Illegal instruction

Illegal instruction

[Carsten Langgaard]

Hi Karel,

I have tried the root-images tar-files: mipselroot-rh7-20010606 and
mipsroot-rh7.
The mipsroot-rh7 (bigendian) root image seem to work fine, but when
I use the mipselroot-rh7-20010606 (littleendian) I get an illegal
instruction.
[cat:179] Illegal instruction 7c010001 at 2ac8b20c ra=00000000.

I'm using a 2.4.3 kernel.
Anyone got an idea ?

/Carsten

--
_    _ ____  ___   Carsten Langgaard   Mailto:carstenl@mips.com
|\  /|||___)(___   MIPS Denmark        Direct: +45 4486 5527
| \/ |||    ____)  Lautrupvang 4B      Switch: +45 4486 5555
  TECHNOLOGIES     2750 Ballerup       Fax...: +45 4486 5556
                   Denmark             http://www.mips.com

Re: Illegal instruction

[Houten K.H.C. van (Karel)]

Carsten Langgaard wrote: 
> I have tried the root-images tar-files: mipselroot-rh7-20010606 and
> mipsroot-rh7.
> The mipsroot-rh7 (bigendian) root image seem to work fine, but when
> I use the mipselroot-rh7-20010606 (littleendian) I get an illegal
> instruction.
> [cat:179] Illegal instruction 7c010001 at 2ac8b20c ra=00000000.
> 
> I'm using a 2.4.3 kernel.
> Anyone got an idea ?

I'm still using 2.4.0-test9 on mipsel, because I've problems with the
newer kernels too.  I hope to test H.J. Lu's toolchain/glibc, but
currently it is much to hot to switch on extra computers :-)

Regards,
-- 
Karel van Houten

----------------------------------------------------------
The box said "Requires Windows 95 or better."
I can't understand why it won't work on my Linux computer. 
----------------------------------------------------------

Re: Illegal instruction

[Maciej W. Rozycki]

On Fri, 6 Jul 2001, Houten K.H.C. van (Karel) wrote:

> I'm still using 2.4.0-test9 on mipsel, because I've problems with the
> newer kernels too.  I hope to test H.J. Lu's toolchain/glibc, but
> currently it is much to hot to switch on extra computers :-)

 Weird, I'm using 2.4.5 for more than a week now on my /240 and it appears
to work fine.  I've built it with my old tools which I'm still using. 

-- 
+  Maciej W. Rozycki, Technical University of Gdansk, Poland   +
+--------------------------------------------------------------+
+        e-mail: macro@ds2.pg.gda.pl, PGP key available        +

Re: Illegal instruction

[Dan Aizenstros]

Hello Carsten,

I have seen the same thing on the same kernel version.

Dan Aizenstros
Software Engineer
V3 Semiconductor Corp.

Carsten Langgaard wrote:
> 
> Hi Karel,
> 
> I have tried the root-images tar-files: mipselroot-rh7-20010606 and
> mipsroot-rh7.
> The mipsroot-rh7 (bigendian) root image seem to work fine, but when
> I use the mipselroot-rh7-20010606 (littleendian) I get an illegal
> instruction.
> [cat:179] Illegal instruction 7c010001 at 2ac8b20c ra=00000000.
> 
> I'm using a 2.4.3 kernel.
> Anyone got an idea ?
> 
> /Carsten
> 
> --
> _    _ ____  ___   Carsten Langgaard   Mailto:carstenl@mips.com
> |\  /|||___)(___   MIPS Denmark        Direct: +45 4486 5527
> | \/ |||    ____)  Lautrupvang 4B      Switch: +45 4486 5555
>   TECHNOLOGIES     2750 Ballerup       Fax...: +45 4486 5556
>                    Denmark             http://www.mips.com

Re: Illegal instruction

[Jun Sun]

[-- Attachment #1: Type: text/plain, Size: 951 bytes --]

Carsten Langgaard wrote:
> 
> Hi Karel,
> 
> I have tried the root-images tar-files: mipselroot-rh7-20010606 and
> mipsroot-rh7.
> The mipsroot-rh7 (bigendian) root image seem to work fine, but when
> I use the mipselroot-rh7-20010606 (littleendian) I get an illegal
> instruction.
> [cat:179] Illegal instruction 7c010001 at 2ac8b20c ra=00000000.
> 
> I'm using a 2.4.3 kernel.
> Anyone got an idea ?
> 
> /Carsten
> 

Is the userland compiled with MIPS I option?  I bet this is the same nasty
MIPS_ATOMIC_SET bug.

Currently, there are three fixes.  One is from Florian, which only takes care
of CPUs with ll/sc.  Another is from Maceij, which introduces a new syscall,
and some changes in glibc.  The thrid is mine, a compromising one.  It takes
care of both ll/sc case and non-llsc case, has minimum change, and practically
makes all programs happy.

Ralf, please take at least one of the fixes.  Even a not-so-good fix is better
than no fix.

Jun

[-- Attachment #2: MIPS_ATOMIC_SET-compromising-fix.010626.010626.pathc --]
[-- Type: text/plain, Size: 3935 bytes --]


This is a compromising fix for sysmips(MIPS_ATOMIC_SET, ...).  It
forces SIGSYS when the return value is a small negative value.  This
limit is OK for glibc.

Jun

diff -Nru linux/arch/mips/kernel/sysmips.c.orig linux/arch/mips/kernel/sysmips.c
--- linux/arch/mips/kernel/sysmips.c.orig	Mon Apr 23 11:32:54 2001
+++ linux/arch/mips/kernel/sysmips.c	Tue Jun 26 11:50:48 2001
@@ -16,7 +16,7 @@
 #include <linux/sched.h>
 #include <linux/string.h>
 #include <linux/utsname.h>
-
+#include <linux/signal.h>
 #include <asm/cachectl.h>
 #include <asm/pgalloc.h>
 #include <asm/sysmips.h>
@@ -75,53 +75,83 @@
 	}
 
 	case MIPS_ATOMIC_SET: {
-#ifdef CONFIG_CPU_HAS_LLSC
-		unsigned int tmp;
+		int *ptr, val, ret, err, tmp;
+		struct siginfo info;
 
-		p = (int *) arg1;
-		errno = verify_area(VERIFY_WRITE, p, sizeof(*p));
-		if (errno)
-			return errno;
-		errno = 0;
-
-		__asm__(".set\tpush\t\t\t# sysmips(MIPS_ATOMIC, ...)\n\t"
-			".set\tmips2\n\t"
-			".set\tnoat\n\t"
-			"1:\tll\t%0, %4\n\t"
-			"move\t$1, %3\n\t"
-			"2:\tsc\t$1, %1\n\t"
-			"beqz\t$1, 1b\n\t"
-			".set\tpop\n\t"
-			".section\t.fixup,\"ax\"\n"
-			"3:\tli\t%2, 1\t\t\t# error\n\t"
-			".previous\n\t"
-			".section\t__ex_table,\"a\"\n\t"
-			".word\t1b, 3b\n\t"
-			".word\t2b, 3b\n\t"
-			".previous\n\t"
-			: "=&r" (tmp), "=o" (* (u32 *) p), "=r" (errno)
-			: "r" (arg2), "o" (* (u32 *) p), "2" (errno)
-			: "$1");
+		ptr = (int *)arg1;
+		val = (int)arg2;
 
-		if (errno)
-			return -EFAULT;
+		/* Don't emulate unaligned accesses. */
+		if ((int)ptr & 3) {
+			info.si_signo = SIGBUS;
+                        info.si_code = BUS_ADRALN;
+			goto fault;
+		}
+
+		/* A zero here saves us three instructions. */
+		err = verify_area(VERIFY_WRITE, ptr, 0);
+		if (err) {
+                        info.si_signo = SIGSEGV;
+			info.si_code = SEGV_ACCERR;
+			goto fault;
+		}
 
