SA_SIGINFO in Linux 2.4.x

DISCUSSION Aneesh Kumar K.V 4 participants 5 messages Aug 2, 2002 1 message not in the archive

lore

Share & tools

lore

A message that is not in the archive (sent off-list or not yet received): [email protected]

Aneesh Kumar K.V

author

2002-08-02 09:48 UTC #1 lore raw

Quoted 9 lines · Camm Maguire

Greetings! The 2.4.x kernels on alpha don't appear to be filling in the si_addr element of the siginfo_t structure when a signal handler is setup with SA_SIGINFO. Is this right? Any other way to get this address in the handler?
Take care,
Signature · 1 line
Camm Maguire			     			[email protected]

Hi,

Try the following patch for arch/alpha/mm/fault.c

-aneesh

fault.c.n renamed from fault.c +21 -3

@@ -88,6 +88,7 @@
 	struct mm_struct *mm = current->mm;
 	unsigned int fixup;
 	int fault;
+	siginfo_t info;
 
 	/* As of EV6, a load into $31/$f31 is a prefetch, and never faults
 	   (or is suppressed by the PALcode).  Support that for older CPUs

@@ -108,6 +109,8 @@
 	if (!mm || in_interrupt())
 		goto no_context;
 
+	info.si_code = SEGV_MAPERR;
+
 #ifdef CONFIG_ALPHA_LARGE_VMALLOC
 	if (address >= TASK_SIZE)
 		goto vmalloc_fault;

@@ -128,6 +131,7 @@
  * we can handle it..
  */
 good_area:
+	info.si_code = SEGV_ACCERR;
 	if (cause < 0) {
 		if (!(vma->vm_flags & VM_EXEC))
 			goto bad_area;

@@ -164,7 +168,12 @@
 	up_read(&mm->mmap_sem);
 
 	if (user_mode(regs)) {
-		force_sig(SIGSEGV, current);
+			
+		info.si_signo = SIGSEGV;
+		info.si_errno = 0;
+		/* info.si_code has been set above */
+		info.si_addr = (void *)address;
+		force_sig_info(SIGSEGV,&info,current);
 		return;
 	}

@@ -212,7 +221,12 @@
 	 * Send a sigbus, regardless of whether we were in kernel
 	 * or user mode.
 	 */
-	force_sig(SIGBUS, current);
+	info.si_signo = SIGBUS;
+	info.si_errno = 0;
+	/* not sure si_code value here  */
+	info.si_code =  BUS_ADRERR;
+	info.si_addr = (void *)address;
+	force_sig_info(SIGBUS,&info,current);
 	if (!user_mode(regs))
 		goto no_context;
 	return;

@@ -220,7 +234,11 @@
 #ifdef CONFIG_ALPHA_LARGE_VMALLOC
 vmalloc_fault:
 	if (user_mode(regs)) {
-		force_sig(SIGSEGV, current);
+		info.si_signo = SIGSEGV;
+		info.si_errno = 0;
+		/* info.si_code has been set above */
+		info.si_addr = (void *)address;
+		force_sig_info(SIGSEGV,&info,current);
 		return;
 	} else {
 		/* Synchronize this task's top level page-table

Daniel Barlow

2002-08-02 12:38 UTC #2 lore raw

Quoted 4 lines · Camm Maguire

Greetings! The 2.4.x kernels on alpha don't appear to be filling in the si_addr element of the siginfo_t structure when a signal handler is setup with SA_SIGINFO. Is this right? Any other way to get this address in the handler?

You may be able to use the third argument to the signal handler:
given a handler declared as (int n, siginfo_t *info,struct ucontext *context), look at context->uc_mcontext.sc_traparg_a0

SBCL has been doing this for a few months now and nobody has yet
complained that it's broken for them. Look for arch_get_bad_addr
in http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/sbcl/sbcl/src/runtime/alpha-arch.c?rev=1.14&content-type=text/vnd.viewcvs-markup

-dan

Signature · 1 line

  http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources

Camm Maguire

2002-08-02 15:52 UTC #3 lore raw

Greetings, and thanks for the suggestion! But it doesn't seem to work for me: (Linux alpha 2.4.17)

=============================================================================
Program received signal SIGSEGV, Segmentation fault. 0x1200a1a7c in read_object (in=0x120a61758) at read.d:531
531 token->st.st_fillp = length;
(gdb)
p/x token
p/x token
$3 = 0x120745f18
(gdb) c
c
Continuing.

