Merge of Wu's GSOC 09 branch (src.20090525.r4372.wu)

Main changes:
- COW optimization for safecopy.
- safemap, a grant-based interface for sharing memory regions between processes.
- Integration with safemap and complete rework of DS, supporting new data types
  natively (labels, memory ranges, memory mapped ranges).
- For further information:
  http://wiki.minix3.org/en/SummerOfCode2009/MemoryGrants

Additional changes not included in the original Wu's branch:
- Fixed unhandled case in VM when using COW optimization for safecopy in case
  of a block that has already been shared as SMAP.
- Better interface and naming scheme for sys_saferevmap and ds_retrieve_map
  calls.
- Better input checking in syslib: check for page alignment when creating
  memory mapping grants.
- DS notifies subscribers when an entry is deleted.
- Documented the behavior of indirect grants in case of memory mapping.
- Test suite in /usr/src/test/safeperf|safecopy|safemap|ds/* reworked
  and extended.
- Minor fixes and general cleanup.
- TO-DO: Grant ids should be generated and managed the way endpoints are to make
sure grant slots are never misreused.

drivers/audio/framework/audio_fw.c

@@ -1027,11 +1027,11 @@ int pci_func;
 	u32_t u32;
 	message m;
 
-	r= ds_retrieve_u32("amddev", &u32);
+	r= ds_retrieve_label_num("amddev", &u32);
 	if (r != OK)
 	{
 #if 0
-		printf("tell_dev: ds_retrieve_u32 failed for 'amddev': %d\n",
+		printf("tell_dev: ds_retrieve_label_num failed for 'amddev': %d\n",
 			r);
 #endif
 		return;

drivers/e1000/e1000.c

@@ -157,9 +157,9 @@ PRIVATE int sef_cb_init_fresh(int type, sef_init_info_t *info)
     memset(e1000_table, 0, sizeof(e1000_table));
 
     /* Perform calibration. */
-    if((r = micro_delay_calibrate()) != OK)
+    if((r = tsc_calibrate()) != OK)
     {
-        panic("e1000", "rmicro_delay_calibrate failed", r);
+        panic("e1000", "tsc_calibrate failed", r);
     }
     /* Try to notify inet that we are present (again) */
     if ((r = ds_retrieve_u32("inet", &tasknr)) == OK)

drivers/fxp/fxp.c

@@ -414,15 +414,15 @@ PRIVATE int sef_cb_init_fresh(int type, sef_init_info_t *info)
 
 	memset(fxp_table, 0, ft);
 
-	if((r=micro_delay_calibrate()) != OK)
+	if((r=tsc_calibrate()) != OK)
-		panic("FXP","rmicro_delay_calibrate failed", r);
+		panic("FXP","tsc_calibrate failed", r);
 
 	/* Try to notify inet that we are present (again) */
-	r= ds_retrieve_u32("inet", &tasknr);
+	r= ds_retrieve_label_num("inet", &tasknr);
 	if (r == OK)
 		notify(tasknr);
 	else if (r != ESRCH)
-		printf("fxp: ds_retrieve_u32 failed for 'inet': %d\n", r);
+		printf("fxp: ds_retrieve_label_num failed for 'inet': %d\n", r);
 
 	return(OK);
 }
@@ -3013,12 +3013,12 @@ int pci_func;
 	u32_t u32;
 	message m;
 
-	r= ds_retrieve_u32("amddev", &u32);
+	r= ds_retrieve_label_num("amddev", &u32);
 	if (r != OK)
 	{
 #if 0
 		printf(
-		"fxp`tell_dev: ds_retrieve_u32 failed for 'amddev': %d\n",
+		"fxp`tell_dev: ds_retrieve_label_num failed for 'amddev': %d\n",
 			r);
 #endif
 		return;

drivers/orinoco/orinoco.c

@@ -367,11 +367,11 @@ PRIVATE int sef_cb_init_fresh(int type, sef_init_info_t *info)
 	/* Try to notify INET that we are present (again). If INET cannot
 	 * be found, assume this is the first time we started and INET is
 	 * not yet alive. */
-	r = ds_retrieve_u32("inet", &inet_proc_nr);
+	r = ds_retrieve_label_num("inet", &inet_proc_nr);
 	if (r == OK) 
 		notify(inet_proc_nr);
 	else if (r != ESRCH)
-		printf("orinoco: ds_retrieve_u32 failed for 'inet': %d\n", r);
+		printf("orinoco: ds_retrieve_label_num failed for 'inet': %d\n", r);
 
 	return(OK);
 }

drivers/rtl8139/rtl8139.c

@@ -336,11 +336,11 @@ PRIVATE int sef_cb_init_fresh(int type, sef_init_info_t *info)
 	 * be found, assume this is the first time we started and INET is
 	 * not yet alive.
 	 */
-	r= ds_retrieve_u32("inet", &inet_proc_nr);
+	r= ds_retrieve_label_num("inet", &inet_proc_nr);
 	if (r == OK)
 		notify(inet_proc_nr);
 	else if (r != ESRCH)
-		printf("rtl8139: ds_retrieve_u32 failed for 'inet': %d\n", r);
+		printf("rtl8139: ds_retrieve_label_num failed for 'inet': %d\n", r);
 
 	return(OK);
 }
@@ -3017,12 +3017,12 @@ int pci_func;
 	u32_t u32;
 	message m;
 
-	r= ds_retrieve_u32("amddev", &u32);
+	r= ds_retrieve_label_num("amddev", &u32);
 	if (r != OK)
 	{
 #if 0
 		printf(
-		"rtl8139`tell_dev: ds_retrieve_u32 failed for 'amddev': %d\n",
+		"rtl8139`tell_dev: ds_retrieve_label_num failed for 'amddev': %d\n",
 			r);
 #endif
 		return;

drivers/ti1225/ti1225.c

@@ -117,8 +117,8 @@ PRIVATE int sef_cb_init_fresh(int type, sef_init_info_t *info)
 	(progname=strrchr(env_argv[0],'/')) ? progname++
 		: (progname=env_argv[0]);
 
-	if((r=micro_delay_calibrate()) != OK)
+	if((r=tsc_calibrate()) != OK)
-		panic("ti1225", "micro_delay_calibrate failed", r);
+		panic("ti1225", "tsc_calibrate failed", r);
 
 	debug= 0;
 	while (c= getopt(env_argc, env_argv, "d?"), c != -1)

include/minix/com.h

@@ -350,8 +350,11 @@
 
 #  define SYS_VTIMER     (KERNEL_CALL + 45)	/* sys_vtimer() */
 #  define SYS_RUNCTL     (KERNEL_CALL + 46)	/* sys_runctl() */
+#  define SYS_SAFEMAP	 (KERNEL_CALL + 47)	/* sys_safemap() */
+#  define SYS_SAFEREVMAP (KERNEL_CALL + 48)	/* sys_saferevmap() sys_saferevmap2() */
+#  define SYS_SAFEUNMAP	 (KERNEL_CALL + 49)	/* sys_safeunmap() */
 
-#define NR_SYS_CALLS	47	/* number of system calls */ 
+#define NR_SYS_CALLS	50	/* number of system calls */ 
 #define SYS_CALL_MASK_SIZE BITMAP_CHUNKS(NR_SYS_CALLS)
 
 /* Field names for SYS_MEMSET. */
@@ -536,6 +539,15 @@
 #define VSCP_VEC_ADDR	m2_p1	/* start of vector */
 #define VSCP_VEC_SIZE	m2_l2	/* elements in vector */
 
+/* Field names for SYS_SAFEMAPs */
+#define SMAP_EP		m2_i1
+#define SMAP_GID	m2_i2
+#define SMAP_OFFSET	m2_i3
+#define SMAP_SEG	m2_p1
+#define SMAP_ADDRESS	m2_l1
+#define SMAP_BYTES	m2_l2
+#define SMAP_FLAG	m2_s1
+
 /* Field names for SYS_SPROF, _CPROF, _PROFBUF. */
 #define PROF_ACTION    m7_i1    /* start/stop/reset/get */
 #define PROF_MEM_SIZE  m7_i2    /* available memory for data */ 
@@ -556,11 +568,13 @@
 #define SVMCTL_PF_WHO		m1_i1	/* GET_PAGEFAULT reply: process ep */
 #define SVMCTL_PF_I386_CR2	m1_i2	/* GET_PAGEFAULT reply: CR2 */
 #define SVMCTL_PF_I386_ERR	m1_i3	/* GET_PAGEFAULT reply: error code */
-#define SVMCTL_MRG_ADDR		m1_p1	/* MEMREQ_GET reply: address */
+#define	SVMCTL_MRG_TARGET	m2_i1	/* MEMREQ_GET reply: target process */
-#define SVMCTL_MRG_LEN		m1_i1	/* MEMREQ_GET reply: length */
+#define	SVMCTL_MRG_ADDR		m2_i2	/* MEMREQ_GET reply: address */
-#define SVMCTL_MRG_WRITE	m1_i2	/* MEMREQ_GET reply: writeflag */
+#define	SVMCTL_MRG_LENGTH	m2_i3	/* MEMREQ_GET reply: length */
-#define SVMCTL_MRG_EP		m1_i3	/* MEMREQ_GET reply: process */
+#define	SVMCTL_MRG_FLAG		m2_s1	/* MEMREQ_GET reply: flag */
-#define SVMCTL_MRG_REQUESTOR	m1_p2	/* MEMREQ_GET reply: requestor */
+#define	SVMCTL_MRG_EP2		m2_l1	/* MEMREQ_GET reply: source process */
+#define	SVMCTL_MRG_ADDR2	m2_l2	/* MEMREQ_GET reply: source address */
+#define SVMCTL_MRG_REQUESTOR	m2_p1	/* MEMREQ_GET reply: requestor */
 #define SVMCTL_MAP_VIR_ADDR	m1_p1
 
 /* Reply message for VMCTL_KERN_PHYSMAP */
@@ -654,19 +668,23 @@
 
 #define DS_RQ_BASE		0x800
 
-#define DS_PUBLISH	(DS_RQ_BASE + 0)	/* publish information */
+#define DS_PUBLISH	(DS_RQ_BASE + 0)	/* publish data */
-#define DS_SUBSCRIBE	(DS_RQ_BASE + 1)	/* subscribe to information */
+#define DS_RETRIEVE	(DS_RQ_BASE + 1)	/* retrieve data by name */
-#define DS_RETRIEVE	(DS_RQ_BASE + 2)	/* retrieve information by name */
+#define DS_SUBSCRIBE	(DS_RQ_BASE + 2)	/* subscribe to data updates */
-#define DS_CHECK	(DS_RQ_BASE + 3)	/* retrieve updated information */
+#define DS_CHECK	(DS_RQ_BASE + 3)	/* retrieve updated data */
+#define DS_DELETE	(DS_RQ_BASE + 4)	/* delete data */
+#define DS_SNAPSHOT	(DS_RQ_BASE + 5)	/* take a snapshot */
+#define DS_RETRIEVE_LABEL  (DS_RQ_BASE + 6)	/* retrieve label's name */
 
-/* DS field names: DS_SUBSCRIBE, DS_PUBLISH, DS_RETRIEVE */
+/* DS field names */
-#  define DS_KEY_GRANT		m2_p1		/* key for the information */
+#  define DS_KEY_GRANT		m2_i1		/* key for the data */
-#  define DS_KEY_LEN		m2_i1		/* length of key incl. '\0' */
+#  define DS_KEY_LEN		m2_s1		/* length of key incl. '\0' */
 #  define DS_FLAGS		m2_i2		/* flags provided by caller */
 
-/* DS_PUBLISH, DS_RETRIEVE */
 #  define DS_VAL		m2_l1		/* data (u32, char *, etc.) */
 #  define DS_VAL_LEN		m2_l2		/* data length */
+#  define DS_NR_SNAPSHOT	m2_i3		/* number of snapshot */
+#  define DS_STRING		m2_i3		/* inline string */
 
 /*===========================================================================*
  *                Miscellaneous messages used by TTY			     *

include/minix/const.h

@@ -79,6 +79,10 @@
 #define CLICK_SHIFT	  12	/* log2 of CLICK_SIZE */
 #endif
 
+/* Click alignment macros. */
+#define CLICK_FLOOR(n)  (((vir_bytes)(n) / CLICK_SIZE) * CLICK_SIZE)
+#define CLICK_CEIL(n)   CLICK_FLOOR((vir_bytes)(n) + CLICK_SIZE-1)
+
 /* Sizes of memory tables. The boot monitor distinguishes three memory areas,
  * namely low mem below 1M, 1M-16M, and mem after 16M. More chunks are needed
  * for DOS MINIX.

include/minix/ds.h

@@ -5,34 +5,68 @@
 
 #include <minix/types.h>
 
-/* DS Flag values. */
+/* Flags. */
-#define DS_IN_USE       0x0001	/* Internal use only. */
+#define DSF_IN_USE		0x001	/* entry is in use */
-#define DS_PUBLIC       0x0002	/* Publically retrievable value. */
+#define DSF_PRIV_RETRIEVE	0x002	/* only owner can retrieve */
-#define DS_INITIAL      0x0004	/* On subscription, send initial contents. */
+#define DSF_PRIV_OVERWRITE	0x004	/* only owner can overwrite */
+#define DSF_PRIV_SNAPSHOT	0x004	/* only owner can take a snapshot */
+#define DSF_PRIV_SUBSCRIBE	0x008	/* only owner can subscribe */
+#define DSF_TYPE_U32		0x010	/* u32 data type */
+#define DSF_TYPE_STR		0x020	/* string data type */
+#define DSF_TYPE_MEM		0x040	/* memory range data type */
+#define DSF_TYPE_MAP		0x080	/* mapped memory range data type */
+#define DSF_TYPE_LABEL		0x100	/* label data type */
 
-/* These type flags are mutually exclusive. Give as args to ds_subscribe. */
+#define DSF_MASK_TYPE		0xFF0	/* mask for type flags. */
-#define DS_TYPE_U32     0x0100
+#define DSF_MASK_INTERNAL	0xFFF	/* mask for internal flags. */
-#define DS_TYPE_STR     0x0200
+
-#define DS_TYPE_MASK    0xff00	/* All type info. */
+#define DSF_OVERWRITE		0x01000	/* overwrite if entry exists */
+#define DSF_INITIAL		0x02000	/* check subscriptions immediately */
+
+#define DSMF_MAP_MAPPED		0x10000	/* map mapped memory range */
+#define DSMF_COPY_MAPPED	0x20000	/* copy mapped memory range */
+#define DSMF_COPY_SNAPSHOT	0x40000	/* copy snapshot */
 
 /* DS constants. */
-#define DS_MAX_KEYLEN 80        /* Max length for a key, including '\0'. */
+#define DS_MAX_KEYLEN 80        /* Max length of a key, including '\0'. */
-#define DS_MAX_VALLEN 100	/* Max legnth for a str value, incl '\0'. */ 
+#define DS_MAX_STRLEN 16        /* Max length of string, including '\0'. */
 
 /* ds.c */
-_PROTOTYPE( int ds_subscribe, (char *name_regexp, int type, int flags));
 
-/* publish/update item */
+/* U32 */
-_PROTOTYPE( int ds_publish_u32, (char *name, u32_t val));
+_PROTOTYPE( int ds_publish_u32, (const char *name, u32_t val, int flags));
-_PROTOTYPE( int ds_publish_str, (char *name, char *val));
+_PROTOTYPE( int ds_retrieve_u32, (const char *name, u32_t *val));
+_PROTOTYPE( int ds_delete_u32, (const char *ds_name));
 
-/* retrieve item by name + type */
+/* STRING */
-_PROTOTYPE( int ds_retrieve_u32, (char *name, u32_t *val)          );
+_PROTOTYPE( int ds_publish_str, (const char *name, char *val, int flags));
-_PROTOTYPE( int ds_retrieve_str, (char *name, char *val, size_t len));
+_PROTOTYPE( int ds_retrieve_str, (const char *name, char *val, size_t len));
+_PROTOTYPE( int ds_delete_str, (const char *ds_name));
 
-/* retrieve updates for item */
+/* MEM */
-_PROTOTYPE( int ds_check_u32, (char *n, size_t namelen, u32_t *val));
+_PROTOTYPE( int ds_publish_mem, (const char *ds_name, void *vaddr,
-_PROTOTYPE( int ds_check_str, (char *n, size_t namelen, char *v, size_t vlen));
+		size_t length, int flags));
+_PROTOTYPE( int ds_retrieve_mem, (const char *ds_name, char *vaddr,
+		size_t *length));
+_PROTOTYPE( int ds_delete_mem, (const char *ds_name));
+
+/* MAP */
+_PROTOTYPE( int ds_publish_map, (const char *ds_name, void *vaddr,
+		size_t length, int flags));
+_PROTOTYPE( int ds_snapshot_map, (const char *ds_name, int *nr_snapshot));
+_PROTOTYPE( int ds_retrieve_map, (const char *ds_name, char *vaddr,
+		size_t *length, int nr_snapshot, int flags));
+_PROTOTYPE( int ds_delete_map, (const char *ds_name));
+
+/* LABEL */
+_PROTOTYPE( int ds_publish_label, (const char *ds_name, u32_t value,int flags));
+_PROTOTYPE( int ds_retrieve_label_name, (char *ds_name, u32_t num));
+_PROTOTYPE( int ds_retrieve_label_num, (const char *ds_name, u32_t *value));
+_PROTOTYPE( int ds_delete_label, (const char *ds_name));
+
+/* Subscribe and check. */
+_PROTOTYPE( int ds_subscribe, (const char *regex, int flags));
+_PROTOTYPE( int ds_check, (char *ds_name, int *type));
 
 #endif /* _MINIX_DS_H */
 

include/minix/safecopies.h

@@ -49,6 +49,13 @@ struct vscp_vec {
         size_t          v_bytes;        /* no. of bytes */
 };
 
+/* Types on VM invocation. */
+#define VMPTYPE_NONE		0
+#define VMPTYPE_CHECK		1
+#define VMPTYPE_COWMAP		2
+#define VMPTYPE_SMAP		3
+#define VMPTYPE_SUNMAP		4
+
 /* Invalid grant number. */
 #define GRANT_INVALID	-1
 #define GRANT_VALID(g)	((g) > GRANT_INVALID)
@@ -56,6 +63,7 @@ struct vscp_vec {
 /* Operations: any combination is ok. */
 #define CPF_READ	0x000001 /* Granted process may read. */
 #define CPF_WRITE	0x000002 /* Granted process may write. */
+#define CPF_MAP		0x000004 /* Granted process may map. */
 
 /* Internal flags. */
 #define CPF_USED	0x000100 /* Grant slot in use. */

include/minix/syslib.h

@@ -62,7 +62,8 @@ _PROTOTYPE( int sys_vmctl, (endpoint_t who, int param, u32_t value));
 _PROTOTYPE( int sys_vmctl_get_pagefault_i386, (endpoint_t *who, u32_t *cr2, u32_t *err));
 _PROTOTYPE( int sys_vmctl_get_cr3_i386, (endpoint_t who, u32_t *cr3)  );
 _PROTOTYPE( int sys_vmctl_get_memreq, (endpoint_t *who, vir_bytes *mem,
-        vir_bytes *len, int *wrflag, endpoint_t *) );
+        vir_bytes *len, int *wrflag, endpoint_t *who_s, vir_bytes *mem_s,
+        endpoint_t *) );
 _PROTOTYPE( int sys_vmctl_enable_paging, (struct mem_map *));
 
 _PROTOTYPE( int sys_readbios, (phys_bytes address, void *buf, size_t size));
@@ -149,6 +150,14 @@ _PROTOTYPE(int sys_vsafecopy, (struct vscp_vec *copyvec, int elements));
 _PROTOTYPE(int sys_memset, (unsigned long pattern, 
 		phys_bytes base, phys_bytes bytes));
 
+/* Grant-based map functions. */
+_PROTOTYPE(int sys_safemap, (endpoint_t grantor, cp_grant_id_t grant,
+	vir_bytes grant_offset, vir_bytes my_address, size_t bytes, int my_seg,
+	int writable));
+_PROTOTYPE(int sys_saferevmap_gid, (cp_grant_id_t grant));
+_PROTOTYPE(int sys_saferevmap_addr, (vir_bytes addr));
+_PROTOTYPE(int sys_safeunmap, (int my_seg, vir_bytes my_address));
+
 _PROTOTYPE(int sys_umap, (endpoint_t proc_ep, int seg, vir_bytes vir_addr,
 	 vir_bytes bytes, phys_bytes *phys_addr));
 _PROTOTYPE(int sys_umap_data_fb, (endpoint_t proc_ep, vir_bytes vir_addr,

include/minix/sysutil.h

@@ -49,13 +49,15 @@ _PROTOTYPE( void panic, (char *who, char *mess, int num));
 _PROTOTYPE( int getuptime, (clock_t *ticks));
 _PROTOTYPE( int getuptime2, (clock_t *ticks, time_t *boottime));
 _PROTOTYPE( int tickdelay, (clock_t ticks));
-_PROTOTYPE( int micro_delay_calibrate, (void));
+_PROTOTYPE( int tsc_calibrate, (void));
 _PROTOTYPE( u32_t sys_hz, (void));
 _PROTOTYPE( double getidle, (void));
 _PROTOTYPE( void util_stacktrace, (void));
 _PROTOTYPE( void util_nstrcat, (char *str, unsigned long n) );
 _PROTOTYPE( void util_stacktrace_strcat, (char *));
 _PROTOTYPE( int micro_delay, (u32_t micros));
+_PROTOTYPE( u32_t tsc_64_to_micros, (u64_t tsc));
+_PROTOTYPE( u32_t tsc_to_micros, (u32_t low, u32_t high));
 _PROTOTYPE( u32_t micros_to_ticks, (u32_t micros));
 _PROTOTYPE( void ser_putc, (char c));
 _PROTOTYPE( void get_randomness, (struct k_randomness *, int));

include/unistd.h

@@ -44,12 +44,13 @@
 #define SI_PROC_TAB	   2	/* copy of entire process table */
 #define SI_DMAP_TAB	   3	/* get device <-> driver mappings */
 #define SI_MEM_ALLOC	   4	/* get memory allocation data */
-#define SI_DATA_STORE	   5	/* get copy of data store */
+#define SI_DATA_STORE	   5	/* get copy of data store mappings */
-#define SI_LOADINFO	   6	/* get copy of load average structure */
+#define SI_SUBSCRIPTION	   6	/* get copy of data store subscriptions */
-#define SI_KPROC_TAB	   7	/* copy of kernel process table */
+#define SI_LOADINFO	   7	/* get copy of load average structure */
-#define SI_CALL_STATS	   8	/* system call statistics */
+#define SI_KPROC_TAB	   8	/* copy of kernel process table */
-#define SI_PCI_INFO	   9	/* get kernel info via PM */
+#define SI_CALL_STATS	   9	/* system call statistics */
-#define SI_PROCPUB_TAB	  10	/* copy of public entries of process table */
+#define SI_PCI_INFO	   10	/* get kernel info via PM */
+#define SI_PROCPUB_TAB	   11	/* copy of public entries of process table */
 