-		/* We're skipping error handling etc.  */
-		if (current->ptrace & PT_TRACESYS)
-			syscall_trace();
-
-		((struct pt_regs *)&cmd)->regs[2] = tmp;
-		((struct pt_regs *)&cmd)->regs[7] = 0;
-
-		__asm__ __volatile__(
-			"move\t$29, %0\n\t"
-			"j\to32_ret_from_sys_call"
-			: /* No outputs */
-			: "r" (&cmd));
-		/* Unreached */
+#ifdef CONFIG_CPU_HAS_LLSC
+		__asm__(".set   mips2\n\t"
+			"1:\n\t"
+			"ll     %0,%5\n\t"
+	                ".set   push\n\t"
+	                ".set   noreorder\n\t"
+	                "beq    %0,%4,3f\n\t"
+	                " move  %3,%4\n"
+	                ".set   pop\n\t"
+	                "2:\n\t"
+	                "sc     %3,%1\n\t"
+	                "beqz   %3,1b\n\t"
+	                "3:\n\t"
+	                ".set   mips0\n\t"
+	                ".section .fixup,\"ax\"\n"
+	                "4:\n\t"
+	                "li     %2,%7\n\t"
+	                "j      3b\n\t"
+	                ".previous\n\t"
+	                ".section __ex_table,\"a\"\n\t"
+	                ".word  1b,4b\n\t"
+	                ".word  2b,4b\n\t"
+	                ".previous"
+	                : "=&r" (ret), "=R" (*ptr), "=r" (err), "=&r" (tmp)
+	                : "r" (val), "1" (*ptr), "2" (0), "i" (-EFAULT));
 #else
-	printk("sys_sysmips(MIPS_ATOMIC_SET, ...) not ready for !CONFIG_CPU_HAS_LLSC\n");
+		save_and_cli(tmp);
+		err = __get_user(ret, ptr);
+		if (ret != val)
+		err |= __put_user(val, ptr);    /* No fault
+                                                   unless unwriteable. */
+		restore_flags(tmp);
 #endif
+
+		if (err) {
+                        info.si_signo = SIGSEGV;
+			info.si_code = SEGV_MAPERR;
+			goto fault;
+		}
+			
+		if ( (ret < 0) && (ret >= -EMAXERRNO) ) {
+                        info.si_signo = SIGSYS;
+			info.si_code = 0;
+			goto fault;
+		}
+			
+		return ret;
+
+fault:
+		/* Go back to SYSCALL. */
+		((struct pt_regs *)&cmd)->cp0_epc -= 4;
+
+		info.si_addr = (void *)((struct pt_regs *)&cmd)->cp0_epc;
+		force_sig_info(info.si_signo, &info, current);
+
+		return 0;
 	}
 
 	case MIPS_FIXADE:

[-- Attachment #3: flo-fast-sysmips.0105X.X.patch --]
[-- Type: text/plain, Size: 4595 bytes --]

diff -Nur linux.orig/arch/mips/kernel/Makefile linux/arch/mips/kernel/Makefile
--- linux.orig/arch/mips/kernel/Makefile        Mon Apr  9 00:23:08 2001
+++ linux/arch/mips/kernel/Makefile     Mon Apr  9 00:23:34 2001
@@ -20,7 +20,7 @@
 obj-y                          += branch.o process.o signal.o entry.o \
                                   traps.o ptrace.o vm86.o ioport.o reset.o \
                                   semaphore.o setup.o syscall.o sysmips.o \
-                                  ipc.o scall_o32.o unaligned.o
+                                  ipc.o scall_o32.o unaligned.o fast-sysmips.o
 obj-$(CONFIG_MODULES)          += mips_ksyms.o

 ifdef CONFIG_CPU_R3000
@@ -69,5 +69,6 @@

 entry.o: entry.S
 head.o: head.S
+fast-sysmips.o: fast-sysmips.S

 include $(TOPDIR)/Rules.make
diff -Nur linux.orig/arch/mips/kernel/fast-sysmips.S linux/arch/mips/kernel/fast-sysmips.S
--- linux.orig/arch/mips/kernel/fast-sysmips.S  Thu Jan  1 01:00:00 1970
+++ linux/arch/mips/kernel/fast-sysmips.S       Mon Apr  9 00:28:20 2001
@@ -0,0 +1,85 @@
+/*
+ * MIPS_ATOMIC_SET asm implementation for ll/sc capable cpus
+ *
+ * This file is subject to the terms and conditions of the GNU General Public
+ * License.  See the file "COPYING" in the main directory of this archive
+ * for more details.
+ *
+ * Copyright (C) 2001 Florian Lohoff <flo@rfc822.org>
+ *
+ */
+#include <asm/asm.h>
+#include <asm/mipsregs.h>
+#include <asm/regdef.h>
+#include <asm/stackframe.h>
+#include <asm/isadep.h>
+#include <asm/unistd.h>
+#include <asm/sysmips.h>
+#include <asm/offset.h>
+#include <asm/errno.h>
+
+#define PT_TRACESYS     0x00000002
+
+       EXPORT(fast_sysmips)
+
+       .set    noreorder
+
+       li      t0, MIPS_ATOMIC_SET
+       beq     a0, t0, 1f
+        nop
+       j       sys_sysmips
+        nop
+
+1:
+
+       # a0 - MIPS_ATOMIC_SET
+       # a1 - mem ptr
+       # a2 - value
+
+       addiu   sp, sp, -8                      # Reserve space
+       sw      a0, (sp)                        # Save arg0
+
+       addiu   a0, a1, 4                       # addr+size
+       ori     v0, a1, 4                       # addr | size
+       lw      v1, THREAD_CURDS(gp)            # current->thread.current_ds
+       or      v0, v0, a0                      # addr | size | (addr+size)
+       and     v1, v1, v0                      # (mask)&(addr | size | (addr+size)
+       bltz    v1, 5f
+        nop
+
+2:
+       ll      v0, (a1)
+       move    t0, a2
+       sc      t0, (a1)
+       beqz    t0, 2b
+        nop
+
+       sw      v0, PT_R2+8(sp)                 # Result value
+       sw      zero, PT_R7+8(sp)               # Success indicator
+
+       lw      t0, TASK_PTRACE(gp)             # syscall tracing enabled?
+       andi    t0, PT_TRACESYS
+       bnez    t0, 3f
+        nop
+
+4:
+       lw      a0, (sp)                        # Restore arg0
+       addiu   sp, sp, 8                       # Restore sp
+
+       j       o32_ret_from_sys_call
+        nop
+
+3:
+       sw      ra, 4(sp)
+       jal     syscall_trace
+        nop
+       lw      ra, 4(sp)
+       j       4b
+        nop
+
+5:
+       lw      a0, (sp)                        # Restore arg0
+       addiu   sp, sp, 8                       # Restore sp
+       j       ra
+        li     v0, -EFAULT
+
diff -Nur linux.orig/arch/mips/kernel/irix5sys.h linux/arch/mips/kernel/irix5sys.h
--- linux.orig/arch/mips/kernel/irix5sys.h      Mon Apr  9 00:16:29 2001
+++ linux/arch/mips/kernel/irix5sys.h   Sun Apr  8 21:21:16 2001
@@ -69,7 +69,7 @@
 SYS(irix_getgid, 0)                    /* 1047  getgid()              V*/
 SYS(irix_unimp, 0)                     /* 1048  (XXX IRIX 4 ssig)     V*/
 SYS(irix_msgsys, 6)                    /* 1049  sys_msgsys            V*/
-SYS(sys_sysmips, 4)                    /* 1050  sysmips()            HV*/
+SYS(fast_sysmips, 4)                   /* 1050  sysmips()            HV*/
 SYS(irix_unimp, 0)                     /* 1051  XXX sysacct()        IV*/
 SYS(irix_shmsys, 5)                    /* 1052  sys_shmsys            V*/
 SYS(irix_semsys, 0)                    /* 1053  sys_semsys            V*/
diff -Nur linux.orig/arch/mips/kernel/syscalls.h linux/arch/mips/kernel/syscalls.h
--- linux.orig/arch/mips/kernel/syscalls.h      Mon Apr  9 00:16:30 2001
+++ linux/arch/mips/kernel/syscalls.h   Sun Apr  8 21:21:43 2001
@@ -163,7 +163,7 @@
 SYS(sys_writev, 3)
 SYS(sys_cacheflush, 3)
 SYS(sys_cachectl, 3)
-SYS(sys_sysmips, 4)
+SYS(fast_sysmips, 4)
 SYS(sys_ni_syscall, 0)                         /* 4150 */
 SYS(sys_getsid, 1)
 SYS(sys_fdatasync, 0)

Re: Illegal instruction

[Carsten Langgaard]

Thanks, that did the trick.
Now I would like to try to install H.J. Lu's RedHat7.1 RPM packages.
If I just do a:
    rpm -Uvh --root /mnt/harddisk *.rpm
then it fails on a lot of dependencies. If I try to force the install like this:
    rpm -Uvh --nodeps --force --root /mnt/harddisk *.rpm
then it actually install a lot of the RPMs, but still some fails, because it actually need
then dependencies.
If I try to redo the above it seems to go better the second time, because a lot of the
dependencies is solved (installed).
Is this really the way to install a set of RPMs like this or is there a better way?

