svn commit: r208651 - head/sys/i386/xen
Author: alc
Date: Sun May 30 04:44:32 2010
New Revision: 208651
URL: http://svn.freebsd.org/changeset/base/208651
Log:
Merge various changes from i386/i386/pmap.c:
The remaining, unmerged portions of r175404
Retire PMAP_DIAGNOSTIC. Any useful diagnostics that were conditionally
compiled under PMAP_DIAGNOSTIC are now KASSERT()s. (Note: The kernel
option DIAGNOSTIC still disables inlining of certain pmap functions.)
Eliminate dead code from pmap_enter(). This code implemented an assertion.
On i386, an equivalent check is already implemented. However, on amd64,
a small change is required to implement an equivalent check.
Eliminate \n from a nearby panic string.
Use KASSERT() to reimplement pmap_copy()'s two assertions.
Merge portions of r177659
To date, we have assumed that the TLB will only set the PG_M bit in a
PTE if that PTE has the PG_RW bit set. However, this assumption does
not hold on recent processors from Intel. For example, consider a PTE
that has the PG_RW bit set but the PG_M bit clear. Suppose this PTE
is cached in the TLB and later the PG_RW bit is cleared in the PTE,
but the corresponding TLB entry is not (yet) invalidated.
Historically, upon a write access using this (stale) TLB entry, the
TLB would observe that the PG_RW bit had been cleared and initiate a
page fault, aborting the setting of the PG_M bit in the PTE. Now,
however, P4- and Core2-family processors will set the PG_M bit before
observing that the PG_RW bit is clear and initiating a page fault. In
other words, the write does not occur but the PG_M bit is still set.
The real impact of this difference is not that great. Specifically,
we should no longer assert that any PTE with the PG_M bit set must
also have the PG_RW bit set, and we should ignore the state of the
PG_M bit unless the PG_RW bit is set.
r208609
Defer freeing any page table pages in pmap_remove_all() until after the
page queues lock is released. This may reduce the amount of time that the
page queues lock is held by pmap_remove_all().
r208645
When I pushed down the page queues lock into pmap_is_modified(), I created
an ordering dependence: A pmap operation that clears PG_WRITEABLE and calls
vm_page_dirty() must perform the call first. Otherwise, pmap_is_modified()
could return FALSE without acquiring the page queues lock because the page
is not (currently) writeable, and the caller to pmap_is_modified() might
believe that the page's dirty field is clear because it has not seen the
effect of the vm_page_dirty() call.
When I pushed down the page queues lock into pmap_is_modified(), I
overlooked one place where this ordering dependence is violated:
pmap_enter(). In a rare situation pmap_enter() can be called to replace a
dirty mapping to one page with a mapping to another page. (I say rare
because replacements generally occur as a result of a copy-on-write fault,
and so the old page is not dirty.) This change delays clearing PG_WRITEABLE
until after vm_page_dirty() has been called.
Fixing the ordering dependency also makes it easy to introduce a small
optimization: When pmap_enter() used to replace a mapping to one page with a
mapping to another page, it freed the pv entry for the first mapping and
later called the pv entry allocator for the new mapping. Now, pmap_enter()
attempts to recycle the old pv entry, saving two calls to the pv entry
allocator.
There is no point in setting PG_WRITEABLE on unmanaged pages, so don't.
Update a comment to reflect this.
Tidy up the variable declarations at the start of pmap_enter().
Modified:
head/sys/i386/xen/pmap.c
Modified: head/sys/i386/xen/pmap.c
==
--- head/sys/i386/xen/pmap.cSun May 30 03:45:41 2010(r208650)
+++ head/sys/i386/xen/pmap.cSun May 30 04:44:32 2010(r208651)
@@ -103,8 +103,6 @@ __FBSDID("$FreeBSD$");
* and to when physical maps must be made correct.
*/
-#define PMAP_DIAGNOSTIC
-
#include "opt_cpu.h"
#include "opt_pmap.h"
#include "opt_msgbuf.h"
@@ -168,13 +166,11 @@ __FBSDID("$FreeBSD$");
#define PMAP_SHPGPERPROC 200
#endif
-#if defined(DIAGNOSTIC)
-#define PMAP_DIAGNOSTIC
-#endif
+#define DIAGNOSTIC
-#if !defined(PMAP_DIAGNOSTIC)
+#if !defined(DIAGNOSTIC)
#ifdef __GNUC_GNU_INLINE__
-#define PMAP_INLINEinline
+#define PMAP_INLINE__attribute__((__gnu_inline__)) inline
#else
#define PMAP_INLINEextern inline
#endif
@@ -298,6 +294,9 @@ SYSCTL_ULONG(_vm_pmap_pde, OID_AUTO, map
static voidfree_pv_entry(pmap_t pmap, pv_entry_t pv);
static pv_entry_t get_pv_entry(pmap_t locked_pmap, int try);
+static voidpmap_pvh_free(struct md_page *pvh, pmap_t pmap, vm_offset_t va);
+static pv_entry_t pmap_pvh_remove(struct md_page *
Re: svn commit: r208649 - head/usr.sbin/newsyslog
Hi gordon. On Sat, 29 May 2010 22:55:59 + (UTC) Gordon Tetlow wrote: > Log: > Add file include processing for newsyslog. > Format for the include line in /etc/newsyslog.conf is: >/etc/defaults/newsyslog.conf Oohh my god, that's great! Do you have any works to support following? 1. include from /usr/local/etc/newsyslog.d/* 2. not supported glob pattern is OK. I want to add and remove to /etc/newsyslog.conf, safe and simlply like: echo include ports > /etc/newsyslog.conf sed -i "" "/include ports/d" /etc/newsyslog.conf And 3. same to syslog.conf -- Norikatsu Shigemura ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208650 - head/sys/dev/fdc
Author: jkim
Date: Sun May 30 03:45:41 2010
New Revision: 208650
URL: http://svn.freebsd.org/changeset/base/208650
Log:
Fix a regression from r200554, which broke fdc(4) attachment to acpi(4).
