreturn_t ubd_intr(int irq, void *dev)
@@ -789,7 +801,7 @@ static int ubd_open_dev(struct ubd *ubd_dev)
 
 	if((fd == -ENOENT) && create_cow){
 		fd = create_cow_file(ubd_dev->file, ubd_dev->cow.file,
-					  ubd_dev->openflags, 1 << 9, PAGE_SIZE,
+					  ubd_dev->openflags, SECTOR_SIZE, PAGE_SIZE,
 					  &ubd_dev->cow.bitmap_offset,
 					  &ubd_dev->cow.bitmap_len,
 					  &ubd_dev->cow.data_offset);
@@ -830,6 +842,14 @@ static int ubd_open_dev(struct ubd *ubd_dev)
 		if(err < 0) goto error;
 		ubd_dev->cow.fd = err;
 	}
+	if (ubd_dev->no_trim == 0) {
+		ubd_dev->queue->limits.discard_granularity = SECTOR_SIZE;
+		ubd_dev->queue->limits.discard_alignment = SECTOR_SIZE;
+		blk_queue_max_discard_sectors(ubd_dev->queue, UBD_MAX_REQUEST);
+		blk_queue_max_write_zeroes_sectors(ubd_dev->queue, UBD_MAX_REQUEST);
+		blk_queue_flag_set(QUEUE_FLAG_DISCARD, ubd_dev->queue);
+	}
+	blk_queue_flag_set(QUEUE_FLAG_NONROT, ubd_dev->queue);
 	return 0;
  error:
 	os_close_file(ubd_dev->fd);
@@ -882,7 +902,7 @@ static int ubd_disk_register(int major, u64 size, int unit,
 	return 0;
 }
 
-#define ROUND_BLOCK(n) ((n + ((1 << 9) - 1)) & (-1 << 9))
+#define ROUND_BLOCK(n) ((n + (SECTOR_SIZE - 1)) & (-SECTOR_SIZE))
 
 static const struct blk_mq_ops ubd_mq_ops = {
 	.queue_rq = ubd_queue_rq,
@@ -1234,10 +1254,10 @@ static void cowify_bitmap(__u64 io_offset, int length, unsigned long *cow_mask,
 			  __u64 bitmap_offset, unsigned long *bitmap_words,
 			  __u64 bitmap_len)
 {
-	__u64 sector = io_offset >> 9;
+	__u64 sector = io_offset >> SECTOR_SHIFT;
 	int i, update_bitmap = 0;
 
-	for(i = 0; i < length >> 9; i++){
+	for (i = 0; i < length >> SECTOR_SHIFT; i++) {
 		if(cow_mask != NULL)
 			ubd_set_bit(i, (unsigned char *) cow_mask);
 		if(ubd_test_bit(sector + i, (unsigned char *) bitmap))
@@ -1271,14 +1291,14 @@ static void cowify_bitmap(__u64 io_offset, int length, unsigned long *cow_mask,
 static void cowify_req(struct io_thread_req *req, unsigned long *bitmap,
 		       __u64 bitmap_offset, __u64 bitmap_len)
 {
-	__u64 sector = req->offset >> 9;
+	__u64 sector = req->offset >> SECTOR_SHIFT;
 	int i;
 
-	if(req->length > (sizeof(req->sector_mask) * 8) << 9)
+	if (req->length > (sizeof(req->sector_mask) * 8) << SECTOR_SHIFT)
 		panic("Operation too long");
 
-	if(req->op == UBD_READ) {
-		for(i = 0; i < req->length >> 9; i++){
+	if (req_op(req->req) == REQ_OP_READ) {
+		for (i = 0; i < req->length >> SECTOR_SHIFT; i++) {
 			if(ubd_test_bit(sector + i, (unsigned char *) bitmap))
 				ubd_set_bit(i, (unsigned char *)
 					    &req->sector_mask);
@@ -1307,68 +1327,86 @@ static int ubd_queue_one_vec(struct blk_mq_hw_ctx *hctx, struct request *req,
 		io_req->fds[0] = dev->fd;
 	io_req->error = 0;
 
