[PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Jeff Dike]

>From Al Viro - this turns the tt mode remapping of the binary into arch 
code.

Signed-off-by: Jeff Dike <jdike@addtoit.com>

Index: linux-2.6.12-rc/arch/um/Makefile
===================================================================
--- linux-2.6.12-rc.orig/arch/um/Makefile	2005-06-02 17:04:01.000000000 -0400
+++ linux-2.6.12-rc/arch/um/Makefile	2005-06-06 11:36:06.000000000 -0400
@@ -119,7 +119,7 @@ ifndef START
   START = $$(($(TOP_ADDR) - $(SIZE)))
 endif
 
-CPPFLAGS_vmlinux.lds = $(shell echo -U$(SUBARCH) \
+CPPFLAGS_vmlinux.lds = $(shell echo -U$(SUBARCH) -I$(SYS_DIR) \
 	-DSTART=$(START) -DELF_ARCH=$(ELF_ARCH) \
 	-DELF_FORMAT=\"$(ELF_FORMAT)\" $(CPP_MODE-y) \
 	-DKERNEL_STACK_SIZE=$(STACK_SIZE))
Index: linux-2.6.12-rc/arch/um/include/sysdep-i386/unmap.lds
===================================================================
--- linux-2.6.12-rc.orig/arch/um/include/sysdep-i386/unmap.lds	2005-06-06 05:54:35.815704808 -0400
+++ linux-2.6.12-rc/arch/um/include/sysdep-i386/unmap.lds	2005-06-02 17:04:44.000000000 -0400
@@ -0,0 +1,14 @@
+  .thread_private : {
+    __start_thread_private = .;
+    errno = .;
+    . += 4;
+    arch/um/sys-i386/unmap_fin.o (.bss)
+    __end_thread_private = .;
+  }
+  . = ALIGN(4096);
+  .remap : { arch/um/sys-i386/unmap_fin.o (.text) }
+
+  /* We want it only if we are in MODE_TT. In both cases, however, when MODE_TT
+   * is off the resulting binary segfaults.*/
+
+  . = ALIGN(4096);		/* Init code and data */
Index: linux-2.6.12-rc/arch/um/include/sysdep-x86_64/unmap.lds
===================================================================
--- linux-2.6.12-rc.orig/arch/um/include/sysdep-x86_64/unmap.lds	2005-06-06 05:54:35.815704808 -0400
+++ linux-2.6.12-rc/arch/um/include/sysdep-x86_64/unmap.lds	2005-06-02 17:04:44.000000000 -0400
@@ -0,0 +1,6 @@
+  . = ALIGN(4096);		/* Init code and data */
+  .remap : {
+	arch/um/sys-x86_64/unmap_fin.o (.bss)
+	arch/um/sys-x86_64/unmap_fin.o (.text)
+  }
+  . = ALIGN(4096);		/* Init code and data */
Index: linux-2.6.12-rc/arch/um/kernel/tt/Makefile
===================================================================
--- linux-2.6.12-rc.orig/arch/um/kernel/tt/Makefile	2005-06-02 17:04:01.000000000 -0400
+++ linux-2.6.12-rc/arch/um/kernel/tt/Makefile	2005-06-06 11:52:11.000000000 -0400
@@ -3,10 +3,6 @@
 # Licensed under the GPL
 #
 
-extra-y := unmap_fin.o
-targets := unmap.o
-clean-files := unmap_tmp.o
-
 obj-y = exec_kern.o exec_user.o gdb.o ksyms.o mem.o mem_user.o process_kern.o \
 	syscall_kern.o syscall_user.o time.o tlb.o tracer.o trap_user.o \
 	uaccess.o uaccess_user.o
@@ -16,14 +12,3 @@ obj-$(CONFIG_PT_PROXY) += gdb_kern.o ptp
 USER_OBJS := gdb.o time.o tracer.o
 