Breakpoint 1, memprotect_handler (sig=544593856, code=4831823984,

    scp=0x1200a0684, addr=0x11fffc810 "�O\177 \001") at sgbc.c:1170

1170 memprotect_handler(int sig, long code, struct sigcontext *scp, char *addr) {
(gdb) p/x *(struct ucontext *)scp
p/x *(struct ucontext *)scp
$4 = {uc_flags = 0x23bdd86c27ba005c, uc_link = 0xb42f002847e00401,

  __uc_osf_sigmask = 0xa45d9d20b3ef0030, uc_stack = {
    ss_sp = 0xa45d9d20a43d9d20, ss_flags = 0xa4620000, 
    ss_size = 0xa43d9088b4410000}, uc_mcontext = {
    sc_onstack = 0x9d4100008d4f0038, sc_mask = 0x894f0040a43d80c0, 
    sc_pc = 0xa43d9a4099410000, sc_ps = 0x99410000894f0044, sc_regs = {
      0x894f0048a43dba50, 0xa45da75099410000, 0xa45da750a43da750, 
      0x40611522a4620000, 0xa02f0030b4410000, 0xa43dba68e420000a, 
      0xa43db6d8b3e10000, 0xa6010000a45d9718, 0xa77da9f8a6220000, 
      0x27ba005c6b5b72c0, 0xa42f002823bdd7d0, 0xc3e0000047e10400, 
      0xa75e000047ef041e, 0xa5fe0010a53e0008, 0x6bfa800123de0060, 
      0x47ff041f2ffe0000, 0x47ff041f2ffe0000, 0x27bb005c2ffe0000, 
      0xb7fef00023bdd790, 0xb75e000023dee840, 0x47fe040fb5fe0008, 
      0xa43d9088b60f0010, 0x9d4f00288d410000, 0x89410000a43d80c0, 
      0xa43d9a40994f0030, 0x994f003489410000, 0x89410000a43dba50, 
      0xa43d9f40994f0038, 0x994f003c89410000, 0xa4410000a43da750, 
      0x8d4f002847e20403, 0x404114029d430000}, 
    sc_ownedfp = 0xb3ef0020b4410000, sc_fpregs = {0xa04f0020a43d9f40, 
      0x404109a2a0210000, 0xc3e00018f4400001, 0x47e20401a04f0020, 
      0x4022052140201441, 0x41e8140340201642, 0xa45db65840620401, 
      0x47e40403a08f0020, 0x4064052340601443, 0x4044040240601644, 
      0x8d6200088d420000, 0x9d4100008d820010, 0x9d8100109d610008, 
      0x40203002a02f0020, 0xc3ffffe2b04f0020, 0x89410000a43d99d0,

---Type <return> to continue, or q <return> to quit---

      0xd35ffd79994f17b0, 0xa4410000a43d9d20, 0xb42f17b820220138, 
      0xa44f17b8a43d9290, 0x404103a2a4210000, 0xa77dbaa8f4400004, 
      0x27ba005c6b5b41e4, 0xa42f17b823bdd660, 0x8d420000a45d8b38, 
      0xa42f17b89d410110, 0x8d420000a45da748, 0xa42f17b89d410118, 
      0x2c610120205f0002, 0x4864004320810120, 0x4443040248440162, 
      0xa42f17b83c410120}, sc_fpcr = 0xa0620000a45daab8, 
    sc_fp_control = 0x2c61012147e30402, sc_reserved1 = 0x4864004320810121, 
    sc_reserved2 = 0x4443040248440162, sc_ssize = 0xa42f17b83c410121, 
    sc_sbase = 0xb4410128a45da8e0, sc_traparg_a0 = 0xa45d8b40a42f17b8, 
    sc_traparg_a1 = 0x9d4101308d420000, sc_traparg_a2 = 0x8d4f17b8a43d9d20, 
    sc_fp_trap_pc = 0xa60f17b89d410000, 
    sc_fp_trigger_sum = 0x6b5b4000a77db808, 
    sc_fp_trigger_inst = 0x23bdd5bc27ba005c}, uc_sigmask = {__val = {
      0xc3ffffc8c3e00001, 0xa0410000a43dba68, 0x203f0001e4400003, 
      0xc3e0001eb02f0024, 0xa77d99c8a60f0010, 0x27ba005c6b5b416e, 
      0x47e0040123bdd588, 0xa43da750b42f0018, 0x47e20403a4410000, 
      0x9d4300008d4f0018, 0xb441000040411402, 0xa0410000a43d9f40, 
      0xa60f0018ec40000c, 0x6b5b5b15a77d94b8, 0x23bdd54427ba005c, 
      0xa47da75047e00401}}}