-	if (req_op(req) == REQ_OP_FLUSH) {
-		io_req->op = UBD_FLUSH;
-	} else {
-		io_req->fds[1] = dev->fd;
-		io_req->cow_offset = -1;
-		io_req->offset = off;
-		io_req->length = bvec->bv_len;
-		io_req->sector_mask = 0;
-		io_req->op = rq_data_dir(req) == READ ? UBD_READ : UBD_WRITE;
-		io_req->offsets[0] = 0;
-		io_req->offsets[1] = dev->cow.data_offset;
+	if (bvec != NULL) {
 		io_req->buffer = page_address(bvec->bv_page) + bvec->bv_offset;
-		io_req->sectorsize = 1 << 9;
-
-		if (dev->cow.file) {
-			cowify_req(io_req, dev->cow.bitmap,
-				   dev->cow.bitmap_offset, dev->cow.bitmap_len);
-		}
+		io_req->length = bvec->bv_len;
+	} else {
+		io_req->buffer = NULL;
+		io_req->length = blk_rq_bytes(req);
 	}
 
+	io_req->sectorsize = SECTOR_SIZE;
+	io_req->fds[1] = dev->fd;
+	io_req->cow_offset = -1;
+	io_req->offset = off;
+	io_req->sector_mask = 0;
+	io_req->offsets[0] = 0;
+	io_req->offsets[1] = dev->cow.data_offset;
+
+	if (dev->cow.file)
+		cowify_req(io_req, dev->cow.bitmap,
+			   dev->cow.bitmap_offset, dev->cow.bitmap_len);
+
 	ret = os_write_file(thread_fd, &io_req, sizeof(io_req));
 	if (ret != sizeof(io_req)) {
 		if (ret != -EAGAIN)
 			pr_err("write to io thread failed: %d\n", -ret);
 		kfree(io_req);
 	}
-
 	return ret;
 }
 
+static int queue_rw_req(struct blk_mq_hw_ctx *hctx, struct request *req)
+{
+	struct req_iterator iter;
+	struct bio_vec bvec;
+	int ret;
+	u64 off = (u64)blk_rq_pos(req) << SECTOR_SHIFT;
+
+	rq_for_each_segment(bvec, req, iter) {
+		ret = ubd_queue_one_vec(hctx, req, off, &bvec);
+		if (ret < 0)
+			return ret;
+		off += bvec.bv_len;
+	}
+	return 0;
+}
+
 static blk_status_t ubd_queue_rq(struct blk_mq_hw_ctx *hctx,
 				 const struct blk_mq_queue_data *bd)
 {
 	struct ubd *ubd_dev = hctx->queue->queuedata;
 	struct request *req = bd->rq;
-	int ret = 0;
+	int ret = 0, res = BLK_STS_OK;
 
 	blk_mq_start_request(req);
 
 	spin_lock_irq(&ubd_dev->lock);
 
-	if (req_op(req) == REQ_OP_FLUSH) {
+	switch (req_op(req)) {
+	/* operations with no lentgth/offset arguments */
+	case REQ_OP_FLUSH:
 		ret = ubd_queue_one_vec(hctx, req, 0, NULL);
-	} else {
-		struct req_iterator iter;
-		struct bio_vec bvec;
-		u64 off = (u64)blk_rq_pos(req) << 9;
-
-		rq_for_each_segment(bvec, req, iter) {
-			ret = ubd_queue_one_vec(hctx, req, off, &bvec);
-			if (ret < 0)
-				goto out;
-			off += bvec.bv_len;
-		}
+		break;
+	case REQ_OP_READ:
+	case REQ_OP_WRITE:
+		ret = queue_rw_req(hctx, req);
+		break;
+	case REQ_OP_DISCARD:
+	case REQ_OP_WRITE_ZEROES:
+		ret = ubd_queue_one_vec(hctx, req, (u64)blk_rq_pos(req) << 9, NULL);
+		break;
+	default:
+		WARN_ON_ONCE(1);
+		res = BLK_STS_NOTSUPP;
 	}
-out:
+
 	spin_unlock_irq(&ubd_dev->lock);
 
 	if (ret < 0)
 		blk_mq_requeue_request(req, true);
 