 include arch/um/scripts/Makefile.rules
-
-UNMAP_CFLAGS := $(patsubst -pg -DPROFILING,,$(USER_CFLAGS))
-UNMAP_CFLAGS := $(patsubst -fprofile-arcs -ftest-coverage,,$(UNMAP_CFLAGS))
-
-#XXX: partially copied from arch/um/scripts/Makefile.rules
-$(obj)/unmap.o: c_flags = -Wp,-MD,$(depfile) $(UNMAP_CFLAGS)
-
-$(obj)/unmap_fin.o : $(obj)/unmap.o
-	$(LD) -r -o $(obj)/unmap_tmp.o $< $(shell $(CC) -print-file-name=libc.a)
-	$(OBJCOPY) $(obj)/unmap_tmp.o $@ -G switcheroo
-
Index: linux-2.6.12-rc/arch/um/kernel/tt/unmap.c
===================================================================
--- linux-2.6.12-rc.orig/arch/um/kernel/tt/unmap.c	2005-06-02 17:04:01.000000000 -0400
+++ linux-2.6.12-rc/arch/um/kernel/tt/unmap.c	2005-06-06 05:54:35.815704808 -0400
@@ -1,31 +0,0 @@
-/* 
- * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
- * Licensed under the GPL
- */
-
-#include <sys/mman.h>
-
-int switcheroo(int fd, int prot, void *from, void *to, int size)
-{
-	if(munmap(to, size) < 0){
-		return(-1);
-	}
-	if(mmap(to, size, prot,	MAP_SHARED | MAP_FIXED, fd, 0) != to){
-		return(-1);
-	}
-	if(munmap(from, size) < 0){
-		return(-1);
-	}
-	return(0);
-}
-
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */
Index: linux-2.6.12-rc/arch/um/kernel/uml.lds.S
===================================================================
--- linux-2.6.12-rc.orig/arch/um/kernel/uml.lds.S	2005-06-02 17:04:01.000000000 -0400
+++ linux-2.6.12-rc/arch/um/kernel/uml.lds.S	2005-06-06 11:52:10.000000000 -0400
@@ -15,20 +15,7 @@ SECTIONS
    * is remapped.*/
   __binary_start = .;
 #ifdef MODE_TT
-  .thread_private : {
-    __start_thread_private = .;
-    errno = .;
-    . += 4;
-    arch/um/kernel/tt/unmap_fin.o (.data)
-    __end_thread_private = .;
-  }
-  . = ALIGN(4096);
-  .remap : { arch/um/kernel/tt/unmap_fin.o (.text) }
-
-  /* We want it only if we are in MODE_TT. In both cases, however, when MODE_TT
-   * is off the resulting binary segfaults.*/
-
-  . = ALIGN(4096);		/* Init code and data */
+#include <unmap.lds>
 #endif
 
   _stext = .;
Index: linux-2.6.12-rc/arch/um/scripts/Makefile.rules
===================================================================
--- linux-2.6.12-rc.orig/arch/um/scripts/Makefile.rules	2005-06-06 12:14:26.000000000 -0400
+++ linux-2.6.12-rc/arch/um/scripts/Makefile.rules	2005-06-06 12:15:45.000000000 -0400
@@ -10,6 +10,11 @@ USER_OBJS := $(foreach file,$(USER_OBJS)
 $(USER_OBJS) : c_flags = -Wp,-MD,$(depfile) $(USER_CFLAGS) \
 	$(CFLAGS_$(notdir $@))
 
+# The stubs and unmap.o can't try to call mcount or update basic block data
+define unprofile
+	$(patsubst -pg,,$(patsubst -fprofile-arcs -ftest-coverage,,$(1)))
+endef
+
 quiet_cmd_make_link = SYMLINK $@
 cmd_make_link       = ln -sf $(srctree)/arch/$(SUBARCH)/$($(notdir $@)-dir)/$(notdir $@) $@
 
Index: linux-2.6.12-rc/arch/um/sys-i386/Makefile
===================================================================
--- linux-2.6.12-rc.orig/arch/um/sys-i386/Makefile	2005-06-02 17:04:01.000000000 -0400
+++ linux-2.6.12-rc/arch/um/sys-i386/Makefile	2005-06-06 11:52:10.000000000 -0400
@@ -2,6 +2,8 @@ obj-y = bitops.o bugs.o checksum.o delay
 	ptrace_user.o semaphore.o signal.o sigcontext.o syscalls.o sysrq.o \
 	sys_call_table.o
 
+extra-$(CONFIG_MODE_TT) := unmap_fin.o
+
 obj-$(CONFIG_HIGHMEM) += highmem.o
 obj-$(CONFIG_MODULES) += module.o
 
@@ -16,4 +18,6 @@ semaphore.c-dir = kernel
 highmem.c-dir = mm
 module.c-dir = kernel
 
+$(obj)/unmap_fin.o : _c_flags = $(call unprofile,$(CFLAGS))
+
 subdir- := util
Index: linux-2.6.12-rc/arch/um/sys-i386/unmap_fin.c
===================================================================
--- linux-2.6.12-rc.orig/arch/um/sys-i386/unmap_fin.c	2005-06-06 05:54:35.815704808 -0400
+++ linux-2.6.12-rc/arch/um/sys-i386/unmap_fin.c	2005-06-03 15:36:17.000000000 -0400
@@ -0,0 +1,25 @@
+/* 
+ * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
+ * Licensed under the GPL
+ */
+
+#include <linux/mman.h>
+#include <asm/unistd.h>
+
+static int errno;
+
+static inline _syscall2(int,munmap,void *,start,size_t,len)
+static inline _syscall6(void *,mmap2,void *,addr,size_t,len,int,prot,int,flags,int,fd,off_t,offset)
+int switcheroo(int fd, int prot, void *from, void *to, int size)
+{
+        if(munmap(to, size) < 0){
+                return(-1);
+        }
+        if(mmap2(to, size, prot, MAP_SHARED | MAP_FIXED, fd, 0) != to){
+                return(-1);
+        }
+        if(munmap(from, size) < 0){
+                return(-1);
+        }
+        return(0);
+}
Index: linux-2.6.12-rc/arch/um/sys-x86_64/Makefile
===================================================================
--- linux-2.6.12-rc.orig/arch/um/sys-x86_64/Makefile	2005-06-02 17:04:01.000000000 -0400
+++ linux-2.6.12-rc/arch/um/sys-x86_64/Makefile	2005-06-06 11:52:10.000000000 -0400
@@ -4,6 +4,8 @@
 # Licensed under the GPL
 #
 
