mailing.openbsd.tech

truss

Fri, 05 Sep 2008 17:01:13 PDT

This diff makes some changes to ktracing and introduces a new
fktrace() syscall (like NetBSD) allowing for a truss-like program.

Included is also a diff that makes a few modifications to kdump
to support the new feature and links a new ktruss program to it.

Installation instructions:
- install this patch:
# cd /usr/src && patch < ktruss.diff
- rebuild syscalls:
# cd /sys/kern && sh makesyscalls.sh syscalls.conf syscalls.master
- build, install and boot into new kernel:
see http://www.openbsd.org/faq/faq5.html#BldKernel
- install new headers:
# cd /usr/src/include
# make prereq
# make includes
- update libc:
# cd /usr/src/lib/libc
# rm -rf /usr/obj/lib/libc
# make obj && make depend && make && make install
- update kdump:
# cd /usr/src/usr.bin/kdump
# rm -rf /usr/obj/usr.bin/kdump
# make obj && make depend && make && make install

Confession: some ideas I stole from NetBSD.

Index: lib/libc/sys/Makefile.inc
===================================================================
RCS file: /cvs/src/lib/libc/sys/Makefile.inc,v
retrieving revision 1.85
diff -u -p -r1.85 Makefile.inc
--- lib/libc/sys/Makefile.inc 22 Mar 2008 19:18:42 -0000 1.85
+++ lib/libc/sys/Makefile.inc 5 Sep 2008 22:25:46 -0000
@@ -38,7 +38,7 @@ ASM= accept.o access.o acct.o adjfreq.o
chown.o chroot.o clock_gettime.o clock_settime.o clock_getres.o \
close.o closefrom.o connect.o dup.o dup2.o execve.o fchdir.o \
fchflags.o fchmod.o fchown.o fcntl.o fhopen.o fhstat.o fhstatfs.o \
- flock.o fpathconf.o fstat.o \
+ fktrace.o flock.o fpathconf.o fstat.o \
fstatfs.o fsync.o futimes.o getdirentries.o getegid.o geteuid.o \
getfh.o getfsstat.o getgid.o getgroups.o getitimer.o getpeereid.o \
getpeername.o \
Index: sys/kern/kern_exec.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_exec.c,v
retrieving revision 1.106
diff -u -p -r1.106 kern_exec.c
--- sys/kern/kern_exec.c 13 Feb 2008 19:31:22 -0000 1.106
+++ sys/kern/kern_exec.c 5 Sep 2008 22:26:05 -0000
@@ -499,10 +499,8 @@ sys_execve(struct proc *p, void *v, regi
* If process is being ktraced, turn off - unless
* root set it.
*/
- if (p->p_tracep && !(p->p_traceflag & KTRFAC_ROOT)) {
- p->p_traceflag = 0;
- ktrsettracevnode(p, NULL);
- }
+ if (p->p_tracep && !(p->p_traceflag & KTRFAC_ROOT))
+ ktrcleartracefile(p);
#endif
p->p_ucred = crcopy(cred);
if (attr.va_mode & VSUID)
Index: sys/kern/kern_exit.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_exit.c,v
retrieving revision 1.73
diff -u -p -r1.73 kern_exit.c
--- sys/kern/kern_exit.c 11 May 2008 23:54:40 -0000 1.73
+++ sys/kern/kern_exit.c 5 Sep 2008 22:26:05 -0000
@@ -224,9 +224,8 @@ exit1(struct proc *p, int rv, int flags)
/*
* release trace file
*/
- p->p_traceflag = 0; /* don't trace the vrele() */
if (p->p_tracep)
- ktrsettracevnode(p, NULL);
+ ktrcleartracefile(p);
#endif
#if NSYSTRACE > 0
if (ISSET(p->p_flag, P_SYSTRACE))
Index: sys/kern/kern_fork.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_fork.c,v
retrieving revision 1.95
diff -u -p -r1.95 kern_fork.c
--- sys/kern/kern_fork.c 11 May 2008 23:54:40 -0000 1.95
+++ sys/kern/kern_fork.c 5 Sep 2008 22:26:05 -0000
@@ -327,8 +327,7 @@ fork1(struct proc *p1, int exitsig, int
*/
if (p1->p_traceflag & KTRFAC_INHERIT) {
p2->p_traceflag = p1->p_traceflag;
- if ((p2->p_tracep = p1->p_tracep) != NULL)
- VREF(p2->p_tracep);
+ ktrsettracefile(p2, p1->p_tracep);
}
#endif

Index: sys/kern/kern_ktrace.c
===================================================================
RCS file: /cvs/src/sys/kern/kern_ktrace.c,v
retrieving revision 1.43
diff -u -p -r1.43 kern_ktrace.c
--- sys/kern/kern_ktrace.c 22 May 2008 17:04:59 -0000 1.43
+++ sys/kern/kern_ktrace.c 5 Sep 2008 22:26:05 -0000
@@ -39,6 +39,7 @@
#include
#include
#include
+#include
#include
#include
#include
@@ -53,31 +54,42 @@
#include

void ktrinitheader(struct ktr_header *, struct proc *, int);
-int ktrops(struct proc *, struct proc *, int, int, struct vnode *);
-int ktrsetchildren(struct proc *, struct proc *, int, int,
- struct vnode *);
+int ktrops(struct proc *, struct proc *, int, int, struct file *);
+int ktrsetchildren(struct proc *, struct proc *, int, int, struct file *);
int ktrwrite(struct proc *, struct ktr_header *);
int ktrcanset(struct proc *, struct proc *);
+int ktrcommon(struct proc *, struct file *, int, int, pid_t);

/*
- * Change the trace vnode in a correct way (to avoid races).
+ * Change the trace file in a correct way (to avoid races).
*/
void
-ktrsettracevnode(struct proc *p, struct vnode *newvp)
+ktrsettracefile(struct proc *p, struct file *fp)
{
- struct vnode *vp;
+ struct file *oldfp;

- if (p->p_tracep == newvp) /* avoid work */
+ if (p->p_tracep && p->p_tracep == fp) /* avoid work */
return;

- if (newvp != NULL)
- VREF(newvp);
+ FREF(fp);

- vp = p->p_tracep;
- p->p_tracep = newvp;
+ oldfp = p->p_tracep;
+ p->p_tracep = fp;

- if (vp != NULL)
- vrele(vp);
+ if (oldfp)
+ closef(oldfp, p);
+}
+
+void
+ktrcleartracefile(struct proc *p)
+{
+ struct file *fp;
+
+ p->p_traceflag = 0; /* prevent tracing of vrele */
+ fp = p->p_tracep;
+ p->p_tracep = NULL;
+ if (fp)
+ closef(fp, p);
}

void
@@ -87,7 +99,7 @@ ktrinitheader(struct ktr_header *kth, st
kth->ktr_type = type;
microtime(&kth->ktr_time);
kth->ktr_pid = p->p_pid;
- bcopy(p->p_comm, kth->ktr_comm, MAXCOMLEN);
+ bcopy(p->p_comm, kth->ktr_comm, MAXCOMLEN);
}

void
@@ -232,7 +244,6 @@ ktrgenio(struct proc *p, int fd, enum ui

free(ktp, M_TEMP);
p->p_traceflag &= ~KTRFAC_ACTIVE;
-
}

void
@@ -280,7 +291,7 @@ ktrcsw(struct proc *p, int out, int user
*/
/* ARGSUSED */
int
-sys_ktrace(struct proc *curp, void *v, register_t *retval)
+sys_ktrace(struct proc *p, void *v, register_t *retval)
{
struct sys_ktrace_args /* {
syscallarg(const char *) fname;
@@ -288,122 +299,174 @@ sys_ktrace(struct proc *curp, void *v, r
syscallarg(int) facs;
syscallarg(pid_t) pid;
} */ *uap = v;
- struct vnode *vp = NULL;
- struct proc *p = NULL;
- struct pgrp *pg;
- int facs = SCARG(uap, facs) & ~((unsigned) KTRFAC_ROOT);
- int ops = KTROP(SCARG(uap, ops));
- int descend = SCARG(uap, ops) & KTRFLAG_DESCEND;
- int ret = 0;
- int error = 0;
+ struct file *fp = NULL;
+ struct vnode *vp;
struct nameidata nd;
+ int fd, error, ops = SCARG(uap, ops);

- curp->p_traceflag |= KTRFAC_ACTIVE;
- if (ops != KTROP_CLEAR) {
+ p->p_traceflag |= KTRFAC_ACTIVE;
+ if (KTROP(ops) == KTROP_SET) {
/*
- * an operation which requires a file argument.
+ * The process file table is used to store the
+ * trace file struct until we return.
*/
- NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, fname),
- curp);
- if ((error = vn_open(&nd, FREAD|FWRITE|O_NOFOLLOW, 0)) != 0) {
- curp->p_traceflag &= ~KTRFAC_ACTIVE;
- return (error);
+ fdplock(p->p_fd);
+ if ((error = falloc(p, &fp, &fd)) != 0) {
+ fdpunlock(p->p_fd);
+ goto done;
}
- vp = nd.ni_vp;