-	return BLK_STS_OK;
+	return res;
 }
 
 static int ubd_getgeo(struct block_device *bdev, struct hd_geometry *geo)
@@ -1413,39 +1451,60 @@ static int ubd_ioctl(struct block_device *bdev, fmode_t mode,
 	return -EINVAL;
 }
 
+static int map_error(int error_code)
+{
+	switch (error_code) {
+	case 0:
+		return BLK_STS_OK;
+	case ENOSYS:
+	case EOPNOTSUPP:
+		return BLK_STS_NOTSUPP;
+	case ENOSPC:
+		return BLK_STS_NOSPC;
+	}
+	return BLK_STS_IOERR;
+}
+
+/*
+ * Everything from here onwards *IS NOT PART OF THE KERNEL*
+ *
+ * The following functions are part of UML hypervisor code.
+ * All functions from here onwards are executed as a helper
+ * thread and are not allowed to execute any kernel functions.
+ *
+ * Any communication must occur strictly via shared memory and IPC.
+ *
+ * Do not add printks, locks, kernel memory operations, etc - it
+ * will result in unpredictable behaviour and/or crashes.
+ */
+
 static int update_bitmap(struct io_thread_req *req)
 {
 	int n;
 
 	if(req->cow_offset == -1)
-		return 0;
+		return map_error(0);
 
 	n = os_pwrite_file(req->fds[1], &req->bitmap_words,
 			  sizeof(req->bitmap_words), req->cow_offset);
-	if(n != sizeof(req->bitmap_words)){
-		printk("do_io - bitmap update failed, err = %d fd = %d\n", -n,
-		       req->fds[1]);
-		return 1;
-	}
+	if (n != sizeof(req->bitmap_words))
+		return map_error(-n);
 
-	return 0;
+	return map_error(0);
 }
 
 static void do_io(struct io_thread_req *req)
 {
-	char *buf;
+	char *buf = NULL;
 	unsigned long len;
 	int n, nsectors, start, end, bit;
 	__u64 off;
 
-	if (req->op == UBD_FLUSH) {
+	/* FLUSH is really a special case, we cannot "case" it with others */
+
+	if (req_op(req->req) == REQ_OP_FLUSH) {
 		/* fds[0] is always either the rw image or our cow file */
-		n = os_sync_file(req->fds[0]);
-		if (n != 0) {
-			printk("do_io - sync failed err = %d "
-			       "fd = %d\n", -n, req->fds[0]);
-			req->error = 1;
-		}
+		req->error = map_error(-os_sync_file(req->fds[0]));
 		return;
 	}
 
@@ -1462,30 +1521,42 @@ static void do_io(struct io_thread_req *req)
 		off = req->offset + req->offsets[bit] +
 			start * req->sectorsize;
 		len = (end - start) * req->sectorsize;
-		buf = &req->buffer[start * req->sectorsize];
+		if (req->buffer != NULL)
+			buf = &req->buffer[start * req->sectorsize];
 
-		if(req->op == UBD_READ){
+		switch (req_op(req->req)) {
+		case REQ_OP_READ:
 			n = 0;
 			do {
 				buf = &buf[n];
 				len -= n;
 				n = os_pread_file(req->fds[bit], buf, len, off);
 				if (n < 0) {
-					printk("do_io - read failed, err = %d "
-					       "fd = %d\n", -n, req->fds[bit]);
-					req->error = 1;
+					req->error = map_error(-n);
 					return;
 				}
 			} while((n < len) && (n != 0));
 			if (n < len) memset(&buf[n], 0, len - n);
-		} else {
+			break;
+		case REQ_OP_WRITE:
 			n = os_pwrite_file(req->fds[bit], buf, len, off);
 			if(n != len){
-				printk("do_io - write failed err = %d "
-				       "fd = %d\n", -n, req->fds[bit]);
-				req->error = 1;
+				req->error = map_error(-n);
+				return;
+			}
+			break;
+		case REQ_OP_DISCARD:
+		case REQ_OP_WRITE_ZEROES:
+			n = os_falloc_punch(req->fds[bit], off, len);
+			if (n) {
+				req->error = map_error(-n);
 				return;
 			}
+			break;
+		default:
+			WARN_ON_ONCE(1);
+			req->error = BLK_STS_NOTSUPP;
+			return;
 		}
 