+extra-$(CONFIG_MODE_TT) := unmap_fin.o
+
 #XXX: why into lib-y?
 lib-y = bitops.o bugs.o csum-partial.o delay.o fault.o mem.o memcpy.o \
 	ptrace.o ptrace_user.o semaphore.o sigcontext.o signal.o \
@@ -29,3 +31,5 @@ thunk.S-dir = lib
 module.c-dir = kernel
 
 subdir- := util
+
+$(obj)/unmap_fin.o: _c_flags = $(call unprofile,$(CFLAGS))
Index: linux-2.6.12-rc/arch/um/sys-x86_64/unmap_fin.c
===================================================================
--- linux-2.6.12-rc.orig/arch/um/sys-x86_64/unmap_fin.c	2005-06-06 05:54:35.815704808 -0400
+++ linux-2.6.12-rc/arch/um/sys-x86_64/unmap_fin.c	2005-06-02 17:04:44.000000000 -0400
@@ -0,0 +1,25 @@
+/* 
+ * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
+ * Licensed under the GPL
+ */
+
+#include <linux/mman.h>
+#include <asm/unistd.h>
+
+static int errno;
+
+static inline _syscall2(int,munmap,void *,start,size_t,len)
+static inline _syscall6(void *,mmap,void *,addr,size_t,len,int,prot,int,flags,int,fd,off_t,offset)
+int switcheroo(int fd, int prot, void *from, void *to, int size)
+{
+        if(munmap(to, size) < 0){
+                return(-1);
+        }
+        if(mmap(to, size, prot, MAP_SHARED | MAP_FIXED, fd, 0) != to){
+                return(-1);
+        }
+        if(munmap(from, size) < 0){
+                return(-1);
+        }
+        return(0);
+}

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Blaisorblade]

On Monday 06 June 2005 22:08, Jeff Dike wrote:
> From Al Viro - this turns the tt mode remapping of the binary into arch
> code.
NACK at all, definitely, don't apply this one please. This patch:

1) On i386 does not fix the problem it was supposed to fix when I originately 
sent the first version (i.e. avoiding to create a .thread_private section to 
allow linking against NPTL glibc). It's done on x86_64 and forgot on i386.
2) Splitting the linker script for subarchs is definitely not needed.
3) This removes the fix (done through objcopy -G switcheroo) to a link time 
conflict happening on some weird glibc combinations.

I'll merge this work when it's ready.
> Signed-off-by: Jeff Dike <jdike@addtoit.com>
-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

		
___________________________________ 
Yahoo! Messenger: chiamate gratuite in tutto il mondo 
http://it.beta.messenger.yahoo.com

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Al Viro]

On Tue, Jun 07, 2005 at 01:05:19AM +0200, Blaisorblade wrote:
> On Monday 06 June 2005 22:08, Jeff Dike wrote:
> > From Al Viro - this turns the tt mode remapping of the binary into arch
> > code.
> NACK at all, definitely, don't apply this one please. This patch:
> 
> 1) On i386 does not fix the problem it was supposed to fix when I originately 
> sent the first version (i.e. avoiding to create a .thread_private section to 
> allow linking against NPTL glibc). It's done on x86_64 and forgot on i386.

True.  i386 still assumes non-NPTL (as it is on the box I'm working on -
such setups *do* exist).

> 2) Splitting the linker script for subarchs is definitely not needed.

Per-subarch - perhaps not.  Per-glibc-type - definitely needed.

> 3) This removes the fix (done through objcopy -G switcheroo) to a link time 
> conflict happening on some weird glibc combinations.

*What* link-time conflict?  We don't link libc into switcheroo anymore.

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Blaisorblade]

On Tuesday 07 June 2005 01:53, Al Viro wrote:
> On Tue, Jun 07, 2005 at 01:05:19AM +0200, Blaisorblade wrote:
> > On Monday 06 June 2005 22:08, Jeff Dike wrote:
> > > From Al Viro - this turns the tt mode remapping of the binary into arch
> > > code.
> >
> > NACK at all, definitely, don't apply this one please. This patch:
> >
> > 1) On i386 does not fix the problem it was supposed to fix when I
> > originately sent the first version (i.e. avoiding to create a
> > .thread_private section to allow linking against NPTL glibc). It's done
> > on x86_64 and forgot on i386.
>
> True.  i386 still assumes non-NPTL (as it is on the box I'm working on -
> such setups *do* exist).
Yes, it's the most common one, and it's even the setup for my box currently.
> > 2) Splitting the linker script for subarchs is definitely not needed.
>
> Per-subarch - perhaps not.  Per-glibc-type - definitely needed.
No, because the setup for NPTL glibc works also on non-NPTL one. Actually, to 
be exact, I've tested it *only* on normal glibc. I'm still waiting to get 
some testing in NPTL environments, but I expect it to work.
> > 3) This removes the fix (done through objcopy -G switcheroo) to a link
> > time conflict happening on some weird glibc combinations.