- VOP_UNLOCK(vp, 0, curp);
+ NDINIT(&nd, LOOKUP, FOLLOW, UIO_USERSPACE, SCARG(uap, fname), p);
+ if ((error = vn_open(&nd, FWRITE, 0)) != 0)
+ goto done;
+ vp = nd.ni_vp;
+ VOP_UNLOCK(vp, 0, p);
if (vp->v_type != VREG) {
- (void) vn_close(vp, FREAD|FWRITE, curp->p_ucred, curp);
- curp->p_traceflag &= ~KTRFAC_ACTIVE;
- return (EACCES);
+ error = EACCES;
+ goto done;
}
+ fp->f_flag = FWRITE;
+ fp->f_type = DTYPE_VNODE;
+ fp->f_ops = &vnops;
+ fp->f_data = vp;
+ FILE_SET_MATURE(fp);
+ }
+ error = ktrcommon(p, fp, SCARG(uap, ops), SCARG(uap, facs),
+ SCARG(uap, pid));
+ done:
+ if (fp) {
+ fdremove(p->p_fd, fd);
+ fdpunlock(p->p_fd);
+ }
+ p->p_traceflag &= ~KTRFAC_ACTIVE;
+ return (error);
+}
+
+/*
+ * fktrace system call
+ */
+/* ARGSUSED */
+int
+sys_fktrace(struct proc *p, void *v, register_t *retval)
+{
+ struct sys_fktrace_args /* {
+ syscallarg(int) fd;
+ syscallarg(int) ops;
+ syscallarg(int) facs;
+ syscallarg(pid_t) pid;
+ } */ *uap = v;
+ struct file *fp = NULL;
+ int error;
+
+ p->p_traceflag |= KTRFAC_ACTIVE;
+ if ((fp = fd_getfile(p->p_fd, SCARG(uap, fd))) == NULL) {
+ error = EBADF;
+ goto done;
}
+ if ((fp->f_flag & FWRITE) == 0)
+ error = EACCES;
+ else
+ error = ktrcommon(p, fp, SCARG(uap, ops),
+ SCARG(uap, facs), SCARG(uap, pid));
+ if (KTROP(SCARG(uap, ops)) == KTROP_SET && !error)
+ fp->f_count++;
+done:
+ p->p_traceflag &= ~KTRFAC_ACTIVE;
+ return (error);
+}
+
+/*
+ * Do initialization common to both sys_ktrace() and sys_fktrace().
+ * Expects a file pointer and descriptor index to the tracefile.
+ */
+int
+ktrcommon(struct proc *curp, struct file *fp, int ops, int facs, pid_t pid)
+{
+ int descend, ret = 0, error = 0;
+ struct pgrp *pg;
+ struct proc *p;
+
+ facs &= ~KTRFAC_ROOT;
+ descend = ops & KTRFLAG_DESCEND;
+ ops = KTROP(ops);
+
+ if (ops != KTROP_CLEAR && ops != KTROP_CLEARFILE && ops != KTROP_SET)
+ return (EINVAL);
+ if (ops == KTROP_CLEAR || ops == KTROP_CLEARFILE)
+ fp = p->p_tracep;
+
/*
- * Clear all uses of the tracefile
+ * Clear all uses of the tracefile.
*/
if (ops == KTROP_CLEARFILE) {
- LIST_FOREACH(p, &allproc, p_list) {
- if (p->p_tracep == vp) {
- if (ktrcanset(curp, p)) {
- p->p_traceflag = 0;
- ktrsettracevnode(p, NULL);
- } else
+ LIST_FOREACH(p, &allproc, p_list)
+ if (p->p_tracep->f_data == fp->f_data) {
+ if (ktrcanset(curp, p))
+ ktrcleartracefile(p);
+ else
error = EPERM;
}
- }
- goto done;
+ return (error);
}
/*
* need something to (un)trace (XXX - why is this here?)
*/
- if (!facs) {
- error = EINVAL;
- goto done;
- }
- /*
+ if (!facs)
+ return (EINVAL);
+ /*
* do it
*/
- if (SCARG(uap, pid) < 0) {
+ if (pid < 0) {
/*
* by process group
*/
- pg = pgfind(-SCARG(uap, pid));
- if (pg == NULL) {
- error = ESRCH;
- goto done;
- }
+ pg = pgfind(-pid);
+ if (pg == NULL)
+ return (ESRCH);
LIST_FOREACH(p, &pg->pg_members, p_pglist)
if (descend)
- ret |= ktrsetchildren(curp, p, ops, facs, vp);
- else
- ret |= ktrops(curp, p, ops, facs, vp);
-
+ ret |= ktrsetchildren(curp, p, ops, facs, fp);
+ else
+ ret |= ktrops(curp, p, ops, facs, fp);
} else {
/*
* by pid
*/
- p = pfind(SCARG(uap, pid));
- if (p == NULL) {
- error = ESRCH;
- goto done;
- }
+ p = pfind(pid);
+ if (p == NULL)
+ return (ESRCH);
if (descend)
- ret |= ktrsetchildren(curp, p, ops, facs, vp);
+ ret |= ktrsetchildren(curp, p, ops, facs, fp);
else
- ret |= ktrops(curp, p, ops, facs, vp);
+ ret |= ktrops(curp, p, ops, facs, fp);
}
if (!ret)
error = EPERM;
-done:
- if (vp != NULL)
- (void) vn_close(vp, FWRITE, curp->p_ucred, curp);
- curp->p_traceflag &= ~KTRFAC_ACTIVE;
return (error);
}

int
-ktrops(struct proc *curp, struct proc *p, int ops, int facs, struct vnode *vp)
+ktrops(struct proc *curp, struct proc *p, int ops, int facs, struct file *fp)
{

if (!ktrcanset(curp, p))
return (0);
+
if (ops == KTROP_SET) {
- ktrsettracevnode(p, vp);
+ ktrsettracefile(p, fp);
p->p_traceflag |= facs;
if (curp->p_ucred->cr_uid == 0)
p->p_traceflag |= KTRFAC_ROOT;
- } else {
+ } else {
/* KTROP_CLEAR */
- if (((p->p_traceflag &= ~facs) & KTRFAC_MASK) == 0) {
+ if (((p->p_traceflag &= ~facs) & KTRFAC_MASK) == 0)
/* no more tracing */
- p->p_traceflag = 0;
- ktrsettracevnode(p, NULL);
- }
+ ktrcleartracefile(p);
}

/*
* Emit an emulation record, every time there is a ktrace
- * change/attach request.
+ * change/attach request.
*/
if (KTRPOINT(p, KTR_EMUL))
ktremul(p, p->p_emul->e_name);
@@ -413,14 +476,14 @@ ktrops(struct proc *curp, struct proc *p

int
ktrsetchildren(struct proc *curp, struct proc *top, int ops, int facs,
- struct vnode *vp)
+ struct file *fp)
{
struct proc *p;
int ret = 0;

p = top;
for (;;) {
- ret |= ktrops(curp, p, ops, facs, vp);
+ ret |= ktrops(curp, p, ops, facs, fp);
/*
* If this process has children, descend to them next,
* otherwise do any siblings, and if done with this level,
@@ -447,10 +510,11 @@ ktrwrite(struct proc *p, struct ktr_head
struct uio auio;
struct iovec aiov[2];
int error;
- struct vnode *vp = p->p_tracep;
+ struct file *fp = p->p_tracep;

- if (vp == NULL)
+ if (fp == NULL)
return 0;
+
auio.uio_iov = &aiov[0];
auio.uio_offset = 0;
auio.uio_segflg = UIO_SYSSPACE;
@@ -466,31 +530,27 @@ ktrwrite(struct proc *p, struct ktr_head
aiov[1].iov_len = kth->ktr_len;
auio.uio_resid += kth->ktr_len;
}
- vn_lock(vp, LK_EXCLUSIVE | LK_RETRY, p);
- error = VOP_WRITE(vp, &auio, IO_UNIT|IO_APPEND, p->p_ucred);
- VOP_UNLOCK(vp, 0, p);
+ error = (*fp->f_ops->fo_write)(fp, &fp->f_offset, &auio, fp->f_cred);
if (!error)
- return 0;
+ return (0);
+
/*
- * If error encountered, give up tracing on this vnode.
+ * If an error is encountered, give up tracing to this file.
*/
log(LOG_NOTICE, "ktrace write failed, errno %%d, tracing stopped\n",
error);
- LIST_FOREACH(p, &allproc, p_list) {
- if (p->p_tracep == vp) {
- p->p_traceflag = 0;
- ktrsettracevnode(p, NULL);
- }
- }
-
- return error;
+ LIST_FOREACH(p, &allproc, p_list)
+ if (p->p_tracep &&
+ p->p_tracep->f_data == fp->f_data)
+ ktrcleartracefile(p);
+ return (error);
}

/*
* Return true if caller has permission to set the ktracing state
* of target. Essentially, the target can't possess any
* more permissions than the caller. KTRFAC_ROOT signifies that
- * root previously set the tracing status on the target process, and
+ * root previously set the tracing status on the target process, and
* so, only root may further change it.
*
* TODO: check groups. use caller effective gid.
@@ -513,4 +573,4 @@ ktrcanset(struct proc *callp, struct pro
return (0);
}

-#endif
+#endif /* KTRACE */
Index: sys/kern/syscalls.master
===================================================================
RCS file: /cvs/src/sys/kern/syscalls.master,v
retrieving revision 1.90
diff -u -p -r1.90 syscalls.master
--- sys/kern/syscalls.master 16 Mar 2008 19:42:57 -0000 1.90
+++ sys/kern/syscalls.master 5 Sep 2008 22:26:06 -0000
@@ -618,3 +618,7 @@
308 STD { int sys_fstatfs(int fd, struct statfs *buf); }
309 STD { int sys_fhstatfs(const fhandle_t *fhp, \
struct statfs *buf); }
+
+#ifdef KTRACE
+310 STD { int sys_fktrace(int fd, int ops, int facs, pid_t pid); }
+#endif
Index: sys/sys/ktrace.h
===================================================================
RCS file: /cvs/src/sys/sys/ktrace.h,v
retrieving revision 1.9
diff -u -p -r1.9 ktrace.h
--- sys/sys/ktrace.h 17 May 2006 02:11:25 -0000 1.9
+++ sys/sys/ktrace.h 5 Sep 2008 22:26:07 -0000
@@ -158,6 +158,7 @@ struct ktr_csw {
#include

__BEGIN_DECLS
+int fktrace(int, int, int, pid_t);
int ktrace(const char *, int, int, pid_t);
__END_DECLS

@@ -171,6 +172,7 @@ void ktrpsig(struct proc *, int, sig_t,
void ktrsyscall(struct proc *, register_t, size_t, register_t []);
void ktrsysret(struct proc *, register_t, int, register_t);