 		start = end;
@@ -1520,11 +1591,6 @@ int io_thread(void *arg)
 			if (n == -EAGAIN) {
 				ubd_read_poll(-1);
 				continue;
-			} else {
-				printk("io_thread - read failed, fd = %d, "
-				       "err = %d,"
-				       "reminder = %d\n",
-				       kernel_fd, -n, io_remainder_size);
 			}
 		}
 
@@ -1539,11 +1605,6 @@ int io_thread(void *arg)
 			res = os_write_file(kernel_fd, ((char *) io_req_buffer) + written, n);
 			if (res >= 0) {
 				written += res;
-			} else {
-				if (res != -EAGAIN) {
-					printk("io_thread - write failed, fd = %d, "
-					       "err = %d\n", kernel_fd, -n);
-				}
 			}
 			if (written < n) {
 				ubd_write_poll(-1);

arch/um/drivers/vector_user.c

@@ -25,11 +25,13 @@
 #include <linux/if_packet.h>
 #include <sys/socket.h>
 #include <sys/wait.h>
+#include <sys/uio.h>
 #include <linux/virtio_net.h>
 #include <netdb.h>
 #include <stdlib.h>
 #include <os.h>
 #include <um_malloc.h>
+#include <sys/uio.h>
 #include "vector_user.h"
 
 #define ID_GRE 0

arch/um/include/asm/Kbuild

@@ -10,9 +10,7 @@ generic-y += exec.h
 generic-y += extable.h
 generic-y += ftrace.h
 generic-y += futex.h
-generic-y += hardirq.h
 generic-y += hw_irq.h
-generic-y += io.h
 generic-y += irq_regs.h
 generic-y += irq_work.h
 generic-y += kdebug.h

arch/um/include/asm/pgtable.h

@@ -197,12 +197,17 @@ static inline pte_t pte_mkold(pte_t pte)
 
 static inline pte_t pte_wrprotect(pte_t pte)
 { 
-	pte_clear_bits(pte, _PAGE_RW);
+	if (likely(pte_get_bits(pte, _PAGE_RW)))
+		pte_clear_bits(pte, _PAGE_RW);
+	else
+		return pte;
 	return(pte_mknewprot(pte)); 
 }
 
 static inline pte_t pte_mkread(pte_t pte)
 { 
+	if (unlikely(pte_get_bits(pte, _PAGE_USER)))
+		return pte;
 	pte_set_bits(pte, _PAGE_USER);
 	return(pte_mknewprot(pte)); 
 }
@@ -221,6 +226,8 @@ static inline pte_t pte_mkyoung(pte_t pte)
 
 static inline pte_t pte_mkwrite(pte_t pte)	
 {
+	if (unlikely(pte_get_bits(pte,  _PAGE_RW)))
+		return pte;
 	pte_set_bits(pte, _PAGE_RW);
 	return(pte_mknewprot(pte)); 
 }

arch/um/include/shared/irq_user.h

@@ -31,7 +31,6 @@ struct irq_fd {
 struct siginfo;
 extern void sigio_handler(int sig, struct siginfo *unused_si, struct uml_pt_regs *regs);
 extern void free_irq_by_fd(int fd);
-extern void reactivate_fd(int fd, int irqnum);
 extern void deactivate_fd(int fd, int irqnum);
 extern int deactivate_all_fds(void);
 extern int activate_ipi(int fd, int pid);

arch/um/include/shared/os.h

@@ -175,6 +175,7 @@ extern int os_fchange_dir(int fd);
 extern unsigned os_major(unsigned long long dev);
 extern unsigned os_minor(unsigned long long dev);
 extern unsigned long long os_makedev(unsigned major, unsigned minor);
+extern int os_falloc_punch(int fd, unsigned long long offset, int count);
 