 /* NULL must be defined in <unistd.h> according to POSIX Sec. 2.7.1. */
 #define NULL    ((void *)0)

kernel/arch/i386/memory.c

@@ -613,10 +613,11 @@ PUBLIC int vm_suspend(struct proc *caller, struct proc *target,
 	util_stacktrace_strcat(caller->p_vmrequest.stacktrace);
 #endif
 
-	caller->p_vmrequest.writeflag = 1;
+	caller->p_vmrequest.req_type = VMPTYPE_CHECK;
-	caller->p_vmrequest.start = linaddr;
+	caller->p_vmrequest.target = target->p_endpoint;
-	caller->p_vmrequest.length = len;
+	caller->p_vmrequest.params.check.start = linaddr;
-	caller->p_vmrequest.who = target->p_endpoint;
+	caller->p_vmrequest.params.check.length = len;
+	caller->p_vmrequest.params.check.writeflag = 1;
 	caller->p_vmrequest.type = type;
 							
 	/* Connect caller on vmrequest wait queue. */	

kernel/arch/i386/protect.c

@@ -137,12 +137,7 @@ PUBLIC void prot_init(void)
   /* Click-round kernel. */
   if(kinfo.data_base % CLICK_SIZE)
 	minix_panic("kinfo.data_base not aligned", NO_NUM);
-  kinfo.data_size = ((kinfo.data_size+CLICK_SIZE-1)/CLICK_SIZE) * CLICK_SIZE;
+  kinfo.data_size = (phys_bytes) (CLICK_CEIL(kinfo.data_size));
-
-  /* Click-round kernel. */
-  if(kinfo.data_base % CLICK_SIZE)
-	minix_panic("kinfo.data_base not aligned", NO_NUM);
-  kinfo.data_size = ((kinfo.data_size+CLICK_SIZE-1)/CLICK_SIZE) * CLICK_SIZE;
 
   /* Build gdt and idt pointers in GDT where the BIOS expects them. */
   dtp= (struct desctableptr_s *) &gdt[GDT_INDEX];

kernel/main.c

@@ -154,13 +154,14 @@ PUBLIC void main()
 
 	/* Convert addresses to clicks and build process memory map */
 	text_base = e_hdr.a_syms >> CLICK_SHIFT;
-	text_clicks = (e_hdr.a_text + CLICK_SIZE-1) >> CLICK_SHIFT;
+	text_clicks = (vir_clicks) (CLICK_CEIL(e_hdr.a_text) >> CLICK_SHIFT);
-	data_clicks = (e_hdr.a_data+e_hdr.a_bss + CLICK_SIZE-1) >> CLICK_SHIFT;
+	data_clicks = (vir_clicks) (CLICK_CEIL(e_hdr.a_data
-	st_clicks= (e_hdr.a_total + CLICK_SIZE-1) >> CLICK_SHIFT;
+		+ e_hdr.a_bss) >> CLICK_SHIFT);
+	st_clicks = (vir_clicks) (CLICK_CEIL(e_hdr.a_total) >> CLICK_SHIFT);
 	if (!(e_hdr.a_flags & A_SEP))
 	{
-		data_clicks= (e_hdr.a_text+e_hdr.a_data+e_hdr.a_bss +
+		data_clicks = (vir_clicks) (CLICK_CEIL(e_hdr.a_text +
-			CLICK_SIZE-1) >> CLICK_SHIFT;
+			e_hdr.a_data + e_hdr.a_bss) >> CLICK_SHIFT);
 		text_clicks = 0;	   /* common I&D */
 	}
 	rp->p_memmap[T].mem_phys = text_base;

kernel/proc.h

@@ -61,7 +61,7 @@ struct proc {
    * memory that isn't present, VM has to fix it. Until it has asked
    * what needs to be done and fixed it, save necessary state here.
    *
-   * The requester gets a copy of its request message in reqmsg and gets
+   * The requestor gets a copy of its request message in reqmsg and gets
    * VMREQUEST set.
    */
   struct {
@@ -70,6 +70,8 @@ struct proc {
 #define VMSTYPE_SYS_NONE	0
 #define VMSTYPE_KERNELCALL	1
 #define VMSTYPE_DELIVERMSG	2
+#define VMSTYPE_MAP		3
+
 	int		type;		/* suspended operation */
 	union {
 		/* VMSTYPE_SYS_MESSAGE */
@@ -77,10 +79,20 @@ struct proc {
 	} saved;
 
 	/* Parameters of request to VM */
+	int		req_type;
+	endpoint_t	target;
+	union {
+		struct {
 			vir_bytes 	start, length;	/* memory range */
 			u8_t		writeflag;	/* nonzero for write access */
-	endpoint_t	who;
+		} check;
-
+		struct {
+			char		writeflag;
+			endpoint_t	ep_s;
+			vir_bytes	vir_s, vir_d;
+			vir_bytes	length;
+		} map;
+	} params;
 	/* VM result when available */
 	int		vmresult;
 

kernel/proto.h

@@ -89,6 +89,12 @@ _PROTOTYPE( void check_runqueues_f, (char *file, int line) );
 _PROTOTYPE( char *rtsflagstr, (int flags) );
 _PROTOTYPE( char *miscflagstr, (int flags) );
 
+/* system/do_safemap.c */
+_PROTOTYPE( int map_invoke_vm, (int req_type,
+		endpoint_t end_d, int seg_d, vir_bytes off_d,
+		endpoint_t end_s, int seg_s, vir_bytes off_s,
+		size_t size, int flag));
+
 /* system/do_safecopy.c */
 _PROTOTYPE( int verify_grant, (endpoint_t, endpoint_t, cp_grant_id_t, vir_bytes,
 	int, vir_bytes, vir_bytes *, endpoint_t *));

kernel/system.c

@@ -208,6 +208,11 @@ PRIVATE void initialize(void)
   map(SYS_SAFECOPYTO, do_safecopy);	/* copy with pre-granted permission */
   map(SYS_VSAFECOPY, do_vsafecopy);	/* vectored safecopy */
 
+  /* Mapping. */
+  map(SYS_SAFEMAP, do_safemap);		/* map pages from other process */
+  map(SYS_SAFEREVMAP, do_saferevmap);	/* grantor revokes the map grant */
+  map(SYS_SAFEUNMAP, do_safeunmap);	/* requestor unmaps the mapped pages */
+
   /* Clock functionality. */
   map(SYS_TIMES, do_times);		/* get uptime and process times */
   map(SYS_SETALARM, do_setalarm);	/* schedule a synchronous alarm */
@@ -263,10 +268,6 @@ int priv_id;				/* privilege id */
   rc->p_priv = sp;			    /* assign new slot */
   rc->p_priv->s_proc_nr = proc_nr(rc);	    /* set association */
 
-  /* Clear some fields */
-  sp->s_asyntab= -1;
-  sp->s_asynsize= 0;
-
   return(OK);
 }
 
@@ -482,9 +483,11 @@ register struct proc *rc;		/* slot of process to clean up */
   if (priv(rc)->s_flags & SYS_PROC)
   {
 	if (priv(rc)->s_asynsize) {
+#if 0
 		kprintf("clear_endpoint: clearing s_asynsize of %s / %d\n",
 			rc->p_name, rc->p_endpoint);
 		proc_stacktrace(rc);
+#endif
 	}
 	priv(rc)->s_asynsize= 0;
   }

kernel/system.h

@@ -181,6 +181,10 @@ _PROTOTYPE( int do_setgrant, (message *m_ptr) );
 _PROTOTYPE( int do_readbios, (message *m_ptr) );	
 _PROTOTYPE( int do_mapdma, (message *m_ptr) );	
 
+_PROTOTYPE( int do_safemap, (message *m_ptr) );	
+_PROTOTYPE( int do_saferevmap, (message *m_ptr) );	
+_PROTOTYPE( int do_safeunmap, (message *m_ptr) );	
+
 _PROTOTYPE( int do_sprofile, (message *m_ptr) );
 #if ! SPROFILE
 #define do_sprofile do_unused

kernel/system/Makefile

@@ -44,6 +44,7 @@ OBJECTS	= \
 	$(SYSTEM)(do_privctl.o) \
 	$(SYSTEM)(do_segctl.o) \
 	$(SYSTEM)(do_safecopy.o) \
+	$(SYSTEM)(do_safemap.o) \
 	$(SYSTEM)(do_sysctl.o) \
 	$(SYSTEM)(do_getksig.o) \
 	$(SYSTEM)(do_endksig.o) \
@@ -155,6 +156,9 @@ $(SYSTEM)(do_privctl.o):	do_privctl.c
 $(SYSTEM)(do_safecopy.o):	do_safecopy.c
 	$(CC) do_safecopy.c
 
+$(SYSTEM)(do_safemap.o):	do_safemap.c
+	$(CC) do_safemap.c
+
 $(SYSTEM)(do_sysctl.o):	do_sysctl.c
 	$(CC) do_sysctl.c
 

kernel/system/do_fork.c

@@ -49,12 +49,6 @@ register message *m_ptr;	/* pointer to request message */
 	return EINVAL;
   }
 
-  /* memory becomes readonly */
-  if (priv(rpp)->s_asynsize > 0) {
-	printf("kernel: process with waiting asynsend table can't fork\n");
-	return EINVAL;
-  }
-
   map_ptr= (struct mem_map *) m_ptr->PR_MEM_PTR;
 
   /* Copy parent 'proc' struct to child. And reinitialize some fields. */

kernel/system/do_privctl.c

@@ -101,6 +101,8 @@ message *m_ptr;			/* pointer to request message */
 	      priv(rp)->s_notify_pending.chunk[i] = 0;	/* - notifications */
 	priv(rp)->s_int_pending = 0;			/* - interrupts */
 	sigemptyset(&priv(rp)->s_sig_pending);		/* - signals */
+	priv(rp)->s_asyntab= -1;			/* - asynsends */
+	priv(rp)->s_asynsize= 0;
 
 	/* Set defaults for privilege bitmaps. */
 	priv(rp)->s_flags= DEF_SYS_F;           /* privilege flags */

kernel/system/do_safecopy.c

@@ -24,6 +24,8 @@
 
 #define MEM_TOP 0xFFFFFFFFUL
 
+#define USE_COW_SAFECOPY 0
+
 FORWARD _PROTOTYPE(int safecopy, (endpoint_t, endpoint_t, cp_grant_id_t, int, int, size_t, vir_bytes, vir_bytes, int));
 
 #define HASGRANTTABLE(gr) \
@@ -229,9 +231,10 @@ int access;			/* CPF_READ for a copy from granter to grantee, CPF_WRITE
 {
 	static struct vir_addr v_src, v_dst;
 	static vir_bytes v_offset;
-	int r;
 	endpoint_t new_granter, *src, *dst;
 	struct proc *granter_p;
+	vir_bytes size;
+	int r;
 
 	/* See if there is a reasonable grant table. */
 	if(!(granter_p = endpoint_lookup(granter))) return EINVAL;
@@ -279,8 +282,48 @@ int access;			/* CPF_READ for a copy from granter to grantee, CPF_WRITE
 	}
 
 	/* Do the regular copy. */
+#if USE_COW_SAFECOPY
+	if(v_offset % CLICK_SIZE != addr % CLICK_SIZE || bytes < CLICK_SIZE) {
+		/* Give up on COW immediately when offsets are not aligned
+		 * or we are copying less than a page.
+		 */
 		return virtual_copy_vmcheck(&v_src, &v_dst, bytes);
+	}
 
+	if((size = v_offset % CLICK_SIZE) != 0) {
+		/* Normal copy for everything before the first page boundary. */
+		size = CLICK_SIZE - size;
+		r = virtual_copy_vmcheck(&v_src, &v_dst, size);
+		if(r != OK)
+			return r;
+		v_src.offset += size;
+		v_dst.offset += size;
+		bytes -= size;
+	}
+	if((size = bytes / CLICK_SIZE) != 0) {
+		/* Use COW optimization when copying entire pages. */
+		size *= CLICK_SIZE;
+		r = map_invoke_vm(VMPTYPE_COWMAP,
+			v_dst.proc_nr_e, v_dst.segment, v_dst.offset,
+			v_src.proc_nr_e, v_src.segment, v_src.offset,
+			size, 0);
+		if(r != OK)
+			return r;
+		v_src.offset += size;
+		v_dst.offset += size;
+		bytes -= size;
+	}
+	if(bytes != 0) {
+		/* Normal copy for everything after the last page boundary. */
+		r = virtual_copy_vmcheck(&v_src, &v_dst, bytes);
+		if(r != OK)
+			return r;
+	}
+
+	return OK;
+#else
+	return virtual_copy_vmcheck(&v_src, &v_dst, bytes);
+#endif
 }
 
 /*===========================================================================*

kernel/system/do_safemap.c

@@ -0,0 +1,285 @@
+/* The kernel call implemented in this file:
+ *   m_type:	SYS_SAFEMAP or SYS_SAFEREVMAP or SYS_SAFEUNMAP
+ *
+ * The parameters for this kernel call are:
+ *	SMAP_EP		endpoint of the grantor
+ *	SMAP_GID	grant id
+ *	SMAP_OFFSET	offset of the grant space
+ *	SMAP_SEG	segment
+ *	SMAP_ADDRESS	address
+ *    	SMAP_BYTES	bytes to be copied
+ *    	SMAP_FLAG	access, writable map or not?
+ */
+
+#include <minix/type.h>
+#include <minix/safecopies.h>
+
+#include "../system.h"
+#include "../vm.h"
+
+
+struct map_info_s {
+	int flag;
+
+	/* Grantor. */
+	endpoint_t grantor;
+	int gid;
+	vir_bytes offset;
+	vir_bytes address_Dseg; /* seg always is D */
+
+	/* Grantee. */
+	endpoint_t grantee;
+	int seg;
+	vir_bytes address;
+
+	/* Length. */
+	vir_bytes bytes;
+};
+
+#define MAX_MAP_INFO 20
+static struct map_info_s map_info[MAX_MAP_INFO];
+
+/*===========================================================================*
+ *				add_info				     *
+ *===========================================================================*/
+static int add_info(endpoint_t grantor, endpoint_t grantee, int gid,
+		vir_bytes offset, vir_bytes address_Dseg,
+		int seg, vir_bytes address, vir_bytes bytes)
+{
+	int i;
+
+	for(i = 0; i < MAX_MAP_INFO; i++) {
+		if(map_info[i].flag == 0)
+			break;
+	}
+	if(i == MAX_MAP_INFO)
+		return EBUSY;
+
+	map_info[i].flag = 1;
+	map_info[i].grantor = grantor;
+	map_info[i].grantee = grantee;
+	map_info[i].gid = gid;
+	map_info[i].address_Dseg = address_Dseg;
+	map_info[i].offset = offset;
+	map_info[i].seg = seg;
+	map_info[i].address = address;
+	map_info[i].bytes = bytes;
+
+	return OK;
+}
+
+/*===========================================================================*
+ *				get_revoke_info				     *
+ *===========================================================================*/
+static struct map_info_s *get_revoke_info(endpoint_t grantor, int flag, int arg)
+{
+	int i;
+	for(i = 0; i < MAX_MAP_INFO; i++) {
+		if(map_info[i].flag == 1
+			&& map_info[i].grantor == grantor
+			&& (flag ? (map_info[i].gid == arg)
+				 : (map_info[i].address_Dseg == arg)))
+			return &map_info[i];
+	}
+
+	return NULL;
+}
+
+/*===========================================================================*
+ *				get_unmap_info				     *
+ *===========================================================================*/
+static struct map_info_s *get_unmap_info(endpoint_t grantee, int seg,
+	vir_bytes address)
+{
+	int i;
+	for(i = 0; i < MAX_MAP_INFO; i++) {
+		if(map_info[i].flag == 1
+			&& map_info[i].grantee == grantee
+			&& map_info[i].seg == seg
+			&& map_info[i].address == address)
+			return &map_info[i];
+	}
+
+	return NULL;
+}
+
+/*===========================================================================*
+ *				clear_info				     *
+ *===========================================================================*/
+static int clear_info(struct map_info_s *p)
+{
+	p->flag = 0;
+
+	return 0;
+}
+
+/*===========================================================================*
+ *				map_invoke_vm				     *
+ *===========================================================================*/
+PUBLIC int map_invoke_vm(int req_type, /* VMPTYPE_... COWMAP, SMAP, SUNMAP */
+		endpoint_t end_d, int seg_d, vir_bytes off_d,
+		endpoint_t end_s, int seg_s, vir_bytes off_s,
+		size_t size, int flag)
+{
+	struct proc *caller, *src, *dst;
+	vir_bytes lin_src, lin_dst;
+
+	src = endpoint_lookup(end_s);
+	dst = endpoint_lookup(end_d);
+	caller = endpoint_lookup(who_e);
+
+	lin_src = umap_local(src, seg_s, off_s, size);
+	lin_dst = umap_local(dst, seg_d, off_d, size);
+	if(lin_src == 0 || lin_dst == 0) {
+		kprintf("map_invoke_vm: error in umap_local.\n");
+		return EINVAL;
+	}
+
+	/* Make sure the linear addresses are both page aligned. */
+	if(lin_src % CLICK_SIZE != 0
+		|| lin_dst % CLICK_SIZE != 0) {
+		kprintf("map_invoke_vm: linear addresses not page aligned.\n");
+		return EINVAL;
+	}
+
+	vmassert(!RTS_ISSET(caller, RTS_VMREQUEST));
+	vmassert(!RTS_ISSET(caller, RTS_VMREQTARGET));
+	vmassert(!RTS_ISSET(dst, RTS_VMREQUEST));
+	vmassert(!RTS_ISSET(dst, RTS_VMREQTARGET));
+	RTS_LOCK_SET(caller, RTS_VMREQUEST);
+	RTS_LOCK_SET(dst, RTS_VMREQTARGET);
+
+	/* Map to the destination. */
+	caller->p_vmrequest.req_type = req_type;
+	caller->p_vmrequest.target = end_d;		/* destination proc */
+	caller->p_vmrequest.params.map.vir_d = lin_dst;	/* destination addr */
+	caller->p_vmrequest.params.map.ep_s = end_s;	/* source process */
+	caller->p_vmrequest.params.map.vir_s = lin_src;	/* source address */
+	caller->p_vmrequest.params.map.length = size;
+	caller->p_vmrequest.params.map.writeflag = flag;
+
+	caller->p_vmrequest.type = VMSTYPE_MAP;
+
+	/* Connect caller on vmrequest wait queue. */
+	if(!(caller->p_vmrequest.nextrequestor = vmrequest))
+		lock_notify(SYSTEM, VM_PROC_NR);
+	vmrequest = caller;
+
+	return OK;
+}
+
+/*===========================================================================*
+ *				do_safemap				     *
+ *===========================================================================*/
+PUBLIC int do_safemap(m_ptr)
+register message *m_ptr;
+{
+	endpoint_t grantor	= m_ptr->SMAP_EP;
+	cp_grant_id_t gid	= m_ptr->SMAP_GID;
+	vir_bytes offset	= m_ptr->SMAP_OFFSET;
+	int seg			= (int)m_ptr->SMAP_SEG;
+	vir_bytes address	= m_ptr->SMAP_ADDRESS;
+	vir_bytes bytes		= m_ptr->SMAP_BYTES;
+	int flag		= m_ptr->SMAP_FLAG;
+
+	vir_bytes offset_result;
+	endpoint_t new_grantor;
+	int r;
+	int access = CPF_MAP | CPF_READ;
+
+	/* Check the grant. We currently support safemap with both direct and
+	 * indirect grants, as verify_grant() stores the original grantor
+	 * transparently in new_grantor below. However, we maintain the original
+	 * semantics associated to indirect grants only here at safemap time.
+	 * After the mapping has been set up, if a process part of the chain
+	 * of trust crashes or exits without revoking the mapping, the mapping
+	 * can no longer be manually or automatically revoked for any of the
+	 * processes lower in the chain. This solution reduces complexity but
+	 * could be improved if we make the assumption that only one process in
+	 * the chain of trust can effectively map the original memory region.
+	 */
+	if(flag != 0)
+		access |= CPF_WRITE;
+	r = verify_grant(grantor, who_e, gid, bytes, access,
+		offset, &offset_result, &new_grantor);
+	if(r != OK) {
+		kprintf("verify_grant for gid %d from %d to %d failed: %d\n",
+			gid, grantor, who_e, r);
+		return r;
+	}
+
+	/* Add map info. */
+	r = add_info(new_grantor, who_e, gid, offset, offset_result, seg,
+		address, bytes);
+	if(r != OK)
+		return r;
+
+	/* Invoke VM. */
+	return map_invoke_vm(VMPTYPE_SMAP,
+		who_e, seg, address, new_grantor, D, offset_result, bytes,flag);
+}
+
+/*===========================================================================*
+ *				safeunmap				     *
+ *===========================================================================*/
+PRIVATE int safeunmap(struct map_info_s *p)
+{
+	vir_bytes offset_result;
+	endpoint_t new_grantor;
+	int r;
+
+	r = verify_grant(p->grantor, p->grantee, p->gid, p->bytes, CPF_MAP,
+		p->offset, &offset_result, &new_grantor);
+	if(r != OK) {
+	    kprintf("safeunmap: error in verify_grant.\n");
+		return r;
+	}
+
+	r = map_invoke_vm(VMPTYPE_SUNMAP,
+		p->grantee, p->seg, p->address,
+		new_grantor, D, offset_result,
+		p->bytes, 0);
+	clear_info(p);
+	if(r != OK) {
+		kprintf("safeunmap: error in map_invoke_vm.\n");
+		return r;
+	}
+	return OK;
+}
+
+/*===========================================================================*
+ *				do_saferevmap				     *
+ *===========================================================================*/
+PUBLIC int do_saferevmap(m_ptr)
+register message *m_ptr;
+{
+	struct map_info_s *p;
+	int flag = m_ptr->SMAP_FLAG;
+	int arg = m_ptr->SMAP_GID; /* gid or address_Dseg */
+	int r;
+
+	while((p = get_revoke_info(who_e, flag, arg)) != NULL) {
+		if((r = safeunmap(p)) != OK)
+			return r;
+	}
+	return OK;
+}
+
+/*===========================================================================*
+ *				do_safeunmap				     *
+ *===========================================================================*/
+PUBLIC int do_safeunmap(m_ptr)
+register message *m_ptr;
+{
+	vir_bytes address = m_ptr->SMAP_ADDRESS;
+	int seg = (int)m_ptr->SMAP_SEG;
+	struct map_info_s *p;
+	int r;
+
+	while((p = get_unmap_info(who_e, seg, address)) != NULL) {
+		if((r = safeunmap(p)) != OK)
+			return r;
+	}
+	return OK;
+}
+

kernel/system/do_vmctl.c

@@ -65,21 +65,51 @@ register message *m_ptr;	/* pointer to request message */
 #endif
 