/Carsten

Jun Sun wrote:

> Carsten Langgaard wrote:
> >
> > Hi Karel,
> >
> > I have tried the root-images tar-files: mipselroot-rh7-20010606 and
> > mipsroot-rh7.
> > The mipsroot-rh7 (bigendian) root image seem to work fine, but when
> > I use the mipselroot-rh7-20010606 (littleendian) I get an illegal
> > instruction.
> > [cat:179] Illegal instruction 7c010001 at 2ac8b20c ra=00000000.
> >
> > I'm using a 2.4.3 kernel.
> > Anyone got an idea ?
> >
> > /Carsten
> >
>
> Is the userland compiled with MIPS I option?  I bet this is the same nasty
> MIPS_ATOMIC_SET bug.
>
> Currently, there are three fixes.  One is from Florian, which only takes care
> of CPUs with ll/sc.  Another is from Maceij, which introduces a new syscall,
> and some changes in glibc.  The thrid is mine, a compromising one.  It takes
> care of both ll/sc case and non-llsc case, has minimum change, and practically
> makes all programs happy.
>
> Ralf, please take at least one of the fixes.  Even a not-so-good fix is better
> than no fix.
>
> Jun
>
>   ------------------------------------------------------------------------
>
> This is a compromising fix for sysmips(MIPS_ATOMIC_SET, ...).  It
> forces SIGSYS when the return value is a small negative value.  This
> limit is OK for glibc.
>
> Jun
>
> diff -Nru linux/arch/mips/kernel/sysmips.c.orig linux/arch/mips/kernel/sysmips.c
> --- linux/arch/mips/kernel/sysmips.c.orig       Mon Apr 23 11:32:54 2001
> +++ linux/arch/mips/kernel/sysmips.c    Tue Jun 26 11:50:48 2001
> @@ -16,7 +16,7 @@
>  #include <linux/sched.h>
>  #include <linux/string.h>
>  #include <linux/utsname.h>
> -
> +#include <linux/signal.h>
>  #include <asm/cachectl.h>
>  #include <asm/pgalloc.h>
>  #include <asm/sysmips.h>
> @@ -75,53 +75,83 @@
>         }
>
>         case MIPS_ATOMIC_SET: {
> -#ifdef CONFIG_CPU_HAS_LLSC
> -               unsigned int tmp;
> +               int *ptr, val, ret, err, tmp;
> +               struct siginfo info;
>
> -               p = (int *) arg1;
> -               errno = verify_area(VERIFY_WRITE, p, sizeof(*p));
> -               if (errno)
> -                       return errno;
> -               errno = 0;
> -
> -               __asm__(".set\tpush\t\t\t# sysmips(MIPS_ATOMIC, ...)\n\t"
> -                       ".set\tmips2\n\t"
> -                       ".set\tnoat\n\t"
> -                       "1:\tll\t%0, %4\n\t"
> -                       "move\t$1, %3\n\t"
> -                       "2:\tsc\t$1, %1\n\t"
> -                       "beqz\t$1, 1b\n\t"
> -                       ".set\tpop\n\t"
> -                       ".section\t.fixup,\"ax\"\n"
> -                       "3:\tli\t%2, 1\t\t\t# error\n\t"
> -                       ".previous\n\t"
> -                       ".section\t__ex_table,\"a\"\n\t"
> -                       ".word\t1b, 3b\n\t"
> -                       ".word\t2b, 3b\n\t"
> -                       ".previous\n\t"
> -                       : "=&r" (tmp), "=o" (* (u32 *) p), "=r" (errno)
> -                       : "r" (arg2), "o" (* (u32 *) p), "2" (errno)
> -                       : "$1");
> +               ptr = (int *)arg1;
> +               val = (int)arg2;
>
> -               if (errno)
> -                       return -EFAULT;
> +               /* Don't emulate unaligned accesses. */
> +               if ((int)ptr & 3) {
> +                       info.si_signo = SIGBUS;
> +                        info.si_code = BUS_ADRALN;
> +                       goto fault;
> +               }
> +
> +               /* A zero here saves us three instructions. */
> +               err = verify_area(VERIFY_WRITE, ptr, 0);
> +               if (err) {
> +                        info.si_signo = SIGSEGV;
> +                       info.si_code = SEGV_ACCERR;
> +                       goto fault;
> +               }
>
> -               /* We're skipping error handling etc.  */
> -               if (current->ptrace & PT_TRACESYS)
> -                       syscall_trace();
> -
> -               ((struct pt_regs *)&cmd)->regs[2] = tmp;
> -               ((struct pt_regs *)&cmd)->regs[7] = 0;
> -
> -               __asm__ __volatile__(
> -                       "move\t$29, %0\n\t"
> -                       "j\to32_ret_from_sys_call"
> -                       : /* No outputs */
> -                       : "r" (&cmd));
> -               /* Unreached */
> +#ifdef CONFIG_CPU_HAS_LLSC
> +               __asm__(".set   mips2\n\t"
> +                       "1:\n\t"
> +                       "ll     %0,%5\n\t"
> +                       ".set   push\n\t"
> +                       ".set   noreorder\n\t"
> +                       "beq    %0,%4,3f\n\t"
> +                       " move  %3,%4\n"
> +                       ".set   pop\n\t"
> +                       "2:\n\t"
> +                       "sc     %3,%1\n\t"
> +                       "beqz   %3,1b\n\t"
> +                       "3:\n\t"
> +                       ".set   mips0\n\t"
> +                       ".section .fixup,\"ax\"\n"
> +                       "4:\n\t"
> +                       "li     %2,%7\n\t"
> +                       "j      3b\n\t"
> +                       ".previous\n\t"
> +                       ".section __ex_table,\"a\"\n\t"
> +                       ".word  1b,4b\n\t"
> +                       ".word  2b,4b\n\t"
> +                       ".previous"
> +                       : "=&r" (ret), "=R" (*ptr), "=r" (err), "=&r" (tmp)
> +                       : "r" (val), "1" (*ptr), "2" (0), "i" (-EFAULT));
>  #else
> -       printk("sys_sysmips(MIPS_ATOMIC_SET, ...) not ready for !CONFIG_CPU_HAS_LLSC\n");
> +               save_and_cli(tmp);
> +               err = __get_user(ret, ptr);
> +               if (ret != val)
> +               err |= __put_user(val, ptr);    /* No fault
> +                                                   unless unwriteable. */
> +               restore_flags(tmp);
>  #endif
> +
> +               if (err) {
> +                        info.si_signo = SIGSEGV;
> +                       info.si_code = SEGV_MAPERR;
> +                       goto fault;
> +               }
> +
> +               if ( (ret < 0) && (ret >= -EMAXERRNO) ) {
> +                        info.si_signo = SIGSYS;
> +                       info.si_code = 0;
> +                       goto fault;
> +               }
> +
> +               return ret;
> +
> +fault:
> +               /* Go back to SYSCALL. */
> +               ((struct pt_regs *)&cmd)->cp0_epc -= 4;
> +
> +               info.si_addr = (void *)((struct pt_regs *)&cmd)->cp0_epc;
> +               force_sig_info(info.si_signo, &info, current);
> +
> +               return 0;
>         }
>
>         case MIPS_FIXADE:
>
>   ------------------------------------------------------------------------
> diff -Nur linux.orig/arch/mips/kernel/Makefile linux/arch/mips/kernel/Makefile
> --- linux.orig/arch/mips/kernel/Makefile        Mon Apr  9 00:23:08 2001
> +++ linux/arch/mips/kernel/Makefile     Mon Apr  9 00:23:34 2001
> @@ -20,7 +20,7 @@
>  obj-y                          += branch.o process.o signal.o entry.o \
>                                    traps.o ptrace.o vm86.o ioport.o reset.o \
>                                    semaphore.o setup.o syscall.o sysmips.o \
> -                                  ipc.o scall_o32.o unaligned.o
> +                                  ipc.o scall_o32.o unaligned.o fast-sysmips.o
>  obj-$(CONFIG_MODULES)          += mips_ksyms.o
>
>  ifdef CONFIG_CPU_R3000
> @@ -69,5 +69,6 @@
>
>  entry.o: entry.S
>  head.o: head.S
> +fast-sysmips.o: fast-sysmips.S
>
>  include $(TOPDIR)/Rules.make
> diff -Nur linux.orig/arch/mips/kernel/fast-sysmips.S linux/arch/mips/kernel/fast-sysmips.S
> --- linux.orig/arch/mips/kernel/fast-sysmips.S  Thu Jan  1 01:00:00 1970
> +++ linux/arch/mips/kernel/fast-sysmips.S       Mon Apr  9 00:28:20 2001
> @@ -0,0 +1,85 @@
> +/*
> + * MIPS_ATOMIC_SET asm implementation for ll/sc capable cpus
> + *
> + * This file is subject to the terms and conditions of the GNU General Public
> + * License.  See the file "COPYING" in the main directory of this archive
> + * for more details.
> + *
> + * Copyright (C) 2001 Florian Lohoff <flo@rfc822.org>
> + *
> + */
> +#include <asm/asm.h>
> +#include <asm/mipsregs.h>
> +#include <asm/regdef.h>
> +#include <asm/stackframe.h>
> +#include <asm/isadep.h>
> +#include <asm/unistd.h>
> +#include <asm/sysmips.h>
> +#include <asm/offset.h>
> +#include <asm/errno.h>
> +
> +#define PT_TRACESYS     0x00000002
> +
> +       EXPORT(fast_sysmips)
> +
> +       .set    noreorder
> +
> +       li      t0, MIPS_ATOMIC_SET
> +       beq     a0, t0, 1f
> +        nop
> +       j       sys_sysmips
> +        nop
> +
> +1:
> +
> +       # a0 - MIPS_ATOMIC_SET
> +       # a1 - mem ptr
> +       # a2 - value
> +
> +       addiu   sp, sp, -8                      # Reserve space
> +       sw      a0, (sp)                        # Save arg0
> +
> +       addiu   a0, a1, 4                       # addr+size
> +       ori     v0, a1, 4                       # addr | size
> +       lw      v1, THREAD_CURDS(gp)            # current->thread.current_ds
> +       or      v0, v0, a0                      # addr | size | (addr+size)
> +       and     v1, v1, v0                      # (mask)&(addr | size | (addr+size)
> +       bltz    v1, 5f
> +        nop
> +
> +2:
> +       ll      v0, (a1)
> +       move    t0, a2
> +       sc      t0, (a1)
> +       beqz    t0, 2b
> +        nop
> +
> +       sw      v0, PT_R2+8(sp)                 # Result value
> +       sw      zero, PT_R7+8(sp)               # Success indicator
> +
> +       lw      t0, TASK_PTRACE(gp)             # syscall tracing enabled?
> +       andi    t0, PT_TRACESYS
> +       bnez    t0, 3f
> +        nop
> +
> +4:
> +       lw      a0, (sp)                        # Restore arg0
> +       addiu   sp, sp, 8                       # Restore sp
> +
> +       j       o32_ret_from_sys_call
> +        nop
> +
> +3:
> +       sw      ra, 4(sp)
> +       jal     syscall_trace
> +        nop
> +       lw      ra, 4(sp)
> +       j       4b
> +        nop
> +
> +5:
> +       lw      a0, (sp)                        # Restore arg0
> +       addiu   sp, sp, 8                       # Restore sp
> +       j       ra
> +        li     v0, -EFAULT
> +
> diff -Nur linux.orig/arch/mips/kernel/irix5sys.h linux/arch/mips/kernel/irix5sys.h
> --- linux.orig/arch/mips/kernel/irix5sys.h      Mon Apr  9 00:16:29 2001
> +++ linux/arch/mips/kernel/irix5sys.h   Sun Apr  8 21:21:16 2001
> @@ -69,7 +69,7 @@
>  SYS(irix_getgid, 0)                    /* 1047  getgid()              V*/
>  SYS(irix_unimp, 0)                     /* 1048  (XXX IRIX 4 ssig)     V*/
>  SYS(irix_msgsys, 6)                    /* 1049  sys_msgsys            V*/
> -SYS(sys_sysmips, 4)                    /* 1050  sysmips()            HV*/
> +SYS(fast_sysmips, 4)                   /* 1050  sysmips()            HV*/
>  SYS(irix_unimp, 0)                     /* 1051  XXX sysacct()        IV*/
>  SYS(irix_shmsys, 5)                    /* 1052  sys_shmsys            V*/
>  SYS(irix_semsys, 0)                    /* 1053  sys_semsys            V*/
> diff -Nur linux.orig/arch/mips/kernel/syscalls.h linux/arch/mips/kernel/syscalls.h
> --- linux.orig/arch/mips/kernel/syscalls.h      Mon Apr  9 00:16:30 2001
> +++ linux/arch/mips/kernel/syscalls.h   Sun Apr  8 21:21:43 2001
> @@ -163,7 +163,7 @@
>  SYS(sys_writev, 3)
>  SYS(sys_cacheflush, 3)
>  SYS(sys_cachectl, 3)
> -SYS(sys_sysmips, 4)
> +SYS(fast_sysmips, 4)
>  SYS(sys_ni_syscall, 0)                         /* 4150 */
>  SYS(sys_getsid, 1)
>  SYS(sys_fdatasync, 0)