 /* start_up.c */
 extern void os_early_checks(void);

arch/um/kernel/irq.c

@@ -350,11 +350,6 @@ static void free_irq_by_irq_and_dev(unsigned int irq, void *dev)
 }
 
 
-void reactivate_fd(int fd, int irqnum)
-{
-	/** NOP - we do auto-EOI now **/
-}
-
 void deactivate_fd(int fd, int irqnum)
 {
 	struct irq_entry *to_free;
@@ -449,7 +444,6 @@ int um_request_irq(unsigned int irq, int fd, int type,
 }
 
 EXPORT_SYMBOL(um_request_irq);
-EXPORT_SYMBOL(reactivate_fd);
 
 /*
  * irq_chip must define at least enable/disable and ack when

arch/um/kernel/sigio.c

@@ -16,7 +16,6 @@ static irqreturn_t sigio_interrupt(int irq, void *data)
 	char c;
 
 	os_read_file(sigio_irq_fd, &c, sizeof(c));
-	reactivate_fd(sigio_irq_fd, SIGIO_WRITE_IRQ);
 	return IRQ_HANDLED;
 }
 

arch/um/kernel/skas/uaccess.c

@@ -62,27 +62,28 @@ static int do_op_one_page(unsigned long addr, int len, int is_write,
 	jmp_buf buf;
 	struct page *page;
 	pte_t *pte;
-	int n, faulted;
+	int n;
 
 	pte = maybe_map(addr, is_write);
 	if (pte == NULL)
 		return -1;
 
 	page = pte_page(*pte);
+#ifdef CONFIG_64BIT
+	pagefault_disable();
+	addr = (unsigned long) page_address(page) +
+		(addr & ~PAGE_MASK);
+#else
 	addr = (unsigned long) kmap_atomic(page) +
 		(addr & ~PAGE_MASK);
+#endif
+	n = (*op)(addr, len, arg);
 
-	current->thread.fault_catcher = &buf;
-
-	faulted = UML_SETJMP(&buf);
-	if (faulted == 0)
-		n = (*op)(addr, len, arg);
-	else
-		n = -1;
-
-	current->thread.fault_catcher = NULL;
-
+#ifdef CONFIG_64BIT
+	pagefault_enable();
+#else
 	kunmap_atomic((void *)addr);
+#endif
 
 	return n;
 }

arch/um/kernel/tlb.c

@@ -37,17 +37,19 @@ struct host_vm_change {
 			} mprotect;
 		} u;
 	} ops[1];
+	int userspace;
 	int index;
-	struct mm_id *id;
+	struct mm_struct *mm;
 	void *data;
 	int force;
 };
 
-#define INIT_HVC(mm, force) \
+#define INIT_HVC(mm, force, userspace) \
 	((struct host_vm_change) \
 	 { .ops		= { { .type = NONE } },	\
-	   .id		= &mm->context.id, \
+	   .mm		= mm, \
        	   .data	= NULL, \
+	   .userspace	= userspace, \
 	   .index	= 0, \
 	   .force	= force })
 