 		/* Reply with request fields. */
-		m_ptr->SVMCTL_MRG_ADDR = (char *) rp->p_vmrequest.start;
+		switch(rp->p_vmrequest.req_type) {
-		m_ptr->SVMCTL_MRG_LEN = rp->p_vmrequest.length;
+		case VMPTYPE_CHECK:
-		m_ptr->SVMCTL_MRG_WRITE = rp->p_vmrequest.writeflag;
+			m_ptr->SVMCTL_MRG_TARGET	=
-		m_ptr->SVMCTL_MRG_EP = rp->p_vmrequest.who;
+				rp->p_vmrequest.target;
-		m_ptr->SVMCTL_MRG_REQUESTOR = (void *) rp->p_endpoint;
+			m_ptr->SVMCTL_MRG_ADDR		=
+				rp->p_vmrequest.params.check.start;
+			m_ptr->SVMCTL_MRG_LENGTH	=
+				rp->p_vmrequest.params.check.length;
+			m_ptr->SVMCTL_MRG_FLAG		=
+				rp->p_vmrequest.params.check.writeflag;
+			m_ptr->SVMCTL_MRG_REQUESTOR	=
+				(void *) rp->p_endpoint;
+			break;
+		case VMPTYPE_COWMAP:
+		case VMPTYPE_SMAP:
+		case VMPTYPE_SUNMAP:
+			m_ptr->SVMCTL_MRG_TARGET	=
+				rp->p_vmrequest.target;
+			m_ptr->SVMCTL_MRG_ADDR		=
+				rp->p_vmrequest.params.map.vir_d;
+			m_ptr->SVMCTL_MRG_EP2		=
+				rp->p_vmrequest.params.map.ep_s;
+			m_ptr->SVMCTL_MRG_ADDR2		=
+				rp->p_vmrequest.params.map.vir_s;
+			m_ptr->SVMCTL_MRG_LENGTH	=
+				rp->p_vmrequest.params.map.length;
+			m_ptr->SVMCTL_MRG_FLAG		=
+				rp->p_vmrequest.params.map.writeflag;
+			m_ptr->SVMCTL_MRG_REQUESTOR	=
+				(void *) rp->p_endpoint;
+			break;
+		default:
+			minix_panic("VMREQUEST wrong type", NO_NUM);
+		}
+
 		rp->p_vmrequest.vmresult = VMSUSPEND;
 
 		/* Remove from request chain. */
 		vmrequest = vmrequest->p_vmrequest.nextrequestor;
 
-		return OK;
+		return rp->p_vmrequest.req_type;
 	case VMCTL_MEMREQ_REPLY:
 		vmassert(RTS_ISSET(p, RTS_VMREQUEST));
 		vmassert(p->p_vmrequest.vmresult == VMSUSPEND);
-  		okendpt(p->p_vmrequest.who, &proc_nr);
+  		okendpt(p->p_vmrequest.target, &proc_nr);
 		target = proc_addr(proc_nr);
 		p->p_vmrequest.vmresult = m_ptr->SVMCTL_VALUE;
 		vmassert(p->p_vmrequest.vmresult != VMSUSPEND);
@@ -94,21 +124,28 @@ register message *m_ptr;	/* pointer to request message */
 			p->p_vmrequest.start + p->p_vmrequest.length,
 			p->p_vmrequest.writeflag, p->p_vmrequest.stacktrace);
 		printf("type %d\n", p->p_vmrequest.type);
+#endif
 
 		vmassert(RTS_ISSET(target, RTS_VMREQTARGET));
 		RTS_LOCK_UNSET(target, RTS_VMREQTARGET);
-#endif
 
-		if(p->p_vmrequest.type == VMSTYPE_KERNELCALL) {
+		switch(p->p_vmrequest.type) {
+		case VMSTYPE_KERNELCALL:
 			/* Put on restart chain. */
 			p->p_vmrequest.nextrestart = vmrestart;
 			vmrestart = p;
-		} else if(p->p_vmrequest.type == VMSTYPE_DELIVERMSG) {
+			break;
+		case VMSTYPE_DELIVERMSG:
 			vmassert(p->p_misc_flags & MF_DELIVERMSG);
 			vmassert(p == target);
 			vmassert(RTS_ISSET(p, RTS_VMREQUEST));
 			RTS_LOCK_UNSET(p, RTS_VMREQUEST);
-		} else {
+			break;
+		case VMSTYPE_MAP:
+			vmassert(RTS_ISSET(p, RTS_VMREQUEST));
+			RTS_LOCK_UNSET(p, RTS_VMREQUEST);
+			break;
+		default:
 #if DEBUG_VMASSERT
 			printf("suspended with stack: %s\n",
 				p->p_vmrequest.stacktrace);

lib/Makefile

@@ -6,10 +6,10 @@ all-gnu:
 
 makefiles: Makefile
 Makedepend-ack Makedepend-gnu: 
-	sh generate.sh . obj-ack/ obj-gnu
+	sh ./generate.sh . obj-ack obj-gnu
 
 Makefile: Makefile.in Makedepend-ack Makedepend-gnu
-	sh generate.sh . obj-ack/ obj-gnu
+	sh ./generate.sh . obj-ack obj-gnu
 	@echo
 	@echo *Attention*
 	@echo Makefile is regenerated... rerun command to see changes
@@ -17,47 +17,47 @@ Makefile: Makefile.in Makedepend-ack Makedepend-gnu
 	@echo
 
 all-ack: makefiles
-	mkdir -p obj-ack//./ansi
+	mkdir -p obj-ack/./ansi
 	cd ansi && $(MAKE) $@
-	mkdir -p obj-ack//./curses
+	mkdir -p obj-ack/./curses
 	cd curses && $(MAKE) $@
-	mkdir -p obj-ack//./dummy
+	mkdir -p obj-ack/./dummy
 	cd dummy && $(MAKE) $@
-	mkdir -p obj-ack//./editline
+	mkdir -p obj-ack/./editline
 	cd editline && $(MAKE) $@
-	mkdir -p obj-ack//./end
+	mkdir -p obj-ack/./end
 	cd end && $(MAKE) $@
-	mkdir -p obj-ack//./ip
+	mkdir -p obj-ack/./ip
 	cd ip && $(MAKE) $@
-	mkdir -p obj-ack//./math
+	mkdir -p obj-ack/./math
 	cd math && $(MAKE) $@
-	mkdir -p obj-ack//./other
+	mkdir -p obj-ack/./other
 	cd other && $(MAKE) $@
-	mkdir -p obj-ack//./posix
+	mkdir -p obj-ack/./posix
 	cd posix && $(MAKE) $@
-	mkdir -p obj-ack//./regex
+	mkdir -p obj-ack/./regex
 	cd regex && $(MAKE) $@
-	mkdir -p obj-ack//./stdio
+	mkdir -p obj-ack/./stdio
 	cd stdio && $(MAKE) $@
-	mkdir -p obj-ack//./stdtime
+	mkdir -p obj-ack/./stdtime
 	cd stdtime && $(MAKE) $@
-	mkdir -p obj-ack//./syscall
+	mkdir -p obj-ack/./syscall
 	cd syscall && $(MAKE) $@
-	mkdir -p obj-ack//./syslib
+	mkdir -p obj-ack/./syslib
 	cd syslib && $(MAKE) $@
-	mkdir -p obj-ack//./util
+	mkdir -p obj-ack/./util
 	cd util && $(MAKE) $@
-	mkdir -p obj-ack//./sysutil
+	mkdir -p obj-ack/./sysutil
 	cd sysutil && $(MAKE) $@
-	mkdir -p obj-ack//./sysvipc
+	mkdir -p obj-ack/./sysvipc
 	cd sysvipc && $(MAKE) $@
-	mkdir -p obj-ack//./timers
+	mkdir -p obj-ack/./timers
 	cd timers && $(MAKE) $@
-	mkdir -p obj-ack//./i386
+	mkdir -p obj-ack/./i386
 	cd i386 && $(MAKE) $@
-	mkdir -p obj-ack//./ack
+	mkdir -p obj-ack/./ack
 	cd ack && $(MAKE) $@
-	mkdir -p obj-ack//./gnu
+	mkdir -p obj-ack/./gnu
 	cd gnu && $(MAKE) $@
 
 all-gnu: makefiles
@@ -150,56 +150,56 @@ makefiles: ack/Makefile
 makefiles: gnu/Makefile
 
 ansi/Makefile: ansi/Makefile.in
-	cd ansi && sh ../generate.sh ./ansi ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd ansi && sh .././generate.sh ./ansi ../obj-ack ../obj-gnu && $(MAKE) makefiles
 curses/Makefile: curses/Makefile.in
-	cd curses && sh ../generate.sh ./curses ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd curses && sh .././generate.sh ./curses ../obj-ack ../obj-gnu && $(MAKE) makefiles
 dummy/Makefile: dummy/Makefile.in
-	cd dummy && sh ../generate.sh ./dummy ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd dummy && sh .././generate.sh ./dummy ../obj-ack ../obj-gnu && $(MAKE) makefiles
 editline/Makefile: editline/Makefile.in
-	cd editline && sh ../generate.sh ./editline ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd editline && sh .././generate.sh ./editline ../obj-ack ../obj-gnu && $(MAKE) makefiles
 end/Makefile: end/Makefile.in
-	cd end && sh ../generate.sh ./end ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd end && sh .././generate.sh ./end ../obj-ack ../obj-gnu && $(MAKE) makefiles
 ip/Makefile: ip/Makefile.in
-	cd ip && sh ../generate.sh ./ip ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd ip && sh .././generate.sh ./ip ../obj-ack ../obj-gnu && $(MAKE) makefiles
 math/Makefile: math/Makefile.in
-	cd math && sh ../generate.sh ./math ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd math && sh .././generate.sh ./math ../obj-ack ../obj-gnu && $(MAKE) makefiles
 other/Makefile: other/Makefile.in
-	cd other && sh ../generate.sh ./other ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd other && sh .././generate.sh ./other ../obj-ack ../obj-gnu && $(MAKE) makefiles
 posix/Makefile: posix/Makefile.in
-	cd posix && sh ../generate.sh ./posix ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd posix && sh .././generate.sh ./posix ../obj-ack ../obj-gnu && $(MAKE) makefiles
 regex/Makefile: regex/Makefile.in
-	cd regex && sh ../generate.sh ./regex ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd regex && sh .././generate.sh ./regex ../obj-ack ../obj-gnu && $(MAKE) makefiles
 stdio/Makefile: stdio/Makefile.in
-	cd stdio && sh ../generate.sh ./stdio ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd stdio && sh .././generate.sh ./stdio ../obj-ack ../obj-gnu && $(MAKE) makefiles
 stdtime/Makefile: stdtime/Makefile.in
-	cd stdtime && sh ../generate.sh ./stdtime ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd stdtime && sh .././generate.sh ./stdtime ../obj-ack ../obj-gnu && $(MAKE) makefiles
 syscall/Makefile: syscall/Makefile.in
-	cd syscall && sh ../generate.sh ./syscall ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd syscall && sh .././generate.sh ./syscall ../obj-ack ../obj-gnu && $(MAKE) makefiles
 syslib/Makefile: syslib/Makefile.in
-	cd syslib && sh ../generate.sh ./syslib ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd syslib && sh .././generate.sh ./syslib ../obj-ack ../obj-gnu && $(MAKE) makefiles
 util/Makefile: util/Makefile.in
-	cd util && sh ../generate.sh ./util ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd util && sh .././generate.sh ./util ../obj-ack ../obj-gnu && $(MAKE) makefiles
 sysutil/Makefile: sysutil/Makefile.in
-	cd sysutil && sh ../generate.sh ./sysutil ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd sysutil && sh .././generate.sh ./sysutil ../obj-ack ../obj-gnu && $(MAKE) makefiles
 sysvipc/Makefile: sysvipc/Makefile.in
-	cd sysvipc && sh ../generate.sh ./sysvipc ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd sysvipc && sh .././generate.sh ./sysvipc ../obj-ack ../obj-gnu && $(MAKE) makefiles
 timers/Makefile: timers/Makefile.in
-	cd timers && sh ../generate.sh ./timers ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd timers && sh .././generate.sh ./timers ../obj-ack ../obj-gnu && $(MAKE) makefiles
 i386/Makefile: i386/Makefile.in
-	cd i386 && sh ../generate.sh ./i386 ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd i386 && sh .././generate.sh ./i386 ../obj-ack ../obj-gnu && $(MAKE) makefiles
 ack/Makefile: ack/Makefile.in
-	cd ack && sh ../generate.sh ./ack ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd ack && sh .././generate.sh ./ack ../obj-ack ../obj-gnu && $(MAKE) makefiles
 gnu/Makefile: gnu/Makefile.in
-	cd gnu && sh ../generate.sh ./gnu ../obj-ack/ ../obj-gnu && $(MAKE) makefiles
+	cd gnu && sh .././generate.sh ./gnu ../obj-ack ../obj-gnu && $(MAKE) makefiles
 
 clean::
-	rm -f obj-ack//./*
+	rm -f obj-ack/./*
 	rm -f obj-gnu/./*
 
 install: install-ack
 
 install-ack: all-ack
-	cp obj-ack//*.[ao] /usr/lib/i386
+	cp obj-ack/*.[ao] /usr/lib/i386
 
 install-gnu: all-gnu
 	cp obj-gnu/*.[ao] /usr/gnu/lib

lib/syslib/Makefile.in

@@ -52,6 +52,7 @@ libsys_FILES=" \
 	sys_readbios.c \
 	sys_runctl.c \
 	sys_safecopy.c \
+	sys_safemap.c \
 	sys_sysctl.c \
 	sys_vsafecopy.c \
 	sys_profbuf.c \

lib/syslib/ds.c

@@ -4,260 +4,257 @@
 
 #include "syslib.h"
 
-#define GRANTBAD -1001
+static message m;
 
-int
+PRIVATE int do_invoke_ds(int type, const char *ds_name)
-ds_subscribe(ds_name_regexp, type, flags)
-char *ds_name_regexp;
-int type;
-int flags;
 {
-	int r;
+	cp_grant_id_t g_key;
-	message m;
+	size_t len_key;
-	cp_grant_id_t g;
+	int access, r;
-	size_t len;
 
-	len = strlen(ds_name_regexp)+1;
+	if(type == DS_CHECK || type == DS_RETRIEVE_LABEL) {
-	g = cpf_grant_direct(DS_PROC_NR,
+		len_key = DS_MAX_KEYLEN;
-		(vir_bytes) ds_name_regexp, len, CPF_READ);
+		access = CPF_WRITE;
+	} else {
+		len_key = strlen(ds_name)+1;
+		access = CPF_READ;
+	}
 
-	if(!GRANT_VALID(g)) 
+	/* Grant for key. */
-		return GRANTBAD;
+	g_key = cpf_grant_direct(DS_PROC_NR, (vir_bytes) ds_name,
+		len_key, access);
+	if(!GRANT_VALID(g_key)) 
+		return errno;
 
-	flags &= DS_INITIAL;
+	m.DS_KEY_GRANT = g_key;
+	m.DS_KEY_LEN = len_key;
 
-	m.DS_KEY_GRANT = (char *) g;
+	r = _taskcall(DS_PROC_NR, type, &m);
-	m.DS_KEY_LEN = len;
-	m.DS_FLAGS = flags | (type & DS_TYPE_MASK);
-
-	r = _taskcall(DS_PROC_NR, DS_SUBSCRIBE, &m);
-
-	cpf_revoke(g);
 
+	cpf_revoke(g_key);
 	return r;
 }
 
-int ds_publish_u32(ds_name, value)
+int ds_publish_label(const char *ds_name, u32_t value, int flags)
-char *ds_name;
-u32_t value;
 {
-	int r;
-	message m;
-	cp_grant_id_t g;
-	size_t len;
-
-	len = strlen(ds_name)+1;
-	g = cpf_grant_direct(DS_PROC_NR,
-		(vir_bytes) ds_name, len, CPF_READ);
-
-	if(!GRANT_VALID(g)) 
-		return GRANTBAD;
-
-	m.DS_KEY_GRANT = (char *) g;
-	m.DS_KEY_LEN = len;
-	m.DS_FLAGS = DS_TYPE_U32;
 	m.DS_VAL = value;
-	m.DS_VAL_LEN = sizeof(value);
+	m.DS_FLAGS = DSF_TYPE_LABEL | flags;
+	return do_invoke_ds(DS_PUBLISH, ds_name);
+}
 
-	r = _taskcall(DS_PROC_NR, DS_PUBLISH, &m);
+int ds_publish_u32(const char *ds_name, u32_t value, int flags)
+{
+	m.DS_VAL = value;
+	m.DS_FLAGS = DSF_TYPE_U32 | flags;
+	return do_invoke_ds(DS_PUBLISH, ds_name);
+}
 
-	cpf_revoke(g);
+int ds_publish_str(const char *ds_name, char *value, int flags)
+{
+	if(strlen(value) >= DS_MAX_STRLEN)
+		return EINVAL;
+	strcpy((char *)(&m.DS_STRING), value);
+	m.DS_FLAGS = DSF_TYPE_STR | flags;
+	return do_invoke_ds(DS_PUBLISH, ds_name);
+}
+
+int ds_publish_mem(const char *ds_name, void *vaddr, size_t length, int flags)
+{
+	cp_grant_id_t gid;
+	int r;
+
+	/* Grant for memory range. */
+	gid = cpf_grant_direct(DS_PROC_NR, (vir_bytes)vaddr, length, CPF_READ);
+	if(!GRANT_VALID(gid))
+		return errno;
+
+	m.DS_VAL = gid;
+	m.DS_VAL_LEN = length;
+	m.DS_FLAGS = DSF_TYPE_MEM | flags;
+
+	r = do_invoke_ds(DS_PUBLISH, ds_name);
+	cpf_revoke(gid);
 
 	return r;
 }
 
-int ds_publish_str(ds_name, value)
+int ds_publish_map(const char *ds_name, void *vaddr, size_t length, int flags)
-char *ds_name;
-char *value;
 {
+	cp_grant_id_t gid;
 	int r;
-	message m;
-	cp_grant_id_t g_key, g_str;
-	size_t len_key, len_str;
 
-	/* Grant for key. */
+	if(((vir_bytes)vaddr % CLICK_SIZE != 0) || (length % CLICK_SIZE != 0))
-	len_key = strlen(ds_name)+1;
+		return EINVAL;
-	g_key = cpf_grant_direct(DS_PROC_NR,
-		(vir_bytes) ds_name, len_key, CPF_READ);
-	if(!GRANT_VALID(g_key)) 
-		return GRANTBAD;
 
-	/* Grant for value. */
+	/* Grant for mapped memory range. */
-	len_str = strlen(value)+1;
+	gid = cpf_grant_direct(DS_PROC_NR, (vir_bytes)vaddr, length,
-	g_str = cpf_grant_direct(DS_PROC_NR,
+			CPF_READ | CPF_MAP);
-		(vir_bytes) value, len_str, CPF_READ);
+	if(!GRANT_VALID(gid))
+		return errno;
 