-void ktrsettracevnode(struct proc *, struct vnode *);
+void ktrcleartracefile(struct proc *);
+void ktrsettracefile(struct proc *, struct file *);

#endif /* !_KERNEL */
Index: sys/sys/proc.h
===================================================================
RCS file: /cvs/src/sys/sys/proc.h,v
retrieving revision 1.103
diff -u -p -r1.103 proc.h
--- sys/sys/proc.h 5 May 2008 15:37:41 -0000 1.103
+++ sys/sys/proc.h 5 Sep 2008 22:26:07 -0000
@@ -206,7 +206,7 @@ struct proc {
u_quad_t p_iticks; /* Statclock hits processing intr. */

int p_traceflag; /* Kernel trace points. */
- struct vnode *p_tracep; /* Trace to vnode. */
+ struct file *p_tracep; /* Trace data to file. */

void *p_systrace; /* Back pointer to systrace */

Index: usr.bin/kdump/Makefile
===================================================================
RCS file: /cvs/src/usr.bin/kdump/Makefile,v
retrieving revision 1.16
diff -u -p -r1.16 Makefile
--- usr.bin/kdump/Makefile 31 Jul 2005 10:11:19 -0000 1.16
+++ usr.bin/kdump/Makefile 5 Sep 2008 22:26:08 -0000
@@ -3,7 +3,9 @@
LDSTATIC=${STATIC}
PROG= kdump
CFLAGS+=-I${.CURDIR}/../ktrace -I${.CURDIR} -I${.CURDIR}/../../sys
-SRCS= kdump.c ioctl.c subr.c
+SRCS= kdump.c ioctl.c subr.c ktruss.c
+LINKS+= ${BINDIR}/kdump ${BINDIR}/ktruss
+MAN= kdump.1 ktruss.1
.PATH: ${.CURDIR}/../ktrace
CLEANFILES+=ioctl.c

Index: usr.bin/kdump/kdump.c
===================================================================
RCS file: /cvs/src/usr.bin/kdump/kdump.c,v
retrieving revision 1.39
diff -u -p -r1.39 kdump.c
--- usr.bin/kdump/kdump.c 29 May 2007 02:01:03 -0000 1.39
+++ usr.bin/kdump/kdump.c 5 Sep 2008 22:26:08 -0000
@@ -163,6 +163,18 @@ static void setemul(const char *);
static void usage(void);

int
+is_ktruss(const char *name)
+{
+ const char *s;
+
+ if ((s = strrchr(name, '/')) != NULL)
+ name = s;
+ if (strcmp(name, "ktruss") == 0)
+ return (1);
+ return (0);
+}
+
+int
main(int argc, char *argv[])
{
int ch, silent;
@@ -172,6 +184,9 @@ main(int argc, char *argv[])

current = &emulations[0]; /* native */

+ if (is_ktruss(argv[0]))
+ return (ktruss_main(argc, argv));
+
while ((ch = getopt(argc, argv, "e:f:dlm:nRp:Tt:xX")) != -1)
switch (ch) {
case 'e':
@@ -217,12 +232,21 @@ main(int argc, char *argv[])
}
if (argc > optind)
usage();
+ if (!freopen(tracefile, "r", stdin))
+ err(1, "%%s", tracefile);
+ dump(trpoints);
+ exit(0);
+}
+
+void
+dump(int trpoints)
+{
+ int ktrlen, size, silent;
+ void *m;

m = malloc(size = 1025);
if (m == NULL)
err(1, NULL);
- if (!freopen(tracefile, "r", stdin))
- err(1, "%%s", tracefile);
while (fread_tail(&ktr_header, sizeof(struct ktr_header), 1)) {
silent = 0;
if (pid != -1 && pid != ktr_header.ktr_pid)
@@ -271,7 +295,6 @@ main(int argc, char *argv[])
if (tail)
(void)fflush(stdout);
}
- exit(0);
}

static int
Index: usr.bin/kdump/kdump.h
===================================================================
RCS file: /cvs/src/usr.bin/kdump/kdump.h,v
retrieving revision 1.1
diff -u -p -r1.1 kdump.h
--- usr.bin/kdump/kdump.h 2 Jul 2003 20:54:17 -0000 1.1
+++ usr.bin/kdump/kdump.h 5 Sep 2008 22:26:08 -0000
@@ -1,3 +1,4 @@
+/* $OpenBSD$ */
/*
* Copyright (c) 2003 Theo de Raadt openbsd.org>
*
@@ -15,3 +16,7 @@
*/

const char *ioctlname(unsigned long);
+int ktruss_main(int, char *[]);
+void dump(int);
+
+extern int timestamp;
Index: usr.bin/kdump/ktruss.1
===================================================================
RCS file: usr.bin/kdump/ktruss.1
diff -N usr.bin/kdump/ktruss.1
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ usr.bin/kdump/ktruss.1 5 Sep 2008 22:26:08 -0000
@@ -0,0 +1,73 @@
+.\" $OpenBSD$
+.\" Written by Jared Yanovich openbsd.org>
+.\" Public domain, 2008/09/05.
+.Dd September 5, 2008
+.Dt KTRUSS 1
+.Os
+.Sh NAME
+.Nm ktruss
+.Nd trace system calls
+.Sh SYNOPSIS
+.Bk -words
+.Nm ktruss
+.Op Fl Ddf
+.Op Fl o Ar file
+.Ar command
+.Op Ar argument ...
+.Nm ktruss
+.Op Fl Ddf
+.Op Fl o Ar file
+.Fl p Ar pid
+.Ek
+.Sh DESCRIPTION
+The
+.Nm
+utility traces the system calls made by a process.
+.Nm
+can attach to a live process or start a new one and attach to it
+immediately.
+.Pp
+This following options are available:
+.Bl -tag -width indent
+.It Fl D
+Display the elapsed time between each system call.
+.It Fl d
+Display an absolute date at the time of each system call.
+.It Fl f
+Follow any
+.Xr fork 2
+operations that the given process makes and trace the child processes
+generated by such operations.
+When attaching to live processes (see
+.Fl p ) ,
+tracing will descend to include any child processes of the target
+process as well.
+.It Fl o Ar file
+Send the output to the given
+.Ar file
+instead of to standard output by default.
+.It Fl p Ar pid
+Attach to and trace the live process with the given
+.Ar pid
+instead of executing
+.Xr fork 2
+and tracing a new process.
+.El
+.Sh EXAMPLES
+Display all system calls made in the command
+.Ic echo hello :
+.Pp
+.Dl $ ktruss echo hello
+.Pp
+Attach to the process with PID 52, tracing any system calls it
+performs:
+.Pp
+.Dl $ ktruss -p 52
+.Pp
+.Sh SEE ALSO
+.Xr ktrace 1 ,
+.Xr ps 1 ,
+.Xr fork 2
+.Sh BUGS
+Although the information is displayed in real-time, it is by no means an
+accurate depiction of when actual events are occurring in the process.
Index: usr.bin/kdump/ktruss.c
===================================================================
RCS file: usr.bin/kdump/ktruss.c
diff -N usr.bin/kdump/ktruss.c
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ usr.bin/kdump/ktruss.c 5 Sep 2008 22:26:08 -0000
@@ -0,0 +1,115 @@
+/* $OpenBSD$ */
+/*
+ * Written by Jared Yanovich openbsd.org>
+ * Public domain, 2008/09/05.
+ */
+
+#include
+#include
+#include
+
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include "ktrace.h"
+#include "kdump.h"
+
+int show_argv;
+int show_stats;
+int show_envp;
+const char *outfn;
+
+void ktruss_usage(void);
+
+int
+ktruss_main(int argc, char *argv[])
+{
+ int ch, trpoints, ops, pidset, pid, fds[2];
+ const char *errstr;
+ FILE *infp;
+
+ trpoints = DEF_POINTS;
+ ops = pid = pidset = 0;
+ while ((ch = getopt(argc, argv, "Ddfo:p:")) != -1) {
+ switch (ch) {
+ case 'D':
+ timestamp = 1; /* relative */
+ break;
+ case 'd':
+ timestamp = 2; /* absolute */
+ break;
+ case 'f':
+ trpoints |= KTRFAC_INHERIT;
+ ops |= KTRFLAG_DESCEND;
+ break;
+ case 'o':
+ outfn = optarg;
+ break;
+ case 'p':
+ pid = strtonum(optarg, 1, INT_MAX, &errstr);
+ if (errstr != NULL)
+ errx(1, "%%s: %%s", optarg, errstr);
+ pidset = 1;
+ break;
+ default:
+ ktruss_usage();
+ /* NOTREACHED */
+ }
+ }
+ argc -= optind;
+ argv += optind;
+
+ if ((pidset && argc) ||
+ (!pidset && !argc))
+ ktruss_usage();
+
+ /* Setup input as stdin for dump(). */
+ if (pipe(fds) == -1)
+ err(1, "pipe");
+
+ if (pidset) {
+ if (fktrace(fds[1], ops, trpoints, pid) == -1)
+ err(1, "fktrace");
+ } else {
+ switch (fork()) {
+ case -1:
+ err(1, "fork");
+ /* NOTREACHED */
+ case 0:
+ pid = getpid();
+ if (fktrace(fds[1], ops, trpoints, pid) == -1)
+ err(1, "fktrace");
+ close(fds[1]);
+ execvp(argv[0], &argv[0]);
+ err(1, "execve");
+ /* NOTREACHED */
+ }
+ }
+ /* Set up output file. */
+ if (outfn != NULL && freopen(outfn, "w", stdout) == NULL)
+ err(1, "%%s", outfn);
+
+ if ((infp = fdopen(fds[0], "r")) == NULL)
+ err(1, "fdopen");
+ if (dup2(fds[0], STDIN_FILENO) == -1)
+ err(1, "dup2");
+ close(fds[1]);
+ dump(trpoints);
+ exit(0);
+}
+
+void
+ktruss_usage(void)
+{
+ extern char *__progname;
+
+ fprintf(stderr,
+ "usage: %%s [-Ddf] [-o file] command [argument ...]\n"
+ " %%s [-Ddf] [-o file] -p pid\n",
+ __progname, __progname);
+ exit(1);
+}

Posted In: mailing.openbsd.tech

4 Comments

Da li zaista znate sve o veli

Fri, 05 Sep 2008 13:27:28 PDT

Top Shop

Top e-revija: 26 l 5. septembar 2008.