> *What* link-time conflict?  We don't link libc into switcheroo anymore.
Hmm, yes, I noted this... and maybe it could even be good (it makes sense, at 
least). Probably you are right on this, too.

P.S: is it only me or you've sent about 20 copies of your last message?
-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

	

	
		
___________________________________ 
Yahoo! Mail: gratis 1GB per i messaggi e allegati da 10MB 
http://mail.yahoo.it

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Al Viro]

On Tue, Jun 07, 2005 at 02:56:36AM +0200, Blaisorblade wrote:
> > Per-subarch - perhaps not.  Per-glibc-type - definitely needed.
> No, because the setup for NPTL glibc works also on non-NPTL one. Actually, to 
> be exact, I've tested it *only* on normal glibc. I'm still waiting to get 
> some testing in NPTL environments, but I expect it to work.

Now, that is interesting.  Which script are you using for i386 and which
libc version does it work with?

> P.S: is it only me or you've sent about 20 copies of your last message?

Headers?

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Blaisorblade]

[-- Attachment #1: Type: text/plain, Size: 1266 bytes --]

On Tuesday 07 June 2005 02:59, Al Viro wrote:
> On Tue, Jun 07, 2005 at 02:56:36AM +0200, Blaisorblade wrote:
> > > Per-subarch - perhaps not.  Per-glibc-type - definitely needed.
> >
> > No, because the setup for NPTL glibc works also on non-NPTL one.
> > Actually, to be exact, I've tested it *only* on normal glibc. I'm still
> > waiting to get some testing in NPTL environments, but I expect it to
> > work.
>
> Now, that is interesting.  Which script are you using for i386 and which
> libc version does it work with?
The one I sent originally didn't work because I had been lazy (used .data 
instead of .bss), but only for that. Well, there were another zillion of 
problems, but now it works.

> > P.S: is it only me or you've sent about 20 copies of your last message?
>
> Headers?
The Message-ID seem to match, so guess it's my Kmail (never seen such a 
problem, though).

Message-ID: <20050606235321.GJ29811@parcelfarce.linux.theplanet.co.uk>
References: <200506062008.j56K89YA008957@ccure.user-mode-linux.org> 
<200506070105.20422.blaisorblade@yahoo.it>

-- 
Inform me of my mistakes, so I can keep imitating Homer Simpson's "Doh!".
Paolo Giarrusso, aka Blaisorblade (Skype ID "PaoloGiarrusso", ICQ 215621894)
http://www.user-mode-linux.org/~blaisorblade

[-- Attachment #2: uml-link-tt-mode-against-nptl.patch --]
[-- Type: text/x-diff, Size: 9200 bytes --]


From: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>, Al Viro <viro@parcelfarce.linux.theplanet.co.uk>

To make sure switcheroo() can execute when we remap all the executable image,
we used a trick to make it use a local copy of errno... this trick does not
work with NPTL glibc, only with LinuxThreads, so use another (simpler) one to
make it work anyway.

Hopefully, a lot improved thanks to merging with the version of Al Viro (which
had his part of problems, though, i.e. removing a fix to another bug and not
fixing the problem on i386).

Signed-off-by: Paolo 'Blaisorblade' Giarrusso <blaisorblade@yahoo.it>
---

 linux-2.6.git-paolo/arch/um/Makefile               |    2 -
 linux-2.6.git-paolo/arch/um/kernel/tt/Makefile     |   15 ----------
 linux-2.6.git-paolo/arch/um/kernel/uml.lds.S       |   15 ++--------
 linux-2.6.git-paolo/arch/um/scripts/Makefile.rules |    6 ++++
 linux-2.6.git-paolo/arch/um/scripts/Makefile.unmap |   22 ++++++++++++++
 linux-2.6.git-paolo/arch/um/sys-i386/Makefile      |    2 +
 linux-2.6.git-paolo/arch/um/sys-i386/unmap.c       |   25 ++++++++++++++++
 linux-2.6.git-paolo/arch/um/sys-x86_64/Makefile    |    2 +
 linux-2.6.git-paolo/arch/um/sys-x86_64/unmap.c     |   25 ++++++++++++++++
 linux-2.6.git/arch/um/kernel/tt/unmap.c            |   31 ---------------------
 10 files changed, 86 insertions(+), 59 deletions(-)