@@ -68,18 +70,40 @@ static int do_ops(struct host_vm_change *hvc, int end,
 		op = &hvc->ops[i];
 		switch (op->type) {
 		case MMAP:
-			ret = map(hvc->id, op->u.mmap.addr, op->u.mmap.len,
-				  op->u.mmap.prot, op->u.mmap.fd,
-				  op->u.mmap.offset, finished, &hvc->data);
+			if (hvc->userspace)
+				ret = map(&hvc->mm->context.id, op->u.mmap.addr,
+					  op->u.mmap.len, op->u.mmap.prot,
+					  op->u.mmap.fd,
+					  op->u.mmap.offset, finished,
+					  &hvc->data);
+			else
+				map_memory(op->u.mmap.addr, op->u.mmap.offset,
+					   op->u.mmap.len, 1, 1, 1);
 			break;
 		case MUNMAP:
-			ret = unmap(hvc->id, op->u.munmap.addr,
-				    op->u.munmap.len, finished, &hvc->data);
+			if (hvc->userspace)
+				ret = unmap(&hvc->mm->context.id,
+					    op->u.munmap.addr,
+					    op->u.munmap.len, finished,
+					    &hvc->data);
+			else
+				ret = os_unmap_memory(
+					(void *) op->u.munmap.addr,
+						      op->u.munmap.len);
+
 			break;
 		case MPROTECT:
-			ret = protect(hvc->id, op->u.mprotect.addr,
-				      op->u.mprotect.len, op->u.mprotect.prot,
-				      finished, &hvc->data);
+			if (hvc->userspace)
+				ret = protect(&hvc->mm->context.id,
+					      op->u.mprotect.addr,
+					      op->u.mprotect.len,
+					      op->u.mprotect.prot,
+					      finished, &hvc->data);
+			else
+				ret = os_protect_memory(
+					(void *) op->u.mprotect.addr,
+							op->u.mprotect.len,
+							1, 1, 1);
 			break;
 		default:
 			printk(KERN_ERR "Unknown op type %d in do_ops\n",
@@ -100,9 +124,12 @@ static int add_mmap(unsigned long virt, unsigned long phys, unsigned long len,
 {
 	__u64 offset;
 	struct host_vm_op *last;
-	int fd, ret = 0;
+	int fd = -1, ret = 0;
 
-	fd = phys_mapping(phys, &offset);
+	if (hvc->userspace)
+		fd = phys_mapping(phys, &offset);
+	else
+		offset = phys;
 	if (hvc->index != 0) {
 		last = &hvc->ops[hvc->index - 1];
 		if ((last->type == MMAP) &&
@@ -215,10 +242,11 @@ static inline int update_pte_range(pmd_t *pmd, unsigned long addr,
 		prot = ((r ? UM_PROT_READ : 0) | (w ? UM_PROT_WRITE : 0) |
 			(x ? UM_PROT_EXEC : 0));
 		if (hvc->force || pte_newpage(*pte)) {
-			if (pte_present(*pte))
-				ret = add_mmap(addr, pte_val(*pte) & PAGE_MASK,
-					       PAGE_SIZE, prot, hvc);
-			else
+			if (pte_present(*pte)) {
+				if (pte_newpage(*pte))
+					ret = add_mmap(addr, pte_val(*pte) & PAGE_MASK,
+						       PAGE_SIZE, prot, hvc);
+			} else
 				ret = add_munmap(addr, PAGE_SIZE, hvc);
 		} else if (pte_newprot(*pte))
 			ret = add_mprotect(addr, PAGE_SIZE, prot, hvc);
@@ -277,9 +305,9 @@ void fix_range_common(struct mm_struct *mm, unsigned long start_addr,
 	pgd_t *pgd;
 	struct host_vm_change hvc;
 	unsigned long addr = start_addr, next;
-	int ret = 0;
+	int ret = 0, userspace = 1;
 
-	hvc = INIT_HVC(mm, force);
+	hvc = INIT_HVC(mm, force, userspace);
 	pgd = pgd_offset(mm, addr);
 	do {
 		next = pgd_addr_end(addr, end_addr);
@@ -314,9 +342,11 @@ static int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)
 	pmd_t *pmd;
 	pte_t *pte;
 	unsigned long addr, last;
-	int updated = 0, err;
+	int updated = 0, err = 0, force = 0, userspace = 0;
+	struct host_vm_change hvc;
 