Najbolja praktiD na reE!enja i saveti za bolji E>ivot

PoD etna l Budi fit l Lepota l Zdravlje l Kuhinja i domaDinstvo

Zabava i deca l Carstvo igraD aka l Knjige

Top Shop

HIT TV proizvodi!

ab rocket

Ab Rocket

steam mop

H20
Steam Mop

leg magic

Massage Cusshion

massage cushion

Kleen
Kut

turbo maximus

Turbo Maximus

Da li znate baE! sve o veliD inama?

"Penjanje uz stepenice" za lepE!i izgled

5 trikova za prijatniju kancelariju

Znanje se nagraDuje: Nagrade za favorite

Dovek je jedno od biDa na planeti, D iji bmuE>jakb ima najveDi
muE!ki polni organ spram svoje veliD ine.

Kada pobrojimo najpopularnije fitnes sprave - meDu njima De se sigurno
naDi i steper.

Kako se odeDate na svom radnom mestu? Da li D esto imate glavobolje,
oseDate da vam je zaguE!ljivo...

U kvizu "Moj favorit" koji je Top Shop organizovao, dodeljene su vredne
nagrade...

Sve o veliD inama

Da li znate baE! sve o veliD inama?

Dovek je jedno od biDa na planeti, D iji bmuE>jakb ima najveDi
polni organ spram svoje veliD ine... ZaE!to su muE!karci D esto
nezadovoljni? E ta E>ene E>ele? Ko uspostavlja standarde po pitanju
veliD ine? Postoji viE!e kriterijuma utvrDivanja veliD ine ljudskog
penisa... [viE!e...]

Septembarska uE!teda!

Power Juicer Express + POKLON

Celluless

Pancake Puffs

Jucer i opener
Cena: 9.990 RSD

Celluless
Cena: 3.590 RSD

Turbo Maximus Queen
Cena: 1.890 RSD

Sokovnik i POKLON - elektriD ni otvaraD za konzerve!

JoE! samo 24 komada po 10%% niE>oj ceni! Oslobodite se celulita.

Za uE!topke savE!enog izgleda, od sada po niE>oj ceni!

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

Steper

"Penjanje uz stepenice" za lepE!i izgled

Kada pobrojimo najpopularnije fitnes sprave - bilo za veE>banje kod
kuDe, bilo u fitnes klubovima - meDu njima Demo se sigurno prisetiti i
stepera. VeE>be na steperu, po principu bbeskonaD nog penjanja uz
stepeniceb pomaE>u lepE!em izgledu, troE!enju suviE!ne energije i
kalorija. [viE!e...]

Lep izgled - zdrav E>ivot

Steper sa tegovima

Air Climber

Slim Stepper

Steper sa tegovima
Cena: 5.990 RSD

Air Climber
Cena: 8.990 RSD

Slim Stepper
Cena: 2.990 RSD

NOVO! Steper koji ima dodatke za akupresuru, tegove, disk za struk.

Vakuumski steper kombinuje kardio i veE>be za trbuE!njake.

VeE>bajte kod kuDe, ne bacajte novac na skupe treninge.

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

VoDe

5 trikova za prijatniju kancelariju

Kako se odeDate na svom radnom mestu? Da li D esto imate glavobolje,
oseDate da vam je zaguE!ljivo ili postajete anksiozni? MoE>da baE! mesto
na koje dolazite dan za danom b kancelarija b ima tako negativan
uticaj na vas. Pa...NaD inite je prijatnijim mestom za boravak...
ProD itajte kako. [viE!e...]

Obradujte sebe i svoju porodicu

Tweeze

Kleen Kut

Joy BOX

Tweeze
Cena: 2.190 RSD

Kleen Kut
Cena: 2.190 RSD

Joy Box
Cena: 4.890 RSD

Pinceta koja precizno hvata i potpuno uklanja neE>eljenje dlake!

ElektriD ni, vodootporni brijaD na baterijeza muE!karce.

Pevajte, gledajte filmove, sluE!ajte muziku, igrajte igrice porodiD no.

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

Dobitnici

Nagrade za favorite

Svi koji redovno poseDuju e-Topshop web stranicu, znaju da pored velikog
izbora proizvoda i tekstova iz raznih oblasti E>ivota, povremeno
organizujemo i nagradne kvizove.Jedan od takvih bio je i junski kviz
bMoj Favoritb posveDen Evropskom prvenstvu u fudbalu Euro 2008... [viE!e...]

Za praktiD an i udoban E>ivot

Total Sauna

Roto Duster

Snap'n'slice

Total Sauna
Cena: 8.990 RSD

Roto Duster
Cena: 2.490 RSD

Snap n Slice
Cena: 3.490 RSD

Mala portabl sauna koju moE>ete koristiti u udobnosti doma!

DistaD koji radi na baterije i D isti statiD kim elektricitetom.

Za dekorativno i brzo seckanje svih vrsta namirnica..

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

[ViE!e l PoruD i odmah]

Ovu elektronsku poE!tu primate, ukoliko ste svojevoljno ostavili svoju
e-mail adresu na nekom od sajtova Top Shop-a, uD estvovali u naE!oj
poklon igri ili nagradnom kvizu ili se prijavili za e-D asopis Top Shop-a
ili nekog od nasih brendova.

Ponude date u ovom e-mailu vaE>e iskljuD ivo za porudE>bine upuDene
putem Interneta ili broja telefona 021 489 26 60.

Ukoliko ne E>elite viE!e da primate naE!e elektronske poruke, za
odjavljivanje sa naE!e e-mailing liste, kliknite ovde.

Studio Moderna d.o.o., Laze NanD iDa 50, 21000 Novi Sad, Tel: 021 489 26
60, Fax: 021 489 26 08,
E-mail: info@news.e-topshop.tv

[IMAGE]If you would no longer like to receive our emails please
unsubscribe by clicking here.

Posted In: mailing.openbsd.tech

no comments

cwm(1) diff

Fri, 05 Sep 2008 11:00:32 PDT

Hello,

In cwm, if you press M-/ C-a to see a list of all windows, the previous
title of each window is displayed instead of the current one.

Regards,
Tim

Index: search.c
===================================================================
RCS file: /cvs/xenocara/app/cwm/search.c,v
retrieving revision 1.12
diff -p -u -r1.12 search.c
--- search.c 11 Jul 2008 14:21:28 -0000 1.12
+++ search.c 5 Sep 2008 15:44:31 -0000
@@ -132,7 +132,7 @@ search_print_client(struct menu *mi, int
flag = '&';

if (list)
- cc->matchname = TAILQ_FIRST(&cc->nameq)->name;
+ cc->matchname = cc->name;

snprintf(mi->print, sizeof(mi->print), "%%c%%s", flag, cc->matchname);

Posted In: mailing.openbsd.tech

no comments

upgrade44.html diif

Thu, 04 Sep 2008 20:00:22 PDT

Hi,

One user is missing in the upgrade 4.4 file.

Here is the diff for it, but looks like it break the line where is
shouldn't.

only chnages is one line:

+useradd -u93 -g=uid -c"YP to LDAP Daemon" -d/var/empty
-s/sbin/nologin _ypldap

Index: upgrade44.html
===================================================================
RCS file: /cvs/www/faq/upgrade44.html,v
retrieving revision 1.10
diff -u -p -r1.10 upgrade44.html
--- upgrade44.html 3 Sep 2008 10:57:33 -0000 1.10
+++ upgrade44.html 5 Sep 2008 02:56:12 -0000
@@ -347,6 +347,7 @@ Create the user and group using
href="http://www.openbsd.org/cgi-bin/man.cgi?query=useradd&sektion=8">useradd(8)


  useradd -u92 -g=uid -c"IPv6 Router Advertisement Daemon" 
-d/var/empty -s/sbin/nologin _rtadvd
+useradd -u93 -g=uid -c"YP to LDAP Daemon" -d/var/empty 
-s/sbin/nologin _ypldap

Posted In: mailing.openbsd.tech

1 Comment

another dc(4) diff needs testing.

Thu, 04 Sep 2008 18:25:14 PDT

The following diff needs tesing if you have a dc(4) NIC using
The ADMtek Centaur, Conexant or Davicom chipsets. Please
check to make sure link state checking works properly. This
can resolve issues with CARP/trunk and some other things.

Please provide a dmesg.

Index: dc.c
===================================================================
RCS file: /cvs/src/sys/dev/ic/dc.c,v
retrieving revision 1.99
diff -u -p -r1.99 dc.c
--- dc.c 26 Nov 2007 17:45:14 -0000 1.99
+++ dc.c 29 Aug 2008 21:24:22 -0000
@@ -2410,8 +2417,12 @@ dc_tick(xsc)
if (sc->dc_link == 0)
mii_tick(mii);
} else {
- r = CSR_READ_4(sc, DC_ISR);
- if ((r & DC_ISR_RX_STATE) == DC_RXSTATE_WAIT &&
+ /*
+ * For NICs which never report DC_RXSTATE_WAIT, we
+ * have to bite the bullet...
+ */
+ if ((DC_HAS_BROKEN_RXSTATE(sc) || (CSR_READ_4(sc,
+ DC_ISR) & DC_ISR_RX_STATE) == DC_RXSTATE_WAIT) &&
sc->dc_cdata.dc_tx_cnt == 0 && !DC_IS_ASIX(sc)) {
mii_tick(mii);
if (!(mii->mii_media_status & IFM_ACTIVE))

--
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.