(gdb)

=============================================================================

Quoted 23 lines · Daniel Barlow

Quoted 4 lines · Camm Maguire
Greetings! The 2.4.x kernels on alpha don't appear to be filling in the si_addr element of the siginfo_t structure when a signal handler is setup with SA_SIGINFO. Is this right? Any other way to get this address in the handler?
You may be able to use the third argument to the signal handler:
given a handler declared as (int n, siginfo_t *info,struct ucontext *context), look at context->uc_mcontext.sc_traparg_a0
SBCL has been doing this for a few months now and nobody has yet
complained that it's broken for them. Look for arch_get_bad_addr
in http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/sbcl/sbcl/src/runtime/alpha-arch.c?rev=1.14&content-type=text/vnd.viewcvs-markup
-dan
Signature · 1 line
  http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources 

Signature · 3 lines

Camm Maguire			     			[email protected]
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah

Camm Maguire

2002-08-02 15:58 UTC #4 lore raw

Greetings! Please forgive the last message -- it *does* work if you use SA_SIGINFO in sigaction. I was used to x86 where they put the context on the stack in any case.

Take care,

Quoted 107 lines · Camm Maguire

Greetings, and thanks for the suggestion! But it doesn't seem to work for me: (Linux alpha 2.4.17)
=============================================================================
Program received signal SIGSEGV, Segmentation fault. 0x1200a1a7c in read_object (in=0x120a61758) at read.d:531
531 token->st.st_fillp = length;
(gdb)
p/x token
p/x token
$3 = 0x120745f18
(gdb) c
c
Continuing.
Breakpoint 1, memprotect_handler (sig=544593856, code=4831823984,
    scp=0x1200a0684, addr=0x11fffc810 "�O\177 \001") at sgbc.c:1170
1170 memprotect_handler(int sig, long code, struct sigcontext *scp, char *addr) {
(gdb) p/x *(struct ucontext *)scp
p/x *(struct ucontext *)scp
$4 = {uc_flags = 0x23bdd86c27ba005c, uc_link = 0xb42f002847e00401,
  __uc_osf_sigmask = 0xa45d9d20b3ef0030, uc_stack = {
    ss_sp = 0xa45d9d20a43d9d20, ss_flags = 0xa4620000, 
    ss_size = 0xa43d9088b4410000}, uc_mcontext = {
    sc_onstack = 0x9d4100008d4f0038, sc_mask = 0x894f0040a43d80c0, 
    sc_pc = 0xa43d9a4099410000, sc_ps = 0x99410000894f0044, sc_regs = {
      0x894f0048a43dba50, 0xa45da75099410000, 0xa45da750a43da750, 
      0x40611522a4620000, 0xa02f0030b4410000, 0xa43dba68e420000a, 
      0xa43db6d8b3e10000, 0xa6010000a45d9718, 0xa77da9f8a6220000, 
      0x27ba005c6b5b72c0, 0xa42f002823bdd7d0, 0xc3e0000047e10400, 
      0xa75e000047ef041e, 0xa5fe0010a53e0008, 0x6bfa800123de0060, 
      0x47ff041f2ffe0000, 0x47ff041f2ffe0000, 0x27bb005c2ffe0000, 
      0xb7fef00023bdd790, 0xb75e000023dee840, 0x47fe040fb5fe0008, 
      0xa43d9088b60f0010, 0x9d4f00288d410000, 0x89410000a43d80c0, 
      0xa43d9a40994f0030, 0x994f003489410000, 0x89410000a43dba50, 
      0xa43d9f40994f0038, 0x994f003c89410000, 0xa4410000a43da750, 
      0x8d4f002847e20403, 0x404114029d430000}, 
    sc_ownedfp = 0xb3ef0020b4410000, sc_fpregs = {0xa04f0020a43d9f40, 
      0x404109a2a0210000, 0xc3e00018f4400001, 0x47e20401a04f0020, 
      0x4022052140201441, 0x41e8140340201642, 0xa45db65840620401, 
      0x47e40403a08f0020, 0x4064052340601443, 0x4044040240601644, 