-	if(!GRANT_VALID(g_str))  {
+	m.DS_VAL = gid;
-		cpf_revoke(g_key);
+	m.DS_VAL_LEN = length;
-		return GRANTBAD;
+	m.DS_FLAGS = DSF_TYPE_MAP | flags;
-	}
 
-	m.DS_KEY_GRANT = (char *) g_key;
+	r = do_invoke_ds(DS_PUBLISH, ds_name);
-	m.DS_KEY_LEN = len_key;
+	cpf_revoke(gid);
-	m.DS_FLAGS = DS_TYPE_STR;
-	m.DS_VAL = g_str;
-	m.DS_VAL_LEN = len_str;
-
-	r = _taskcall(DS_PROC_NR, DS_PUBLISH, &m);
-
-	cpf_revoke(g_key);
-	cpf_revoke(g_str);
 
 	return r;
 }
 
-int ds_retrieve_u32(ds_name, value)
+int ds_snapshot_map(const char *ds_name, int *nr_snapshot)
-char *ds_name;
-u32_t *value;
 {
 	int r;
-	message m;
+	r = do_invoke_ds(DS_SNAPSHOT, ds_name);
-	cp_grant_id_t g_key;
+	*nr_snapshot = m.DS_NR_SNAPSHOT;
-	size_t len_key;
+	return r;
+}
 
-	/* Grant for key. */
+int ds_retrieve_label_name(char *ds_name, u32_t num)
-	len_key = strlen(ds_name)+1;
+{
-	g_key = cpf_grant_direct(DS_PROC_NR,
+	int r;
-		(vir_bytes) ds_name, len_key, CPF_READ);
+	m.DS_VAL = num;
-	if(!GRANT_VALID(g_key)) 
+	r = do_invoke_ds(DS_RETRIEVE_LABEL, ds_name);
-		return GRANTBAD;
+	return r;
+}
 
-	/* Do request. */
+int ds_retrieve_label_num(const char *ds_name, u32_t *value)
-	m.DS_KEY_GRANT = (char *) g_key;
+{
-	m.DS_KEY_LEN = len_key;
+	int r;
-	m.DS_FLAGS = DS_TYPE_U32;
+	m.DS_FLAGS = DSF_TYPE_LABEL;
-
+	r = do_invoke_ds(DS_RETRIEVE, ds_name);
-	r = _taskcall(DS_PROC_NR, DS_RETRIEVE, &m);
-
-	cpf_revoke(g_key);
-
-	if(r == OK) {
-		/* Assign u32 value. */
 	*value = m.DS_VAL;
-	}
-
 	return r;
 }
 
-int ds_retrieve_str(ds_name, value, len_str)
+int ds_retrieve_u32(const char *ds_name, u32_t *value)
-char *ds_name;
-char *value;
-size_t len_str;
 {
 	int r;
-	message m;
+	m.DS_FLAGS = DSF_TYPE_U32;
-	cp_grant_id_t g_key, g_str;
+	r = do_invoke_ds(DS_RETRIEVE, ds_name);
-	size_t len_key;
-
-	/* Grant for key. */
-	len_key = strlen(ds_name)+1;
-	g_key = cpf_grant_direct(DS_PROC_NR,
-		(vir_bytes) ds_name, len_key, CPF_READ);
-	if(!GRANT_VALID(g_key)) 
-		return GRANTBAD;
-
-	/* Grant for value. */
-	g_str = cpf_grant_direct(DS_PROC_NR,
-		(vir_bytes) value, len_str, CPF_WRITE);
-
-	if(!GRANT_VALID(g_str))  {
-		cpf_revoke(g_key);
-		return GRANTBAD;
-	}
-
-	/* Do request. */
-
-	m.DS_KEY_GRANT = (char *) g_key;
-	m.DS_KEY_LEN = len_key;
-	m.DS_FLAGS = DS_TYPE_STR;
-	m.DS_VAL = g_str;
-	m.DS_VAL_LEN = len_str;
-
-	r = _taskcall(DS_PROC_NR, DS_RETRIEVE, &m);
-
-	cpf_revoke(g_key);
-	cpf_revoke(g_str);
-
-	return r;
-}
-
-int ds_check_str(ds_key, len_key, value, len_str)
-char *ds_key;
-size_t len_key;
-char *value;
-size_t len_str;
-{
-	int r;
-	message m;
-	cp_grant_id_t g_key, g_str;
-
-	if(len_key < 1 || len_str < 1) return -1002;
-
-	/* Grant for key. */
-	g_key = cpf_grant_direct(DS_PROC_NR,
-		(vir_bytes) ds_key, len_key, CPF_WRITE);
-	if(!GRANT_VALID(g_key)) 
-		return GRANTBAD;
-
-	/* Grant for value. */
-	g_str = cpf_grant_direct(DS_PROC_NR,
-		(vir_bytes) value, len_str, CPF_WRITE);
-
-	if(!GRANT_VALID(g_str))  {
-		cpf_revoke(g_key);
-		return GRANTBAD;
-	}
-
-	/* Do request. */
-
-	m.DS_KEY_GRANT = (char *) g_key;
-	m.DS_KEY_LEN = len_key;
-	m.DS_FLAGS = DS_TYPE_STR;
-	m.DS_VAL = g_str;
-	m.DS_VAL_LEN = len_str;
-
-	r = _taskcall(DS_PROC_NR, DS_CHECK, &m);
-
-	cpf_revoke(g_key);
-	cpf_revoke(g_str);
-
-	ds_key[len_key-1] = '\0';
-	value[len_str-1] = '\0';
-
-	return r;
-}
-
-int ds_check_u32(ds_key, len_key, value)
-char *ds_key;
-size_t len_key;
-u32_t *value;
-{
-	int r;
-	message m;
-	cp_grant_id_t g_key;
-
-	if(len_key < 1) return -1;
-
-	/* Grant for key. */
-	g_key = cpf_grant_direct(DS_PROC_NR,
-		(vir_bytes) ds_key, len_key, CPF_WRITE);
-	if(!GRANT_VALID(g_key)) 
-		return GRANTBAD;
-
-	/* Do request. */
-	m.DS_KEY_GRANT = (char *) g_key;
-	m.DS_KEY_LEN = len_key;
-	m.DS_FLAGS = DS_TYPE_U32;
-
-	r = _taskcall(DS_PROC_NR, DS_CHECK, &m);
-
-	cpf_revoke(g_key);
-
-	ds_key[len_key-1] = '\0';
-
-	/* Assign u32 value. */
 	*value = m.DS_VAL;
+	return r;
+}
+
+int ds_retrieve_str(const char *ds_name, char *value, size_t len_str)
+{
+	int r;
+	m.DS_FLAGS = DSF_TYPE_STR;
+	r = do_invoke_ds(DS_RETRIEVE, ds_name);
+	strncpy(value, (char *)(&m.DS_STRING), DS_MAX_STRLEN);
+	value[DS_MAX_STRLEN - 1] = '\0';
+	return r;
+}
+
+int ds_retrieve_mem(const char *ds_name, char *vaddr, size_t *length)
+{
+	cp_grant_id_t gid;
+	int r;
+
+	/* Grant for memory range. */
+	gid = cpf_grant_direct(DS_PROC_NR, (vir_bytes)vaddr, *length, CPF_WRITE);
+	if(!GRANT_VALID(gid))
+		return errno;
+
+	m.DS_VAL = gid;
+	m.DS_VAL_LEN = *length;
+	m.DS_FLAGS = DSF_TYPE_MEM;
+	r = do_invoke_ds(DS_RETRIEVE, ds_name);
+	*length = m.DS_VAL_LEN;
+	cpf_revoke(gid);
 
 	return r;
 }
 
+int ds_retrieve_map(const char *ds_name, char *vaddr, size_t *length,
+		int nr_snapshot, int flags)
+{
+	cp_grant_id_t gid;
+	int r;
+
+	/* Map a mapped memory range. */
+	if(flags & DSMF_MAP_MAPPED) {
+		/* Request DS to grant. */
+		m.DS_FLAGS = DSF_TYPE_MAP | DSMF_MAP_MAPPED;
+		r = do_invoke_ds(DS_RETRIEVE, ds_name);
+		if(r != OK)
+			return r;
+
+		/* Do the safemap. */
+		if(*length > m.DS_VAL_LEN)
+			*length = m.DS_VAL_LEN;
+		*length = (size_t) CLICK_FLOOR(*length);
+		r = sys_safemap(DS_PROC_NR, m.DS_VAL, 0,
+				(vir_bytes)vaddr, *length, D, 0);
+
+	/* Copy mapped memory range or a snapshot. */
+	} else if(flags & (DSMF_COPY_MAPPED|DSMF_COPY_SNAPSHOT)) {
+		/* Grant for memory range first. */
+		gid = cpf_grant_direct(DS_PROC_NR, (vir_bytes)vaddr,
+				*length, CPF_WRITE);
+		if(!GRANT_VALID(gid))
+			return errno;
+
+		m.DS_VAL = gid;
+		m.DS_VAL_LEN = *length;
+		if(flags & DSMF_COPY_MAPPED) {
+			m.DS_FLAGS = DSF_TYPE_MAP | DSMF_COPY_MAPPED;
+		}
+		else {
+			m.DS_NR_SNAPSHOT = nr_snapshot;
+			m.DS_FLAGS = DSF_TYPE_MAP | DSMF_COPY_SNAPSHOT;
+		}
+		r = do_invoke_ds(DS_RETRIEVE, ds_name);
+		*length = m.DS_VAL_LEN;
+		cpf_revoke(gid);
+	}
+	else {
+		return EINVAL;
+	}
+
+	return r;
+}
+
+int ds_delete_u32(const char *ds_name)
+{
+	m.DS_FLAGS = DSF_TYPE_U32;
+	return do_invoke_ds(DS_DELETE, ds_name);
+}
+
+int ds_delete_str(const char *ds_name)
+{
+	m.DS_FLAGS = DSF_TYPE_STR;
+	return do_invoke_ds(DS_DELETE, ds_name);
+}
+
+int ds_delete_mem(const char *ds_name)
+{
+	m.DS_FLAGS = DSF_TYPE_MEM;
+	return do_invoke_ds(DS_DELETE, ds_name);
+}
+
+int ds_delete_map(const char *ds_name)
+{
+	m.DS_FLAGS = DSF_TYPE_MAP;
+	return do_invoke_ds(DS_DELETE, ds_name);
+}
+
+int ds_delete_label(const char *ds_name)
+{
+	m.DS_FLAGS = DSF_TYPE_LABEL;
+	return do_invoke_ds(DS_DELETE, ds_name);
+}
+
+int ds_subscribe(const char *regexp, int flags)
+{
+	m.DS_FLAGS = flags;
+	return do_invoke_ds(DS_SUBSCRIBE, regexp);
+}
+
+int ds_check(char *ds_key, int *type)
+{
+	int r;
+	r = do_invoke_ds(DS_CHECK, ds_key);
+	*type = m.DS_FLAGS;
+	return r;
+}

lib/syslib/pci_del_acl.c

@@ -21,7 +21,7 @@ endpoint_t proc_ep;
 
 	if (pci_procnr == ANY)
 	{
-		r= ds_retrieve_u32("pci", &u32);
+		r= ds_retrieve_label_num("pci", &u32);
 		if (r != 0)
 		{
 			panic("syslib/" __FILE__,

lib/syslib/pci_init1.c

@@ -22,9 +22,9 @@ char *name;
 	size_t len;
 	message m;
 
-	r= ds_retrieve_u32("pci", &u32);
+	r= ds_retrieve_label_num("pci", &u32);
 	if (r != 0)
-		panic("syslib/" __FILE__, "pci_init1: ds_retrieve_u32 failed for 'pci'", r);
+		panic("syslib/" __FILE__, "pci_init1: ds_retrieve_label_num failed for 'pci'", r);
 	pci_procnr= u32;
 
 	m.m_type= BUSC_PCI_INIT;

lib/syslib/pci_set_acl.c

@@ -22,11 +22,11 @@ struct rs_pci *rs_pci;
 
 	if (pci_procnr == ANY)
 	{
-		r= ds_retrieve_u32("pci", &u32);
+		r= ds_retrieve_label_num("pci", &u32);
 		if (r != 0)
 		{
 			panic("syslib/" __FILE__,
-				"pci_set_acl: ds_retrieve_u32 failed for 'pci'",
+				"pci_set_acl: ds_retrieve_label_num failed for 'pci'",
 				r);
 		}
 		pci_procnr = u32;

lib/syslib/safecopies.c

@@ -16,7 +16,7 @@
 #include <string.h>
 
 #define ACCESS_CHECK(a) { 			\
-	if((a) & ~(CPF_READ|CPF_WRITE)) {	\
+	if((a) & ~(CPF_READ|CPF_WRITE|CPF_MAP)) {	\
 		errno = EINVAL;			\
 		return -1;			\
 	}					\
@@ -37,6 +37,13 @@
 	}							\
    }
 
+#define CLICK_ALIGNMENT_CHECK(addr, bytes) {				      \
+	if(((vir_bytes)(addr) % CLICK_SIZE != 0)			      \
+		|| ((vir_bytes)(bytes) % CLICK_SIZE != 0)) {		      \
+		return EINVAL;						      \
+	}								      \
+   }
+
 #define NR_STATIC_GRANTS 2
 PRIVATE cp_grant_t static_grants[NR_STATIC_GRANTS];
 PRIVATE cp_grant_t *grants = NULL;
@@ -206,8 +213,16 @@ PUBLIC int
 cpf_revoke(cp_grant_id_t g)
 {
 /* Revoke previously granted access, identified by grant id. */
+	int r;
 	GID_CHECK_USED(g);
 
+	/* If this grant is for memory mapping, revoke the mapping first. */
+	if(grants[g].cp_flags & CPF_MAP) {
+		r = sys_saferevmap_gid(g);
+		if(r != 0)
+			return r;
+	}
+
 	/* Make grant invalid by setting flags to 0, clearing CPF_USED.
 	 * This invalidates the grant.
 	 */
@@ -262,6 +277,12 @@ int access;
 	GID_CHECK(gid);
 	ACCESS_CHECK(access);
 
+	/* Check click alignment in case of memory mapping grant. */
+	if(access & CPF_MAP) {
+		CLICK_ALIGNMENT_CHECK(addr, bytes);
+	}
+
+	/* Fill in new slot data. */
 	grants[gid].cp_flags = access | CPF_DIRECT | CPF_USED | CPF_VALID;
 	grants[gid].cp_u.cp_direct.cp_who_to = who;
 	grants[gid].cp_u.cp_direct.cp_start = addr;
@@ -298,6 +319,11 @@ int access;
 	GID_CHECK(gid);
 	ACCESS_CHECK(access);
 
+	/* Check click alignment in case of memory mapping grant. */
+	if(access & CPF_MAP) {
+		CLICK_ALIGNMENT_CHECK(addr, bytes);
+	}
+
 	/* Fill in new slot data. */
 	grants[gid].cp_flags = CPF_USED | CPF_MAGIC | CPF_VALID | access;
 	grants[gid].cp_u.cp_magic.cp_who_to = who_to;

lib/syslib/sys_safemap.c

@@ -0,0 +1,72 @@
+
+#include "syslib.h"
+
+#include <minix/safecopies.h>
+
+/*===========================================================================*
+ *				sys_safemap				     *
+ *===========================================================================*/
+PUBLIC int sys_safemap(endpoint_t grantor, cp_grant_id_t grant,
+	vir_bytes grant_offset, vir_bytes my_address,
+	size_t bytes, int my_seg, int writable)
+{
+/* Map a block of data for which the other process has previously
+ * granted permission. 
+ */
+
+	message copy_mess;
+
+	copy_mess.SMAP_EP = grantor;
+	copy_mess.SMAP_GID = grant;
+	copy_mess.SMAP_OFFSET = grant_offset;
+	copy_mess.SMAP_SEG = (void*) my_seg;
+	copy_mess.SMAP_ADDRESS = my_address;
+	copy_mess.SMAP_BYTES = bytes;
+	copy_mess.SMAP_FLAG = writable;
+
+	return(_taskcall(SYSTASK, SYS_SAFEMAP, &copy_mess));
+
+}
+
+/*===========================================================================*
+ *			     sys_saferevmap_gid				     *
+ *===========================================================================*/
+PUBLIC int sys_saferevmap_gid(cp_grant_id_t grant)
+{
+/* Grantor revokes safemap by grant id. */
+	message copy_mess;
+
+	copy_mess.SMAP_FLAG = 1;
+	copy_mess.SMAP_GID = grant;
+
+	return(_taskcall(SYSTASK, SYS_SAFEREVMAP, &copy_mess));
+}
+
+/*===========================================================================*
+ *			    sys_saferevmap_addr				     *
+ *===========================================================================*/
+PUBLIC int sys_saferevmap_addr(vir_bytes addr)
+{
+/* Grantor revokes safemap by address. */
+	message copy_mess;
+
+	copy_mess.SMAP_FLAG = 0;
+	copy_mess.SMAP_GID = addr;
+
+	return(_taskcall(SYSTASK, SYS_SAFEREVMAP, &copy_mess));
+}
+
+/*===========================================================================*
+ *				sys_safeunmap				     *
+ *===========================================================================*/
+PUBLIC int sys_safeunmap(int my_seg, vir_bytes my_address)
+{
+/* Requestor unmaps safemap. */
+	message copy_mess;
+
+	copy_mess.SMAP_SEG = (void*) my_seg;
+	copy_mess.SMAP_ADDRESS = my_address;
+
+	return(_taskcall(SYSTASK, SYS_SAFEUNMAP, &copy_mess));
+}
+

lib/syslib/sys_vmctl.c

@@ -43,7 +43,8 @@ PUBLIC int sys_vmctl_get_cr3_i386(endpoint_t who, u32_t *cr3)
 }
 
 PUBLIC int sys_vmctl_get_memreq(endpoint_t *who, vir_bytes *mem,
-	vir_bytes *len, int *wrflag, endpoint_t *requestor)
+        vir_bytes *len, int *wrflag, endpoint_t *who_s, vir_bytes *mem_s,
+        endpoint_t *requestor)
 {
   message m;
   int r;
@@ -51,11 +52,13 @@ PUBLIC int sys_vmctl_get_memreq(endpoint_t *who, vir_bytes *mem,
   m.SVMCTL_WHO = SELF;
   m.SVMCTL_PARAM = VMCTL_MEMREQ_GET;
   r = _taskcall(SYSTASK, SYS_VMCTL, &m);
-  if(r == OK) {
+  if(r >= 0) {
-	*who = m.SVMCTL_MRG_EP;
+	*who = m.SVMCTL_MRG_TARGET;
-	*mem = (vir_bytes) m.SVMCTL_MRG_ADDR;
+	*mem = m.SVMCTL_MRG_ADDR;
-	*len = m.SVMCTL_MRG_LEN;
+	*len = m.SVMCTL_MRG_LENGTH;
-	*wrflag = m.SVMCTL_MRG_WRITE;
+	*wrflag = m.SVMCTL_MRG_FLAG;
+	*who_s = m.SVMCTL_MRG_EP2;
+	*mem_s = m.SVMCTL_MRG_ADDR2;
 	*requestor = (endpoint_t) m.SVMCTL_MRG_REQUESTOR;
   }
   return r;

lib/sysutil/Makefile.in

@@ -18,7 +18,7 @@ libsys_FILES=" \
 	env_panic.c \
 	env_prefix.c \
 	fkey_ctl.c \
-	micro_delay.c \
+	tsc_util.c \
 	report.c \
 	taskcall.c \
 	read_tsc.s \

lib/sysutil/tsc_util.c

@@ -17,15 +17,15 @@
 #define CALIBRATE 						\
 	if(!calibrated) {					\
 		int r;						\
-		if((r=micro_delay_calibrate()) != OK)		\
+		if((r=tsc_calibrate()) != OK)			\
-			panic(__FILE__, "micro_delay: calibrate failed\n", r); \
+			panic(__FILE__, "calibrate failed\n", r); \
 	}
 
 static u32_t calib_tsc, Hz = 0;
 static int calibrated = 0;
 
 int
-micro_delay_calibrate(void)
+tsc_calibrate(void)
 {
 	u64_t start, end, diff;
 	struct tms tms;
@@ -50,11 +50,11 @@ micro_delay_calibrate(void)
 	diff = sub64(end, start);
 	if(ex64hi(diff) != 0)
 	  panic(__FILE__,
-		"micro_delay_calibrate: CALIBRATE_TICKS too high "
+		"tsc_calibrate: CALIBRATE_TICKS too high "
 			"for TSC frequency\n", NO_NUM);
 	calib_tsc = ex64lo(diff);
 #if 0
-	printf("micro_delay_calibrate: "
+	printf("tsc_calibrate: "
 		"%lu cycles/%d ticks of %d Hz; %lu cycles/s\n",
 			calib_tsc, CALIBRATE_TICKS(Hz), Hz,
 			div64u(mul64u(calib_tsc, Hz), CALIBRATE_TICKS(Hz)));
@@ -94,3 +94,23 @@ micro_delay(u32_t micros)
 	return OK;
 }
 