I removed too many lines and a wrong pointer was accidentally passed down.
Tested by:Scott Allendorf (scott-allendorf at uiowa dot edu), kib
MFC after:3 days
Modified:
head/sys/dev/fdc/fdc_acpi.c
Modified: head/sys/dev/fdc/fdc_acpi.c
==
--- head/sys/dev/fdc/fdc_acpi.c Sat May 29 22:55:59 2010(r208649)
+++ head/sys/dev/fdc/fdc_acpi.c Sun May 30 03:45:41 2010(r208650)
@@ -96,6 +96,7 @@ fdc_acpi_attach(device_t dev)
{
struct fdc_data *sc;
ACPI_BUFFER buf;
+ ACPI_OBJECT *obj;
device_t bus;
int error;
@@ -131,7 +132,8 @@ fdc_acpi_attach(device_t dev)
}
/* Add fd child devices as specified. */
- error = fdc_acpi_probe_children(bus, dev, buf.Pointer);
+ obj = buf.Pointer;
+ error = fdc_acpi_probe_children(bus, dev, obj->Buffer.Pointer);
out:
if (buf.Pointer)
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208649 - head/usr.sbin/newsyslog
Author: gordon
Date: Sat May 29 22:55:59 2010
New Revision: 208649
URL: http://svn.freebsd.org/changeset/base/208649
Log:
Add file include processing for newsyslog.
Format for the include line in /etc/newsyslog.conf is:
/etc/defaults/newsyslog.conf
Other notes of interest:
Globbing is supported in statements.
Properly detect circular include loop dependencies.
Reviewed by: gad@
Approved by: wes@ (mentor)
MFC after:2 months
Modified:
head/usr.sbin/newsyslog/newsyslog.c
head/usr.sbin/newsyslog/newsyslog.conf.5
Modified: head/usr.sbin/newsyslog/newsyslog.c
==
--- head/usr.sbin/newsyslog/newsyslog.c Sat May 29 22:52:17 2010
(r208648)
+++ head/usr.sbin/newsyslog/newsyslog.c Sat May 29 22:55:59 2010
(r208649)
@@ -111,6 +111,7 @@ __FBSDID("$FreeBSD$");
#defineDEFAULT_MARKER ""
#defineDEBUG_MARKER""
+#defineINCLUDE_MARKER ""
struct conf_entry {
STAILQ_ENTRY(conf_entry) cf_nextp;
@@ -149,6 +150,11 @@ struct zipwork_entry {
char zw_fname[1]; /* the file to compress */
};
+struct include_entry {
+ STAILQ_ENTRY(include_entry) inc_nextp;
+ const char *file; /* Name of file to process */
+};
+
typedef enum {
FREE_ENT, KEEP_ENT
} fk_entry;
@@ -156,6 +162,7 @@ typedef enum {
STAILQ_HEAD(cflist, conf_entry);
SLIST_HEAD(swlisthead, sigwork_entry) swhead = SLIST_HEAD_INITIALIZER(swhead);
SLIST_HEAD(zwlisthead, zipwork_entry) zwhead = SLIST_HEAD_INITIALIZER(zwhead);
+STAILQ_HEAD(ilist, include_entry);
int dbg_at_times; /* -D Show details of 'trim_at' code */
@@ -189,10 +196,12 @@ char hostname[MAXHOSTNAMELEN];/* hostna
static struct cflist *get_worklist(char **files);
static void parse_file(FILE *cf, struct cflist *work_p, struct cflist *glob_p,
- struct conf_entry *defconf_p);
+ struct conf_entry *defconf_p, struct ilist *inclist);
+static void add_to_queue(const char *fname, struct ilist *inclist);
static char *sob(char *p);
static char *son(char *p);
static int isnumberstr(const char *);
+static int isglobstr(const char *);
static char *missing_field(char *p, char *errline);
static void change_attrs(const char *, const struct conf_entry *);
static fk_entry do_entry(struct conf_entry *);
@@ -731,13 +740,15 @@ static struct cflist *
get_worklist(char **files)
{
FILE *f;
- const char *fname;
char **given;
- struct cflist *filelist, *globlist, *cmdlist;
+ struct cflist *cmdlist, *filelist, *globlist;
struct conf_entry *defconf, *dupent, *ent;
+ struct ilist inclist;
+ struct include_entry *inc;
int gmatch, fnres;
defconf = NULL;
+ STAILQ_INIT(&inclist);
filelist = malloc(sizeof(struct cflist));
if (filelist == NULL)
@@ -748,21 +759,29 @@ get_worklist(char **files)
err(1, "malloc of globlist");
STAILQ_INIT(globlist);
- fname = conf;
- if (fname == NULL)
- fname = _PATH_CONF;
+ inc = malloc(sizeof(struct include_entry));
+ if (inc == NULL)
+ err(1, "malloc of inc");
+ inc->file = conf;
+ if (inc->file == NULL)
+ inc->file = _PATH_CONF;
+ STAILQ_INSERT_TAIL(&inclist, inc, inc_nextp);
+
+ STAILQ_FOREACH(inc, &inclist, inc_nextp) {
+ if (strcmp(inc->file, "-") != 0)
+ f = fopen(inc->file, "r");
+ else {
+ f = stdin;
+ inc->file = "";
+ }
+ if (!f)
+ err(1, "%s", inc->file);
- if (strcmp(fname, "-") != 0)