Posted In: mailing.openbsd.tech

no comments

pf optimizations, help us test them

Wed, 03 Sep 2008 06:11:52 PDT

just before the release we had to disable the pf optimizations related
to storing a state key pointer in the pkthdr mbufs. these help saving
state lookups and thus help performance notably.

please tess the following diff which re-enables that stuff. it is
tested well for regular forwarding setups, tho more never hurts, of
course. we need testing with more complicated setups. especially any
form of tunnels - ipsec, gif, gre, pppoe, tun, and things like
route-to.

i pretty much expect everything to work fine now. at the same time i
expect that the tunnel interfaces except ipsec do not behave the way
intended yet, which you should only notice by messages like

pf: state key linking mismatch! dir=OUT, if=dc0,
stored af=2, a0: 1.2.3.4:123, a1: 5.6.7.8:8500, proto=17,
found af=2, a0: 9.8.3.5:154, a1: 1.23.42.2:243, proto=17.

(on one line, wrapped for readability)
in the dmesg buffer or /var/log/messages

stored and found must be the exact reverse of each other, so the two
af= and proto= must be the same, stored a0 must be the same as found
a1, and stored a1 = found a0. the message is only printed if that is
NOT the case, and we need to find out why. probably related to some
kind of tunneling. so if you see these messages, please report them to
me and include ifconfig -A output.

Index: pf.c
===================================================================
RCS file: /cvs/src/sys/net/pf.c,v
retrieving revision 1.618
diff -u -p -r1.618 pf.c
--- pf.c 3 Sep 2008 12:57:19 -0000 1.618
+++ pf.c 3 Sep 2008 12:57:57 -0000
@@ -5817,10 +5817,8 @@ done:
if ((s && s->tag) || r->rtableid)
pf_tag_packet(m, s ? s->tag : 0, r->rtableid);

-#if 0
if (dir == PF_IN && s && s->key[PF_SK_STACK])
m->m_pkthdr.pf.statekey = s->key[PF_SK_STACK];
-#endif

#ifdef ALTQ
if (action == PF_PASS && r->qid) {
@@ -6198,10 +6196,8 @@ done:
if ((s && s->tag) || r->rtableid)
pf_tag_packet(m, s ? s->tag : 0, r->rtableid);

-#if 0
if (dir == PF_IN && s && s->key[PF_SK_STACK])
m->m_pkthdr.pf.statekey = s->key[PF_SK_STACK];
-#endif

#ifdef ALTQ
if (action == PF_PASS && r->qid) {

Posted In: mailing.openbsd.tech

no comments

pf (missing ioctl in manpage) & relayd

Wed, 03 Sep 2008 05:36:18 PDT

Hi tech (and patch submitter)!

I'm not sure if this is the right list but since this is somewhat
technical I'll give it a try..

I'm playing around with relayd and stumbled on the problem with sticky-
address and source-track entries keeping clients from the working
servers.

This should be fixed in upcoming 4.4, but judging from the diff
(http://www.openbsd.org/cgi-bin/cvsweb/src/usr.sbin/relayd/pfe_filter.c.diff?
r1=1.26&r2=1.27
) it seem to use DIOCCLRSRCNODES instead of the, in my head, more
appropriate DIOCKILLSRCNODES.

Is there any reason for clearing out ALL source tracking entries
instead of just the ones related to the host which just taken down?
One reason could be that DIOCKILLSRCNODES doesn't actually see to be
documented in pf.4, so it might just have been missed..

I guess the missing entry in the manpage is a "bug" at least :)
However what I'm wondering is if the relayd's usage of CLR instead of
KILL is a design decision or if the submitter (mkoc at prime.pl) was
just unaware of the existanse of kill?

Clearing out all source tracking entries works I guess, but in my eyes
it would be neater if just the affected ones were removed, right?

If I'm totally wrong here, I hope you can excuse me, I'm quite new to
openbsd and the load balancing part.

Thanks
Johan Strvm

Posted In: mailing.openbsd.tech

5 Comments

[patch] nForce AGP support

Tue, 02 Sep 2008 22:51:41 PDT

Hi.

I wanted to give the DRI stuff that oga@ has been working so hard on
a try. The X server found radeondrm(4) fine after I enabled it via
config(8) but then complained that agp(4) wasn't working. X was right;
this machine has the NVIDIA nForce2 core chipset, and there's no support
for agp(4) on that. A search through the archives of this list turned up
a posting from March of 2006 in which Dimitry Andric provided a patch to
add that support.

The following is that patch, but tweaked so that it applies to current.
My thanks to Mr. Andric and the original developers for the heavy lifting;
all I had to do was tweak things to get the patch to apply and compile.

With this patch, and radeondrm(4) enabled, I now get:

agp0 at vga1: aperture at 0xe0000000, size 0x4000000
radeondrm0 at vga1 ATI Radeon AP 9600(0), 1.29.0 20080613

That's an ATI Radeon 9600 PRO AGP card in the AGP slot on this mainboard,
which is an MSI K7N2 Delta-L (MS-6570).

I've put a full dmesg after the patch. Tested on i386 only; I see that
agp(4) is also in the amd64 kernel, but I have no amd64 system.

X reports that DRI is there and glxgears(1) and the GL modes in xlock(1)
all run much faster. Marathon and marathoninfinity (from packages)
both start up OK and look nice, but I haven't really had time to play.
Celestia (also from packages) throws a software fallback warning from the
Mesa r300 driver and then crashes with a propolice-caught stack overflow.
I haven't tried any other GL programs yet.

patch:
------------------------------------------------------------------------------
Index: sys/arch/i386/conf/files.i386
===================================================================
RCS file: /cvs/src/sys/arch/i386/conf/files.i386,v
retrieving revision 1.182
diff -u -r1.182 files.i386
--- sys/arch/i386/conf/files.i386 8 Jul 2008 05:22:00 -0000 1.182
+++ sys/arch/i386/conf/files.i386 1 Sep 2008 02:41:24 -0000
@@ -101,7 +101,7 @@
file arch/i386/pci/pcic_pci_machdep.c pcic_pci

# AGP devices
-device agp: agp_ali, agp_amd, agp_i810, agp_intel, agp_sis, agp_via
+device agp: agp_ali, agp_amd, agp_i810, agp_intel, agp_nvidia, agp_sis, agp_via
attach agp at agpbus
file arch/i386/pci/agp_machdep.c agp

Index: sys/dev/pci/agp.c
===================================================================
RCS file: /cvs/src/sys/dev/pci/agp.c,v
retrieving revision 1.25
diff -u -r1.25 agp.c
--- sys/dev/pci/agp.c 17 Aug 2008 21:31:38 -0000 1.25
+++ sys/dev/pci/agp.c 1 Sep 2008 05:44:14 -0000
@@ -52,6 +52,7 @@
#include "agp_apple.h"
#include "agp_i810.h"
#include "agp_intel.h"
+#include "agp_nvidia.h"
#include "agp_sis.h"
#include "agp_via.h"