+u32_t tsc_64_to_micros(u64_t tsc)
+{
+	return tsc_to_micros(ex64lo(tsc), ex64hi(tsc));
+}
+
+u32_t tsc_to_micros(u32_t low, u32_t high)
+{
+	u32_t micros;
+
+	if(high) {
+		return 0;
+	}
+	CALIBRATE;
+
+	micros = (div64u(mul64u(low, MICROHZ * CALIBRATE_TICKS(Hz)),
+		calib_tsc)/Hz);
+
+	return micros;
+}
+

servers/ds/inc.h

@@ -7,12 +7,14 @@
 #include <sys/types.h>
 #include <limits.h>
 #include <errno.h>
+#include <regex.h>
 
 #include <minix/callnr.h>
 #include <minix/config.h>
 #include <minix/type.h>
 #include <minix/const.h>
 #include <minix/com.h>
+#include <minix/ds.h>
 #include <minix/syslib.h>
 #include <minix/sysutil.h>
 #include <minix/keymap.h>
@@ -27,5 +29,4 @@
 
 #include "proto.h"
 #include "glo.h"
-#include "store.h"
 

servers/ds/main.c

@@ -71,12 +71,21 @@ PUBLIC int main(int argc, char **argv)
       case DS_RETRIEVE:
 	  result = do_retrieve(&m);
 	  break;
+      case DS_RETRIEVE_LABEL:
+	  result = do_retrieve_label(&m);
+	  break;
+      case DS_DELETE:
+	  result = do_delete(&m);
+	  break;
       case DS_SUBSCRIBE:
 	  result = do_subscribe(&m);
 	  break;
       case DS_CHECK:
 	  result = do_check(&m);
 	  break;
+      case DS_SNAPSHOT:
+	  result = do_snapshot(&m);
+	  break;
       case GETSYSINFO:
 	  result = do_getsysinfo(&m);
 	  break;

servers/ds/proto.h

@@ -6,8 +6,11 @@ _PROTOTYPE(int main, (int argc, char **argv));
 /* store.c */
 _PROTOTYPE(int do_publish, (message *m_ptr));
 _PROTOTYPE(int do_retrieve, (message *m_ptr));
+_PROTOTYPE(int do_retrieve_label, (message *m_ptr));
 _PROTOTYPE(int do_subscribe, (message *m_ptr));
 _PROTOTYPE(int do_check, (message *m_ptr));
+_PROTOTYPE(int do_delete, (message *m_ptr));
+_PROTOTYPE(int do_snapshot, (message *m_ptr));
 _PROTOTYPE(int do_getsysinfo, (message *m_ptr));
 _PROTOTYPE(int sef_cb_init_fresh, (int type, sef_init_info_t *info));
 _PROTOTYPE(int map_service, (struct rprocpub *rpub));

servers/ds/store.h

@@ -1,33 +1,46 @@
-/* Type definitions for the Data Store Server. */
+#ifndef _DS_STORE_H_
+#define _DS_STORE_H_
 
+/* Type definitions for the Data Store Server. */
 #include <sys/types.h>
 #include <minix/sys_config.h>
 #include <minix/ds.h>
 #include <minix/bitmap.h>
 #include <regex.h>
 
-/* Constants for the Data Store Server. */
+#define NR_DS_KEYS	64	/* number of entries */
-#define NR_DS_KEYS               64	/* reserve space for so many items */
+#define NR_DS_SUBS	(4*_NR_SYS_PROCS)	/* number of subscriptions */
-#define NR_DS_SUBS   (4*_NR_SYS_PROCS)	/* .. and so many subscriptions */
+#define NR_DS_SNAPSHOT	5	/* number of snapshots */
-
-/* Types. */
 
+/* Base 'class' for the following 3 structs. */
 struct data_store {
-  int ds_flags;			/* flags for this store, includes type info */
+	int	flags;
-  char ds_key[DS_MAX_KEYLEN];	/* key to lookup information */
+	char	key[DS_MAX_KEYLEN];	/* key to lookup information */
-  union {
+	char	owner[DS_MAX_KEYLEN];
-    u32_t ds_val_u32;			/* u32 data (DS_TYPE_U32) */
-    char  ds_val_str[DS_MAX_VALLEN];	/* string data (DS_TYPE_STR) */
-  } ds_val;
 
-  /* out of date subscribers. */
+	union {
-  bitchunk_t ds_old_subs[BITMAP_CHUNKS(NR_DS_SUBS)];	
+		unsigned u32;
+		char string[DS_MAX_STRLEN];
+		struct {
+			void *data;
+			size_t length;
+			size_t reallen;
+		} mem;
+		struct dsi_map {
+			void *data;
+			size_t length;
+			void *realpointer;
+			void *snapshots[NR_DS_SNAPSHOT];
+			int sindex;
+		} map;
+	} u;
 };
 
 struct subscription {
-  int sub_flags;		/* flags for this subscription */
+	int		flags;
-  regex_t sub_regex;		/* regular expression agains keys */
+	char		owner[DS_MAX_KEYLEN];
-  endpoint_t sub_owner;		/* who is subscribed */
+	regex_t		regex;
+	bitchunk_t	old_subs[BITMAP_CHUNKS(NR_DS_KEYS)];	
 };
 
-
+#endif /* _DS_STORE_H_ */

servers/inet/inet.c

@@ -257,9 +257,9 @@ PRIVATE int sef_cb_init_fresh(int type, sef_init_info_t *info)
 	init_rand256(randbits);
 
 	/* Our new identity as a server. */
-	r= ds_retrieve_u32("inet", &tasknr);
+	r= ds_retrieve_label_num("inet", &tasknr);
 	if (r != OK)
-		ip_panic(("inet: ds_retrieve_u32 failed for 'inet': %d", r));
+		ip_panic(("inet: ds_retrieve_label_num failed for 'inet': %d", r));
 	this_proc= tasknr;
 
 	/* Register the device group. */

servers/inet/mnx_eth.c

@@ -99,10 +99,10 @@ PUBLIC void osdep_eth_init()
 		}
 		eth_port->etp_osdep.etp_rd_vec_grant= gid;
 
-		r= ds_retrieve_u32(ecp->ec_task, &tasknr);
+		r= ds_retrieve_label_num(ecp->ec_task, &tasknr);
 		if (r != OK && r != ESRCH)
 		{
-			printf("inet: ds_retrieve_u32 failed for '%s': %d\n",
+			printf("inet: ds_retrieve_label_num failed for '%s': %d\n",
 				ecp->ec_task, r);
 		}
 		if (r != OK)

servers/is/dmp_ds.c

@@ -1,66 +1,54 @@
-/* This file contains procedures to dump DS data structures.
- *
- * The entry points into this file are
- *   data_store_dmp:   	display DS data store contents 
- *
- * Created:
- *   Oct 18, 2005:	by Jorrit N. Herder
- */
-
 #include "inc.h"
 #include "../ds/store.h"
 
-PUBLIC struct data_store store[NR_DS_KEYS];
+PRIVATE struct data_store ds_store[NR_DS_KEYS];
 
-FORWARD _PROTOTYPE( char *s_flags_str, (int flags)		);
-
-/*===========================================================================*
- *				data_store_dmp				     *
- *===========================================================================*/
 PUBLIC void data_store_dmp()
 {
-  struct data_store *dsp;
+  struct data_store *p;
-  int i,j, n=0, s;
+  static int prev_i = 0;
-  static int prev_i=0;
+  int r, i, n = 0;
 
-
+  if((r=getsysinfo(DS_PROC_NR, SI_DATA_STORE, ds_store)) != OK) {
-  printf("Data Store (DS) contents dump\n");
+	printf("Couldn't talk to DS: %d.\n", r);
-
-  if((s=getsysinfo(DS_PROC_NR, SI_DATA_STORE, store)) != OK) {
-	printf("Couldn't talk to DS: %d.\n", s);
 	return;
   }
 
-  printf("slot key                  type value\n");
+  printf("Data store contents:\n");
+  printf("-slot- ------key------ -----owner----- ---type--- ----value---\n");
+  for(i = prev_i; i < NR_DS_KEYS; i++) {
+	p = &ds_store[i];
+	if(!(p->flags & DSF_IN_USE))
+		continue;
 
-  for (i=prev_i; i<NR_DS_KEYS; i++) {
+	printf("%6d %-15s %-15s ", i, p->key, p->owner);
-  	dsp = &store[i];
+	switch(p->flags & DSF_MASK_TYPE) {
-  	if (! dsp->ds_flags & DS_IN_USE) continue;
+	case DSF_TYPE_U32:
-  	if (++n > 22) break;
+		printf("%-10s %12u\n", "U32", p->u.u32);
-  	printf("%3d  %-20s ",
+		break;
-		i, dsp->ds_key);
+	case DSF_TYPE_STR:
-	if(dsp->ds_flags & DS_TYPE_U32) {
+		printf("%-10s %12s\n", "STR", p->u.string);
-		printf("u32  %lu\n", dsp->ds_val.ds_val_u32);
+		break;
-	} else if(dsp->ds_flags & DS_TYPE_STR) {
+	case DSF_TYPE_MEM:
-		printf("str  \"%s\"\n", dsp->ds_val.ds_val_str);
+		printf("%-10s %12u\n", "MEM", p->u.mem.length);
-	} else {
+		break;
-		printf("Bogus type\n");
+	case DSF_TYPE_MAP:
+		printf("%-10s %9u/%3u\n", "MAP", p->u.map.length,
+			p->u.map.sindex);
+		break;
+	case DSF_TYPE_LABEL:
+		printf("%-10s %12u\n", "LABEL", p->u.u32);
+		break;
+	default:
+		return;
 	}
+
+	if(n++ == 21)
+		break;
   }
+
   if (i >= NR_DS_KEYS) i = 0;
   else printf("--more--\r");
   prev_i = i;
 }
 
-
-PRIVATE char *s_flags_str(int flags)
-{
-	static char str[5];
-	str[0] = (flags & DS_IN_USE) ? 'U' : '-';
-	str[1] = (flags & DS_PUBLIC) ? 'P' : '-';
-	str[2] = '-';
-	str[3] = '\0';
-
-	return(str);
-}
-

servers/mfs/mount.c

@@ -49,11 +49,12 @@ PUBLIC int fs_readsuper()
 	return(EINVAL);
   }
 
-  r = ds_retrieve_u32(fs_dev_label, &tasknr);
+  r= ds_retrieve_label_num(fs_dev_label, &tasknr);
-  if (r != OK) {
+  if (r != OK)
-	printf("%s:%d fs_readsuper: ds_retrieve_u32 failed for '%s': %d\n",
+  {
-	       __FILE__, __LINE__, fs_dev_label, r);
+	printf("mfs:fs_readsuper: ds_retrieve_label_num failed for '%s': %d\n",
-	return(EINVAL);
+		fs_dev_label, r);
+	return EINVAL;
   }
 
   driver_e = tasknr;

servers/rs/memory.c

@@ -45,8 +45,8 @@ PRIVATE int check_mem_available(char *new_brksize)
   mem_sp = &proc.p_memmap[S];            /* pointer to stack segment map */
 
   /* Compute how many clicks the data segment is to become. */
-  data_clicks = (vir_clicks) ( ((long) new_brksize + CLICK_SIZE - 1) >>
+  data_clicks = (vir_clicks) (CLICK_CEIL(new_brksize) >> CLICK_SHIFT)
-      CLICK_SHIFT) - mem_dp->mem_vir;
+	- mem_dp->mem_vir;
 
   /* See if stack size has gone negative (i.e., sp too close to 0xFFFF...) */
   base_of_stack = (long) mem_sp->mem_vir + (long) mem_sp->mem_len;
@@ -62,7 +62,7 @@ PRIVATE int check_mem_available(char *new_brksize)
 
   /* Add a safety margin for future stack growth. Impossible to do right. */
 #define SAFETY_BYTES  (384 * sizeof(char *))
-#define SAFETY_CLICKS ((SAFETY_BYTES + CLICK_SIZE - 1) / CLICK_SIZE)
+#define SAFETY_CLICKS ((vir_clicks) (CLICK_CEIL(SAFETY_BYTES) >> CLICK_SHIFT))
   gap_base = mem_dp->mem_vir + data_clicks + SAFETY_CLICKS;
   if (sp_lower < gap_base)
   {

servers/rs/service.c

@@ -112,7 +112,7 @@ PRIVATE void print_usage(char *app_name, char *problem)
   fprintf(stderr, "Warning, %s\n", problem);
   fprintf(stderr, "Usage:\n");
   fprintf(stderr,
-  "    %s [-c] (up|run) <binary> [%s <args>] [%s <special>] [%s <ticks>]\n", 
+  "    %s [-c -r] (up|run) <binary> [%s <args>] [%s <special>] [%s <ticks>]\n", 
 	app_name, ARG_ARGS, ARG_DEV, ARG_PERIOD);
   fprintf(stderr, "    %s down label\n", app_name);
   fprintf(stderr, "    %s refresh label\n", app_name);
@@ -729,6 +729,9 @@ struct
 	{ "GETINFO",		SYS_GETINFO },
 	{ "SAFECOPYFROM",	SYS_SAFECOPYFROM },
 	{ "SAFECOPYTO",		SYS_SAFECOPYTO },
+	{ "SAFEMAP",		SYS_SAFEMAP },
+	{ "SAFEREVMAP",		SYS_SAFEREVMAP },
+	{ "SAFEUNMAP",		SYS_SAFEUNMAP },
 	{ "VSAFECOPY",		SYS_VSAFECOPY },
 	{ "SETGRANT",		SYS_SETGRANT },
 	{ "READBIOS",		SYS_READBIOS },

servers/rs/table.c

@@ -15,7 +15,8 @@
  */
 #define FS_KC   SYS_KILL, SYS_VIRCOPY, SYS_SAFECOPYFROM, SYS_SAFECOPYTO, \
     SYS_UMAP, SYS_GETINFO, SYS_EXIT, SYS_TIMES, SYS_SETALARM, \
-    SYS_PRIVCTL, SYS_TRACE , SYS_SETGRANT, SYS_PROFBUF, SYS_SYSCTL
+    SYS_PRIVCTL, SYS_TRACE , SYS_SETGRANT, SYS_PROFBUF, SYS_SYSCTL, \
+    SYS_SAFEMAP, SYS_SAFEREVMAP, SYS_SAFEUNMAP
 #define DRV_KC	FS_KC, SYS_SEGCTL, SYS_IRQCTL, SYS_INT86, SYS_DEVIO, \
     SYS_SDEVIO, SYS_VDEVIO, SYS_SETGRANT, SYS_PROFBUF, SYS_SYSCTL
 

servers/rs/utility.c

@@ -45,7 +45,7 @@ struct rproc *rp;				/* pointer to process slot */
   rpub = rp->r_pub;
 
   /* Register its label with DS. */
-  s= ds_publish_u32(rpub->label, rpub->endpoint);
+  s= ds_publish_label(rpub->label, rpub->endpoint, DSF_OVERWRITE);
   if (s != OK) {
       return s;
   }

servers/vfs/dmap.c

@@ -96,7 +96,7 @@ PUBLIC int do_mapdriver()
 
 	label[label_len]= '\0';
 
-	r= ds_retrieve_u32(label, &tasknr);
+	r= ds_retrieve_label_num(label, &tasknr);
 	if (r != OK)
 	{
 		printf("vfs:do_mapdriver: ds doesn't know '%s'\n", label);

servers/vfs/mount.c

@@ -96,7 +96,7 @@ PUBLIC int do_mount()
 
 	mount_label[sizeof(mount_label)-1] = 0;
 
-	r = ds_retrieve_u32(mount_label, &fs_e);
+	r = ds_retrieve_label_num(mount_label, &fs_e);
 	if (r != OK) return(r);
 
 	if (isokendpt(fs_e, &proc_nr) != OK) return(EINVAL);

servers/vm/Makefile

@@ -5,7 +5,7 @@ include /etc/make.conf
 
 OBJ = main.o alloc.o utility.o exec.o exit.o fork.o break.o \
 	signal.o vfs.o mmap.o slaballoc.o region.o pagefaults.o addravl.o \
-	physravl.o rs.o queryexit.o
+	physravl.o rs.o queryexit.o map_mem.o
 ARCHOBJ =  $(ARCH)/vm.o $(ARCH)/pagetable.o $(ARCH)/arch_pagefaults.o $(ARCH)/util.o 
 
 CPPFLAGS=-I../../kernel/arch/$(ARCH)/include -I$(ARCH)

servers/vm/break.c

@@ -96,7 +96,7 @@ vir_bytes sp;			/* new value of sp */
 
   /* Add a safety margin for future stack growth. Impossible to do right. */
 #define SAFETY_BYTES  (384 * sizeof(char *))
-#define SAFETY_CLICKS ((SAFETY_BYTES + CLICK_SIZE - 1) / CLICK_SIZE)
+#define SAFETY_CLICKS ((vir_clicks) (CLICK_CEIL(SAFETY_BYTES) >> CLICK_SHIFT))
   gap_base = mem_dp->mem_vir + data_clicks + SAFETY_CLICKS;
   if (sp_lower < gap_base)
   {
@@ -165,7 +165,7 @@ vir_bytes v;
 	vir_clicks new_clicks;
 	int r;
 
-	new_clicks = (vir_clicks) ( ((long) v + CLICK_SIZE - 1) >> CLICK_SHIFT);
+	new_clicks = (vir_clicks) (CLICK_CEIL(v) >> CLICK_SHIFT);
 	if (new_clicks < vmp->vm_arch.vm_seg[D].mem_vir) {
 		printf("VM: real_brk failed because new_clicks too high: %d\n",
 			new_clicks);

servers/vm/exec.c

@@ -110,10 +110,10 @@ SANITYCHECK(SCL_DETAIL);
 	}
 
 	/* Check to see if segment sizes are feasible. */
-	tc = ((unsigned long) args.text_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+	tc = (vir_clicks) (CLICK_CEIL(args.text_bytes) >> CLICK_SHIFT);
-	dc = (args.data_bytes+args.bss_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+	dc = (vir_clicks) (CLICK_CEIL(args.data_bytes+args.bss_bytes) >> CLICK_SHIFT);
-	totc = (args.tot_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+	totc = (vir_clicks) (CLICK_CEIL(args.tot_bytes) >> CLICK_SHIFT);
-	sc = (args.args_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+	sc = (vir_clicks) (CLICK_CEIL(args.args_bytes) >> CLICK_SHIFT);
 	if (dc >= totc) {
 		printf("VM: newmem: no stack?\n");
 		return(ENOEXEC); /* stack must be at least 1 click */
@@ -204,10 +204,10 @@ vir_bytes *stack_top;		/* top of process stack */
    * and stack occupies an integral number of clicks, starting at click
    * boundary.  The data and bss parts are run together with no space.
    */
-  text_clicks = ((unsigned long) text_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+  text_clicks = (vir_clicks) (CLICK_CEIL(text_bytes) >> CLICK_SHIFT);
-  data_clicks = (data_bytes + bss_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+  data_clicks = (vir_clicks) (CLICK_CEIL(data_bytes + bss_bytes) >> CLICK_SHIFT);
-  stack_clicks = (stk_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+  stack_clicks = (vir_clicks) (CLICK_CEIL(stk_bytes) >> CLICK_SHIFT);
-  tot_clicks = (tot_bytes + CLICK_SIZE - 1) >> CLICK_SHIFT;
+  tot_clicks = (vir_clicks) (CLICK_CEIL(tot_bytes) >> CLICK_SHIFT);
   gap_clicks = tot_clicks - data_clicks - stack_clicks;
   if ( (int) gap_clicks < 0) {
 	printf("VM: new_mem: no gap?\n");

servers/vm/i386/pagetable.c

@@ -2,8 +2,6 @@
 #define _SYSTEM 1
 #define _POSIX_SOURCE 1
 
-#define VERBOSE 0
-
 #include <minix/callnr.h>
 #include <minix/com.h>
 #include <minix/config.h>

servers/vm/main.c

@@ -1,8 +1,6 @@
 
 #define _SYSTEM 1
 
-#define VERBOSE 0
-
 #include <minix/callnr.h>
 #include <minix/com.h>
 #include <minix/config.h>