- f = fopen(fname, "r");
- else {
- f = stdin;
- fname = "";
+ if (verbose)
+ printf("Processing %s\n", inc->file);
+ parse_file(f, filelist, globlist, defconf, &inclist);
+ (void) fclose(f);
}
- if (!f)
- err(1, "%s", fname);
-
- parse_file(f, filelist, globlist, defconf);
- (void) fclose(f);
/*
* All config-file information has been read in and turned into
@@ -965,14 +984,16 @@ expand_globs(struct cflist *work_p, stru
*/
static void
parse_file(FILE *cf, struct cflist *work_p, struct cflist *glob_p,
-struct conf_entry *defconf_p)
+struct conf_entry *defconf_p, struct ilist *inclist)
{
char line[BUFSIZ], *parse, *q;
char *cp, *errline, *group;
struct conf_entry *working;
struct passwd *pwd;
struct group *grp;
+ glob_t pglob;
int eol, ptm_opts, res, special;
+ size_t i;
errline = NULL;
while (fgets(line, BUFSIZ, cf)) {
@@ -1017,6 +1038,37 @@ parse_file(FILE *cf, struct cflist *work
svn commit: r208648 - head/usr.sbin/newsyslog
Author: gordon
Date: Sat May 29 22:52:17 2010
New Revision: 208648
URL: http://svn.freebsd.org/changeset/base/208648
Log:
Convert newsyslog to using queue(3) macros instead of a home rolled version.
Reviewed by: gad@
Approved by: wes@ (mentor)
MFC after:2 months
Modified:
head/usr.sbin/newsyslog/newsyslog.c
Modified: head/usr.sbin/newsyslog/newsyslog.c
==
--- head/usr.sbin/newsyslog/newsyslog.c Sat May 29 20:24:01 2010
(r208647)
+++ head/usr.sbin/newsyslog/newsyslog.c Sat May 29 22:52:17 2010
(r208648)
@@ -113,6 +113,7 @@ __FBSDID("$FreeBSD$");
#defineDEBUG_MARKER""
struct conf_entry {
+ STAILQ_ENTRY(conf_entry) cf_nextp;
char *log; /* Name of the log */
char *pid_file; /* PID file */
char *r_reason; /* The reason this file is being rotated */
@@ -129,7 +130,6 @@ struct conf_entry {
int flags; /* CE_COMPACT, CE_BZCOMPACT, CE_BINARY */
int sig;/* Signal to send */
int def_cfg;/* Using the rule for this file */
- struct conf_entry *next;/* Linked list pointer */
};
struct sigwork_entry {
@@ -153,6 +153,7 @@ typedef enum {
FREE_ENT, KEEP_ENT
} fk_entry;
+STAILQ_HEAD(cflist, conf_entry);
SLIST_HEAD(swlisthead, sigwork_entry) swhead = SLIST_HEAD_INITIALIZER(swhead);
SLIST_HEAD(zwlisthead, zipwork_entry) zwhead = SLIST_HEAD_INITIALIZER(zwhead);
@@ -186,9 +187,9 @@ char daytime[DAYTIME_LEN]; /* The curren
* used for rotation-tracking messages. */
char hostname[MAXHOSTNAMELEN]; /* hostname */
-static struct conf_entry *get_worklist(char **files);
-static void parse_file(FILE *cf, const char *cfname, struct conf_entry
**work_p,
- struct conf_entry **glob_p, struct conf_entry **defconf_p);
+static struct cflist *get_worklist(char **files);
+static void parse_file(FILE *cf, struct cflist *work_p, struct cflist *glob_p,
+ struct conf_entry *defconf_p);
static char *sob(char *p);
static char *son(char *p);
static int isnumberstr(const char *);
@@ -205,9 +206,8 @@ static struct zipwork_entry *
sigwork_entry *, int, const char *);
static void set_swpid(struct sigwork_entry *, const struct conf_entry *);
static int sizefile(const char *);
-static void expand_globs(struct conf_entry **work_p,
- struct conf_entry **glob_p);
-static void free_clist(struct conf_entry **firstent);
+static void expand_globs(struct cflist *work_p, struct cflist *glob_p);
+static void free_clist(struct cflist *list);
static void free_entry(struct conf_entry *ent);
static struct conf_entry *init_entry(const char *fname,
struct conf_entry *src_entry);
@@ -233,8 +233,8 @@ static void createlog(const struct conf_
int
main(int argc, char **argv)
{
- fk_entry free_or_keep;
- struct conf_entry *p, *q;
+ struct cflist *worklist;
+ struct conf_entry *p;
struct sigwork_entry *stmp;
struct zipwork_entry *ztmp;
@@ -247,18 +247,17 @@ main(int argc, char **argv)
if (needroot && getuid() && geteuid())
errx(1, "must have root privs");
- p = q = get_worklist(argv);
+ worklist = get_worklist(argv);
/*
* Rotate all the files which need to be rotated. Note that
* some users have *hundreds* of entries in newsyslog.conf!