diff -L arch/um/kernel/tt/unmap.c -puN arch/um/kernel/tt/unmap.c~uml-link-tt-mode-against-nptl /dev/null
--- linux-2.6.git/arch/um/kernel/tt/unmap.c
+++ /dev/null	2005-03-04 21:12:55.000000000 +0100
@@ -1,31 +0,0 @@
-/* 
- * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
- * Licensed under the GPL
- */
-
-#include <sys/mman.h>
-
-int switcheroo(int fd, int prot, void *from, void *to, int size)
-{
-	if(munmap(to, size) < 0){
-		return(-1);
-	}
-	if(mmap(to, size, prot,	MAP_SHARED | MAP_FIXED, fd, 0) != to){
-		return(-1);
-	}
-	if(munmap(from, size) < 0){
-		return(-1);
-	}
-	return(0);
-}
-
-/*
- * Overrides for Emacs so that we follow Linus's tabbing style.
- * Emacs will notice this stuff at the end of the file and automatically
- * adjust the settings for this buffer only.  This must remain at the end
- * of the file.
- * ---------------------------------------------------------------------------
- * Local variables:
- * c-file-style: "linux"
- * End:
- */
diff -puN arch/um/kernel/uml.lds.S~uml-link-tt-mode-against-nptl arch/um/kernel/uml.lds.S
--- linux-2.6.git/arch/um/kernel/uml.lds.S~uml-link-tt-mode-against-nptl	2005-06-07 02:48:03.000000000 +0200
+++ linux-2.6.git-paolo/arch/um/kernel/uml.lds.S	2005-06-07 02:48:03.000000000 +0200
@@ -14,19 +14,10 @@ SECTIONS
   /* Used in arch/um/kernel/mem.c. Any memory between START and __binary_start
    * is remapped.*/
   __binary_start = .;
-#ifdef MODE_TT
-  .thread_private : {
-    __start_thread_private = .;
-    errno = .;
-    . += 4;
-    arch/um/kernel/tt/unmap_fin.o (.data)
-    __end_thread_private = .;
-  }
-  . = ALIGN(4096);
-  .remap : { arch/um/kernel/tt/unmap_fin.o (.text) }
 
-  /* We want it only if we are in MODE_TT. In both cases, however, when MODE_TT
-   * is off the resulting binary segfaults.*/
+#ifdef MODE_TT
+  .remap_data : { arch/um/sys-SUBARCH/unmap_fin.o (.data .bss) }
+  .remap : { arch/um/sys-SUBARCH/unmap_fin.o (.text) }
 
   . = ALIGN(4096);		/* Init code and data */
 #endif
diff -puN /dev/null arch/um/sys-i386/unmap.c
--- /dev/null	2005-03-04 21:12:55.000000000 +0100
+++ linux-2.6.git-paolo/arch/um/sys-i386/unmap.c	2005-06-07 02:48:03.000000000 +0200
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
+ * Licensed under the GPL
+ */
+
+#include <linux/mman.h>
+#include <asm/unistd.h>
+
+static int errno;
+
+static inline _syscall2(int,munmap,void *,start,size_t,len)
+static inline _syscall6(void *,mmap2,void *,addr,size_t,len,int,prot,int,flags,int,fd,off_t,offset)
+int switcheroo(int fd, int prot, void *from, void *to, int size)
+{
+	if(munmap(to, size) < 0){
+		return(-1);
+	}
+	if(mmap2(to, size, prot, MAP_SHARED | MAP_FIXED, fd, 0) != to){
+		return(-1);
+	}
+	if(munmap(from, size) < 0){
+		return(-1);
+	}
+	return(0);
+}
diff -puN /dev/null arch/um/sys-x86_64/unmap.c
--- /dev/null	2005-03-04 21:12:55.000000000 +0100
+++ linux-2.6.git-paolo/arch/um/sys-x86_64/unmap.c	2005-06-07 02:48:03.000000000 +0200
@@ -0,0 +1,25 @@
+/*
+ * Copyright (C) 2000 Jeff Dike (jdike@karaya.com)
+ * Licensed under the GPL
+ */
+
+#include <linux/mman.h>
+#include <asm/unistd.h>
+
+static int errno;
+
+static inline _syscall2(int,munmap,void *,start,size_t,len)
+static inline _syscall6(void *,mmap,void *,addr,size_t,len,int,prot,int,flags,int,fd,off_t,offset)
+int switcheroo(int fd, int prot, void *from, void *to, int size)
+{
+	if(munmap(to, size) < 0){
+		return(-1);
+	}
+	if(mmap(to, size, prot, MAP_SHARED | MAP_FIXED, fd, 0) != to){
+		return(-1);
+	}
+	if(munmap(from, size) < 0){
+		return(-1);
+	}
+	return(0);
+}
diff -puN arch/um/kernel/tt/Makefile~uml-link-tt-mode-against-nptl arch/um/kernel/tt/Makefile
--- linux-2.6.git/arch/um/kernel/tt/Makefile~uml-link-tt-mode-against-nptl	2005-06-07 02:48:03.000000000 +0200
+++ linux-2.6.git-paolo/arch/um/kernel/tt/Makefile	2005-06-07 02:48:03.000000000 +0200
@@ -3,10 +3,6 @@
 # Licensed under the GPL
 #
 