servers/vm/map_mem.c

@@ -0,0 +1,330 @@
+
+#define _SYSTEM 1
+
+#include <minix/com.h>
+#include <minix/callnr.h>
+#include <minix/type.h>
+#include <minix/config.h>
+#include <minix/const.h>
+#include <minix/sysutil.h>
+#include <minix/syslib.h>
+
+#include <sys/mman.h>
+
+#include <limits.h>
+#include <string.h>
+#include <errno.h>
+#include <assert.h>
+#include <stdint.h>
+#include <memory.h>
+
+#include "vm.h"
+#include "proto.h"
+#include "util.h"
+#include "glo.h"
+#include "region.h"
+#include "sanitycheck.h"
+
+/*===========================================================================*
+ *				split_phys				     *
+ *===========================================================================*/
+PRIVATE int split_phys(struct phys_region *pr, vir_bytes point)
+{
+	struct phys_region *newpr, *q, *prev;
+	struct phys_block *newpb;
+	struct phys_block *pb = pr->ph;
+/* Split the phys region into 2 parts by @point. */
+
+	if(pr->offset >= point || pr->offset + pb->length <= point)
+		return OK;
+	if(!SLABALLOC(newpb))
+		return ENOMEM;
+
+	/* Split phys block. */
+	*newpb = *pb;
+	pb->length = point - pr->offset;
+	newpb->length -= pb->length;
+	newpb->phys += pb->length;
+
+	/* Split phys regions in a list. */
+	for(q = pb->firstregion; q; q = q->next_ph_list) {
+		if(!SLABALLOC(newpr))
+			return ENOMEM;
+
+		*newpr = *q;
+		newpr->ph = newpb;
+		newpr->offset += pb->length;
+
+		/* Link to the vir region's phys region list. */
+		physr_insert(newpr->parent->phys, newpr);
+
+		/* Link to the next_ph_list. */
+		if(q == pb->firstregion) {
+			newpb->firstregion = newpr;
+			prev = newpr;
+		} else {
+			prev->next_ph_list = newpr;
+			prev = newpr;
+		}
+	}
+	prev->next_ph_list = NULL;
+
+	return OK;
+}
+
+/*===========================================================================*
+ *				rm_phys_regions				     *
+ *===========================================================================*/
+PRIVATE void rm_phys_regions(struct vir_region *region,
+	vir_bytes begin, vir_bytes length)
+{
+/* Remove all phys regions between @begin and @begin+length.
+ *
+ * Don't update the page table, because we will update it at map_memory()
+ * later.
+ */
+	struct phys_region *pr;
+	physr_iter iter;
+
+	physr_start_iter(region->phys, &iter, begin, AVL_GREATER_EQUAL);
+	while((pr = physr_get_iter(&iter)) && pr->offset < begin + length) {
+		pb_unreferenced(region, pr);
+		physr_remove(region->phys, pr->offset);
+		physr_start_iter(region->phys, &iter, begin,
+			AVL_GREATER_EQUAL);
+		SLABFREE(pr);
+	}
+}
+
+/*===========================================================================*
+ *				clean_phys_regions			     *
+ *===========================================================================*/
+PRIVATE void clean_phys_regions(struct vir_region *region,
+	vir_bytes offset, vir_bytes length)
+{
+/* Consider @offset as the start address and @offset+length as the end address.
+ * If there are phys regions crossing the start address or the end address,
+ * split them into 2 parts.
+ *
+ * We assume that the phys regions are listed in order and don't overlap.
+ */
+	struct phys_region *pr;
+	physr_iter iter;
+
+	physr_start_iter_least(region->phys, &iter);
+	while((pr = physr_get_iter(&iter))) {
+		/* If this phys region crosses the start address, split it. */
+		if(pr->offset < offset
+			&& pr->offset + pr->ph->length > offset) {
+			split_phys(pr, offset);
+			physr_start_iter_least(region->phys, &iter);
+		}
+		/* If this phys region crosses the end address, split it. */
+		else if(pr->offset < offset + length
+			&& pr->offset + pr->ph->length > offset + length) {
+			split_phys(pr, offset + length);
+			physr_start_iter_least(region->phys, &iter);
+		}
+		else {
+			physr_incr_iter(&iter);
+		}
+	}
+}
+
+/*===========================================================================*
+ *				do_map_memory				     *
+ *===========================================================================*/
+PRIVATE int do_map_memory(struct vmproc *vms, struct vmproc *vmd,
+	struct vir_region *vrs, struct vir_region *vrd,
+	vir_bytes offset_s, vir_bytes offset_d,
+	vir_bytes length, int flag)
+{
+	struct phys_region *prs;
+	struct phys_region *newphysr;
+	struct phys_block *pb;
+	physr_iter iter;
+	u32_t pt_flag = PTF_PRESENT | PTF_USER;
+	vir_bytes end;
+
+	/* Search for the first phys region in the source process. */
+	physr_start_iter(vrs->phys, &iter, offset_s, AVL_EQUAL);
+	prs = physr_get_iter(&iter);
+	if(!prs)
+		vm_panic("map_memory: no aligned phys region.", 0);
+
+	/* flag: 0 -> read-only
+	 *       1 -> writable
+	 *      -1 -> share as COW, so read-only
+	 */
+	if(flag > 0)
+		pt_flag |= PTF_WRITE;
+
+	/* Map phys blocks in the source process to the destination process. */
+	end = offset_d + length;
+	while((prs = physr_get_iter(&iter)) && offset_d < end) {
+		/* If a SMAP share was requested but the phys block has already
+		 * been shared as COW, copy the block for the source phys region
+		 * first.
+		 */
+		pb = prs->ph;
+		if(flag >= 0 && pb->refcount > 1
+			&& pb->share_flag == PBSH_COW) {
+			map_copy_ph_block(vms, vrs, prs);
+			pb = prs->ph;
+		}
+
+		/* Allocate a new phys region. */
+		if(!SLABALLOC(newphysr))
+			return ENOMEM;
+
+		/* Set and link the new phys region to the block. */
+		newphysr->ph = pb;
+		newphysr->offset = offset_d;
+		newphysr->parent = vrd;
+		newphysr->next_ph_list = pb->firstregion;
+		pb->firstregion = newphysr;
+		physr_insert(newphysr->parent->phys, newphysr);
+		pb->refcount++;
+
+		/* If a COW share was requested but the phys block has already
+		 * been shared as SMAP, give up on COW and copy the block for
+		 * the destination phys region now.
+		 */
+		if(flag < 0 && pb->refcount > 1
+			&& pb->share_flag == PBSH_SMAP) {
+			map_copy_ph_block(vmd, vrd, newphysr);
+		}
+		else {
+			/* See if this is a COW share or SMAP share. */
+			if(flag < 0) {			/* COW share */
+				pb->share_flag = PBSH_COW;
+				/* Update the page table for the src process. */
+				pt_writemap(&vms->vm_pt, offset_s + vrs->vaddr,
+					pb->phys, pb->length,
+					pt_flag, WMF_OVERWRITE);
+			}
+			else {				/* SMAP share */
+				pb->share_flag = PBSH_SMAP;
+			}
+			/* Update the page table for the destination process. */
+			pt_writemap(&vmd->vm_pt, offset_d + vrd->vaddr,
+				pb->phys, pb->length, pt_flag, WMF_OVERWRITE);
+		}
+
+		physr_incr_iter(&iter);
+		offset_d += pb->length;
+		offset_s += pb->length;
+	}
+}
+
+/*===========================================================================*
+ *				map_memory				     *
+ *===========================================================================*/
+PUBLIC int map_memory(endpoint_t sour, endpoint_t dest,
+	vir_bytes virt_s, vir_bytes virt_d, vir_bytes length, int flag)
+{
+/* This is the entry point. This function will be called by handle_memory() when
+ * VM recieves a map-memory request.
+ */
+	struct vmproc *vms, *vmd;
+	struct vir_region *vrs, *vrd;
+	struct phys_region *prs, *prd;
+	physr_iter iters, iterd;
+	vir_bytes offset_s, offset_d;
+	int p;
+
+	if(vm_isokendpt(sour, &p) != OK)
+		vm_panic("handle_memory: endpoint wrong", sour);
+	vms = &vmproc[p];
+	if(vm_isokendpt(dest, &p) != OK)
+		vm_panic("handle_memory: endpoint wrong", dest);
+	vmd = &vmproc[p];
+
+	vrs = map_lookup(vms, virt_s);
+	vm_assert(vrs);
+	vrd = map_lookup(vmd, virt_d);
+	vm_assert(vrd);
+
+	/* Linear address -> offset from start of vir region. */
+	offset_s = virt_s - vrs->vaddr;
+	offset_d = virt_d - vrd->vaddr;
+
+	/* Make sure that the range in the source process has been mapped
+	 * to physical memory.
+	 */
+	map_handle_memory(vms, vrs, offset_s, length, 0);
+
+	/* Prepare work. */
+	#define map_printregion(x, y) (x = x)
+	#define printf(x, y, z) (z = z)
+	printf("before clean with offset, length: %d, %d\n", offset_s, length);
+	map_printregion(vms, vrs);
+	clean_phys_regions(vrs, offset_s, length);
+	printf("after clean with offset, length: %d, %d\n", offset_s, length);
+	map_printregion(vms, vrs);
+
+	printf("before clean with offset, length: %d, %d\n", offset_d, length);
+	map_printregion(vmd, vrd);
+	clean_phys_regions(vrd, offset_d, length);
+	printf("after clean with offset, length: %d, %d\n", offset_d, length);
+	map_printregion(vmd, vrd);
+
+	rm_phys_regions(vrd, offset_d, length);
+	printf("after rm with offset, length: %d, %d\n", offset_d, length);
+	map_printregion(vmd, vrd);
+
+	/* Map memory. */
+	do_map_memory(vms, vmd, vrs, vrd, offset_s, offset_d, length, flag);
+	printf("after map (dst) with offset, length: %d, %d\n", offset_d, length);
+	map_printregion(vmd, vrd);
+	#undef map_printregion
+	#undef printf
+
+	return OK;
+}
+
+/*===========================================================================*
+ *				unmap_memory				     *
+ *===========================================================================*/
+PUBLIC int unmap_memory(endpoint_t sour, endpoint_t dest,
+	vir_bytes virt_s, vir_bytes virt_d, vir_bytes length, int flag)
+{
+	struct vmproc *vmd;
+	struct vir_region *vrd;
+	struct phys_region *pr;
+	struct phys_block *pb;
+	physr_iter iter;
+	vir_bytes off, end;
+	int p;
+
+	/* Use information on the destination process to unmap. */
+	if(vm_isokendpt(dest, &p) != OK)
+		vm_panic("handle_memory: endpoint wrong", dest);
+	vmd = &vmproc[p];
+
+	vrd = map_lookup(vmd, virt_d);
+	vm_assert(vrd);
+
+	/* Search for the first phys region in the destination process. */
+	off = virt_d - vrd->vaddr;
+	physr_start_iter(vrd->phys, &iter, off, AVL_EQUAL);
+	pr = physr_get_iter(&iter);
+	if(!pr)
+		vm_panic("map_memory: no aligned phys region.", 0);
+
+	/* Copy the phys block now rather than doing COW. */
+	end = off + length;
+	while((pr = physr_get_iter(&iter)) && off < end) {
+		pb = pr->ph;
+		vm_assert(pb->refcount > 1);
+		vm_assert(pb->share_flag == PBSH_SMAP);
+
+		map_copy_ph_block(vmd, vrd, pr);
+
+		physr_incr_iter(&iter);
+		off += pb->length;
+	}
+
+	return OK;
+}
+

servers/vm/mmap.c

@@ -1,8 +1,6 @@
 
 #define _SYSTEM 1
 
-#define VERBOSE 0
-
 #include <minix/callnr.h>
 #include <minix/com.h>
 #include <minix/config.h>

servers/vm/pagefaults.c

@@ -131,22 +131,38 @@ PUBLIC void do_pagefaults(void)
  *===========================================================================*/
 PUBLIC void do_memory(void)
 {
-	int r, s;
+	endpoint_t who, who_s, requestor;
-	endpoint_t who, requestor;
+	vir_bytes mem, mem_s;
-	vir_bytes mem;
 	vir_bytes len;
 	int wrflag;
 
-	while((r=sys_vmctl_get_memreq(&who, &mem, &len, &wrflag, &requestor))
+	while(1) {
-	  == OK) {
 		int p, r = OK;
 		struct vmproc *vmp;
 
+		r = sys_vmctl_get_memreq(&who, &mem, &len, &wrflag, &who_s,
+			&mem_s, &requestor);
+
+		switch(r) {
+		case VMPTYPE_CHECK:
 			if(vm_isokendpt(who, &p) != OK)
-			vm_panic("do_memory: endpoint wrong", who);
+				vm_panic("do_memory: bad endpoint", who);
 			vmp = &vmproc[p];
 
 			r = handle_memory(vmp, mem, len, wrflag);
+			break;
+		case VMPTYPE_COWMAP:
+			r = map_memory(who_s, who, mem_s, mem, len, -1);
+			break;
+		case VMPTYPE_SMAP:
+			r = map_memory(who_s, who, mem_s, mem, len, wrflag);
+			break;
+		case VMPTYPE_SUNMAP:
+			r = unmap_memory(who_s, who, mem_s, mem, len, wrflag);
+			break;
+		default:
+			return;
+		}
 
 		if(sys_vmctl(requestor, VMCTL_MEMREQ_REPLY, r) != OK)
 			vm_panic("do_memory: sys_vmctl failed", r);

servers/vm/proto.h

@@ -4,6 +4,8 @@ struct vmproc;
 struct stat;
 struct mem_map;
 struct memory;
+struct vir_region;
+struct phys_region;
 
 #include <minix/ipc.h>
 #include <minix/endpoint.h>
@@ -76,6 +78,12 @@ _PROTOTYPE( int vfs_open, (struct vmproc *for_who, callback_t callback,
 _PROTOTYPE( int vfs_close, (struct vmproc *for_who, callback_t callback,
 	int fd));
 
+/* map_mem.c */
+_PROTOTYPE( int map_memory, (endpoint_t sour, endpoint_t dest,
+	vir_bytes virt_s, vir_bytes virt_d, vir_bytes length, int flag));
+_PROTOTYPE( int unmap_memory, (endpoint_t sour, endpoint_t dest,
+	vir_bytes virt_s, vir_bytes virt_d, vir_bytes length, int flag));
+
 /* mmap.c */
 _PROTOTYPE(int do_mmap, (message *msg)					);
 _PROTOTYPE(int do_munmap, (message *msg)				);
@@ -156,6 +164,10 @@ _PROTOTYPE(int map_remap, (struct vmproc *dvmp, vir_bytes da, size_t size,
 _PROTOTYPE(int map_get_phys, (struct vmproc *vmp, vir_bytes addr, phys_bytes *r));
 _PROTOTYPE(int map_get_ref, (struct vmproc *vmp, vir_bytes addr, u8_t *cnt));
 
+_PROTOTYPE(int map_copy_ph_block, (struct vmproc *vmp,
+	struct vir_region *region, struct phys_region *ph));
+_PROTOTYPE(void pb_unreferenced, (struct vir_region *region,
+	struct phys_region *pr));
 #if SANITYCHECKS
 _PROTOTYPE(void map_sanitycheck,(char *file, int line));
 #endif

servers/vm/queryexit.c

@@ -1,8 +1,6 @@
 
 #define _SYSTEM 1
 
-#define VERBOSE 0
-
 #include <minix/callnr.h>
 #include <minix/com.h>
 #include <minix/config.h>

servers/vm/region.c

@@ -40,8 +40,6 @@ FORWARD _PROTOTYPE(struct phys_region *map_new_physblock, (struct vmproc *vmp,
 FORWARD _PROTOTYPE(int map_ph_writept, (struct vmproc *vmp, struct vir_region *vr,
 	struct phys_region *pr));
 
-FORWARD _PROTOTYPE(int map_copy_ph_block, (struct vmproc *vmp, struct vir_region *region, struct phys_region *ph));
-
 FORWARD _PROTOTYPE(struct vir_region *map_copy_region, (struct vmproc *vmp, struct vir_region *vr));
 
 PRIVATE char *map_name(struct vir_region *vr)
@@ -422,7 +420,7 @@ USE(newregion,
 /*===========================================================================*
  *				pb_unreferenced				     *
  *===========================================================================*/
-void pb_unreferenced(struct vir_region *region, struct phys_region *pr)
+PUBLIC void pb_unreferenced(struct vir_region *region, struct phys_region *pr)
 {
 	struct phys_block *pb;
 	int remap = 0;
@@ -460,20 +458,6 @@ void pb_unreferenced(struct vir_region *region, struct phys_region *pr)
 			vm_panic("strange phys flags", NO_NUM);
 		}
 		SLABFREE(pb);
-	} else if(WRITABLE(region, pb)) {
-		/* If a writable piece of physical memory is now only
-		 * referenced once, map it writable right away instead of
-		 * waiting for a page fault.
-		 */
-			vm_assert(pb);
-			vm_assert(pb->firstregion);
-			vm_assert(!pb->firstregion->next_ph_list);
-			vm_assert(pb->firstregion->ph == pb);
-			vm_assert(pb->firstregion->ph == pb);
-			if(map_ph_writept(pb->firstregion->parent->parent,
-				pb->firstregion->parent, pb->firstregion) != OK) {
-				vm_panic("pb_unreferenced: writept", NO_NUM);
-			}
 	}
 }
 
@@ -709,7 +693,7 @@ phys_bytes what_mem;
 /*===========================================================================*
  *				map_copy_ph_block			     *
  *===========================================================================*/
-PRIVATE int map_copy_ph_block(vmp, region, ph)
+PUBLIC int map_copy_ph_block(vmp, region, ph)
 struct vmproc *vmp;
 struct vir_region *region;
 struct phys_region *ph;
@@ -815,10 +799,14 @@ int write;
 			r = map_ph_writept(vmp, region, ph);
 			if(r != OK)
 				printf("map_ph_writept failed\n");
+		} else {
+			if(ph->ph->refcount > 0
+				&& ph->ph->share_flag != PBSH_COW) {
+				printf("VM: write RO mapped pages.\n");
+				return EFAULT;
 			} else {
 				r = map_copy_ph_block(vmp, region, ph);
-			if(r != OK)
+			}
-				printf("map_copy_ph_block failed\n");
 		}
 	} else {
 		/* Pagefault in non-existing block. Map in new block. */
@@ -959,9 +947,12 @@ int write;
 	SANITYCHECK(SCL_FUNCTIONS);
 
 	if(changes < 1) {
+#if VERBOSE
 		printf("region start at 0x%lx offset 0x%lx len 0x%lx write %d\n", 
 			region->vaddr, offset, length, write);
-		vm_panic("no changes in map_handle_memory", NO_NUM);
+		printf("no changes in map_handle_memory\n");
+#endif
+		return EFAULT;
 	}
 
 #if SANITYCHECKS
@@ -1065,11 +1056,19 @@ PUBLIC int map_writept(struct vmproc *vmp)
 		physr_iter iter;
 		physr_start_iter_least(vr->phys, &iter);
 		while((ph = physr_get_iter(&iter))) {
+			physr_incr_iter(&iter);
+
+			/* If this phys block is shared as SMAP, then do
+			 * not update the page table. */
+			if(ph->ph->refcount > 1
+				&& ph->ph->share_flag == PBSH_SMAP) {
+				continue;
+			}
+
 			if((r=map_ph_writept(vmp, vr, ph)) != OK) {
 				printf("VM: map_writept: failed\n");
 				return r;
 			}
-			physr_incr_iter(&iter);
 		}
 	}
 
@@ -1125,6 +1124,14 @@ struct vmproc *src;
 			USE(pb, pb->refcount++;);
 			vm_assert(pb->refcount > 1);
 
+			/* If the phys block has been shared as SMAP,
+			 * do the regular copy. */
+			if(pb->refcount > 2 && pb->share_flag == PBSH_SMAP) {
+				map_copy_ph_block(dst, newvr, new_ph);
+			} else {
+				pb->share_flag = PBSH_COW;
+			}
+
 			/* Get next new physregion */
 			physr_incr_iter(&iter_orig);
 			physr_incr_iter(&iter_new);

servers/vm/region.h

@@ -23,6 +23,9 @@ struct phys_block {
 	vir_bytes		length;	/* no. of contiguous bytes */
 	phys_bytes		phys;	/* physical memory */
 	u8_t			refcount;	/* Refcount of these pages */
+#define PBSH_COW	1
+#define PBSH_SMAP	2
+	u8_t			share_flag;	/* PBSH_COW or PBSH_SMAP */
 