*/
- while (p) {
- free_or_keep = do_entry(p);
- p = p->next;
- if (free_or_keep == FREE_ENT)
- free_entry(q);
- q = p;
+ while (!STAILQ_EMPTY(worklist)) {
+ p = STAILQ_FIRST(worklist);
+ STAILQ_REMOVE_HEAD(worklist, cf_nextp);
+ if (do_entry(p) == FREE_ENT)
+ free_entry(p);
}
/*
@@ -365,7 +364,6 @@ init_entry(const char *fname, struct con
tempwork->sig = SIGHUP;
tempwork->def_cfg = 0;
}
- tempwork->next = NULL;
return (tempwork);
}
@@ -403,21 +401,18 @@ free_entry(struct conf_entry *ent)
}
static void
-free_clist(struct conf_entry **firstent)
+free_clist(struct cflist *list)
{
- struct conf_entry *ent, *nextent;
-
- if (firstent == NULL)
- return; /* There is nothing to do. */
+ struct conf_entry *ent;
- ent = *firstent;
- firstent = NULL;
-
- while (ent) {
- nextent = ent->next;
+ while (!STAILQ_EMPTY(list)) {
+ ent = STAILQ_FIRST(list);
+ STAILQ_REMOVE_HEAD(list, cf_nextp);
free_entry(ent);
- ent = nextent;
}
+
+ free(list);
+ list = NULL;
}
static fk_entry
@@ -732,17 +727,26 @@ usage(void)
* Parse a configuration file and
Re: svn commit: r208644 - head/sys/dev/ath/ath_hal/ar5211
On 29 May 2010, at 17:14, Rui Paulo wrote: > Author: rpaulo > Date: Sat May 29 16:14:02 2010 > New Revision: 208644 > URL: http://svn.freebsd.org/changeset/base/208644 > > Log: > Due to the way HALDEBUG() is defined, we need to add curly brackets > when using it as a sole if clause instruction. > While there, fix 'const static' typo. > > Submitted by:Arnaud Lacombe This should read a...@netbsd.org. Regards, -- Rui Paulo ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208647 - in head: sbin/mount share/man/man5
Author: jilles Date: Sat May 29 20:24:01 2010 New Revision: 208647 URL: http://svn.freebsd.org/changeset/base/208647 Log: Clarify devfs manpages slightly. mount(8): add xref to devfs(5) devfs(5): change example to something more likely to be useful (it is not necessary to mount a devfs on /dev manually, but for chroots/jails it is often needed), mention since when devfs is preferred to device nodes on ufs PR: 146600 MFC after:2 weeks Modified: head/sbin/mount/mount.8 head/share/man/man5/devfs.5 Modified: head/sbin/mount/mount.8 == --- head/sbin/mount/mount.8 Sat May 29 18:26:44 2010(r208646) +++ head/sbin/mount/mount.8 Sat May 29 20:24:01 2010(r208647) @@ -525,6 +525,7 @@ support for a particular file system mig .Xr nmount 2 , .Xr acl 3 , .Xr mac 4 , +.Xr devfs 5 , .Xr ext2fs 5 , .Xr fstab 5 , .Xr procfs 5 , Modified: head/share/man/man5/devfs.5 == --- head/share/man/man5/devfs.5 Sat May 29 18:26:44 2010(r208646) +++ head/share/man/man5/devfs.5 Sat May 29 20:24:01 2010(r208647) @@ -38,7 +38,7 @@ .\" .\" $FreeBSD$ .\" -.Dd December 14, 1996 +.Dd May 29, 2010 .Dt DEVFS 5 .Os .Sh NAME @@ -80,9 +80,9 @@ mount point. To mount a .Nm volume located on -.Pa /dev : +.Pa /mychroot/dev : .Pp -.Dl "mount -t devfs devfs /dev" +.Dl "mount -t devfs devfs /mychroot/dev" .Sh SEE ALSO .Xr devfs 8 , .Xr mount 8 @@ -90,7 +90,9 @@ volume located on The .Nm file system first appeared in -.Fx 2.0 . +.Fx 2.0 +and became the preferred method for accessing devices in +.Fx 5.0 . The .Nm manual page first appeared in ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208646 - head/sys/ia64/ia64
Author: alc Date: Sat May 29 18:26:44 2010 New Revision: 208646 URL: http://svn.freebsd.org/changeset/base/208646 Log: Don't set PG_WRITEABLE in pmap_enter() unless the page is managed. Modified: head/sys/ia64/ia64/pmap.c Modified: head/sys/ia64/ia64/pmap.c == --- head/sys/ia64/ia64/pmap.c Sat May 29 17:10:45 2010(r208645) +++ head/sys/ia64/ia64/pmap.c Sat May 29 18:26:44 2010(r208646) @@ -1595,7 +1595,7 @@ validate: if (icache_inval) ia64_sync_icache(va, PAGE_SIZE); - if ((prot & VM_PROT_WRITE) != 0) + if ((prot & VM_PROT_WRITE) != 0 && managed) vm_page_flag_set(m, PG_WRITEABLE); vm_page_unlock_queues(); pmap_switch(oldpmap); ___ svn-src-head@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208645 - in head/sys: amd64/amd64 i386/i386 vm
Author: alc
Date: Sat May 29 17:10:45 2010
New Revision: 208645
URL: http://svn.freebsd.org/changeset/base/208645
Log:
When I pushed down the page queues lock into pmap_is_modified(), I created
an ordering dependence: A pmap operation that clears PG_WRITEABLE and calls
vm_page_dirty() must perform the call first. Otherwise, pmap_is_modified()
could return FALSE without acquiring the page queues lock because the page
is not (currently) writeable, and the caller to pmap_is_modified() might
believe that the page's dirty field is clear because it has not seen the
effect of the vm_page_dirty() call.