-extra-y := unmap_fin.o
-targets := unmap.o
-clean-files := unmap_tmp.o
-
 obj-y = exec_kern.o exec_user.o gdb.o ksyms.o mem.o mem_user.o process_kern.o \
 	syscall_kern.o syscall_user.o time.o tlb.o tracer.o trap_user.o \
 	uaccess.o uaccess_user.o
@@ -16,14 +12,3 @@ obj-$(CONFIG_PT_PROXY) += gdb_kern.o ptp
 USER_OBJS := gdb.o time.o tracer.o
 
 include arch/um/scripts/Makefile.rules
-
-UNMAP_CFLAGS := $(patsubst -pg -DPROFILING,,$(USER_CFLAGS))
-UNMAP_CFLAGS := $(patsubst -fprofile-arcs -ftest-coverage,,$(UNMAP_CFLAGS))
-
-#XXX: partially copied from arch/um/scripts/Makefile.rules
-$(obj)/unmap.o: c_flags = -Wp,-MD,$(depfile) $(UNMAP_CFLAGS)
-
-$(obj)/unmap_fin.o : $(obj)/unmap.o
-	$(LD) -r -o $(obj)/unmap_tmp.o $< $(shell $(CC) -print-file-name=libc.a)
-	$(OBJCOPY) $(obj)/unmap_tmp.o $@ -G switcheroo
-
diff -puN arch/um/sys-i386/Makefile~uml-link-tt-mode-against-nptl arch/um/sys-i386/Makefile
--- linux-2.6.git/arch/um/sys-i386/Makefile~uml-link-tt-mode-against-nptl	2005-06-07 02:48:03.000000000 +0200
+++ linux-2.6.git-paolo/arch/um/sys-i386/Makefile	2005-06-07 02:48:03.000000000 +0200
@@ -17,3 +17,5 @@ highmem.c-dir = mm
 module.c-dir = kernel
 
 subdir- := util
+
+include arch/um/scripts/Makefile.unmap
diff -puN arch/um/sys-x86_64/Makefile~uml-link-tt-mode-against-nptl arch/um/sys-x86_64/Makefile
--- linux-2.6.git/arch/um/sys-x86_64/Makefile~uml-link-tt-mode-against-nptl	2005-06-07 02:48:03.000000000 +0200
+++ linux-2.6.git-paolo/arch/um/sys-x86_64/Makefile	2005-06-07 02:48:03.000000000 +0200
@@ -29,3 +29,5 @@ thunk.S-dir = lib
 module.c-dir = kernel
 
 subdir- := util
+
+include arch/um/scripts/Makefile.unmap
diff -puN /dev/null arch/um/scripts/Makefile.unmap
--- /dev/null	2005-03-04 21:12:55.000000000 +0100
+++ linux-2.6.git-paolo/arch/um/scripts/Makefile.unmap	2005-06-07 03:05:56.000000000 +0200
@@ -0,0 +1,22 @@
+clean-files += unmap_tmp.o unmap_fin.o unmap.o
+
+ifdef CONFIG_MODE_TT
+
+#Always build unmap_fin.o
+extra-y += unmap_fin.o
+#Do dependency tracking for unmap.o (it will be always built, but won't get the tracking unless we use this).
+targets += unmap.o
+
+#XXX: partially copied from arch/um/scripts/Makefile.rules
+$(obj)/unmap.o: _c_flags = $(call unprofile,$(CFLAGS))
+
+quiet_cmd_wrapld = LD      $@
+define cmd_wrapld
+	$(LD) -r -o $(obj)/unmap_tmp.o $< $(shell $(CC) -print-file-name=libc.a); \
+	$(OBJCOPY) $(obj)/unmap_tmp.o $@ -G switcheroo
+endef
+
+$(obj)/unmap_fin.o : $(obj)/unmap.o FORCE
+	$(call if_changed,wrapld)
+
+endif
diff -puN arch/um/Makefile~uml-link-tt-mode-against-nptl arch/um/Makefile
--- linux-2.6.git/arch/um/Makefile~uml-link-tt-mode-against-nptl	2005-06-07 02:48:03.000000000 +0200
+++ linux-2.6.git-paolo/arch/um/Makefile	2005-06-07 02:48:03.000000000 +0200
@@ -118,7 +118,7 @@ endif
 CPPFLAGS_vmlinux.lds = $(shell echo -U$(SUBARCH) \
 	-DSTART=$(START) -DELF_ARCH=$(ELF_ARCH) \
 	-DELF_FORMAT=\"$(ELF_FORMAT)\" $(CPP_MODE-y) \
-	-DKERNEL_STACK_SIZE=$(STACK_SIZE))
+	-DKERNEL_STACK_SIZE=$(STACK_SIZE) -DSUBARCH=$(SUBARCH))
 
 #The wrappers will select whether using "malloc" or the kernel allocator.
 LINK_WRAPS = -Wl,--wrap,malloc -Wl,--wrap,free -Wl,--wrap,calloc
diff -puN arch/um/scripts/Makefile.rules~uml-link-tt-mode-against-nptl arch/um/scripts/Makefile.rules
--- linux-2.6.git/arch/um/scripts/Makefile.rules~uml-link-tt-mode-against-nptl	2005-06-07 03:03:42.000000000 +0200
+++ linux-2.6.git-paolo/arch/um/scripts/Makefile.rules	2005-06-07 03:08:55.000000000 +0200
@@ -10,6 +10,12 @@ USER_OBJS := $(foreach file,$(USER_OBJS)
 $(USER_OBJS) : c_flags = -Wp,-MD,$(depfile) $(USER_CFLAGS) \
 	$(CFLAGS_$(notdir $@))
 