@@ -97,6 +98,10 @@
#if NAGP_INTEL > 0
{ PCI_VENDOR_INTEL, -1, agp_intel_attach },
#endif
+#if NAGP_NVIDIA > 0
+ { PCI_VENDOR_NVIDIA,PCI_PRODUCT_NVIDIA_NFORCE_PCHB,agp_nvidia_attach },
+ { PCI_VENDOR_NVIDIA,PCI_PRODUCT_NVIDIA_NFORCE2_PCHB,agp_nvidia_attach },
+#endif
#if NAGP_SIS > 0
{ PCI_VENDOR_SIS, -1, agp_sis_attach },
#endif
Index: sys/dev/pci/agp_nvidia.c
===================================================================
RCS file: sys/dev/pci/agp_nvidia.c
diff -N sys/dev/pci/agp_nvidia.c
--- /dev/null 1 Jan 1970 00:00:00 -0000
+++ sys/dev/pci/agp_nvidia.c 1 Sep 2008 05:31:41 -0000
@@ -0,0 +1,392 @@
+/* $OpenBSD$ */
+
+/*-
+ * Copyright (c) 2003 Matthew N. Dodd
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ * notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ * notice, this list of conditions and the following disclaimer in the
+ * documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD: src/sys/pci/agp_nvidia.c,v 1.11 2005/12/20 21:12:26 jhb Exp $
+ */
+
+/*
+ * Written using information gleaned from the
+ * NVIDIA nForce/nForce2 AGPGART Linux Kernel Patch.
+ */
+
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include
+#include
+#include
+#include
+#include
+#include
+
+#include
+#include
+
+struct agp_nvidia_softc {
+ u_int32_t initial_aperture; /* aperture size at startup */
+ struct agp_gatt * gatt;
+
+ pcitag_t tag; /* AGP Controller */
+ pcitag_t mc1_tag; /* Memory Controller */
+ pcitag_t mc2_tag; /* Memory Controller */
+ pcitag_t btag; /* Bridge */
+
+ u_int32_t wbc_mask;
+ int num_dirs;
+ int num_active_entries;
+ off_t pg_offset;
+};
+
+static u_int32_t agp_nvidia_get_aperture(struct agp_softc *);
+static int agp_nvidia_set_aperture(struct agp_softc *, u_int32_t);
+static int agp_nvidia_bind_page(struct agp_softc *, off_t, bus_addr_t);
+static int agp_nvidia_unbind_page(struct agp_softc *, off_t);
+static void agp_nvidia_flush_tlb(struct agp_softc *);
+static int nvidia_init_iorr(u_int32_t, u_int32_t);
+
+struct agp_methods agp_nvidia_methods = {
+ agp_nvidia_get_aperture,
+ agp_nvidia_set_aperture,
+ agp_nvidia_bind_page,
+ agp_nvidia_unbind_page,
+ agp_nvidia_flush_tlb,
+ agp_generic_enable,
+ agp_generic_alloc_memory,
+ agp_generic_free_memory,
+ agp_generic_bind_memory,
+ agp_generic_unbind_memory,
+};
+
+int
+agp_nvidia_attach(struct agp_softc *sc, struct pci_attach_args *pa)
+{
+ struct agp_nvidia_softc *nsc;
+ struct agp_gatt *gatt;
+ u_int32_t apbase;
+ u_int32_t aplimit;
+ pcireg_t temp;
+ int size;
+ int i;
+ int error;
+
+ nsc = malloc(sizeof *nsc, M_AGP, M_NOWAIT);
+ if (nsc == NULL) {
+ printf(": can't allocate chipset-specific softc\n");
+ return (ENOMEM);
+ }
+ memset(nsc, 0, sizeof *nsc);
+ sc->sc_chipc = nsc;
+ sc->sc_methods = &agp_nvidia_methods;
+
+ if (agp_map_aperture(pa, sc, AGP_APBASE, PCI_MAPREG_TYPE_MEM) != 0) {
+ printf(": failed to map aperture\n");
+ free(nsc, M_DEVBUF);
+ sc->sc_chipc = NULL;
+ return (ENXIO);
+ }
+
+ switch (PCI_PRODUCT(pa->pa_id)) {
+ case PCI_PRODUCT_NVIDIA_NFORCE_PCHB:
+ nsc->wbc_mask = 0x00010000;
+ break;
+ case PCI_PRODUCT_NVIDIA_NFORCE2_PCHB:
+ nsc->wbc_mask = 0x80000000;
+ break;
+ default:
+ printf(": unknown product id %%#x\n", PCI_PRODUCT(pa->pa_id));
+ free(nsc, M_DEVBUF);
+ sc->sc_chipc = NULL;
+ return (ENODEV);
+ }
+
+ /* AGP Controller */
+ nsc->tag = pa->pa_tag;
+
+ /* Memory Controller 1 */
+ nsc->mc1_tag = pci_make_tag(pa->pa_pc, pa->pa_bus, 0, 1);
+ /* XXX Validate existence */
+
+ /* Memory Controller 2 */
+ nsc->mc2_tag = pci_make_tag(pa->pa_pc, pa->pa_bus, 0, 2);
+ /* XXX Validate existence */
+
+ /* AGP Host to PCI Bridge */
+ nsc->btag = pci_make_tag(pa->pa_pc, pa->pa_bus, 30, 0);
+ /* XXX Validate existence */
+
+ nsc->initial_aperture = AGP_GET_APERTURE(sc);
+
+ for (;;) {
+ gatt = agp_alloc_gatt(sc);
+ if (gatt != NULL)
+ break;
+ /*
+ * Probably contigmalloc failure. Try reducing the
+ * aperture so that the gatt size reduces.
+ */
+ if (AGP_SET_APERTURE(sc, AGP_GET_APERTURE(sc) / 2)) {
+ free(nsc, M_DEVBUF);
+ sc->sc_chipc = NULL;
+ return (ENOMEM);
+ }
+ }
+ nsc->gatt = gatt;
+
+ apbase = sc->sc_apaddr;
+ aplimit = apbase + AGP_GET_APERTURE(sc) - 1;
+ pci_conf_write(sc->sc_pc, nsc->mc2_tag, AGP_NVIDIA_2_APBASE, apbase);
+ pci_conf_write(sc->sc_pc, nsc->mc2_tag, AGP_NVIDIA_2_APLIMIT, aplimit);
+ pci_conf_write(sc->sc_pc, nsc->btag, AGP_NVIDIA_3_APBASE, apbase);
+ pci_conf_write(sc->sc_pc, nsc->btag, AGP_NVIDIA_3_APLIMIT, aplimit);
+
+ error = nvidia_init_iorr(apbase, AGP_GET_APERTURE(sc));
+ if (error) {
+ printf("Failed to setup IORRs\n");
+ agp_free_gatt(sc, gatt);
+ free(nsc, M_DEVBUF);
+ sc->sc_chipc = NULL;
+ return (ENXIO);
+ }
+
+ /* directory size is 64k */
+ size = AGP_GET_APERTURE(sc) / 1024 / 1024;
+ nsc->num_dirs = size / 64;
+ nsc->num_active_entries = (size == 32) ? 16384 : ((size * 1024) / 4);
+ nsc->pg_offset = 0;
+ if (nsc->num_dirs == 0) {
+ nsc->num_dirs = 1;
+ nsc->num_active_entries /= (64 / size);
+ nsc->pg_offset = (apbase & (64 * 1024 * 1024 - 1) &
+ ~(AGP_GET_APERTURE(sc) - 1)) / PAGE_SIZE;
+ }
+
+ /* (G)ATT Base Address */
+ for (i = 0; i < 8; i++) {
+ pci_conf_write(sc->sc_pc, nsc->mc2_tag, AGP_NVIDIA_2_ATTBASE(i),
+ (nsc->gatt->ag_physical +
+ (i %% nsc->num_dirs) * 64 * 1024) | 1);
+ }
+
+ /* GTLB Control */
+ temp = pci_conf_read(sc->sc_pc, nsc->mc2_tag, AGP_NVIDIA_2_GARTCTRL);
+ pci_conf_write(sc->sc_pc, nsc->mc2_tag, AGP_NVIDIA_2_GARTCTRL, temp | 0x11);
+
+ /* GART Control */
+ temp = pci_conf_read(sc->sc_pc, sc->sc_pcitag, AGP_NVIDIA_0_APSIZE);
+ pci_conf_write(sc->sc_pc, sc->sc_pcitag, AGP_NVIDIA_0_APSIZE, temp | 0x100);
+
+ return (0);
+}
+
+#if 0
+static int
+agp_nvidia_detach(struct agp_softc *sc)
+{
+ struct agp_nvidia_softc *nsc = sc->sc_chipc;
+ int error;
+ pcireg_t temp;
+
+ error = agp_generic_detach(sc);
+ if (error)
+ return (error);
+
+ /* GART Control */
+ temp = pci_conf_read(sc->sc_pc, sc->sc_pcitag, AGP_NVIDIA_0_APSIZE);
+ pci_conf_write(sc->sc_pc, sc->sc_pcitag, AGP_NVIDIA_0_APSIZE, temp & ~0x100);
+
+ /* GTLB Control */
+ temp = pci_conf_read(sc->sc_pc, nsc->mc2_tag, AGP_NVIDIA_2_GARTCTRL);
+ pci_conf_write(sc->sc_pc, nsc->mc2_tag, AGP_NVIDIA_2_GARTCTRL, temp & ~0x11);
+
+ /* Put the aperture back the way it started. */
+ AGP_SET_APERTURE(sc, nsc->initial_aperture);
+
+ /* restore iorr for previous aperture size */
+ nvidia_init_iorr(sc->sc_apaddr, nsc->initial_aperture);
+
+ agp_free_gatt(sc, nsc->gatt);
+
+ return (0);
+}
+#endif
+
+static u_int32_t
+agp_nvidia_get_aperture(struct agp_softc *sc)
+{
+ pcireg_t reg;
+
+ reg = pci_conf_read(sc->sc_pc, sc->sc_pcitag, AGP_NVIDIA_0_APSIZE);
+ reg &= 0xf;
+ switch (reg) {
+ case 0: return (512 * 1024 * 1024); break;
+ case 8: return (256 * 1024 * 1024); break;
+ case 12: return (128 * 1024 * 1024); break;
+ case 14: return (64 * 1024 * 1024); break;
+ case 15: return (32 * 1024 * 1024); break;
+ default:
+ printf("Invalid aperture setting 0x%%x", reg);
+ return 0;
+ }
+}
+
+static int
+agp_nvidia_set_aperture(struct agp_softc *sc, u_int32_t aperture)
+{
+ u_int8_t val;
+ u_int8_t key;
+
+ switch (aperture) {
+ case (512 * 1024 * 1024): key = 0; break;
+ case (256 * 1024 * 1024): key = 8; break;
+ case (128 * 1024 * 1024): key = 12; break;
+ case (64 * 1024 * 1024): key = 14; break;
+ case (32 * 1024 * 1024): key = 15; break;
+ default:
+ printf("Invalid aperture size (%%dMb)\n", aperture / 1024 / 1024);
+ return (EINVAL);
+ }
+ val = pci_conf_read(sc->sc_pc, sc->sc_pcitag, AGP_NVIDIA_0_APSIZE);
+ val &= ~0xf;
+ val |= key;
+ pci_conf_write(sc->sc_pc, sc->sc_pcitag, AGP_NVIDIA_0_APSIZE, val);
+
+ return (0);
+}
+
+static int
+agp_nvidia_bind_page(struct agp_softc *sc, off_t offset, bus_addr_t physical)
+{
+ struct agp_nvidia_softc *nsc = sc->sc_chipc;
+ u_int32_t index;
+
+ if (offset < 0 || offset >= (nsc->gatt->ag_entries << AGP_PAGE_SHIFT))
+ return (EINVAL);
+
+ index = (nsc->pg_offset + offset) >> AGP_PAGE_SHIFT;
+ nsc->gatt->ag_virtual[index] = physical | 1;
+
+ return (0);
+}
+
+static int
+agp_nvidia_unbind_page(struct agp_softc *sc, off_t offset)
+{
+ struct agp_nvidia_softc *nsc = sc->sc_chipc;
+ u_int32_t index;
+
+ if (offset < 0 || offset >= (nsc->gatt->ag_entries << AGP_PAGE_SHIFT))
+ return (EINVAL);
+
+ index = (nsc->pg_offset + offset) >> AGP_PAGE_SHIFT;
+ nsc->gatt->ag_virtual[index] = 0;
+
+ return (0);
+}
+
+static void
+agp_nvidia_flush_tlb(struct agp_softc *sc)
+{
+ struct agp_nvidia_softc *nsc = sc->sc_chipc;
+ u_int32_t wbc_reg, temp;
+ volatile u_int32_t *ag_virtual;
+ int i;
+
+ if (nsc->wbc_mask) {
+ wbc_reg = pci_conf_read(sc->sc_pc, nsc->mc1_tag, AGP_NVIDIA_1_WBC);
+ wbc_reg |= nsc->wbc_mask;
+ pci_conf_write(sc->sc_pc, nsc->mc1_tag, AGP_NVIDIA_1_WBC, wbc_reg);
+
+ /* Wait no more than 3 seconds. */
+ for (i = 0; i < 3000; i++) {
+ wbc_reg = pci_conf_read(sc->sc_pc, nsc->mc1_tag, AGP_NVIDIA_1_WBC);
+ if ((nsc->wbc_mask & wbc_reg) == 0)
+ break;
+ else
+ DELAY(1000);
+ }
+ if (i == 3000)
+ printf("TLB flush took more than 3 seconds.\n");
+ }
+
+ ag_virtual = (volatile u_int32_t *)nsc->gatt->ag_virtual;
+
+ /* Flush TLB entries. */
+ for(i = 0; i < 32 + 1; i++)
+ temp = ag_virtual[i * PAGE_SIZE / sizeof(u_int32_t)];
+ for(i = 0; i < 32 + 1; i++)
+ temp = ag_virtual[i * PAGE_SIZE / sizeof(u_int32_t)];
+}
+
+#define SYSCFG 0xC0010010
+#define IORR_BASE0 0xC0010016
+#define IORR_MASK0 0xC0010017
+#define AMD_K7_NUM_IORR 2
+
+static int
+nvidia_init_iorr(u_int32_t addr, u_int32_t size)
+{
+ quad_t base, mask, sys;
+ u_int32_t iorr_addr, free_iorr_addr;
+
+ /* Find the iorr that is already used for the addr */
+ /* If not found, determine the uppermost available iorr */
+ free_iorr_addr = AMD_K7_NUM_IORR;
+ for(iorr_addr = 0; iorr_addr < AMD_K7_NUM_IORR; iorr_addr++) {
+ base = rdmsr(IORR_BASE0 + 2 * iorr_addr);
+ mask = rdmsr(IORR_MASK0 + 2 * iorr_addr);
+
+ if ((base & 0xfffff000ULL) == (addr & 0xfffff000))
+ break;
+
+ if ((mask & 0x00000800ULL) == 0)
+ free_iorr_addr = iorr_addr;
+ }
+
+ if (iorr_addr >= AMD_K7_NUM_IORR) {
+ iorr_addr = free_iorr_addr;
+ if (iorr_addr >= AMD_K7_NUM_IORR)
+ return (EINVAL);
+ }
+
+ base = (addr & ~0xfff) | 0x18;
+ mask = (0xfULL << 32) | ((~(size - 1)) & 0xfffff000) | 0x800;
+ wrmsr(IORR_BASE0 + 2 * iorr_addr, base);
+ wrmsr(IORR_MASK0 + 2 * iorr_addr, mask);
+
+ sys = rdmsr(SYSCFG);
+ sys |= 0x00100000ULL;
+ wrmsr(SYSCFG, sys);
+
+ return (0);
+}
Index: sys/dev/pci/agpreg.h
===================================================================
RCS file: /cvs/src/sys/dev/pci/agpreg.h,v
retrieving revision 1.10
diff -u -r1.10 agpreg.h
--- sys/dev/pci/agpreg.h 23 Mar 2008 08:36:41 -0000 1.10
+++ sys/dev/pci/agpreg.h 1 Sep 2008 02:45:25 -0000
@@ -266,4 +266,17 @@
#define AGP_G33_PGTBL_SIZE_1M (1U << 8)
#define AGP_G33_PGTBL_SIZE_2M (2U << 8)