--
_    _ ____  ___   Carsten Langgaard   Mailto:carstenl@mips.com
|\  /|||___)(___   MIPS Denmark        Direct: +45 4486 5527
| \/ |||    ____)  Lautrupvang 4B      Switch: +45 4486 5555
  TECHNOLOGIES     2750 Ballerup       Fax...: +45 4486 5556
                   Denmark             http://www.mips.com

Re: Illegal instruction

[H . J . Lu]

On Tue, Jul 10, 2001 at 01:14:12PM +0200, Carsten Langgaard wrote:
> Thanks, that did the trick.
> Now I would like to try to install H.J. Lu's RedHat7.1 RPM packages.
> If I just do a:
>     rpm -Uvh --root /mnt/harddisk *.rpm

Those rpms have to be installed in the right order. I have a set up
to do that. I will see what I can do.


H.J.

Re: Illegal instruction

[Carsten Langgaard]

"H . J . Lu" wrote:

> On Tue, Jul 10, 2001 at 01:14:12PM +0200, Carsten Langgaard wrote:
> > Thanks, that did the trick.
> > Now I would like to try to install H.J. Lu's RedHat7.1 RPM packages.
> > If I just do a:
> >     rpm -Uvh --root /mnt/harddisk *.rpm
>
> Those rpms have to be installed in the right order. I have a set up
> to do that. I will see what I can do.

Thanks, please do.
Another thing, I can see your distribution is lacking kernel header
files, I guess they are needed to do a native compile of the source RPMs.

Are you only cross compiling the RPMs.
It would be great, if you could put together a RPM containing the kernel
header files.

>
> H.J.

--
_    _ ____  ___   Carsten Langgaard   Mailto:carstenl@mips.com
|\  /|||___)(___   MIPS Denmark        Direct: +45 4486 5527
| \/ |||    ____)  Lautrupvang 4B      Switch: +45 4486 5555
  TECHNOLOGIES     2750 Ballerup       Fax...: +45 4486 5556
                   Denmark             http://www.mips.com

Updates on RedHat 7.1/mips

[H . J . Lu]

On Mon, Jul 16, 2001 at 01:26:32PM +0200, Carsten Langgaard wrote:
> "H . J . Lu" wrote:
> > > Now I would like to try to install H.J. Lu's RedHat7.1 RPM packages.
> > > If I just do a:
> > >     rpm -Uvh --root /mnt/harddisk *.rpm
> >
> > Those rpms have to be installed in the right order. I have a set up
> > to do that. I will see what I can do.
> 
> Thanks, please do.

They are in install.tar.bz2 now.