+# The stubs and unmap.o can't try to call mcount or update basic block data
+define unprofile
+	$(patsubst -pg,,$(patsubst -fprofile-arcs -ftest-coverage,,$(1)))
+endef
+
+
 quiet_cmd_make_link = SYMLINK $@
 cmd_make_link       = ln -sf $(srctree)/arch/$(SUBARCH)/$($(notdir $@)-dir)/$(notdir $@) $@
 
_

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Rob Landley]

On Monday 06 June 2005 09:18 pm, Blaisorblade wrote:
> > > P.S: is it only me or you've sent about 20 copies of your last message?
> >
> > Headers?
>
> The Message-ID seem to match, so guess it's my Kmail (never seen such a
> problem, though).
>
> Message-ID: <20050606235321.GJ29811@parcelfarce.linux.theplanet.co.uk>
> References: <200506062008.j56K89YA008957@ccure.user-mode-linux.org>
> <200506070105.20422.blaisorblade@yahoo.it>

I got about 20 copies of that message too.  There's some kind of loop going 
through Intel's servers, the message headers are progressively longer 
variants of:

Return-Path: <user-mode-linux-devel-admin@lists.sourceforge.net>
 Received: from lists-outbound.sourceforge.net (lists-outbound.sourceforge.net 
[66.35.250.225])
        by grelber.thyrsus.com (8.13.1/8.13.1) with ESMTP id j5727DMt031684
        for <rob@landley.net>; Mon, 6 Jun 2005 22:07:13 -0400
 Received: from projects.sourceforge.net (sc8-sf-list1-b.sourceforge.net 
[10.3.1.7])
        by sc8-sf-spam2.sourceforge.net (Postfix) with ESMTP
        id 2E0C5E4AF; Mon,  6 Jun 2005 17:46:15 -0700 (PDT)
 Received: from sc8-sf-mx2-b.sourceforge.net ([10.3.1.12] 
helo=sc8-sf-mx2.sourceforge.net)
        by sc8-sf-list1.sourceforge.net with esmtp (Exim 4.30)
        id 1DfSDQ-0007UB-Rp
        for user-mode-linux-devel@lists.sourceforge.net; Mon, 06 Jun 2005 
17:45:12 -0700
 Received: from fmr19.intel.com ([134.134.136.18] helo=orsfmr004.jf.intel.com)
        by sc8-sf-mx2.sourceforge.net with esmtp (Exim 4.41)
        id 1DfSDO-0003gl-Il
        for user-mode-linux-devel@lists.sourceforge.net; Mon, 06 Jun 2005 
17:45:12 -0700
 Received: from orsfmr101.jf.intel.com (orsfmr101.jf.intel.com [10.7.209.17])
        by orsfmr004.jf.intel.com (8.12.10/8.12.10/d: major-outer.mc,v 1.1 
2004/09/17 17:50:56 root Exp $) with ESMTP id j570j4TU027763;
        Tue, 7 Jun 2005 00:45:04 GMT
 Received: from orsmsxvs041.jf.intel.com (orsmsxvs041.jf.intel.com 
[192.168.65.54])
        by orsfmr101.jf.intel.com (8.12.10/8.12.10/d: major-inner.mc,v 1.2 
2004/09/17 18:05:01 root Exp $) with SMTP id j570dvQF011113;
        Tue, 7 Jun 2005 00:44:50 GMT
 Received: from orsmsx332.amr.corp.intel.com ([192.168.65.60])
 by orsmsxvs041.jf.intel.com (SAVSMTP 3.1.7.47) with SMTP id 
M2005060617432614357
 ; Mon, 06 Jun 2005 17:43:35 -0700
 Received: from mail pickup service by orsmsx332.amr.corp.intel.com with 
Microsoft SMTPSVC;
         Mon, 6 Jun 2005 17:38:55 -0700
 Received: from orsmsxvs040.jf.intel.com ([192.168.65.206]) by 
orsmsx332.amr.corp.intel.com with Microsoft SMTPSVC(6.0.3790.211);
         Mon, 6 Jun 2005 17:11:03 -0700
 Received: from orsfmr100.jf.intel.com ([10.7.209.16])
 by orsmsxvs040.jf.intel.com (SAVSMTP 3.1.7.47) with SMTP id 
M2005060617110225137
 for <tom.l.nguyen@intel.com>; Mon, 06 Jun 2005 17:11:02 -0700
 Received: from orsfmr004.jf.intel.com (orsfmr004.jf.intel.com [10.7.208.20])
        by orsfmr100.jf.intel.com (8.12.10/8.12.10/d: major-inner.mc,v 1.2 
2004/09/17 18:05:01 root Exp $) with ESMTP id j570B2qr012945
        for <tom.l.nguyen@intel.com>; Tue, 7 Jun 2005 00:11:02 GMT
 Received: from vger.kernel.org (vger.kernel.org [12.107.209.244])
        by orsfmr004.jf.intel.com (8.12.10/8.12.10/d: major-outer.mc,v 1.1 
2004/09/17 17:50:56 root Exp $) with ESMTP id j570AlTr014175
        for <tom.l.nguyen@intel.com>; Tue, 7 Jun 2005 00:11:02 GMT
 Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
        id S261787AbVFGABn (ORCPT <rfc822;tom.l.nguyen@intel.com>);
        Mon, 6 Jun 2005 20:01:43 -0400
 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261788AbVFFX5w
        (ORCPT <rfc822;linux-kernel-outgoing>);
        Mon, 6 Jun 2005 19:57:52 -0400
 Received: from parcelfarce.linux.theplanet.co.uk ([195.92.249.252]:56209 
"EHLO
        parcelfarce.linux.theplanet.co.uk") by vger.kernel.org with ESMTP
        id S261778AbVFFXwe (ORCPT <rfc822;linux-kernel@vger.kernel.org>);
        Mon, 6 Jun 2005 19:52:34 -0400
 Received: from viro by parcelfarce.linux.theplanet.co.uk with local (Exim 
4.43)
        id 1DfRPF-0002v4-Sq; Tue, 07 Jun 2005 00:53:21 +0100
 From: Al Viro <viro@parcelfarce.linux.theplanet.co.uk>