 	/* first in list of phys_regions that reference this block */
 	struct phys_region	*firstregion;	

servers/vm/rs.c

@@ -1,8 +1,6 @@
 
 #define _SYSTEM 1
 
-#define VERBOSE 0
-
 #include <minix/callnr.h>
 #include <minix/com.h>
 #include <minix/config.h>

servers/vm/utility.c

@@ -76,8 +76,8 @@ struct memory *mem_chunks;                      /* store mem chunks here */
         base = mem_chunks[i].base;
         size = mem_chunks[i].size;
         limit = base + size;
-        base = (base + CLICK_SIZE-1) & ~(long)(CLICK_SIZE-1);
+        base = (phys_bytes) (CLICK_CEIL(base));
-        limit &= ~(long)(CLICK_SIZE-1);
+        limit = (phys_bytes) (CLICK_FLOOR(limit));
         if (limit <= base) {
                 memp->base = memp->size = 0;
         } else { 

servers/vm/vfs.c

@@ -1,8 +1,6 @@
 
 #define _SYSTEM 1
 
-#define VERBOSE 0
-
 #include <minix/callnr.h>
 #include <minix/com.h>
 #include <minix/config.h>

test/ds/Makefile

@@ -0,0 +1,18 @@
+all: dstest subs
+	chmod +x down run
+
+dstest: dstest.c inc.h
+	cc -o $@ $< -lsys
+
+subs: subs.c inc.h
+	cc -o $@ $< -lsys
+
+run: all
+	sh run
+
+kill:
+	sh down
+
+clean:
+	rm -f dstest subs
+

test/ds/README

@@ -0,0 +1,23 @@
+Test Program for DataServer
+
+How to run
+==========
+
+  1. Type `make run` to prepare and run test.
+  2. When done testing, type `make clean` to clean up.
+
+How the test works
+==================
+
+`dstest` tests the new DS API (excluding ds_subscribe() and ds_check()).
+test_u32, test_str, test_mem, test_label, and test_map test U32, STR, MEM, LABEL
+and MAP type respectively.
+
+Invalid invokation is tested as well. Erroneous conditions are tested only once.
+For example, publishing an entry with same label name, but without 
+DSF_OVERWRITE set, is tested in test_u32 and nowhere else again.
+Type-specific features are always tested.
+
+`subs` tests ds_subscribe() and ds_check(). The server subscribes
+to a U32 type. When `dstest` runs, `subs` catches all the updates.
+

test/ds/down

@@ -0,0 +1,4 @@
+#!/bin/sh
+
+service down dstest
+service down subs

test/ds/dstest.c

@@ -0,0 +1,230 @@
+#include "inc.h"
+
+char *key_u32 = "test_u32";
+char *key_str = "test_str";
+char *key_mem = "test_mem";
+char *key_map = "test_map";
+char *key_label = "test_label";
+
+/*===========================================================================*
+ *				test_u32				     *
+ *===========================================================================*/
+void test_u32(void)
+{
+	int r;
+	unsigned long value;
+
+	/* Publish and retrieve. */
+	r = ds_publish_u32(key_u32, 1234, 0);
+	assert(r == OK);
+	r = ds_retrieve_u32(key_u32, &value);
+	assert(r == OK && value == 1234);
+
+	/* If dstest deletes 'test_u32' immediately after publishing it,
+	 * subs will catch the event, but it can't check it immediately.
+	 * So dstest will sleep 2 seconds to wait for subs to complete. */
+	sleep(2);
+
+	/* Publish again without DSF_OVERWRITE. */
+	r = ds_publish_u32(key_u32, 4321, 0);
+	assert(r == EEXIST);
+
+	/* Publish again with DSF_OVERWRITE to overwrite it. */
+	r = ds_publish_u32(key_u32, 4321, DSF_OVERWRITE);
+	assert(r == OK);
+	r = ds_retrieve_u32(key_u32, &value);
+	assert(r == OK && value == 4321);
+
+	/* Delete. */
+	r = ds_delete_u32(key_u32);
+	assert(r == OK);
+	r = ds_retrieve_u32(key_u32, &value);
+	assert(r == ESRCH);
+
+	printf("DSTEST: U32 test successful!\n");
+}
+
+/*===========================================================================*
+ *				test_str				     *
+ *===========================================================================*/
+void test_str(void)
+{
+	int r;
+	char *string = "little";
+	char *longstring = "verylooooooongstring";
+	char buf[17];
+
+	r = ds_publish_str(key_str, string, 0);
+	assert(r == OK);
+
+	r = ds_retrieve_str(key_str, buf, 0);
+	assert(r == OK && strcmp(string, buf) == 0);
+
+	r = ds_delete_str(key_str);
+	assert(r == OK);
+
+	/* Publish a long string. */
+	r = ds_publish_str(key_str, longstring, 0);
+	assert(r == EINVAL);
+
+	printf("DSTEST: STRING test successful!\n");
+}
+
+/*===========================================================================*
+ *				test_mem				     *
+ *===========================================================================*/
+void test_mem(void)
+{
+	char *string1 = "ok, this is a string";
+	char *string2 = "ok, this is a very looooong string";
+	size_t len1 = strlen(string1) + 1;
+	size_t len2 = strlen(string2) + 1;
+	char buf[100];
+	size_t get_len;
+	int r;
+
+	/* Publish and retrieve. */
+	r = ds_publish_mem(key_mem, string1, len1, 0);
+	assert(r == OK);
+	get_len = 100;
+	r = ds_retrieve_mem(key_mem, buf, &get_len);
+	assert(r == OK && strcmp(string1, buf) == 0);
+	assert(get_len == len1);
+
+	/* Let get_len=8, which is less than strlen(string1). */
+	get_len = 8;
+	r = ds_retrieve_mem(key_mem, buf, &get_len);
+	assert(r == OK && get_len == 8);
+
+	/* Publish again to overwrite with a bigger memory range. */
+	r = ds_publish_mem(key_mem, string2, len2, DSF_OVERWRITE);
+	assert(r == OK);
+
+	get_len = 100;
+	r = ds_retrieve_mem(key_mem, buf, &get_len);
+	assert(r == OK && strcmp(string2, buf) == 0);
+	assert(get_len == len2);
+
+	r = ds_delete_mem(key_mem);
+	assert(r == OK);
+
+	printf("DSTEST: MEM test successful!\n");
+}
+
+/*===========================================================================*
+ *				test_label				     *
+ *===========================================================================*/
+void test_label(void)
+{
+	int r;
+	char get_label[DS_MAX_KEYLEN];
+	unsigned long num;
+
+	/* Publish and retrieve. */
+	r = ds_publish_label(key_label, 1234, 0);
+	assert(r == OK);
+	r = ds_retrieve_label_num(key_label, &num);
+	assert(r == OK && num == 1234);
+
+	/* Here are the differences w.r.t. U32. */
+	r = ds_publish_label("hello", 1234, 0);
+	assert(r == EEXIST);
+	r = ds_retrieve_label_name(get_label, 1234);
+	assert(r == OK && strcmp(key_label, get_label) == 0);
+
+	r = ds_delete_label(key_label);
+	assert(r == OK);
+
+	printf("DSTEST: LABEL test successful!\n");
+}
+
+/*===========================================================================*
+ *				test_map				     *
+ *===========================================================================*/
+void test_map(void)
+{
+	char buf_buf[CLICK_SIZE * 2];
+	char buf_buf2[CLICK_SIZE * 2];
+	char *buf, *buf2;
+	char get_buf[CLICK_SIZE];
+	int *p;
+	volatile int *p2;
+	int *get_p;
+	size_t get_len;
+	int is;
+	int r;
+
+	buf = (char*) CLICK_CEIL(buf_buf);
+	buf2 = (char*) CLICK_CEIL(buf_buf2);
+
+	p = (int *)buf;
+	p2 = (int *)buf2;
+	get_p = (int *)get_buf;
+
+	*p = 1;
+	r = ds_publish_map(key_map, buf, CLICK_SIZE, 0);
+	assert(r == OK);
+
+	r = ds_snapshot_map(key_map, &is);
+	assert(r == OK);
+
+	/* Copy the mapped memory range.
+	 * Set *p=2, then the mapped memory range should change too
+	 * and *get_p should be 2.
+	 */
+	*p = 2;
+	get_len = CLICK_SIZE;
+	r = ds_retrieve_map(key_map, get_buf, &get_len, 0, DSMF_COPY_MAPPED);
+	assert(r == OK && get_len == CLICK_SIZE && *get_p == 2);
+
+	/* Copy snapshot, where *get_p should still be 1. */
+	get_len = CLICK_SIZE;
+	r = ds_retrieve_map(key_map, get_buf, &get_len, is, DSMF_COPY_SNAPSHOT);
+	assert(r == OK && get_len == CLICK_SIZE && *get_p == 1);
+
+	/* Map the mapped memory range to @buf2, then set *p=3, which
+	 * in turn should let *p2=3.
+	 */
+	get_len = CLICK_SIZE;
+	r = ds_retrieve_map(key_map, buf2, &get_len, 0, DSMF_MAP_MAPPED);
+	assert(r == OK && get_len == CLICK_SIZE);
+	*p = 3;
+	assert(*p2 == 3);
+
+	r = ds_delete_map(key_map);
+	assert(r == OK);
+
+	printf("DSTEST: MAP test successful!\n");
+}
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *				main					     *
+ *===========================================================================*/
+int main(void)
+{
+	/* SEF local startup. */
+	sef_local_startup();
+
+	/* Run all the tests. */
+	test_u32();
+	test_str();
+	test_mem();
+	test_map();
+	test_label();
+
+	return 0;
+}
+
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/ds/inc.h

@@ -0,0 +1,15 @@
+#define _MINIX
+#define _SYSTEM
+#include <stdio.h>
+#include <unistd.h>
+#include <assert.h>
+#include <errno.h>
+#include <string.h>
+
+#include <minix/config.h>
+#include <minix/com.h>
+#include <minix/type.h>
+#include <minix/const.h>
+#include <minix/endpoint.h>
+#include <minix/ds.h>
+#include <minix/syslib.h>

test/ds/run

@@ -0,0 +1,6 @@
+#!/bin/sh
+
+PWD=`pwd`
+
+service up ${PWD}/subs -config ${PWD}/system.conf -script ${PWD}/down
+service up ${PWD}/dstest -config ${PWD}/system.conf -script ${PWD}/down

test/ds/subs.c

@@ -0,0 +1,93 @@
+#include "inc.h"
+
+char *key_u32 = "test_u32";
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *				main					     *
+ *===========================================================================*/
+int main(void)
+{
+	int r;
+	message mess;
+	char key[DS_MAX_KEYLEN];
+	int type;
+	unsigned long num;
+	char string[17];
+	char buf[1000];
+	size_t length = 1000;
+
+	/* SEF local startup. */
+	sef_local_startup();
+
+	/* Subscribe. */
+	r = ds_subscribe(key_u32, DSF_INITIAL);
+	if(r != OK && r != EEXIST) {
+		printf("SUBSCRIBER: error in ds_subscribe: %d\n", r);
+		return -1;
+	}
+
+	while(1) {
+		/* Wait for a message. */
+		r = sef_receive(ANY, &mess);
+		if(r != OK) {
+			printf("SUBSCRIBER: sef_receive failed.\n");
+			return 1;
+		}
+		/* Only handle notifications from DS. */
+		if(mess.m_source != DS_PROC_NR)
+			continue;
+
+		/* Check which one was changed. */
+		r = ds_check(key, &type);
+		if(r == ENOENT) {
+			printf("SUBSCRIBER: the key %s was deleted.\n",
+				key);
+			continue;
+		}
+		if(r != OK) {
+			printf("SUBSCRIBER: error in ds_check.\n");
+			continue;
+		}
+
+		/* Retrieve the entry. */
+		printf("SUBSCRIBER: key: %s, ", key);
+		switch(type) {
+		case DSF_TYPE_U32:
+			r = ds_retrieve_u32(key, &num);
+			if(r != OK)
+				printf("error in ds_retrieve_u32.\n");
+			printf("U32: %d\n", num);
+			break;
+		case DSF_TYPE_STR:
+			r = ds_retrieve_str(key, string, 0);
+			if(r != OK)
+				printf("error in ds_retrieve_str.\n");
+			printf("STR: %s\n", string);
+			break;
+		case DSF_TYPE_MEM:
+			r = ds_retrieve_mem(key, buf, &length);
+			if(r != OK)
+				printf("error in ds_retrieve_mem.\n");
+			break;
+		case DSF_TYPE_MAP:
+			break;
+		default:
+			printf("error in type! %d\n", type);
+		}
+	}
+
+	return 0;
+}
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/ds/system.conf

@@ -0,0 +1,25 @@
+
+service dstest
+{
+	system
+		UMAP
+		PRIVCTL
+		SAFEMAP
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};
+
+service subs
+{
+	system
+		UMAP
+		PRIVCTL
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};

test/safecopy/Makefile

@@ -0,0 +1,21 @@
+all: requestor grantor 1fifo 2fifo
+	chmod +x down run
+
+requestor: requestor.c inc.h
+	cc -o $@ $< -lsys
+
+grantor: grantor.c inc.h
+	cc -o $@ $< -lsys
+
+1fifo 2fifo:
+	mkfifo $@
+
+run: all
+	sh run
+
+kill:
+	sh down
+
+clean:
+	rm -f grantor requestor 1fifo 2fifo
+

test/safecopy/README

@@ -0,0 +1,19 @@
+Test Program for Safecopy with or without COW
+
+How to run
+==========
+
+Test safecopy with COW:
+  1. Set USE_COW_SAFECOPY to 1 in /usr/src/kernel/system/do_safecopy.c
+  2. Type `make run` to prepare and run test.
+  3. When done testing, type `make clean` to clean up.
+
+Test safecopy without COW:
+  1. Set USE_COW_SAFECOPY to 0 in /usr/src/kernel/system/do_safecopy.c
+  2. Type `make run` to prepare and run test.
+  3. When done testing, type `make clean` to clean up.
+
+If you want to modify the test buffer size:
+  1. Modify TEST_PAGE_NUM in inc.h
+  2. Modify TEST_PAGE_SHIFT in inc.h
+

test/safecopy/down

@@ -0,0 +1,5 @@
+#!/bin/sh
+
+service down grantor
+service down requestor
+

test/safecopy/grantor.c

@@ -0,0 +1,59 @@
+#include "inc.h"
+
+char buf_buf[BUF_SIZE + CLICK_SIZE];
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *				    main				     *
+ *===========================================================================*/
+int main(int argc, char **argv)
+{
+	endpoint_t ep_self, ep_requestor;
+	int fid_send, fid_get;
+	cp_grant_id_t gid;
+	char *buf;
+	int i;
+
+	/* SEF local startup. */
+	env_setargs(argc, argv);
+	sef_local_startup();
+
+	/* Prepare work. */
+	buf = (char*) CLICK_CEIL(buf_buf);
+	fid_send = open(FIFO_GRANTOR, O_WRONLY);
+	fid_get = open(FIFO_REQUESTOR, O_RDONLY);
+	if(fid_get < 0 || fid_send < 0) {
+		printf("GRANTOR: can't open fifo files.\n");
+		return 1;
+	}
+	buf[0] = BUF_START;
+
+	/* Get the requestor's endpoint. */
+	read(fid_get, &ep_requestor, sizeof(ep_requestor));
+	dprint("GRANTOR: getting requestor's endpoint: %d\n", ep_requestor);
+
+	/* Grant. */
+	gid = cpf_grant_direct(ep_requestor, (long)buf, BUF_SIZE,
+		CPF_READ | CPF_WRITE);
+	ep_self = getprocnr();
+	dprint("GRANTOR: sending my endpoint %d and gid %d\n", ep_self, gid);
+	write(fid_send, &ep_self, sizeof(ep_self));
+	write(fid_send, &gid, sizeof(gid));
+
+	/* Wait till requestor is done. */
+	FIFO_WAIT(fid_get);
+
+	return 0;
+}
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/safecopy/inc.h

@@ -0,0 +1,50 @@
+#define _SYSTEM
+#define _MINIX
+#include <stdio.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+#include <minix/config.h>
+#include <minix/com.h>
+#include <minix/type.h>
+#include <minix/const.h>
+#include <minix/endpoint.h>
+#include <minix/safecopies.h>
+#include <minix/syslib.h>
+#include <minix/sysutil.h>
+#include <errno.h>
+
+/* TEST_PAGE_SHIFT =
+ * log2(CLICK_SIZE * TEST_PAGE_NUM) = CLICK_SHIFT + log2(TEST_PAGE_NUM)
+ */
+#define TEST_PAGE_NUM    8
+#define TEST_PAGE_SHIFT 15
+
+#define BUF_SIZE (TEST_PAGE_NUM * CLICK_SIZE)
+#define BUF_START 100
+
+#define FIFO_REQUESTOR "/usr/src/test/safecopy/1fifo"
+#define FIFO_GRANTOR   "/usr/src/test/safecopy/2fifo"
+
+#define FIFO_WAIT(fid) {                                                      \
+	int a;                                                                \
+	if(read(fid, &a, sizeof(a)) != sizeof(a))                             \
+		panic(__FILE__, "FIFO_WAIT failed", NO_NUM);                  \
+}
+#define FIFO_NOTIFY(fid) {                                                    \
+	int a = 1;                                                            \
+	if(write(fid, &a, sizeof(a)) != sizeof(a))                            \
+		panic(__FILE__, "FIFO_NOTIFY failed", NO_NUM);                \
+}
+
+#define DEBUG 0
+#if DEBUG
+#	define dprint printf
+#else
+#	define dprint (void)
+#endif
+

test/safecopy/requestor.c

@@ -0,0 +1,96 @@
+#include "inc.h"
+
+char buf_buf[BUF_SIZE + CLICK_SIZE];
+
+endpoint_t ep_granter;
+int gid;
+char *buf;
+
+/*===========================================================================*
+ *				    test				     *
+ *===========================================================================*/
+int test(size_t size)
+{
+	u32_t low1, high1;
+	u32_t low2, high2;
+	int r;
+
+	/* Timing. */
+	read_tsc(&high1, &low1);
+	r = sys_safecopyfrom(ep_granter, gid, 0, (long)buf, size, D);
+	read_tsc(&high2, &low2);
+	if(r != OK) {
+		printf("REQUESTOR: error in safecopy: %d\n", r);
+		return r;
+	}
+	printf("REQUESTOR: SAFECOPY 0x%-8x - %d\n", size, low2 -  low1);
+
+	/* Test. */
+	if(buf[0] != BUF_START) {
+		printf("REQUESTOR: error in safecopy!\n");
+		printf("	size, value: %d, %d\n", size, buf[0]);
+		return r;
+	}
+
+	return OK;
+}
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *				    main				     *
+ *===========================================================================*/
+int main(int argc, char **argv)
+{
+	endpoint_t ep_self;
+	int fid_send, fid_get;
+	int i;
+
+	/* SEF local startup. */
+	env_setargs(argc, argv);
+	sef_local_startup();
+
+	/* Prepare work. */
+	buf = (char*) CLICK_CEIL(buf_buf);
+	fid_get = open(FIFO_GRANTOR, O_RDONLY);
+	fid_send = open(FIFO_REQUESTOR, O_WRONLY);
+	if(fid_get < 0 || fid_send < 0) {
+		printf("REQUESTOR: can't open fifo files.\n");
+		return 1;
+	}
+
+	/* Sending the endpoint to the granter, in order to let him
+	 * create the grant.
+	 */
+	ep_self = getprocnr();
+	write(fid_send, &ep_self, sizeof(ep_self));
+	dprint("REQUESTOR: sending my endpoint: %d\n", ep_self);
+
+	/* Getting the granter's endpoint and gid. */
+	read(fid_get, &ep_granter, sizeof(ep_granter));
+	read(fid_get, &gid, sizeof(gid));
+	dprint("REQUESTOR: getting granter's endpoint %d and gid %d\n",
+		ep_granter, gid);
+
+	/* Test SAFECOPY. */
+	for(i = 0; i <= TEST_PAGE_SHIFT; i++) {
+		if(test(1 << i) != OK)
+			break;
+	}
+
+	/* Notify grantor we are done. */
+	FIFO_NOTIFY(fid_send);
+
+	return 0;
+}
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/safecopy/run

@@ -0,0 +1,7 @@
+#!/bin/sh
+
+PWD=`pwd`
+
+service up ${PWD}/grantor -config ${PWD}/system.conf -script ${PWD}/down
+service up ${PWD}/requestor -config ${PWD}/system.conf -script ${PWD}/down
+

test/safecopy/system.conf

@@ -0,0 +1,34 @@
+
+service requestor
+{
+	system
+		SAFECOPYFROM
+		SAFECOPYTO
+		SAFEMAP
+		SAFEREVMAP
+		SAFEUNMAP
+		UMAP
+		PRIVCTL
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};
+
+service grantor
+{
+	system
+		SAFECOPYFROM
+		SAFECOPYTO
+		SAFEMAP
+		SAFEREVMAP
+		SAFEUNMAP
+		UMAP
+		PRIVCTL
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};

test/safemap/Makefile

@@ -0,0 +1,21 @@
+all: requestor grantor 1fifo 2fifo
+	chmod +x down run
+
+requestor: requestor.c inc.h
+	cc -o $@ $< -lsys
+
+grantor: grantor.c inc.h
+	cc -o $@ $< -lsys
+
+1fifo 2fifo:
+	mkfifo $@
+
+run: all
+	sh run
+
+kill:
+	sh down
+
+clean:
+	rm -f grantor requestor 1fifo 2fifo
+

test/safemap/README

@@ -0,0 +1,8 @@
+Test Program for Safemap
+
+How to run
+==========
+
+  1. Type `make run` to prepare and run test.
+  2. When done testing, type `make clean` to clean up.
+

test/safemap/down

@@ -0,0 +1,4 @@
+#!/bin/sh
+
+service down grantor
+service down requestor

test/safemap/grantor.c

@@ -0,0 +1,129 @@
+#include "inc.h"
+
+char buf_buf[BUF_SIZE + CLICK_SIZE];
+
+int fid_send, fid_get;
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *				    main				     *
+ *===========================================================================*/
+int main(int argc, char **argv)
+{
+	endpoint_t ep_self, ep_requestor, ep_child;
+	cp_grant_id_t gid;
+	int i, r, pid;
+	char *buf;
+	int status;
+
+	/* SEF local startup. */
+	env_setargs(argc, argv);
+	sef_local_startup();
+
+	/* Prepare work. */
+	buf = (char*) CLICK_CEIL(buf_buf);
+	fid_send = open(FIFO_GRANTOR, O_WRONLY);
+	fid_get = open(FIFO_REQUESTOR, O_RDONLY);
+	if(fid_get < 0 || fid_send < 0) {
+		printf("GRANTOR: can't open fifo files.\n");
+		return 1;
+	}
+
+	/* Get the requestor's endpoint. */
+	read(fid_get, &ep_requestor, sizeof(ep_requestor));
+	dprint("GRANTOR: getting requestor's endpoint: %d\n", ep_requestor);
+
+	/* Grant. */
+	gid = cpf_grant_direct(ep_requestor, (long)buf, BUF_SIZE,
+		CPF_READ | CPF_WRITE | CPF_MAP);
+	ep_self = getprocnr();
+	dprint("GRANTOR: sending my endpoint %d and gid %d\n", ep_self, gid);
+	write(fid_send, &ep_self, sizeof(ep_self));
+	write(fid_send, &gid, sizeof(gid));
+
+	/* Test MAP. */
+	buf[0] = BUF_START_GRANTOR;
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("GRANTOR", buf[0], BUF_START_REQUESTOR, "MAP");
+
+	/* Test UNMAP. */
+	buf[0] = BUF_START_GRANTOR;
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("GRANTOR", buf[0], BUF_START_GRANTOR, "UNMAP");
+
+	/* Test REVOKE. */
+	r = sys_saferevmap_gid(gid);
+	if(r != OK) {
+		printf("GRANTOR: error in sys_saferevmap_gid: %d\n", r);
+		return 1;
+	}
+	buf[0] = BUF_START_GRANTOR+1;
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("GRANTOR", buf[0], BUF_START_GRANTOR+1, "REVOKE");
+
+	/* Test SMAP_COW. */
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	buf[0] = BUF_START_GRANTOR;
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("GRANTOR", buf[0], BUF_START_GRANTOR, "SMAP_COW");
+
+	/* Test COW_SMAP. */
+	r = sys_saferevmap_gid(gid);
+	if(r != OK) {
+		printf("GRANTOR: error in sys_saferevmap_gid: %d\n", r);
+		return 1;
+	}
+	buf[0] = BUF_START_GRANTOR+1;
+	pid = fork();
+	if(pid < 0) {
+		printf("GRANTOR: error in fork.\n");
+		return 1;
+	}
+	if(pid == 0) {
+		buf[0] = BUF_START_GRANTOR+2;
+		exit(0);
+	}
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	ep_child = getnprocnr(pid);
+	if ((r = sys_privctl(ep_child, SYS_PRIV_SET_USER, NULL)) != OK) {
+		printf("GRANTOR: unable to set privileges: %d\n", r);
+		return 1;
+	}
+	if ((r = sys_privctl(ep_child, SYS_PRIV_ALLOW, NULL)) != OK) {
+		printf("GRANTOR: child process can't run: %d\n", r);
+		return 1;
+	}
+	wait(&status);
+	FIFO_NOTIFY(fid_send);
+	CHECK_TEST("GRANTOR", buf[0], BUF_START_GRANTOR+1, "COW_SMAP");
+
+	/* Test COW_SMAP2 (with COW safecopy). */
+	r = sys_saferevmap_gid(gid);
+	if(r != OK) {
+		printf("GRANTOR: error in sys_saferevmap_gid: %d\n", r);
+		return 1;
+	}
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("GRANTOR", buf[0], BUF_START_REQUESTOR, "COW_SMAP2");
+
+	return 0;
+}
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/safemap/inc.h