> Another thing, I can see your distribution is lacking kernel header
> files, I guess they are needed to do a native compile of the source RPMs.

I added gdb and kernel-headers. I also updated gcc, glibc and binutils.
The toolchain rpms are updated.


H.J.
------
My mini-port of RedHat 7.1 is at

ftp://oss.sgi.com/pub/linux/mips/redhat/7.1/

you should be able to put a small RedHat 7.1 on the mips/mipsel box and
compile the rest of RedHat 7.1 yourselves.

Here are something you should know:

1. The cross compiler hosted on RedHat 7.1/ia32 is provided as a
toolchain rpm. The binary rpms for the mips and mipsel cross compilers
are included. You will need glibc 2.2.3-11 or above to use those
rpms. The glibc x86 binary rpms under RPMS/i386 should be ok.
2. You have to find a way to put those rpms on your machine. I use
network boot and NFS root to do it.
3. install.tar.bz2 has some scripts to prepare NFS root and install
RedHat 7.1 on a hard drive.

Thanks.


H.J.

Re: Updates on RedHat 7.1/mips

[Carsten Langgaard]

"H . J . Lu" wrote:

> On Mon, Jul 16, 2001 at 01:26:32PM +0200, Carsten Langgaard wrote:
> > "H . J . Lu" wrote:
> > > > Now I would like to try to install H.J. Lu's RedHat7.1 RPM packages.
> > > > If I just do a:
> > > >     rpm -Uvh --root /mnt/harddisk *.rpm
> > >
> > > Those rpms have to be installed in the right order. I have a set up
> > > to do that. I will see what I can do.
> >
> > Thanks, please do.
>
> They are in install.tar.bz2 now.

Thanks a lot, that look like something we have been needing for a long time

> > Another thing, I can see your distribution is lacking kernel header
> > files, I guess they are needed to do a native compile of the source RPMs.
>
> I added gdb and kernel-headers. I also updated gcc, glibc and binutils.
> The toolchain rpms are updated.

Now that I'm at it :-)
Could you add a Perl rpm to you distribution, this also seems to be needed to
build the RPMs natively.

I also would like to do a cross build, but how do I do that ?
Do I have to change all the spec files, to set the target arch to MIPS, or is
there a easier and better way to do that ?
I guess if I just install the SRPMs and the cross toolchain, and try to build
out of that I will simply build for the build platform, which is i386.

>
>
> H.J.
> ------
> My mini-port of RedHat 7.1 is at
>
> ftp://oss.sgi.com/pub/linux/mips/redhat/7.1/
>
> you should be able to put a small RedHat 7.1 on the mips/mipsel box and
> compile the rest of RedHat 7.1 yourselves.
>
> Here are something you should know:
>
> 1. The cross compiler hosted on RedHat 7.1/ia32 is provided as a
> toolchain rpm. The binary rpms for the mips and mipsel cross compilers
> are included. You will need glibc 2.2.3-11 or above to use those
> rpms. The glibc x86 binary rpms under RPMS/i386 should be ok.
> 2. You have to find a way to put those rpms on your machine. I use
> network boot and NFS root to do it.
> 3. install.tar.bz2 has some scripts to prepare NFS root and install
> RedHat 7.1 on a hard drive.
>
> Thanks.
>
> H.J.

Re: Updates on RedHat 7.1/mips

[H . J . Lu]

On Tue, Jul 17, 2001 at 09:16:05PM +0200, Carsten Langgaard wrote:
> Could you add a Perl rpm to you distribution, this also seems to be needed to
> build the RPMs natively.

Perl has to be built natively. I uploaded mysql-3.23.36-1.1.src.rpm,
perl-5.6.0-12.1.src.rpm, apache-1.3.19-5.src.rpm,
mod_perl-1.24_01-2.src.rpm, tcsh-6.10-5.src.rpm and
zsh-3.0.8-8.src.rpm. Just installed my RedHat 7.1. Then you can build
perl yourself. You may need to build/install the tcsh rpm first.

> 
> I also would like to do a cross build, but how do I do that ?
> Do I have to change all the spec files, to set the target arch to MIPS, or is
> there a easier and better way to do that ?
> I guess if I just install the SRPMs and the cross toolchain, and try to build
> out of that I will simply build for the build platform, which is i386.

You need my patched `rpm' to do cross build. You also need some
specicial set up. I am enclosng a message of mine. I will see if
I can cleanup my cross build environment and upload it.


H.J.
-----
> they work great! Thanks for all your hard work. Now, I would like
> to be able to build some RPMs as well. I have never cross-built
> RPMs before. I took the SRPMS for make and tried:
> 
>    rpm -ba make.spec --target=mipsel
> 
> This did not do anything and built nothing. Is there a FAQ that
> I can read on cross building RPMS in addition to the documentation
> on http://www.rpm.org/ ? Do you have any quick pointers? Thanks
> again.

Although my source rpms do support cross compile, FYI, all my
mips/mipsel rpms are cross compiled from ia32 machines, the
unpatched rpm, that is /bin/rpm, doesn't support cross compile.
In my RedHat 7.1, there are i386 rpm binaries which support
cross compile. I have a setup to cross build rpms to mips/mipsel.
Make sure your rpm sees

%__ar                   %{_target_prefix}-linux-ar
%__as                   %{_target_prefix}-linux-as
%__cc                   %{_target_prefix}-linux-gcc
%__cxx                  %{_target_prefix}-linux-c++
%__ld                   %{_target_prefix}-linux-ld
%__nm                   %{_target_prefix}-linux-nm
%__objcopy              %{_target_prefix}-linux-objcopy
%__objdump              %{_target_prefix}-linux-objdump
%__ranlib               %{_target_prefix}-linux-ranlib
%__strip                %{_target_prefix}-linux-strip

and do

# rpm --target=mips --define '_target_prefix mips'

or

# rpm --target=mipsel --define '_target_prefix mipsel'

BTW, you alao need to rebuild binutils with

# rpm --define 'ENABLE_ALL_TARGETS 1' -ta binutils-2.11.90.0.23.tar.gz


H.J.

Re: Updates on RedHat 7.1/mips

[Daniel Jacobowitz]

On Tue, Jul 17, 2001 at 12:29:02PM -0700, H . J . Lu wrote:
> On Tue, Jul 17, 2001 at 09:16:05PM +0200, Carsten Langgaard wrote:
> > Could you add a Perl rpm to you distribution, this also seems to be needed to
> > build the RPMs natively.
> 
> Perl has to be built natively. I uploaded mysql-3.23.36-1.1.src.rpm,
> perl-5.6.0-12.1.src.rpm, apache-1.3.19-5.src.rpm,
> mod_perl-1.24_01-2.src.rpm, tcsh-6.10-5.src.rpm and
> zsh-3.0.8-8.src.rpm. Just installed my RedHat 7.1. Then you can build
> perl yourself. You may need to build/install the tcsh rpm first.

It's not yet available for MIPS (later this week), but MontaVista Journeyman
contains the patches to cross-compile Perl.  It's not pretty, though.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: Updates on RedHat 7.1/mips

[H . J . Lu]

On Tue, Jul 17, 2001 at 12:50:27PM -0700, Daniel Jacobowitz wrote:
> > 
> > Perl has to be built natively. I uploaded mysql-3.23.36-1.1.src.rpm,
> > perl-5.6.0-12.1.src.rpm, apache-1.3.19-5.src.rpm,
> > mod_perl-1.24_01-2.src.rpm, tcsh-6.10-5.src.rpm and
> > zsh-3.0.8-8.src.rpm. Just installed my RedHat 7.1. Then you can build
> > perl yourself. You may need to build/install the tcsh rpm first.
> 
> It's not yet available for MIPS (later this week), but MontaVista Journeyman
> contains the patches to cross-compile Perl.  It's not pretty, though.
> 

That is a reason why I didn't bother with cross-compile Perl :-). The
next thing on my todo list is to cross compile XFree86 :-(.


H.J.

