undefined reference to `copy_siginfo_from_user32'

undefined reference to `copy_siginfo_from_user32'

[Giuseppe Sacco]

Hi list,
since a few days, whenever I try to recompile the latest kernel (from git) it always print this error message:

[...]
  AS      arch/mips/lib/strlen_user.o
  AS      arch/mips/lib/strncpy_user.o
  AS      arch/mips/lib/strnlen_user.o
  CC      arch/mips/lib/uncached.o
  AR      arch/mips/lib/lib.a
  LD      vmlinux.o
  MODPOST vmlinux.o
WARNING: modpost: Found 11 section mismatch(es).
To see full details build your kernel with:
'make CONFIG_DEBUG_SECTION_MISMATCH=y'
  GEN     .version
  CHK     include/linux/compile.h
  UPD     include/linux/compile.h
  CC      init/version.o
  LD      init/built-in.o
  LD      .tmp_vmlinux1
kernel/built-in.o: In function `$L261':
ptrace.c:(.text+0x13550): undefined reference to `copy_siginfo_from_user32'
ptrace.c:(.text+0x13550): relocation truncated to fit: R_MIPS_26 against `copy_siginfo_from_user32'
make[1]: *** [.tmp_vmlinux1] Error 1
make[1]: Leaving directory `/usr/local/src/kernel/2.6.25'
make: *** [debian/stamp-build-kernel] Error 2

Any idea about a possible cause/solution?

Thank you very mcuh,
Giuseppe

Re: undefined reference to `copy_siginfo_from_user32'

[Christoph Hellwig]

On Mon, Apr 28, 2008 at 09:23:27PM +0200, Giuseppe Sacco wrote:
> Hi list,
> since a few days, whenever I try to recompile the latest kernel (from git) it always print this error message:

This should be fixed in mainline.  But the right fix would be to switch
mips to the generic compat_ptrace.  And untested (and in fact even
uncompiled) patch ontop of the copy_siginfo_to_user32 posted to the list
a while ago is below to sketch how this should look like:


Index: linux-2.6/arch/mips/kernel/ptrace32.c
===================================================================
--- linux-2.6.orig/arch/mips/kernel/ptrace32.c	2008-04-29 10:54:44.000000000 +0200
+++ linux-2.6/arch/mips/kernel/ptrace32.c	2008-04-29 10:59:17.000000000 +0200
@@ -42,56 +42,17 @@ int ptrace_setregs(struct task_struct *c
 int ptrace_getfpregs(struct task_struct *child, __u32 __user *data);
 int ptrace_setfpregs(struct task_struct *child, __u32 __user *data);
 
+
 /*
  * Tracing a 32-bit process with a 64-bit strace and vice versa will not
  * work.  I don't know how to fix this.
  */
-asmlinkage int sys32_ptrace(int request, int pid, int addr, int data)
+long compat_arch_ptrace(struct task_struct *child, compat_long_t request,
+			compat_ulong_t addr, compat_ulong_t data)
 {
-	struct task_struct *child;
 	int ret;
 
-#if 0
-	printk("ptrace(r=%d,pid=%d,addr=%08lx,data=%08lx)\n",
-	       (int) request, (int) pid, (unsigned long) addr,
-	       (unsigned long) data);
-#endif
-	lock_kernel();
-	if (request == PTRACE_TRACEME) {
-		ret = ptrace_traceme();
-		goto out;
-	}
-
-	child = ptrace_get_task_struct(pid);
-	if (IS_ERR(child)) {
-		ret = PTR_ERR(child);
-		goto out;
-	}
-
-	if (request == PTRACE_ATTACH) {
-		ret = ptrace_attach(child);
-		goto out_tsk;
-	}
-
-	ret = ptrace_check_attach(child, request == PTRACE_KILL);
-	if (ret < 0)
-		goto out_tsk;
-
 	switch (request) {
-	/* when I and D space are separate, these will need to be fixed. */
-	case PTRACE_PEEKTEXT: /* read word at location addr. */
-	case PTRACE_PEEKDATA: {
-		unsigned int tmp;
-		int copied;
-
-		copied = access_process_vm(child, addr, &tmp, sizeof(tmp), 0);
-		ret = -EIO;
-		if (copied != sizeof(tmp))
-			break;
-		ret = put_user(tmp, (unsigned int __user *) (unsigned long) data);
-		break;
-	}
-
 	/*
 	 * Read 4 bytes of the other process' storage
 	 *  data is a pointer specifying where the user wants the
@@ -237,16 +198,6 @@ asmlinkage int sys32_ptrace(int request,
 		break;
 	}
 
-	/* when I and D space are separate, this will have to be fixed. */
-	case PTRACE_POKETEXT: /* write the word at location addr. */
-	case PTRACE_POKEDATA:
-		ret = 0;
-		if (access_process_vm(child, addr, &data, sizeof(data), 1)
-		    == sizeof(data))
-			break;
-		ret = -EIO;
-		break;
-
 	/*
 	 * Write 4 bytes into the other process' storage
 	 *  data is the 4 bytes that the user wants written
@@ -400,24 +351,15 @@ asmlinkage int sys32_ptrace(int request,
 		ret = ptrace_detach(child, data);
 		break;
 
-	case PTRACE_GETEVENTMSG:
-		ret = put_user(child->ptrace_message,
-			       (unsigned int __user *) (unsigned long) data);
-		break;
-
 	case PTRACE_GET_THREAD_AREA_3264:
 		ret = put_user(task_thread_info(child)->tp_value,
 				(unsigned long __user *) (unsigned long) data);
 		break;
 
 	default:
-		ret = ptrace_request(child, request, addr, data);
+		ret = compat_ptrace_request(child, request, addr, data);
 		break;
 	}
 
-out_tsk:
-	put_task_struct(child);
-out:
-	unlock_kernel();
 	return ret;
 }
Index: linux-2.6/include/asm-mips/ptrace.h
===================================================================
--- linux-2.6.orig/include/asm-mips/ptrace.h	2008-04-29 11:00:10.000000000 +0200
+++ linux-2.6/include/asm-mips/ptrace.h	2008-04-29 11:00:19.000000000 +0200
@@ -76,6 +76,8 @@ struct pt_regs {
 #include <linux/linkage.h>
 #include <asm/isadep.h>
 
+#define __ARCH_WANT_COMPAT_SYS_PTRACE
+
 /*
  * Does the process account for user or for system time?
  */

Re: undefined reference to `copy_siginfo_from_user32'

[Giuseppe Sacco]

Il giorno mar, 29/04/2008 alle 11.00 +0200, Christoph Hellwig ha
scritto:
> On Mon, Apr 28, 2008 at 09:23:27PM +0200, Giuseppe Sacco wrote:
> > Hi list,
> > since a few days, whenever I try to recompile the latest kernel (from git) it always print this error message:
> 
> This should be fixed in mainline.  But the right fix would be to switch
> mips to the generic compat_ptrace.  And untested (and in fact even
> uncompiled) patch ontop of the copy_siginfo_to_user32 posted to the list
> a while ago is below to sketch how this should look like:

I cannot apply this patch to the latest kernel from git. Could you
provide a new one?

Thanks a lot,
Giuseppe Sacco

Re: undefined reference to `copy_siginfo_from_user32'

[Christoph Hellwig]

On Thu, May 01, 2008 at 08:11:52PM +0200, Giuseppe Sacco wrote:
> I cannot apply this patch to the latest kernel from git. Could you
> provide a new one?

It still applies against the latest Linus tree, I don't currently
have a mips tree at hand.  But Ralf promised to look into the
compat_sys_ptrace conversion.