When I pushed down the page queues lock into pmap_is_modified(), I
overlooked one place where this ordering dependence is violated:
pmap_enter(). In a rare situation pmap_enter() can be called to replace a
dirty mapping to one page with a mapping to another page. (I say rare
because replacements generally occur as a result of a copy-on-write fault,
and so the old page is not dirty.) This change delays clearing PG_WRITEABLE
until after vm_page_dirty() has been called.
Fixing the ordering dependency also makes it easy to introduce a small
optimization: When pmap_enter() used to replace a mapping to one page with a
mapping to another page, it freed the pv entry for the first mapping and
later called the pv entry allocator for the new mapping. Now, pmap_enter()
attempts to recycle the old pv entry, saving two calls to the pv entry
allocator.
There is no point in setting PG_WRITEABLE on unmanaged pages, so don't.
Update a comment to reflect this.
Tidy up the variable declarations at the start of pmap_enter().
Modified:
head/sys/amd64/amd64/pmap.c
head/sys/i386/i386/pmap.c
head/sys/vm/vm_page.h
Modified: head/sys/amd64/amd64/pmap.c
==
--- head/sys/amd64/amd64/pmap.c Sat May 29 16:14:02 2010(r208644)
+++ head/sys/amd64/amd64/pmap.c Sat May 29 17:10:45 2010(r208645)
@@ -3128,11 +3128,11 @@ void
pmap_enter(pmap_t pmap, vm_offset_t va, vm_prot_t access, vm_page_t m,
vm_prot_t prot, boolean_t wired)
{
- vm_paddr_t pa;
pd_entry_t *pde;
pt_entry_t *pte;
- vm_paddr_t opa;
- pt_entry_t origpte, newpte;
+ pt_entry_t newpte, origpte;
+ pv_entry_t pv;
+ vm_paddr_t opa, pa;
vm_page_t mpte, om;
boolean_t invlva;
@@ -3200,6 +3200,9 @@ pmap_enter(pmap_t pmap, vm_offset_t va,
}
goto validate;
}
+
+ pv = NULL;
+
/*
* Mapping has changed, invalidate old range and fall through to
* handle validating new mapping.
@@ -3209,7 +3212,7 @@ pmap_enter(pmap_t pmap, vm_offset_t va,
pmap->pm_stats.wired_count--;
if (origpte & PG_MANAGED) {
om = PHYS_TO_VM_PAGE(opa);
- pmap_remove_entry(pmap, om, va);
+ pv = pmap_pvh_remove(&om->md, pmap, va);
}
if (mpte != NULL) {
mpte->wire_count--;
@@ -3226,9 +3229,13 @@ pmap_enter(pmap_t pmap, vm_offset_t va,
if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) == 0) {
KASSERT(va < kmi.clean_sva || va >= kmi.clean_eva,
("pmap_enter: managed mapping within the clean submap"));
- pmap_insert_entry(pmap, va, m);
+ if (pv == NULL)
+ pv = get_pv_entry(pmap, FALSE);
+ pv->pv_va = va;
+ TAILQ_INSERT_TAIL(&m->md.pv_list, pv, pv_list);
pa |= PG_MANAGED;
- }
+ } else if (pv != NULL)
+ free_pv_entry(pmap, pv);
/*
* Increment counters
@@ -3243,7 +3250,8 @@ validate:
newpte = (pt_entry_t)(pa | pmap_cache_bits(m->md.pat_mode, 0) | PG_V);
if ((prot & VM_PROT_WRITE) != 0) {
newpte |= PG_RW;
- vm_page_flag_set(m, PG_WRITEABLE);
+ if ((newpte & PG_MANAGED) != 0)
+ vm_page_flag_set(m, PG_WRITEABLE);
}
if ((prot & VM_PROT_EXECUTE) == 0)
newpte |= pg_nx;
@@ -3278,6 +3286,10 @@ validate:
if ((newpte & PG_RW) == 0)
invlva = TRUE;
}
+ if ((origpte & PG_MANAGED) != 0 &&
+ TAILQ_EMPTY(&om->md.pv_list) &&
+ TAILQ_EMPTY(&pa_to_pvh(opa)->pv_list))
+ vm_page_flag_clear(om, PG_WRITEABLE);
if (invlva)
pmap_invalidate_page(pmap, va);
} else
Modified: head/sys/i386/i386/pmap.c
==
--- head/sys/i386/i386/pmap.c Sat May 29 16:14:02 2010(r208
svn commit: r208644 - head/sys/dev/ath/ath_hal/ar5211
Author: rpaulo
Date: Sat May 29 16:14:02 2010
New Revision: 208644
URL: http://svn.freebsd.org/changeset/base/208644
Log:
Due to the way HALDEBUG() is defined, we need to add curly brackets
when using it as a sole if clause instruction.