Re: Updates on RedHat 7.1/mips

[Ilya Volynets]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On Tuesday 17 July 2001 12:57, H . J . Lu wrote:
> On Tue, Jul 17, 2001 at 12:50:27PM -0700, Daniel Jacobowitz wrote:
> > > Perl has to be built natively. I uploaded mysql-3.23.36-1.1.src.rpm,
> > > perl-5.6.0-12.1.src.rpm, apache-1.3.19-5.src.rpm,
> > > mod_perl-1.24_01-2.src.rpm, tcsh-6.10-5.src.rpm and
> > > zsh-3.0.8-8.src.rpm. Just installed my RedHat 7.1. Then you can build
> > > perl yourself. You may need to build/install the tcsh rpm first.
> >
> > It's not yet available for MIPS (later this week), but MontaVista
> > Journeyman contains the patches to cross-compile Perl.  It's not pretty,
> > though.
>
> That is a reason why I didn't bother with cross-compile Perl :-). The
> next thing on my todo list is to cross compile XFree86 :-(.
If you can get cross-compiling full XFree86, you'll be a hero of all times.
It shouldn't be hard to get it with KDriveArchitecture & TinyX enabled.

I was working on it a while ago, and here are few pointers:
Some tools have to be run natively (i.e. xkbcomp), but also need to
be installed on target. I din't find a rule that does both. I think new
rule is needed.

gcc-3.0 crashes when compiling some parts of Xserver and Xlib,
with very obscure bug. Minimal test case I came up with is
~45(!) lines long. Keith filed report to gcc team on my behalf,
but there seems to be no responce. I do not know if your
gcc has same problem, but someone mentioned similar problem
with 2.9x.y series on this list not so long ago.

Hmm... I can't remember any other interesting things at this moment, but if I
will, I'll follow up...

	Ilya.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAjtUm18ACgkQtKh84cA8u2ll+QCfSjexRpQ2H7Qa16JYrun/vvVd
vCQAoIKb1BRxJ/508SWu9uHEP0okhH00
=kWTa
-----END PGP SIGNATURE-----

Re: Updates on RedHat 7.1/mips

[H . J . Lu]

On Tue, Jul 17, 2001 at 01:09:00PM -0700, Ilya Volynets wrote:
> 
> I was working on it a while ago, and here are few pointers:
> Some tools have to be run natively (i.e. xkbcomp), but also need to
> be installed on target. I din't find a rule that does both. I think new
> rule is needed.

I am aware of those. I want to delay it as much as I can.

> 
> gcc-3.0 crashes when compiling some parts of Xserver and Xlib,
> with very obscure bug. Minimal test case I came up with is
> ~45(!) lines long. Keith filed report to gcc team on my behalf,
> but there seems to be no responce. I do not know if your
> gcc has same problem, but someone mentioned similar problem
> with 2.9x.y series on this list not so long ago.
> 

Don't bother with gcc-3.0. I won't use it myself for building X.


H.J.

Re: Updates on RedHat 7.1/mips

[Arnaldo Carvalho de Melo]

Em Tue, Jul 17, 2001 at 01:11:46PM -0700, H . J . Lu escreveu:
> On Tue, Jul 17, 2001 at 01:09:00PM -0700, Ilya Volynets wrote:
> > 
> > I was working on it a while ago, and here are few pointers:
> > Some tools have to be run natively (i.e. xkbcomp), but also need to
> > be installed on target. I din't find a rule that does both. I think new
> > rule is needed.
> 
> I am aware of those. I want to delay it as much as I can.

recent messages to the XFree86 mailing lists has the recipe to x-compile
XFree86, IIRC Keith Packard did the work
 
> > 
> > gcc-3.0 crashes when compiling some parts of Xserver and Xlib,
> > with very obscure bug. Minimal test case I came up with is
> > ~45(!) lines long. Keith filed report to gcc team on my behalf,
> > but there seems to be no responce. I do not know if your
> > gcc has same problem, but someone mentioned similar problem
> > with 2.9x.y series on this list not so long ago.
> > 
> 
> Don't bother with gcc-3.0. I won't use it myself for building X.
> 
> 
> H.J.

-- 


                        - Arnaldo

Re: Updates on RedHat 7.1/mips

[Ilya Volynets]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Those messages are about compiling using TinyX and KDrive, which
works with only minor modifications. Thing is, it is not what is really needed.

On Tuesday 17 July 2001 20:52, Arnaldo Carvalho de Melo wrote:
> Em Tue, Jul 17, 2001 at 01:11:46PM -0700, H . J . Lu escreveu:
> > On Tue, Jul 17, 2001 at 01:09:00PM -0700, Ilya Volynets wrote:
> > > I was working on it a while ago, and here are few pointers:
> > > Some tools have to be run natively (i.e. xkbcomp), but also need to
> > > be installed on target. I din't find a rule that does both. I think new
> > > rule is needed.
> >
> > I am aware of those. I want to delay it as much as I can.
>
> recent messages to the XFree86 mailing lists has the recipe to x-compile
> XFree86, IIRC Keith Packard did the work
>
> > > gcc-3.0 crashes when compiling some parts of Xserver and Xlib,
> > > with very obscure bug. Minimal test case I came up with is
> > > ~45(!) lines long. Keith filed report to gcc team on my behalf,
> > > but there seems to be no responce. I do not know if your
> > > gcc has same problem, but someone mentioned similar problem
> > > with 2.9x.y series on this list not so long ago.
> >
> > Don't bother with gcc-3.0. I won't use it myself for building X.
> >
> >
> > H.J.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.4 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAjtVChMACgkQtKh84cA8u2nRTACgvPDSNX0E62POj2W+QkntYE9l
pzIAnAjjprS4qWgUfELszqT94T+XGV0C
=tJST
-----END PGP SIGNATURE-----

Re: Updates on RedHat 7.1/mips

[Carsten Langgaard]

"H . J . Lu" wrote:

> On Tue, Jul 17, 2001 at 09:16:05PM +0200, Carsten Langgaard wrote:
> > Could you add a Perl rpm to you distribution, this also seems to be needed to
> > build the RPMs natively.
>
> Perl has to be built natively. I uploaded mysql-3.23.36-1.1.src.rpm,
> perl-5.6.0-12.1.src.rpm, apache-1.3.19-5.src.rpm,
> mod_perl-1.24_01-2.src.rpm, tcsh-6.10-5.src.rpm and
> zsh-3.0.8-8.src.rpm. Just installed my RedHat 7.1. Then you can build
> perl yourself. You may need to build/install the tcsh rpm first.

It look like there is a cross dependence, the build of tcsh failed with the
following message:

/var/tmp/rpm-tmp.7250: /usr/bin/perl: No such file or directory
error: Bad exit status from /var/tmp/rpm-tmp.7250 (%build)

So tcsh is needed to build perl and perl is needed to build tcsh :-(

>
> >
> > I also would like to do a cross build, but how do I do that ?
> > Do I have to change all the spec files, to set the target arch to MIPS, or is
> > there a easier and better way to do that ?
> > I guess if I just install the SRPMs and the cross toolchain, and try to build
> > out of that I will simply build for the build platform, which is i386.
>
> You need my patched `rpm' to do cross build. You also need some
> specicial set up. I am enclosng a message of mine. I will see if
> I can cleanup my cross build environment and upload it.
>
> H.J.
> -----
> > they work great! Thanks for all your hard work. Now, I would like
> > to be able to build some RPMs as well. I have never cross-built
> > RPMs before. I took the SRPMS for make and tried:
> >
> >    rpm -ba make.spec --target=mipsel
> >
> > This did not do anything and built nothing. Is there a FAQ that
> > I can read on cross building RPMS in addition to the documentation
> > on http://www.rpm.org/ ? Do you have any quick pointers? Thanks
> > again.
>
> Although my source rpms do support cross compile, FYI, all my
> mips/mipsel rpms are cross compiled from ia32 machines, the
> unpatched rpm, that is /bin/rpm, doesn't support cross compile.
> In my RedHat 7.1, there are i386 rpm binaries which support
> cross compile. I have a setup to cross build rpms to mips/mipsel.
> Make sure your rpm sees
>
> %__ar                   %{_target_prefix}-linux-ar
> %__as                   %{_target_prefix}-linux-as
> %__cc                   %{_target_prefix}-linux-gcc
> %__cxx                  %{_target_prefix}-linux-c++
> %__ld                   %{_target_prefix}-linux-ld
> %__nm                   %{_target_prefix}-linux-nm
> %__objcopy              %{_target_prefix}-linux-objcopy
> %__objdump              %{_target_prefix}-linux-objdump
> %__ranlib               %{_target_prefix}-linux-ranlib
> %__strip                %{_target_prefix}-linux-strip
>

How do I do that ?