      0x8d6200088d420000, 0x9d4100008d820010, 0x9d8100109d610008, 
      0x40203002a02f0020, 0xc3ffffe2b04f0020, 0x89410000a43d99d0, 
---Type <return> to continue, or q <return> to quit---
      0xd35ffd79994f17b0, 0xa4410000a43d9d20, 0xb42f17b820220138, 
      0xa44f17b8a43d9290, 0x404103a2a4210000, 0xa77dbaa8f4400004, 
      0x27ba005c6b5b41e4, 0xa42f17b823bdd660, 0x8d420000a45d8b38, 
      0xa42f17b89d410110, 0x8d420000a45da748, 0xa42f17b89d410118, 
      0x2c610120205f0002, 0x4864004320810120, 0x4443040248440162, 
      0xa42f17b83c410120}, sc_fpcr = 0xa0620000a45daab8, 
    sc_fp_control = 0x2c61012147e30402, sc_reserved1 = 0x4864004320810121, 
    sc_reserved2 = 0x4443040248440162, sc_ssize = 0xa42f17b83c410121, 
    sc_sbase = 0xb4410128a45da8e0, sc_traparg_a0 = 0xa45d8b40a42f17b8, 
    sc_traparg_a1 = 0x9d4101308d420000, sc_traparg_a2 = 0x8d4f17b8a43d9d20, 
    sc_fp_trap_pc = 0xa60f17b89d410000, 
    sc_fp_trigger_sum = 0x6b5b4000a77db808, 
    sc_fp_trigger_inst = 0x23bdd5bc27ba005c}, uc_sigmask = {__val = {
      0xc3ffffc8c3e00001, 0xa0410000a43dba68, 0x203f0001e4400003, 
      0xc3e0001eb02f0024, 0xa77d99c8a60f0010, 0x27ba005c6b5b416e, 
      0x47e0040123bdd588, 0xa43da750b42f0018, 0x47e20403a4410000, 
      0x9d4300008d4f0018, 0xb441000040411402, 0xa0410000a43d9f40, 
      0xa60f0018ec40000c, 0x6b5b5b15a77d94b8, 0x23bdd54427ba005c, 
      0xa47da75047e00401}}}
(gdb)
=============================================================================
Quoted 23 lines · Daniel Barlow
Quoted 4 lines · Camm Maguire
Greetings! The 2.4.x kernels on alpha don't appear to be filling in the si_addr element of the siginfo_t structure when a signal handler is setup with SA_SIGINFO. Is this right? Any other way to get this address in the handler?
You may be able to use the third argument to the signal handler:
given a handler declared as (int n, siginfo_t *info,struct ucontext *context), look at context->uc_mcontext.sc_traparg_a0
SBCL has been doing this for a few months now and nobody has yet
complained that it's broken for them. Look for arch_get_bad_addr
in http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/sbcl/sbcl/src/runtime/alpha-arch.c?rev=1.14&content-type=text/vnd.viewcvs-markup
-dan
Signature · 1 line
  http://ww.telent.net/cliki/ - Link farm for free CL-on-Unix resources 
Signature · 8 lines
Camm Maguire			     			[email protected]
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah

_______________________________________________
Gcl-devel mailing list
[email protected]
http://mail.gnu.org/mailman/listinfo/gcl-devel

Signature · 3 lines

Camm Maguire			     			[email protected]
==========================================================================
"The earth is but one country, and mankind its citizens."  --  Baha'u'llah

Goswin Brederlow

2002-08-02 21:33 UTC #5 lore raw

In reply to Aneesh Kumar K.V

Quoted 8 lines · Aneesh Kumar K.V

Quoted 7 lines · Camm Maguire
Greetings! The 2.4.x kernels on alpha don't appear to be filling in the si_addr element of the siginfo_t structure when a signal handler is setup with SA_SIGINFO. Is this right? Any other way to get this address in the handler?
Take care,

Quoted 7 lines

Quoted 2 lines
Signature · 1 line
Camm Maguire			     			[email protected]
Hi,
Try the following patch for arch/alpha/mm/fault.c

Thats no based on the patch I did a while back, is it?

MfG
Goswin