@@ -0,0 +1,54 @@
+#define _SYSTEM
+#define _MINIX
+#include <stdio.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+#include <minix/config.h>
+#include <minix/com.h>
+#include <minix/type.h>
+#include <minix/const.h>
+#include <minix/endpoint.h>
+#include <minix/safecopies.h>
+#include <minix/syslib.h>
+#include <minix/sysutil.h>
+#include <errno.h>
+
+#define TEST_PAGE_NUM    4
+#define BUF_SIZE (TEST_PAGE_NUM * CLICK_SIZE)
+#define BUF_START_REQUESTOR 10
+#define BUF_START_GRANTOR   20
+
+#define FIFO_REQUESTOR "/usr/src/test/safemap/1fifo"
+#define FIFO_GRANTOR   "/usr/src/test/safemap/2fifo"
+
+#define FIFO_WAIT(fid) {                                                      \
+	int a;                                                                \
+	if(read(fid, &a, sizeof(a)) != sizeof(a))                             \
+		panic(__FILE__, "FIFO_WAIT failed", NO_NUM);                  \
+}
+#define FIFO_NOTIFY(fid) {                                                    \
+	int a = 1;                                                            \
+	if(write(fid, &a, sizeof(a)) != sizeof(a))                            \
+		panic(__FILE__, "FIFO_NOTIFY failed", NO_NUM);                \
+}
+
+#define CHECK_TEST(who, result, expected, test_name) {                        \
+	printf("%-9s: test %s %s\n", who, test_name,                          \
+		(expected == result ? "succeeded" : "failed"));               \
+	if(expected != result) {                                              \
+		exit(1);                                                      \
+	}                                                                     \
+}
+
+#define DEBUG 0
+#if DEBUG
+#	define dprint printf
+#else
+#	define dprint (void)
+#endif
+

test/safemap/requestor.c

@@ -0,0 +1,153 @@
+#include "inc.h"
+
+char buf_buf[BUF_SIZE + CLICK_SIZE];
+
+endpoint_t ep_granter;
+cp_grant_id_t gid;
+char *buf;
+int fid_send, fid_get;
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *				    main				     *
+ *===========================================================================*/
+int main(int argc, char **argv)
+{
+	endpoint_t ep_self, ep_child;
+	size_t size = BUF_SIZE;
+	int i, r, pid;
+	int status;
+
+	/* SEF local startup. */
+	env_setargs(argc, argv);
+	sef_local_startup();
+
+	/* Prepare work. */
+	buf = (char*) CLICK_CEIL(buf_buf);
+	fid_get = open(FIFO_GRANTOR, O_RDONLY);
+	fid_send = open(FIFO_REQUESTOR, O_WRONLY);
+	if(fid_get < 0 || fid_send < 0) {
+		printf("REQUESTOR: can't open fifo files.\n");
+		return 1;
+	}
+
+	/* Send the endpoint to the granter, in order to let him to
+	 * create the grant.
+	 */
+	ep_self = getprocnr();
+	write(fid_send, &ep_self, sizeof(ep_self));
+	dprint("REQUESTOR: sending my endpoint: %d\n", ep_self);
+
+	/* Get the granter's endpoint and gid. */
+	read(fid_get, &ep_granter, sizeof(ep_granter));
+	read(fid_get, &gid, sizeof(gid));
+	dprint("REQUESTOR: getting granter's endpoint %d and gid %d\n",
+		ep_granter, gid);
+
+	/* Test MAP. */
+	FIFO_WAIT(fid_get);
+	r = sys_safemap(ep_granter, gid, 0, (long)buf, size, D, 1);
+	if(r != OK) {
+		printf("REQUESTOR: error in sys_safemap: %d\n", r);
+		return 1;
+	}
+	CHECK_TEST("REQUESTOR", buf[0], BUF_START_GRANTOR, "MAP");
+	buf[0] = BUF_START_REQUESTOR;
+	r = sys_safeunmap(D, (long)buf);
+	if(r != OK) {
+		printf("REQUESTOR: error in sys_safeunmap: %d\n", r);
+		return 1;
+	}
+	FIFO_NOTIFY(fid_send);
+
+	/* Test UNMAP. */
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("REQUESTOR", buf[0], BUF_START_REQUESTOR, "UNMAP");
+	r = sys_safemap(ep_granter, gid, 0, (long)buf, size, D, 1);
+	if(r != 0) {
+		printf("REQUESTOR: error in sys_safemap: %d\n", r);
+		return 1;
+	}
+	FIFO_NOTIFY(fid_send);
+
+	/* Test REVOKE. */
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("REQUESTOR", buf[0], BUF_START_GRANTOR, "REVOKE");
+	buf[0] = BUF_START_REQUESTOR;
+	FIFO_NOTIFY(fid_send);
+
+	/* Test SMAP_COW. */
+	FIFO_WAIT(fid_get);
+	r = sys_safemap(ep_granter, gid, 0, (long)buf, size, D, 1);
+	if(r != OK) {
+		printf("REQUESTOR: error in sys_safemap: %d\n", r);
+		return 1;
+	}
+	buf[0] = BUF_START_REQUESTOR;
+	pid = fork();
+	if(pid < 0) {
+		printf("REQUESTOR: error in fork\n");
+		return 1;
+	}
+	if(pid == 0) {
+		exit(buf[0] != BUF_START_REQUESTOR);
+	}
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	ep_child = getnprocnr(pid);
+	if ((r = sys_privctl(ep_child, SYS_PRIV_SET_USER, NULL)) != OK) {
+		printf("REQUESTOR: unable to set privileges: %d\n", r);
+		return 1;
+	}
+	if ((r = sys_privctl(ep_child, SYS_PRIV_ALLOW, NULL)) != OK) {
+		printf("REQUESTOR: child process can't run: %d\n", r);
+		return 1;
+	}
+	wait(&status);
+	FIFO_NOTIFY(fid_send);
+	CHECK_TEST("REQUESTOR", buf[0], BUF_START_GRANTOR, "SMAP_COW");
+	CHECK_TEST("REQUESTOR", 1, WIFEXITED(status)
+		&& (WEXITSTATUS(status) == 0), "SMAP_COW child");
+
+	/* Test COW_SMAP. */
+	FIFO_WAIT(fid_get);
+	buf[0] = BUF_START_REQUESTOR;
+	r = sys_safemap(ep_granter, gid, 0, (long)buf, size, D, 1);
+	if(r != OK) {
+		printf("REQUESTOR: error in sys_safemap: %d\n", r);
+		return 1;
+	}
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+	CHECK_TEST("REQUESTOR", buf[0], BUF_START_GRANTOR+1, "COW_SMAP");
+
+	/* Test COW_SMAP2 (with COW safecopy). */
+	FIFO_WAIT(fid_get);
+	buf[0] = BUF_START_REQUESTOR;
+	r = sys_safecopyto(ep_granter, gid, 0, (long)buf, size, D);
+	if(r != OK) {
+		printf("REQUESTOR: error in sys_safecopyto: %d\n", r);
+		return 1;
+	}
+	r = sys_safemap(ep_granter, gid, 0, (long)buf, size, D, 1);
+	if(r != OK) {
+		printf("REQUESTOR: error in sys_safemap: %d\n", r);
+		return 1;
+	}
+	FIFO_NOTIFY(fid_send);
+	CHECK_TEST("REQUESTOR", buf[0], BUF_START_REQUESTOR, "COW_SMAP2");
+
+	return 0;
+}
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/safemap/run

@@ -0,0 +1,6 @@
+#!/bin/sh
+
+PWD=`pwd`
+
+service up ${PWD}/grantor -config ${PWD}/system.conf -script ${PWD}/down
+service up ${PWD}/requestor -config ${PWD}/system.conf -script ${PWD}/down

test/safemap/system.conf

@@ -0,0 +1,34 @@
+
+service requestor
+{
+	system
+		SAFECOPYFROM
+		SAFECOPYTO
+		SAFEMAP
+		SAFEREVMAP
+		SAFEUNMAP
+		UMAP
+		PRIVCTL
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};
+
+service grantor
+{
+	system
+		SAFECOPYFROM
+		SAFECOPYTO
+		SAFEMAP
+		SAFEREVMAP
+		SAFEUNMAP
+		UMAP
+		PRIVCTL
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};

test/safeperf/Makefile

@@ -0,0 +1,21 @@
+all: requestor grantor 1fifo 2fifo
+	chmod +x down run
+
+requestor: requestor.c inc.h
+	cc -o $@ $< -lsys
+
+grantor: grantor.c inc.h
+	cc -o $@ $< -lsys
+
+1fifo 2fifo:
+	mkfifo $@
+
+run: all
+	sh run
+
+kill:
+	sh down
+
+clean:
+	rm -f grantor requestor 1fifo 2fifo
+

test/safeperf/README

@@ -0,0 +1,11 @@
+Performance Test Program for Safecopy and Safemap
+
+How to run
+==========
+
+  1. Set USE_COW_SAFECOPY to 1 or 0 in kernel/system/do_safecopy.c
+     to run safecopy tests with or without COW optimization.
+  2. Type `make run` to prepare and run tests. Configuration parameters
+     for performance tests are in ./run.
+  3. When done testing, type `make clean` to clean up.
+

test/safeperf/down

@@ -0,0 +1,5 @@
+#!/bin/sh
+
+service down grantor
+service down requestor
+

test/safeperf/grantor.c

@@ -0,0 +1,62 @@
+#include "inc.h"
+
+char buf_buf[BUF_SIZE + CLICK_SIZE];
+
+int fid_send, fid_get;
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *				    main				     *
+ *===========================================================================*/
+int main(int argc, char **argv)
+{
+	endpoint_t ep_self, ep_requestor, ep_child;
+	cp_grant_id_t gid;
+	int i, r, pid;
+	char *buf;
+	int status;
+
+	/* SEF local startup. */
+	env_setargs(argc, argv);
+	sef_local_startup();
+
+	/* Prepare work. */
+	buf = (char*) CLICK_CEIL(buf_buf);
+	fid_send = open(FIFO_GRANTOR, O_WRONLY);
+	fid_get = open(FIFO_REQUESTOR, O_RDONLY);
+	if(fid_get < 0 || fid_send < 0) {
+		printf("GRANTOR: can't open fifo files.\n");
+		return 1;
+	}
+
+	/* Get the requestor's endpoint. */
+	read(fid_get, &ep_requestor, sizeof(ep_requestor));
+	dprint("GRANTOR: getting requestor's endpoint: %d\n", ep_requestor);
+
+	/* Grant. */
+	gid = cpf_grant_direct(ep_requestor, (long)buf, BUF_SIZE,
+		CPF_READ | CPF_WRITE | CPF_MAP);
+	ep_self = getprocnr();
+	dprint("GRANTOR: sending my endpoint %d and gid %d\n", ep_self, gid);
+	write(fid_send, &ep_self, sizeof(ep_self));
+	write(fid_send, &gid, sizeof(gid));
+
+	/* Test safemap. */
+	buf[0] = 0;
+	FIFO_NOTIFY(fid_send);
+	FIFO_WAIT(fid_get);
+
+	return 0;
+}
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/safeperf/inc.h

@@ -0,0 +1,51 @@
+#define _SYSTEM
+#define _MINIX
+#include <stdio.h>
+#include <unistd.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <signal.h>
+#include <sys/types.h>
+#include <sys/wait.h>
+
+#include <minix/config.h>
+#include <minix/com.h>
+#include <minix/type.h>
+#include <minix/const.h>
+#include <minix/endpoint.h>
+#include <minix/safecopies.h>
+#include <minix/syslib.h>
+#include <minix/sysutil.h>
+#include <errno.h>
+
+#define TEST_PAGE_NUM    50
+#define BUF_SIZE (TEST_PAGE_NUM * CLICK_SIZE)
+
+#define NR_TEST_ITERATIONS 100
+
+#define FIFO_REQUESTOR "/usr/src/test/safeperf/1fifo"
+#define FIFO_GRANTOR   "/usr/src/test/safeperf/2fifo"
+
+#define FIFO_WAIT(fid) {                                                      \
+	int a;                                                                \
+	if(read(fid, &a, sizeof(a)) != sizeof(a))                             \
+		panic(__FILE__, "FIFO_WAIT failed", NO_NUM);                  \
+}
+#define FIFO_NOTIFY(fid) {                                                    \
+	int a = 1;                                                            \
+	if(write(fid, &a, sizeof(a)) != sizeof(a))                            \
+		panic(__FILE__, "FIFO_NOTIFY failed", NO_NUM);                \
+}
+
+#define REPORT_TEST(who, test_name, diff) {                                   \
+	printf("%-9s: test %s took an average of %d.%dus per page\n",         \
+	    who, test_name, (int)diff, ((int)(diff*1000))%1000);              \
+}
+
+#define DEBUG 0
+#if DEBUG
+#	define dprint printf
+#else
+#	define dprint (void)
+#endif
+

test/safeperf/requestor.c

@@ -0,0 +1,166 @@
+#include "inc.h"
+
+char buf_buf[BUF_SIZE + CLICK_SIZE];
+
+endpoint_t ep_granter;
+cp_grant_id_t gid;
+char *buf;
+int fid_send, fid_get;
+
+/* SEF functions and variables. */
+FORWARD _PROTOTYPE( void sef_local_startup, (void) );
+
+/*===========================================================================*
+ *			      read_write_buff				     *
+ *===========================================================================*/
+void read_write_buff(char* buff, int size, int is_write)
+{
+	int i;
+	char c;
+
+	if(size % CLICK_SIZE != 0) {
+		panic("REQUESTOR", "buff_size not page aligned", NO_NUM);
+	}
+	if(is_write) {
+	    for(i=0;i<size;i+=CLICK_SIZE) buff[i] = 1;
+	}
+	else {
+	    for(i=0;i<size;i+=CLICK_SIZE) c = buff[i];
+	}
+}
+
+/*===========================================================================*
+ *				 exit_usage				     *
+ *===========================================================================*/
+void exit_usage(void)
+{
+	printf("Usage: requestor pages=<nr_pages> map=<0|1> write=<0|1>\n");
+	exit(1);
+}
+
+/*===========================================================================*
+ *				    main				     *
+ *===========================================================================*/
+int main(int argc, char **argv)
+{
+	endpoint_t ep_self, ep_child;
+	size_t size = BUF_SIZE;
+	int i, r, pid;
+	int status;
+	u64_t start, end, diff;
+	double micros;
+	char nr_pages_str[10], is_map_str[2], is_write_str[2];
+	int nr_pages, is_map, is_write;
+
+	/* SEF local startup. */
+	env_setargs(argc, argv);
+	sef_local_startup();
+
+	/* Parse the command line. */
+	r = env_get_param("pages", nr_pages_str, sizeof(nr_pages_str));
+	errno = 0;
+	nr_pages = atoi(nr_pages_str);
+	if (r != OK || errno || nr_pages <=0) {
+		exit_usage();
+	}
+	if(nr_pages > TEST_PAGE_NUM) {
+		printf("REQUESTOR: too many pages. Max allowed: %d\n",
+			TEST_PAGE_NUM);
+		exit_usage();
+	}
+	r = env_get_param("map", is_map_str, sizeof(is_map_str));
+	errno = 0;
+	is_map = atoi(is_map_str);
+	if (r != OK || errno || (is_map!=0 && is_map!=1)) {
+		exit_usage();
+	}
+	r = env_get_param("write", is_write_str, sizeof(is_write_str));
+	errno = 0;
+	is_write = atoi(is_write_str);
+	if (r != OK || errno || (is_write!=0 && is_write!=1)) {
+		exit_usage();
+	}
+	printf("REQUESTOR: Running tests with pages=%d map=%d write=%d...\n",
+		nr_pages, is_map, is_write);
+
+	/* Prepare work. */
+	buf = (char*) CLICK_CEIL(buf_buf);
+	fid_get = open(FIFO_GRANTOR, O_RDONLY);
+	fid_send = open(FIFO_REQUESTOR, O_WRONLY);
+	if(fid_get < 0 || fid_send < 0) {
+		printf("REQUESTOR: can't open fifo files.\n");
+		return 1;
+	}
+
+	/* Send the endpoint to the granter, in order to let him to
+	 * create the grant.
+	 */
+	ep_self = getprocnr();
+	write(fid_send, &ep_self, sizeof(ep_self));
+	dprint("REQUESTOR: sending my endpoint: %d\n", ep_self);
+
+	/* Get the granter's endpoint and gid. */
+	read(fid_get, &ep_granter, sizeof(ep_granter));
+	read(fid_get, &gid, sizeof(gid));
+	dprint("REQUESTOR: getting granter's endpoint %d and gid %d\n",
+		ep_granter, gid);
+
+	FIFO_WAIT(fid_get);
+	diff = make64(0, 0);
+
+	if(is_map) {
+		/* Test safemap. */
+		for(i=0;i<NR_TEST_ITERATIONS;i++) {
+			read_tsc_64(&start);
+			r = sys_safemap(ep_granter, gid, 0, (long)buf,
+				nr_pages*CLICK_SIZE, D, 1);
+			if(r != OK) {
+				printf("REQUESTOR: safemap error: %d\n", r);
+				return 1;
+			}
+			read_write_buff(buf, nr_pages*CLICK_SIZE, is_write);
+			read_tsc_64(&end);
+			diff = add64(diff, (sub64(end, start)));
+			r = sys_safeunmap(D, (long)buf);
+			if(r != OK) {
+				printf("REQUESTOR: safeunmap error: %d\n", r);
+				return 1;
+			}
+		}
+		micros = ((double)tsc_64_to_micros(diff))
+			/ (NR_TEST_ITERATIONS*nr_pages);
+		REPORT_TEST("REQUESTOR", "SAFEMAP", micros);
+	}
+	else {
+		/* Test safecopy. */
+		for(i=0;i<NR_TEST_ITERATIONS;i++) {
+			read_tsc_64(&start);
+			r = sys_safecopyfrom(ep_granter, gid, 0, (long)buf,
+				nr_pages*CLICK_SIZE, D);
+			if(r != OK) {
+				printf("REQUESTOR: safecopy error: %d\n", r);
+				return 1;
+			}
+			read_write_buff(buf, nr_pages*CLICK_SIZE, is_write);
+			read_tsc_64(&end);
+			diff = add64(diff, (sub64(end, start)));
+		}
+		micros = ((double)tsc_64_to_micros(diff))
+			/ (NR_TEST_ITERATIONS*nr_pages);
+		REPORT_TEST("REQUESTOR", "SAFECOPY", micros);
+	}
+
+	FIFO_NOTIFY(fid_send);
+
+	return 0;
+}
+
+/*===========================================================================*
+ *			       sef_local_startup			     *
+ *===========================================================================*/
+PRIVATE void sef_local_startup()
+{
+  /* Let SEF perform startup. */
+  sef_startup();
+}
+

test/safeperf/run

@@ -0,0 +1,16 @@
+#!/bin/sh
+
+PAGES="1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50"
+MAP=0
+WRITE=1
+
+PWD=`pwd`
+
+for P in `echo $PAGES`
+do
+  service up ${PWD}/grantor -config ${PWD}/system.conf -script ${PWD}/down
+  service up ${PWD}/requestor -config ${PWD}/system.conf -script ${PWD}/down \
+		-args pages=${P}\ map=${MAP}\ write=${WRITE}
+  sleep 2
+done
+

test/safeperf/system.conf

@@ -0,0 +1,34 @@
+
+service requestor
+{
+	system
+		SAFECOPYFROM
+		SAFECOPYTO
+		SAFEMAP
+		SAFEREVMAP
+		SAFEUNMAP
+		UMAP
+		PRIVCTL
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};
+
+service grantor
+{
+	system
+		SAFECOPYFROM
+		SAFECOPYTO
+		SAFEMAP
+		SAFEREVMAP
+		SAFEUNMAP
+		UMAP
+		PRIVCTL
+		SETGRANT
+		GETINFO
+		SYSCTL
+	;
+	uid	0;
+};