+/*
+ * NVIDIA nForce/nForce2 registers
+ */
+#define AGP_NVIDIA_0_APBASE 0x10
+#define AGP_NVIDIA_0_APSIZE 0x80
+#define AGP_NVIDIA_1_WBC 0xf0
+#define AGP_NVIDIA_2_GARTCTRL 0xd0
+#define AGP_NVIDIA_2_APBASE 0xd8
+#define AGP_NVIDIA_2_APLIMIT 0xdc
+#define AGP_NVIDIA_2_ATTBASE(i) (0xe0 + (i) * 4)
+#define AGP_NVIDIA_3_APBASE 0x50
+#define AGP_NVIDIA_3_APLIMIT 0x54
+
#endif /* !_PCI_AGPREG_H_ */
Index: sys/dev/pci/agpvar.h
===================================================================
RCS file: /cvs/src/sys/dev/pci/agpvar.h,v
retrieving revision 1.14
diff -u -r1.14 agpvar.h
--- sys/dev/pci/agpvar.h 12 Jul 2008 17:31:06 -0000 1.14
+++ sys/dev/pci/agpvar.h 1 Sep 2008 06:05:14 -0000
@@ -207,6 +207,7 @@
int agp_amd_attach(struct agp_softc *, struct pci_attach_args *);
int agp_i810_attach(struct agp_softc *, struct pci_attach_args *);
int agp_intel_attach(struct agp_softc *, struct pci_attach_args *);
+int agp_nvidia_attach(struct agp_softc *, struct pci_attach_args *);
int agp_via_attach(struct agp_softc *, struct pci_attach_args *);
int agp_sis_attach(struct agp_softc *, struct pci_attach_args *);

Index: sys/dev/pci/files.agp
===================================================================
RCS file: /cvs/src/sys/dev/pci/files.agp,v
retrieving revision 1.2
diff -u -r1.2 files.agp
--- sys/dev/pci/files.agp 16 Mar 2008 19:00:28 -0000 1.2
+++ sys/dev/pci/files.agp 1 Sep 2008 03:00:10 -0000
@@ -17,6 +17,9 @@
define agp_intel
file dev/pci/agp_intel.c agp_intel & agpbus needs-flag

+define agp_nvidia
+file dev/pci/agp_nvidia.c agp_nvidia & agpbus needs-flag
+
define agp_sis
file dev/pci/agp_sis.c agp_sis & agpbus needs-flag