>
> and do
>
> # rpm --target=mips --define '_target_prefix mips'
>
> or
>
> # rpm --target=mipsel --define '_target_prefix mipsel'
>
> BTW, you alao need to rebuild binutils with
>
> # rpm --define 'ENABLE_ALL_TARGETS 1' -ta binutils-2.11.90.0.23.tar.gz
>
> H.J.

--
_    _ ____  ___   Carsten Langgaard   Mailto:carstenl@mips.com
|\  /|||___)(___   MIPS Denmark        Direct: +45 4486 5527
| \/ |||    ____)  Lautrupvang 4B      Switch: +45 4486 5555
  TECHNOLOGIES     2750 Ballerup       Fax...: +45 4486 5556
                   Denmark             http://www.mips.com

Re: Updates on RedHat 7.1/mips

[H . J . Lu]

On Wed, Jul 18, 2001 at 09:13:20AM +0200, Carsten Langgaard wrote:
> "H . J . Lu" wrote:
> 
> > On Tue, Jul 17, 2001 at 09:16:05PM +0200, Carsten Langgaard wrote:
> > > Could you add a Perl rpm to you distribution, this also seems to be needed to
> > > build the RPMs natively.
> >
> > Perl has to be built natively. I uploaded mysql-3.23.36-1.1.src.rpm,
> > perl-5.6.0-12.1.src.rpm, apache-1.3.19-5.src.rpm,
> > mod_perl-1.24_01-2.src.rpm, tcsh-6.10-5.src.rpm and
> > zsh-3.0.8-8.src.rpm. Just installed my RedHat 7.1. Then you can build
> > perl yourself. You may need to build/install the tcsh rpm first.
> 
> It look like there is a cross dependence, the build of tcsh failed with the
> following message:
> 
> /var/tmp/rpm-tmp.7250: /usr/bin/perl: No such file or directory
> error: Bad exit status from /var/tmp/rpm-tmp.7250 (%build)
> 
> So tcsh is needed to build perl and perl is needed to build tcsh :-(
> 

It is ok since the tcsh binary has been built. Just install it to
/bin/tcsh and build the perl rpm. You may have to

# rpm --nodeps --force --rebuild perl-xxx.src.rpm

You only have to do those the first time.



H.J.

Re: Updates on RedHat 7.1/mips

[Carsten Langgaard]

"H . J . Lu" wrote:

> On Wed, Jul 18, 2001 at 09:13:20AM +0200, Carsten Langgaard wrote:
> > "H . J . Lu" wrote:
> >
> > > On Tue, Jul 17, 2001 at 09:16:05PM +0200, Carsten Langgaard wrote:
> > > > Could you add a Perl rpm to you distribution, this also seems to be needed to
> > > > build the RPMs natively.
> > >
> > > Perl has to be built natively. I uploaded mysql-3.23.36-1.1.src.rpm,
> > > perl-5.6.0-12.1.src.rpm, apache-1.3.19-5.src.rpm,
> > > mod_perl-1.24_01-2.src.rpm, tcsh-6.10-5.src.rpm and
> > > zsh-3.0.8-8.src.rpm. Just installed my RedHat 7.1. Then you can build
> > > perl yourself. You may need to build/install the tcsh rpm first.
> >
> > It look like there is a cross dependence, the build of tcsh failed with the
> > following message:
> >
> > /var/tmp/rpm-tmp.7250: /usr/bin/perl: No such file or directory
> > error: Bad exit status from /var/tmp/rpm-tmp.7250 (%build)
> >
> > So tcsh is needed to build perl and perl is needed to build tcsh :-(
> >
>
> It is ok since the tcsh binary has been built. Just install it to
> /bin/tcsh and build the perl rpm. You may have to
>
> # rpm --nodeps --force --rebuild perl-xxx.src.rpm
>
> You only have to do those the first time.

I got through the tcsh and perl build, but now it fails on glibc.

make -s -j2 -C iconvdata others
make[2]: Entering directory `/usr/src/redhat/BUILD/glibc-2.2.3/iconvdata'
make[2]: warning: -jN forced in submake: disabling jobserver mode.
In file included from 8bit-gap.c:33,
                 from iso8859-15.c:28:
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:260:
numeric constant with no digits
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:260: parse
error before `nan'
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:260:
numeric constant with no digits
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:260: parse
error before `nan'
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:260: `nan'
undeclared here (not in a function)
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:260:
initializer element is not constant
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:260: (near
initialization for `from_idx[1].idx')
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:261: `nan'
undeclared here (not in a function)
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:261:
initializer element is not constant
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:261: (near
initialization for `from_idx[2].idx')
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:262: `nan'
undeclared here (not in a function)
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:262:
initializer element is not constant
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:262: (near
initialization for `from_idx[3].idx')
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:263: `nan'
undeclared here (not in a function)
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:263:
initializer element is not constant
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:263: (near
initialization for `from_idx[4].idx')
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:264: `nan'
undeclared here (not in a function)
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:264:
initializer element is not constant
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:264: (near
initialization for `from_idx[5].idx')
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:265: `nan'
undeclared here (not in a function)
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:265:
initializer element is not constant
/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.h:265: (near
initialization for `from_idx[6].idx')
make[2]: ***
[/usr/src/redhat/BUILD/glibc-2.2.3/build-mipsel-linux/iconvdata/iso8859-15.os] Error 1

make[2]: *** Waiting for unfinished jobs....
make[2]: Leaving directory `/usr/src/redhat/BUILD/glibc-2.2.3/iconvdata'
make[1]: *** [iconvdata/others] Error 2
make[1]: Leaving directory `/usr/src/redhat/BUILD/glibc-2.2.3'
make: *** [all] Error 2
error: Bad exit status from /var/tmp/rpm-tmp.98017 (%build)


RPM build errors:
    Bad exit status from /var/tmp/rpm-tmp.98017 (%build)


>
> H.J.

--
_    _ ____  ___   Carsten Langgaard   Mailto:carstenl@mips.com
|\  /|||___)(___   MIPS Denmark        Direct: +45 4486 5527
| \/ |||    ____)  Lautrupvang 4B      Switch: +45 4486 5555
  TECHNOLOGIES     2750 Ballerup       Fax...: +45 4486 5556
                   Denmark             http://www.mips.com

Re: Updates on RedHat 7.1/mips

[Seth Mos]

I am having trouble getting the mipsel packages installed using the
installer provided for 7.1.


I have hooked up a second disk to my i386 workstatition that will house
the filesystem for my mipsel box. Once everything is installed I will put
this disk back in my Qube2 for testing. In the worst case scenario I will
need to boot with the 2.0 kernel that came with the cube.

I have the original disk for the Qube2 so it isn't a big problem if it
takes 50 attempts. I have a lot of patience ;-)

Am I correct to assume that I need to use the install.i386.hd on my
workstation to install the mipsel packages on the second harddrive (hdd in
my case)

I have tried this approach using both the Makefile approach and by running
the install.i386.hd script.

Traceback (innermost, last):
  File "./findrpm", line 5, in ?
    import rpm
ImportError: No module named rpm

Does this ring any bells?
The mipsel packages are located in the RPMS directory which is in the same
directory as the installer and I edited the Makefile and changed ROOT and
REDHAT-ROOT

I wonder if anyone can provide a basic root tarbal which I can use to set
up the basics after which i can start adding stuff.

If I get it into a decent state I will probably make a dump of the fs so
other people with a Qube (1 or 2) can easily clone it.

I was working to use a 2.4 kernel on a standard Qube but
that was a big problem since 2.4 does not compile with the compiler that
came with it and the glibc did not have large file support.

A fs that I would like to put on the Qube is XFS which I already use on my
own i386 home machines and production machines at work.
Since XFS is also tested on mip64 Irix (IIRC) ai64 PPC ai32 and alpha
under linux I expect that it will probably more work but it is more
resilient and faster recovering after crashes.

So if anyone can give advice on installing using the toolkit or provide a
bootable dump image or root tarbal I would be most grateful.
The qube2 comes with a boot loader that understands ext2 and loads the
linux kernel it finds on the system it will probably make it easier to
boot afterwards.