While there, fix 'const static' typo.
Submitted by: Arnaud Lacombe
MFC after:1 week
Modified:
head/sys/dev/ath/ath_hal/ar5211/ar5211_reset.c
Modified: head/sys/dev/ath/ath_hal/ar5211/ar5211_reset.c
==
--- head/sys/dev/ath/ath_hal/ar5211/ar5211_reset.c Sat May 29 16:11:51
2010(r208643)
+++ head/sys/dev/ath/ath_hal/ar5211/ar5211_reset.c Sat May 29 16:14:02
2010(r208644)
@@ -46,7 +46,7 @@ typedef struct {
} CHAN_INFO_2GHZ;
#define CI_2GHZ_INDEX_CORRECTION 19
-const static CHAN_INFO_2GHZ chan2GHzData[] = {
+static const CHAN_INFO_2GHZ chan2GHzData[] = {
{ 1, 0x46, 96 }, /* 2312 -19 */
{ 1, 0x46, 97 }, /* 2317 -18 */
{ 1, 0x46, 98 }, /* 2322 -17 */
@@ -926,9 +926,10 @@ ar5211IsNfGood(struct ath_hal *ah, struc
if (!getNoiseFloorThresh(ah, chan, &nfThresh))
return AH_FALSE;
- if (OS_REG_READ(ah, AR_PHY_AGC_CONTROL) & AR_PHY_AGC_CONTROL_NF)
+ if (OS_REG_READ(ah, AR_PHY_AGC_CONTROL) & AR_PHY_AGC_CONTROL_NF) {
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: NF did not complete in calibration window\n",
__func__);
+ }
nf = ar5211GetNoiseFloor(ah);
if (nf > nfThresh) {
HALDEBUG(ah, HAL_DEBUG_ANY,
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208643 - head/sys/dev/ath/ath_hal/ar5210
Author: rpaulo
Date: Sat May 29 16:11:51 2010
New Revision: 208643
URL: http://svn.freebsd.org/changeset/base/208643
Log:
Due to the way HALDEBUG() is defined, we need to add curly brackets when
using it as a sole if clause instruction.
Submitted by: Arnaud Lacombe
MFC after:1 week
Modified:
head/sys/dev/ath/ath_hal/ar5210/ar5210_reset.c
Modified: head/sys/dev/ath/ath_hal/ar5210/ar5210_reset.c
==
--- head/sys/dev/ath/ath_hal/ar5210/ar5210_reset.c Sat May 29 16:10:07
2010(r208642)
+++ head/sys/dev/ath/ath_hal/ar5210/ar5210_reset.c Sat May 29 16:11:51
2010(r208643)
@@ -526,9 +526,10 @@ ar5210PerCalibrationN(struct ath_hal *ah
/* AGC calibration (this was added to make the NF threshold check work)
*/
OS_REG_WRITE(ah, AR_PHY_AGCCTL,
OS_REG_READ(ah, AR_PHY_AGCCTL) | AR_PHY_AGC_CAL);
- if (!ath_hal_wait(ah, AR_PHY_AGCCTL, AR_PHY_AGC_CAL, 0))
+ if (!ath_hal_wait(ah, AR_PHY_AGCCTL, AR_PHY_AGC_CAL, 0)) {
HALDEBUG(ah, HAL_DEBUG_ANY, "%s: AGC calibration timeout\n",
__func__);
+ }
/* Rewrite our AGC values we stored off earlier (return AGC to normal
operation) */
OS_REG_WRITE(ah, 0x9858, reg9858);
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208642 - head/sys/dev/ath/ath_hal
Author: rpaulo
Date: Sat May 29 16:10:07 2010
New Revision: 208642
URL: http://svn.freebsd.org/changeset/base/208642
Log:
Don't shadow the global variable 'version'.