Rob

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Jeff Dike]

On Tue, Jun 07, 2005 at 01:59:58AM +0100, Al Viro wrote:
> > P.S: is it only me or you've sent about 20 copies of your last message?
> 
> Headers?

I got 18 copies.

Something horrible happened at Intel:

Received: from orsfmr005.jf.intel.com (fmr20.intel.com [134.134.136.19])
        by saraswathi.solana.com (8.13.1/8.13.1) with ESMTP id j570cDAU017960 
        for <jdike@addtoit.com>; Mon, 6 Jun 2005 20:38:18 -0400
Received: from orsfmr101.jf.intel.com (orsfmr101.jf.intel.com [10.7.209.17])
        by orsfmr005.jf.intel.com (8.12.10/8.12.10/d: major-outer.mc,v 1.1
+2004/09/17 17:50:56 root Exp $) with ESMTP id j570c60S009128;
        Tue, 7 Jun 2005 00:38:06 GMT
Received: from orsmsxvs041.jf.intel.com (orsmsxvs041.jf.intel.com
+[192.168.65.54])
        by orsfmr101.jf.intel.com (8.12.10/8.12.10/d: major-inner.mc,v 1.2
+2004/09/17 18:05:01 root Exp $) with SMTP id j570c6OK009846;
        Tue, 7 Jun 2005 00:38:06 GMT
Received: from orsmsx332.amr.corp.intel.com ([192.168.65.60])
 by orsmsxvs041.jf.intel.com (SAVSMTP 3.1.7.47) with SMTP id
+M2005060617380513306
 ; Mon, 06 Jun 2005 17:38:05 -0700
Received: from mail pickup service by orsmsx332.amr.corp.intel.com with
+Microsoft SMTPSVC;
         Mon, 6 Jun 2005 17:38:04 -0700
Received: from orsmsxvs041.jf.intel.com ([192.168.65.54]) by
+orsmsx332.amr.corp.intel.com with Microsoft SMTPSVC(6.0.3790.211);
         Mon, 6 Jun 2005 17:11:02 -0700
Received: from orsfmr100.jf.intel.com ([10.7.209.16])
 by orsmsxvs041.jf.intel.com (SAVSMTP 3.1.7.47) with SMTP id
+M2005060617110210382
 for <suresh.b.siddha@intel.com>; Mon, 06 Jun 2005 17:11:02 -0700
Received: from orsfmr004.jf.intel.com (orsfmr004.jf.intel.com [10.7.208.20])
        by orsfmr100.jf.intel.com (8.12.10/8.12.10/d: major-inner.mc,v 1.2
+2004/09/17 18:05:01 root Exp $) with ESMTP id j570B2qr012933
        for <suresh.b.siddha@intel.com>; Tue, 7 Jun 2005 00:11:02 GMT
Received: from vger.kernel.org (vger.kernel.org [12.107.209.244])
        by orsfmr004.jf.intel.com (8.12.10/8.12.10/d: major-outer.mc,v 1.1
+2004/09/17 17:50:56 root Exp $) with ESMTP id j570AlTq014175
        for <suresh.b.siddha@intel.com>; Tue, 7 Jun 2005 00:11:01 GMT

This looks like it was intended for Suresh (hi!), but Intel forwarded it
back to my server.

The others are the same, except they involve different Intel people, some
of whom I know, some I don't :-)

				Jeff

Re: [uml-devel] [PATCH 3/5] UML - Clean up tt mode remapping of UML binary

[Jeff Dike]

On Tue, Jun 07, 2005 at 01:05:19AM +0200, Blaisorblade wrote:
> NACK at all, definitely, don't apply this one please. This patch:
> 
> 1) On i386 does not fix the problem it was supposed to fix when I originately 
> sent the first version (i.e. avoiding to create a .thread_private section to 
> allow linking against NPTL glibc). It's done on x86_64 and forgot on i386.
> 2) Splitting the linker script for subarchs is definitely not needed.
> 3) This removes the fix (done through objcopy -G switcheroo) to a link time 
> conflict happening on some weird glibc combinations.
> 
> I'll merge this work when it's ready.

OK, Andrew, just drop this one for now.

				Jeff