Cheers
Seth

Re: Updates on RedHat 7.1/mips

[H . J . Lu]

On Wed, Jul 18, 2001 at 10:53:22PM +0200, Seth Mos wrote:
> 
> I am having trouble getting the mipsel packages installed using the
> installer provided for 7.1.
> 
> 
> I have hooked up a second disk to my i386 workstatition that will house
> the filesystem for my mipsel box. Once everything is installed I will put
> this disk back in my Qube2 for testing. In the worst case scenario I will
> need to boot with the 2.0 kernel that came with the cube.
> 
> I have the original disk for the Qube2 so it isn't a big problem if it
> takes 50 attempts. I have a lot of patience ;-)
> 
> Am I correct to assume that I need to use the install.i386.hd on my
> workstation to install the mipsel packages on the second harddrive (hdd in
> my case)

No. It should be install.mipsel.nfs. install.mipsel.hd is for the
native mipsel machine install.

> 
> Traceback (innermost, last):
>   File "./findrpm", line 5, in ?
>     import rpm
> ImportError: No module named rpm
> 
> Does this ring any bells?

You need to install the rpm-python rpm.

> The mipsel packages are located in the RPMS directory which is in the same
> directory as the installer and I edited the Makefile and changed ROOT and
> REDHAT-ROOT

Sure.

> 
> I wonder if anyone can provide a basic root tarbal which I can use to set
> up the basics after which i can start adding stuff.
> 

install.mipsel.nfs is used to create a NFS root tree for mipsel. If
you tar it up, you get what you are asking for.



H.J.

Re: Updates on RedHat 7.1/mips

[Seth Mos]

At 17:36 18-7-2001 -0700, H . J . Lu wrote:
>On Wed, Jul 18, 2001 at 10:53:22PM +0200, Seth Mos wrote:
> >
> > I am having trouble getting the mipsel packages installed using the
> > installer provided for 7.1.
> >
> >
> > I have hooked up a second disk to my i386 workstatition that will house
> > the filesystem for my mipsel box. Once everything is installed I will put
> > this disk back in my Qube2 for testing. In the worst case scenario I will
> > need to boot with the 2.0 kernel that came with the cube.
> >
> > I have the original disk for the Qube2 so it isn't a big problem if it
> > takes 50 attempts. I have a lot of patience ;-)
> >
> > Am I correct to assume that I need to use the install.i386.hd on my
> > workstation to install the mipsel packages on the second harddrive (hdd in
> > my case)
>
>No. It should be install.mipsel.nfs. install.mipsel.hd is for the
>native mipsel machine install.

Thank you, it worked :-)

> > Traceback (innermost, last):
> >   File "./findrpm", line 5, in ?
> >     import rpm
> > ImportError: No module named rpm
> >
> > Does this ring any bells?
>
>You need to install the rpm-python rpm.

And again, thank you.

>install.mipsel.nfs is used to create a NFS root tree for mipsel. If
>you tar it up, you get what you are asking for.

Will do once it works.
Everythinhg seems to be installed like it should and I copied the 2.0.34 
cobalt kernel to /boot on the Qube2. I have enabled the serial console of 
the Qube2 and see the firmware scroll along.

BOOTLOADER ramcode: selected partition /dev/hda1
Decompressing done
Executing bootloader kernel...
Jump_to_Real_Kernel: disk error, trying BFD again.
BOOTLOADER ramcode: selected partition /dev/hdc1

etc.. Qubes don't have the second IDE controller so this is nonsense ;-)
Does anybody know what's missing? Do I need to format the ext2 root fs in a 
special way? Do I need to pass on other bootoptions and if so. How do I get 
to a command line prompt of the firmware?
I see that the NetBSD people did that to make it boot another kernel.

I thought that the firmware from the Qubes just boot the vmlinux.gz image 
that it finds in the /boot directory on hda1?

Any help appreciated.

Cheers


--
Seth
Every program has two purposes one for which
it was written and another for which it wasn't
I use the last kind.

Re: Updates on RedHat 7.1/mips

[Gleb O. Raiko]

[-- Attachment #1: Type: text/plain, Size: 486 bytes --]

Hello,

"H . J . Lu" wrote:
> The next thing on my todo list is to cross compile XFree86 :-(.
> 

It has been done already for Xserver only. I use to cross-compile FBDev
XServer for a long time. My host.def is attached here for reference. In
order to cross-compile entire XFree86 cleanly you need to patch
config/util/Imakefile first, it's primary source of problems during
cross-compilation. It may be just workarounded though, if you don't
pretend to be perfectionist.

Regards,
Gleb.

[-- Attachment #2: host.def --]
[-- Type: image/x-xbitmap, Size: 3198 bytes --]

Re: Updates on RedHat 7.1/mips

[Ralf Baechle]

On Wed, Jul 18, 2001 at 09:13:20AM +0200, Carsten Langgaard wrote:

> It look like there is a cross dependence, the build of tcsh failed with the
> following message:
> 
> /var/tmp/rpm-tmp.7250: /usr/bin/perl: No such file or directory
> error: Bad exit status from /var/tmp/rpm-tmp.7250 (pbuild)
> 
> So tcsh is needed to build perl and perl is needed to build tcsh :-(

One of the reasons why I believe that the major Linux distributions are
fundamentally flawed - their build process doesn't account for such
dependencies.  Point for BSD.

The escape from this circular dependency is to build those packages manually
which may require cheating in the configure and build process a bit.  Or
install them from a binary package (which may require --nodeps and --force
to install).  General rule therefore should be to only do such distribution
package builds on systems which have a maximum installation.

  Ralf

Re: Updates on RedHat 7.1/mips

[Daniel Jacobowitz]

On Wed, Jul 18, 2001 at 03:26:31PM +0200, Ralf Baechle wrote:
> On Wed, Jul 18, 2001 at 09:13:20AM +0200, Carsten Langgaard wrote:
> 
> > It look like there is a cross dependence, the build of tcsh failed with the
> > following message:
> > 
> > /var/tmp/rpm-tmp.7250: /usr/bin/perl: No such file or directory
> > error: Bad exit status from /var/tmp/rpm-tmp.7250 (pbuild)
> > 
> > So tcsh is needed to build perl and perl is needed to build tcsh :-(
> 
> One of the reasons why I believe that the major Linux distributions are
> fundamentally flawed - their build process doesn't account for such
> dependencies.  Point for BSD.

Debian has been going to great lengths to fix this.  We're a great deal
of the way along, too - to the point where bringing up our complete
hppa and ia64 distributions (with the exception of a few
platform-specific tool bugs and some gcc-3.0 incompatibilities) was
really relatively painless for the vast majority of packages.  Only a
tiny subset of packages are assumed installed at build time if they
aren't explicitly listed.

> The escape from this circular dependency is to build those packages manually
> which may require cheating in the configure and build process a bit.  Or
> install them from a binary package (which may require --nodeps and --force
> to install).  General rule therefore should be to only do such distribution
> package builds on systems which have a maximum installation.

Although I'm pretty sure that Debian's Perl package builds without the
use of tcsh... I've no idea where this is coming from.

-- 
Daniel Jacobowitz                           Carnegie Mellon University
MontaVista Software                         Debian GNU/Linux Developer

Re: Updates on RedHat 7.1/mips

[Ralf Baechle]

On Wed, Jul 18, 2001 at 09:48:42AM -0700, Daniel Jacobowitz wrote:

> Although I'm pretty sure that Debian's Perl package builds without the
> use of tcsh... I've no idea where this is coming from.

Anyway, Perl is calling csh internally for something but I forgot what it
was; too long that I researched the details.

  Ralf

Re: Updates on RedHat 7.1/mips

[Florian Lohoff]

On Thu, Jul 19, 2001 at 12:23:22AM +0200, Ralf Baechle wrote:
> On Wed, Jul 18, 2001 at 09:48:42AM -0700, Daniel Jacobowitz wrote:
> 
> > Although I'm pretty sure that Debian's Perl package builds without the
> > use of tcsh... I've no idea where this is coming from.
> 
> Anyway, Perl is calling csh internally for something but I forgot what it
> was; too long that I researched the details.

For path globbing perl uses csh or tcsh 

Flo
-- 
Florian Lohoff                  flo@rfc822.org             +49-5201-669912
     Why is it called "common sense" when nobody seems to have any?