 	mm = &init_mm;
+	hvc = INIT_HVC(mm, force, userspace);
 	for (addr = start; addr < end;) {
 		pgd = pgd_offset(mm, addr);
 		if (!pgd_present(*pgd)) {
@@ -325,8 +355,7 @@ static int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)
 				last = end;
 			if (pgd_newpage(*pgd)) {
 				updated = 1;
-				err = os_unmap_memory((void *) addr,
-						      last - addr);
+				err = add_munmap(addr, last - addr, &hvc);
 				if (err < 0)
 					panic("munmap failed, errno = %d\n",
 					      -err);
@@ -342,8 +371,7 @@ static int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)
 				last = end;
 			if (pud_newpage(*pud)) {
 				updated = 1;
-				err = os_unmap_memory((void *) addr,
-						      last - addr);
+				err = add_munmap(addr, last - addr, &hvc);
 				if (err < 0)
 					panic("munmap failed, errno = %d\n",
 					      -err);
@@ -359,8 +387,7 @@ static int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)
 				last = end;
 			if (pmd_newpage(*pmd)) {
 				updated = 1;
-				err = os_unmap_memory((void *) addr,
-						      last - addr);
+				err = add_munmap(addr, last - addr, &hvc);
 				if (err < 0)
 					panic("munmap failed, errno = %d\n",
 					      -err);
@@ -372,22 +399,25 @@ static int flush_tlb_kernel_range_common(unsigned long start, unsigned long end)
 		pte = pte_offset_kernel(pmd, addr);
 		if (!pte_present(*pte) || pte_newpage(*pte)) {
 			updated = 1;
-			err = os_unmap_memory((void *) addr,
-					      PAGE_SIZE);
+			err = add_munmap(addr, PAGE_SIZE, &hvc);
 			if (err < 0)
 				panic("munmap failed, errno = %d\n",
 				      -err);
 			if (pte_present(*pte))
-				map_memory(addr,
-					   pte_val(*pte) & PAGE_MASK,
-					   PAGE_SIZE, 1, 1, 1);
+				err = add_mmap(addr, pte_val(*pte) & PAGE_MASK,
+					       PAGE_SIZE, 0, &hvc);
 		}
 		else if (pte_newprot(*pte)) {
 			updated = 1;
-			os_protect_memory((void *) addr, PAGE_SIZE, 1, 1, 1);
+			err = add_mprotect(addr, PAGE_SIZE, 0, &hvc);
 		}
 		addr += PAGE_SIZE;
 	}
+	if (!err)
+		err = do_ops(&hvc, hvc.index, 1);
+
+	if (err < 0)
+		panic("flush_tlb_kernel failed, errno = %d\n", err);
 	return updated;
 }
 
@@ -491,6 +521,13 @@ pte_t *addr_pte(struct task_struct *task, unsigned long addr)
 
 void flush_tlb_all(void)
 {
+	/*
+	 * Don't bother flushing if this address space is about to be
+	 * destroyed.
+	 */
+	if (atomic_read(&current->mm->mm_users) == 0)
+		return;
+
 	flush_tlb_mm(current->mm);
 }
 
@@ -512,6 +549,13 @@ void __flush_tlb_one(unsigned long addr)
 static void fix_range(struct mm_struct *mm, unsigned long start_addr,
 		      unsigned long end_addr, int force)
 {
+	/*
+	 * Don't bother flushing if this address space is about to be
+	 * destroyed.
+	 */
+	if (atomic_read(&mm->mm_users) == 0)
+		return;
+
 	fix_range_common(mm, start_addr, end_addr, force);
 }
 
@@ -527,13 +571,6 @@ EXPORT_SYMBOL(flush_tlb_range);
 void flush_tlb_mm_range(struct mm_struct *mm, unsigned long start,
 			unsigned long end)
 {
-	/*
-	 * Don't bother flushing if this address space is about to be
-	 * destroyed.
-	 */
-	if (atomic_read(&mm->mm_users) == 0)
-		return;
-
 	fix_range(mm, start, end, 0);
 }
 

arch/um/os-Linux/file.c

@@ -610,3 +610,13 @@ unsigned long long os_makedev(unsigned major, unsigned minor)
 {
 	return makedev(major, minor);
 }
+
+int os_falloc_punch(int fd, unsigned long long offset, int len)
+{
+	int n = fallocate(fd, FALLOC_FL_PUNCH_HOLE|FALLOC_FL_KEEP_SIZE, offset, len);
+
+	if (n < 0)
+		return -errno;
+	return n;
+}
+