Submitted by: Arnaud Lacombe
MFC after:1 week
Modified:
head/sys/dev/ath/ath_hal/ah_eeprom_v1.c
Modified: head/sys/dev/ath/ath_hal/ah_eeprom_v1.c
==
--- head/sys/dev/ath/ath_hal/ah_eeprom_v1.c Sat May 29 13:05:41 2010
(r208641)
+++ head/sys/dev/ath/ath_hal/ah_eeprom_v1.c Sat May 29 16:10:07 2010
(r208642)
@@ -112,7 +112,7 @@ ath_hal_v1EepromAttach(struct ath_hal *a
{
HAL_EEPROM_v1 *ee = AH_PRIVATE(ah)->ah_eeprom;
uint16_t athvals[AR_EEPROM_ATHEROS_MAX];/* XXX off stack */
- uint16_t protect, version, eeval;
+ uint16_t protect, eeprom_version, eeval;
uint32_t sum;
int i, loc;
@@ -138,18 +138,18 @@ ath_hal_v1EepromAttach(struct ath_hal *a
HALDEBUG(ah, HAL_DEBUG_ATTACH, "EEPROM protect 0x%x\n", protect);
/* XXX check proper access before continuing */
- if (!ath_hal_eepromRead(ah, AR_EEPROM_VERSION, &version)) {
+ if (!ath_hal_eepromRead(ah, AR_EEPROM_VERSION, &eeprom_version)) {
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: unable to read EEPROM version\n", __func__);
return HAL_EEREAD;
}
- if (((version>>12) & 0xf) != 1) {
+ if (((eeprom_version>>12) & 0xf) != 1) {
/*
* This code only groks the version 1 EEPROM layout.
*/
HALDEBUG(ah, HAL_DEBUG_ANY,
"%s: unsupported EEPROM version 0x%x found\n",
- __func__, version);
+ __func__, eeprom_version);
return HAL_EEVERSION;
}
@@ -183,7 +183,7 @@ ath_hal_v1EepromAttach(struct ath_hal *a
return HAL_ENOMEM;
}
- ee->ee_version = version;
+ ee->ee_version = eeprom_version;
ee->ee_protect = protect;
ee->ee_antenna = athvals[2];
ee->ee_biasCurrents = athvals[3];
@@ -243,7 +243,7 @@ ath_hal_v1EepromAttach(struct ath_hal *a
}
AH_PRIVATE(ah)->ah_eeprom = ee;
- AH_PRIVATE(ah)->ah_eeversion = version;
+ AH_PRIVATE(ah)->ah_eeversion = eeprom_version;
AH_PRIVATE(ah)->ah_eepromDetach = v1EepromDetach;
AH_PRIVATE(ah)->ah_eepromGet = v1EepromGet;
AH_PRIVATE(ah)->ah_eepromSet = v1EepromSet;
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208639 - head/sys/pc98/pc98
Author: nyan
Date: Sat May 29 09:12:02 2010
New Revision: 208639
URL: http://svn.freebsd.org/changeset/base/208639
Log:
MFi386: the part of revision 181809
Use SEL_KPL macro.
Modified:
head/sys/pc98/pc98/machdep.c
Modified: head/sys/pc98/pc98/machdep.c
==
--- head/sys/pc98/pc98/machdep.cSat May 29 09:07:40 2010
(r208638)
+++ head/sys/pc98/pc98/machdep.cSat May 29 09:12:02 2010
(r208639)
@@ -1439,7 +1439,7 @@ struct soft_segment_descriptor gdt_segs[
{ .ssd_base = 0x0,
.ssd_limit = 0x0,
.ssd_type = 0,
- .ssd_dpl = 0,
+ .ssd_dpl = SEL_KPL,
.ssd_p = 0,
.ssd_xx = 0, .ssd_xx1 = 0,
.ssd_def32 = 0,
@@ -1448,7 +1448,7 @@ struct soft_segment_descriptor gdt_segs[
{ .ssd_base = 0x0,
.ssd_limit = 0xf,
.ssd_type = SDT_MEMRWA,
- .ssd_dpl = 0,
+ .ssd_dpl = SEL_KPL,
.ssd_p = 1,
.ssd_xx = 0, .ssd_xx1 = 0,
.ssd_def32 = 1,
@@ -1475,7 +1475,7 @@ struct soft_segment_descriptor gdt_segs[
{ .ssd_base = 0x0,
.ssd_limit = 0xf,
.ssd_type = SDT_MEMERA,
- .ssd_dpl = 0,
+ .ssd_dpl = SEL_KPL,
.ssd_p = 1,
.ssd_xx = 0, .ssd_xx1 = 0,
.ssd_def32 = 1,
@@ -1484,7 +1484,7 @@ struct soft_segment_descriptor gdt_segs[
{ .ssd_base = 0x0,
.ssd_limit = 0xf,
.ssd_type = SDT_MEMRWA,
- .ssd_dpl = 0,
+ .ssd_dpl = SEL_KPL,
.ssd_p = 1,
.ssd_xx = 0, .ssd_xx1 = 0,
.ssd_def32 = 1,
@@ -1511,7 +1511,7 @@ struct soft_segment_descriptor gdt_segs[
{ .ssd_base = 0x400,
.ssd_limit = 0xf,
.ssd_type = SDT_MEMRWA,
- .ssd_dpl = 0,
+ .ssd_dpl = SEL_KPL,
.ssd_p = 1,
.ssd_xx = 0, .ssd_xx1 = 0,
.ssd_def32 = 1,
___
svn-src-head@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r208638 - head/sys/pc98/pc98
Author: nyan
Date: Sat May 29 09:07:40 2010
New Revision: 208638
URL: http://svn.freebsd.org/changeset/base/208638
Log:
MFi386: revision 178471
- Add an integer argument to idle to indicate how likely we are to wake
from idle over the next tick.