------------------------------------------------------------------------------

dmesg:
------------------------------------------------------------------------------
OpenBSD 4.4-current (GENERIC+nvagp) #0: Mon Sep 1 19:11:49 PDT 2008
clamat@bluesong.XXXXXXXXXXXX:/sys/arch/i386/compile/GENERIC+nvagp
cpu0: AMD Athlon(tm) XP 2800+ ("AuthenticAMD" 686-class, 512KB L2 cache) 2.08 GHz
cpu0: FPU,V86,DE,PSE,TSC,MSR,PAE,MCE,CX8,APIC,SEP,MTRR,PGE,MCA,CMOV,PAT,PSE36,MMX,FXSR,SSE
real mem = 2146988032 (2047MB)
avail mem = 2067615744 (1971MB)
mainbus0 at root
bios0 at mainbus0: AT/286+ BIOS, date 08/02/04, BIOS32 rev. 0 @ 0xfbba0, SMBIOS rev. 2.3 @ 0xf0000 (37 entries)
bios0: vendor Phoenix Technologies, LTD version "6.00 PG" date 08/02/2004
bios0: MICRO-STAR INTERNATIONAL CO., LTD MS-6570
apm0 at bios0: Power Management spec V1.2 (slowidle)
apm0: AC on, battery charge unknown
acpi at bios0 function 0x0 not configured
pcibios0 at bios0: rev 2.1 @ 0xf0000/0xdb04
pcibios0: PCI IRQ Routing Table rev 1.0 @ 0xfda40/192 (10 entries)
pcibios0: PCI Exclusive IRQs: 7 10 11 14 15
pcibios0: no compatible PCI ICU found
pcibios0: Warning, unable to fix up PCI interrupt routing
pcibios0: PCI bus #2 is the last bus
bios0: ROM list: 0xc0000/0xd000 0xd0000/0x4000
cpu0 at mainbus0
pci0 at mainbus0 bus 0: configuration mode 1 (no bios)
pchb0 at pci0 dev 0 function 0 "NVIDIA nForce2 PCI" rev 0xc1
"NVIDIA nForce2" rev 0xc1 at pci0 dev 0 function 1 not configured
"NVIDIA nForce2" rev 0xc1 at pci0 dev 0 function 2 not configured
"NVIDIA nForce2" rev 0xc1 at pci0 dev 0 function 3 not configured
"NVIDIA nForce2" rev 0xc1 at pci0 dev 0 function 4 not configured
"NVIDIA nForce2" rev 0xc1 at pci0 dev 0 function 5 not configured
pcib0 at pci0 dev 1 function 0 "NVIDIA nForce2 ISA" rev 0xa3
nviic0 at pci0 dev 1 function 1 "NVIDIA nForce2 SMBus" rev 0xa2
iic0 at nviic0
iic0: addr 0x1b 06=f0 08=03 09=03 0b=7f 0c=08 0d=08 0e=f0 0f=07 10=0c 11=03 14=55 15=55 17=ff 18=ff 1c=02 20=7f 22=07 26=30 29=ff words 00=00ff 01=00ff 02=00ff 03=00ff 04=00ff 05=00ff 06=f0ff 07=00ff
iic0: addr 0x2f 04=00 06=02 07=00 0c=00 0d=07 0e=84 0f=00 10=ca 11=10 12=00 13=60 14=14 15=62 16=01 17=06 words 00=ffff 01=ffff 02=ffff 03=ffff 04=00ff 05=ffff 06=02ff 07=00ff
spdmem0 at iic0 addr 0x50: 1GB DDR SDRAM non-parity PC3200CL3.0
spdmem1 at iic0 addr 0x51: 1GB DDR SDRAM non-parity PC3200CL3.0
iic1 at nviic0
ohci0 at pci0 dev 2 function 0 "NVIDIA nForce2 USB" rev 0xa3: irq 15, version 1.0, legacy support
ohci1 at pci0 dev 2 function 1 "NVIDIA nForce2 USB" rev 0xa3: irq 7, version 1.0, legacy support
ehci0 at pci0 dev 2 function 2 "NVIDIA nForce2 USB" rev 0xa3: irq 14
usb0 at ehci0: USB revision 2.0
uhub0 at usb0 "NVIDIA EHCI root hub" rev 2.00/1.00 addr 1
ppb0 at pci0 dev 8 function 0 "NVIDIA nForce2 PCI-PCI" rev 0xa3
pci1 at ppb0 bus 1
siop0 at pci1 dev 7 function 0 "Symbios Logic 53c1010-33" rev 0x01: irq 15, using 8K of on-board RAM
scsibus0 at siop0: 16 targets, initiator 7
sd0 at scsibus0 targ 5 lun 0: SCSI3 0/direct fixed
sd0: 34715MB, 512 bytes/sec, 71096640 sec total
sd1 at scsibus0 targ 6 lun 0: SCSI3 0/direct fixed
sd1: 34715MB, 512 bytes/sec, 71096640 sec total
sd2 at scsibus0 targ 15 lun 0: SCSI3 0/direct fixed
sd2: 17501MB, 512 bytes/sec, 35843670 sec total
siop1 at pci1 dev 7 function 1 "Symbios Logic 53c1010-33" rev 0x01: irq 11, using 8K of on-board RAM
scsibus1 at siop1: 16 targets, initiator 7
cd0 at scsibus1 targ 2 lun 0: SCSI2 5/cdrom removable
st0 at scsibus1 targ 4 lun 0: SCSI2 1/sequential removable
vr0 at pci1 dev 8 function 0 "VIA VT6105 RhineIII" rev 0x86: irq 11, address 00:05:5d:80:c4:8d
ukphy0 at vr0 phy 1: Generic IEEE 802.3u media interface, rev. 4: OUI 0x004063, model 0x0034
eap0 at pci1 dev 9 function 0 "Ensoniq AudioPCI" rev 0x00: irq 10
audio0 at eap0
midi0 at eap0:
pciide0 at pci0 dev 9 function 0 "NVIDIA nForce2 IDE" rev 0xa2: DMA, channel 0 configured to compatibility, channel 1 configured to compatibility
pciide0: channel 0 ignored (disabled)
pciide0: channel 1 ignored (disabled)
ppb1 at pci0 dev 30 function 0 "NVIDIA nForce2 AGP" rev 0xc1
pci2 at ppb1 bus 2
vga1 at pci2 dev 0 function 0 "ATI Radeon 9600 Pro" rev 0x00
wsdisplay0 at vga1 mux 1: console (80x25, vt100 emulation)
wsdisplay0: screen 1-5 added (80x25, vt100 emulation)
agp0 at vga1: aperture at 0xe0000000, size 0x4000000
radeondrm0 at vga1 ATI Radeon AP 9600(0), 1.29.0 20080613
"ATI Radeon 9600 Pro Sec" rev 0x00 at pci2 dev 0 function 1 not configured
isa0 at pcib0
isadma0 at isa0
com0 at isa0 port 0x3f8/8 irq 4: ns16550a, 16 byte fifo
com1 at isa0 port 0x2f8/8 irq 3: ns16550a, 16 byte fifo
pckbc0 at isa0 port 0x60/5
pckbd0 at pckbc0 (kbd slot)
pckbc0: using irq 1 for kbd slot
wskbd0 at pckbd0: console keyboard, using wsdisplay0
pmsi0 at pckbc0 (aux slot)
pckbc0: using irq 12 for aux slot
wsmouse0 at pmsi0 mux 0
pcppi0 at isa0 port 0x61
midi1 at pcppi0:
spkr0 at pcppi0
wbsio0 at isa0 port 0x4e/2: W83627HF rev 0x17
lm1 at wbsio0 port 0x290/8: W83627HF
npx0 at isa0 port 0xf0/16: reported by CPUID; using exception 16
fdc0 at isa0 port 0x3f0/6 irq 6 drq 2
fd0 at fdc0 drive 0: 1.44MB 80 cyl, 2 head, 18 sec
usb1 at ohci0: USB revision 1.0
uhub1 at usb1 "NVIDIA OHCI root hub" rev 1.00/1.00 addr 1
usb2 at ohci1: USB revision 1.0
uhub2 at usb2 "NVIDIA OHCI root hub" rev 1.00/1.00 addr 1
biomask ebe5 netmask ebe5 ttymask fbff
mtrr: Pentium Pro MTRR support
uhub3 at uhub0 port 4 "Standard Microsystems product 0x2514" rev 2.00/0.00 addr 2
udsbr0 at uhub2 port 1 "GemTek Corp USB FM Radio" rev 1.00/4.10 addr 2
radio0 at udsbr0
softraid0 at root
siop0: target 5 now using tagged DT 16 bit 80.0 MHz 62 REQ/ACK offset xfers
siop0: target 6 now using tagged DT 16 bit 80.0 MHz 62 REQ/ACK offset xfers
siop0: target 15 now using tagged DT 16 bit 80.0 MHz 62 REQ/ACK offset xfers
root on sd0a swap on sd0b dump on sd0b
info: [drm] Setting GART location based on new memory map
info: [drm] Loading R300 Microcode
info: [drm] Num pipes: 1
info: [drm] writeback test succeeded in 1 usecs
info: [drm] Num pipes: 1
info: [drm] Setting GART location based on new memory map
info: [drm] Loading R300 Microcode
info: [drm] Num pipes: 1
info: [drm] writeback test succeeded in 1 usecs
info: [drm] Num pipes: 1
info: [drm] Setting GART location based on new memory map
info: [drm] Loading R300 Microcode
info: [drm] Num pipes: 1
info: [drm] writeback test succeeded in 1 usecs
siop1: target 4 now using 8 bit 10.0 MHz 31 REQ/ACK offset xfers
umass0 at uhub0 port 1 configuration 1 interface 0 "Genesys Logic USB2.0 card Reader" rev 2.00/91.44 addr 3
umass0: using SCSI over Bulk-Only
scsibus2 at umass0: 2 targets, initiator 0
sd3 at scsibus2 targ 1 lun 0: SCSI0 0/direct removable
sd3: drive offline
sd4 at scsibus2 targ 1 lun 1: SCSI0 0/direct removable
sd4: drive offline
sd5 at scsibus2 targ 1 lun 2: SCSI0 0/direct removable
sd5: drive offline
sd6 at scsibus2 targ 1 lun 3: SCSI0 0/direct removable
sd6: drive offline
sd7 at scsibus2 targ 1 lun 4: SCSI0 0/direct removable
sd7: drive offline
umass1 at uhub3 port 2 configuration 1 interface 0 "LEXAR MEDIA JUMPDRIVE PRO" rev 2.00/0.00 addr 4
umass1: using SCSI over Bulk-Only
scsibus3 at umass1: 2 targets, initiator 0
sd8 at scsibus3 targ 1 lun 0: SCSI2 0/direct removable
sd8: 246MB, 512 bytes/sec, 503808 sec total
sd8 detached
scsibus3 detached
umass1 detached
------------------------------------------------------------------------------

--
Debugging is twice as hard as writing the code in the first place.
Therefore, if you write the code as cleverly as possible, you are, by
definition, not smart enough to debug it. -- Kernighan
-- quoted by Quentyn Taylor

Posted In: mailing.openbsd.tech

2 Comments

Looking for a OpenBSD Kernel-Developer wanted because of a company seeking for a contact to et their devices supported

Tue, 02 Sep 2008 08:59:23 PDT

Hello guys,

I am looking for a developer who would like to get FPGA-Cards (uploading
the firmware and more (they have even PPC cores!)) supported (some can get
plugged into the cardbus) into the official OpenBSD-Kernel.

A company I am in contact with already finished nearly a driver (wich may
needs checks too) but except this they do look for a way to get these
device s supported in GENERIC (and other architectures too as far as I
understood him).

But it would be better if a developer talks to them about all this. :)

Pasted from a e-Mail:

> I would be happy to provide whatever we could to OpenBSD
> developers. I am not sure that our products meet their criteria for
> official inclusion in their distribution kernels.
So that's no real issue I think (also he made some Linux comparsions but
hey, I wont be too critic about this)).

Please send me a mail! :)

Kind regards,
Sebastian

Posted In: mailing.openbsd.tech

2 Comments

Sat, 30 Aug 2008 17:20:17 PDT

Kswxhe `sdhnjmhch h nasw`~yee bhden dk ahgmeq`.
G`j`g m` q`ire: biz-audio.nxt.ru

>Qanpmhj hg 100 @sdhnjmhc dk Ahgmeq`, Lnrhb`vhh h Q`lnp`gbhrh.
3 DVD - 40 USD

>Jnkkejvh hg 21 Bhden-Qelhm`p` ^ph Lnpng`.
Qngd`mhe ahgmeq`, }ttejrhbmne opnbedemhe deknb{u oepecnbnpnb, qrp`rech ahgmeq`,
onb{xemhej`weqrb` l{xkemh, }ttejrhbmne sop`bkemhe ahgmeqnl h r.d.
9 DVD - 50 USD

>Dhqr`mvhnmm{i Nasw`~yhi Jspq Xjnk{ Qbnecn Dek` "Ahgmeq q Msk"
Ondpnam`, onx`cnb` hmqrpsjvh qngd`mh qbnecn ahgmeq` q msk.
10 leqveb naswemh - 30 USD

>Jnkkejvh 17 nasw`~yhu bhdenjspqnb: "Bqe Reumhweqjhe Lnlemr{ Nmk`im
Ahgmeq` b Bhdentnpl`re" Onkm{i nasw`~yhi jnlokejr bhdenjspqnb, jnrnp{e
m`sw`r B`q qngd`b`r| j`weqrbemm{e q`ir{, dek`r| jp`qhbs~ cp`thjs b Photoshop,
qbnandmn nbk`der| g{jnl p`glerjh HTML, qrhklh CSS h lmncne-lmncne
dpscne m` ndmnl DVD-dhqje!
20 USD

Oph g`j`ge ankee ndmncn jnlokejr` - qjhdjh.

Nrop`bkel b k~as~ qrp`ms lhp`!

with the arbiter, and, for aught she knew, assiduously cultivating his

waited for people to come to him. But he always made an exception in
-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-
another, and at length, in order to provide for the security of his
accordingly proclaimed her queen.

Posted In: mailing.openbsd.tech

no comments