- Add a new MD routine, cpu_wake_idle() to wakeup idle threads who are
suspended in cpu specific states. This function can fail and cause the
scheduler to fall back to another mechanism (ipi).
- Implement support for mwait in cpu_idle() on i386/amd64 machines that
support it. mwait is a higher performance way to synchronize cpus
as compared to hlt & ipis.
- Allow selecting the idle routine by name via sysctl machdep.idle. This
replaces machdep.cpu_idle_hlt. Only idle routines supported by the
current machine are permitted.
Modified:
head/sys/pc98/pc98/machdep.c
Modified: head/sys/pc98/pc98/machdep.c
==
--- head/sys/pc98/pc98/machdep.cSat May 29 09:07:25 2010
(r208637)
+++ head/sys/pc98/pc98/machdep.cSat May 29 09:07:40 2010
(r208638)
@@ -1108,6 +1108,7 @@ cpu_est_clockrate(int cpu_id, uint64_t *
return (0);
}
+
/*
* Shutdown the CPU as much as possible
*/
@@ -1118,70 +1119,177 @@ cpu_halt(void)
__asm__ ("hlt");
}
-/*
- * Hook to idle the CPU when possible. In the SMP case we default to
- * off because a halted cpu will not currently pick up a new thread in the
- * run queue until the next timer tick. If turned on this will result in
- * approximately a 4.2% loss in real time performance in buildworld tests
- * (but improves user and sys times oddly enough), and saves approximately
- * 5% in power consumption on an idle machine (tests w/2xCPU 1.1GHz P3).
- *
- * XXX we need to have a cpu mask of idle cpus and generate an IPI or
- * otherwise generate some sort of interrupt to wake up cpus sitting in HLT.
- * Then we can have our cake and eat it too.
- *
- * XXX I'm turning it on for SMP as well by default for now. It seems to
- * help lock contention somewhat, and this is critical for HTT. -Peter
- */
-static int cpu_idle_hlt = 1;
-TUNABLE_INT("machdep.cpu_idle_hlt", &cpu_idle_hlt);
-SYSCTL_INT(_machdep, OID_AUTO, cpu_idle_hlt, CTLFLAG_RW,
-&cpu_idle_hlt, 0, "Idle loop HLT enable");
-
static void
-cpu_idle_default(void)
+cpu_idle_hlt(int busy)
{
/*
-* we must absolutely guarentee that hlt is the
-* absolute next instruction after sti or we
-* introduce a timing window.
+* we must absolutely guarentee that hlt is the next instruction
+* after sti or we introduce a timing window.
*/
- __asm __volatile("sti; hlt");
+ disable_intr();
+ if (sched_runnable())
+ enable_intr();
+ else
+ __asm __volatile("sti; hlt");
}
-/*
- * Note that we have to be careful here to avoid a race between checking
- * sched_runnable() and actually halting. If we don't do this, we may waste
- * the time between calling hlt and the next interrupt even though there
- * is a runnable process.
- */
+static void
+cpu_idle_spin(int busy)
+{
+ return;
+}
+
+void (*cpu_idle_fn)(int) = cpu_idle_hlt;
+
void
cpu_idle(int busy)
{
-
-#ifdef SMP
+#if defined(SMP)
if (mp_grab_cpu_hlt())
return;
#endif
+ cpu_idle_fn(busy);
+}
+
+/*
+ * mwait cpu power states. Lower 4 bits are sub-states.
+ */
+#defineMWAIT_C00xf0
+#defineMWAIT_C10x00
+#defineMWAIT_C20x10
+#defineMWAIT_C30x20
+#defineMWAIT_C40x30
+
+#defineMWAIT_DISABLED 0x0
+#defineMWAIT_WOKEN 0x1
+#defineMWAIT_WAITING 0x2
+
+static void
+cpu_idle_mwait(int busy)
+{
+ int *mwait;
- if (cpu_idle_hlt) {
- disable_intr();
- if (sched_runnable())
- enable_intr();
- else
- (*cpu_idle_hook)();
+ mwait = (int *)PCPU_PTR(monitorbuf);
+ *mwait = MWAIT_WAITING;
+ if (sched_runnable())
+ return;
+ cpu_monitor(mwait, 0, 0);
+ if (*mwait == MWAIT_WAITING)
+ cpu_mwait(0, MWAIT_C1);
+}
+
+static void
+cpu_idle_mwait_hlt(int busy)
+{
+ int *mwait;
+
+ mwait = (int *)PCPU_PTR(monitorbuf);
+ if (busy == 0) {
+ *mwait = MWAIT_DISABLED;
+ cpu_idle_hlt(busy);
+ return;
}
+ *mwait = MWAIT_WAITING;
+ if (sched_runnable())
+ return;
+ cpu_monitor(mwait, 0, 0);
+ if (*mwait == MWAIT_WAITING)
+ cpu_mwait(0, MWAIT_C1);
}
int
cpu_idle_wakeup(int cpu)
{
+ struct pcpu *pcpu;
+ int *mwait;
- return (0);
+ if (cpu_idle_fn == cpu_idle_spin)
+ return (1);
+ if (cpu_idle_fn != cpu_idle_mwait && cpu
