Re: [PATCH] Contribute FreeBSD unwind support (x86_64 and x86)

[Jeff Law]

On 05/27/2015 11:43 AM, John Marino wrote:

On 5/27/2015 18:17, Jeff Law wrote:

Thanks for pointing that out.  Otherwise I'd probably have asked the
redundant question :-)

Installed on the trunk.



Thanks a lot, Jeff!  Very painless this time. :)

By the way, I can't see any reason why this couldn't be backported to
the gcc-5 branch as well.  I don't know the procedure for that though.
The procedure is pretty simple.  Ask Richi, Jakub or Joseph for 
permission.  They typically like to see patches on the trunk for a short 
period of time to help ensure there's no fallout.


jeff

Re: [PATCH] Contribute FreeBSD unwind support (x86_64 and x86)

[John Marino]

On 5/27/2015 18:17, Jeff Law wrote:
> Thanks for pointing that out.  Otherwise I'd probably have asked the
> redundant question :-)
> 
> Installed on the trunk.
> 

Thanks a lot, Jeff!  Very painless this time. :)

By the way, I can't see any reason why this couldn't be backported to
the gcc-5 branch as well.  I don't know the procedure for that though.

Regards,
John

Re: [PATCH] Contribute FreeBSD unwind support (x86_64 and x86)

[Jeff Law]

On 05/20/2015 01:49 PM, John Marino wrote:

I have maintained unwind support for FreeBSD i386 and x86_64 in my
"gnat-aux" repository for many years (I created it).  I've always
intended on contributing it back to GCC, but I never got around to
"proving" it worked until now.
Happens.  I can't count how many things I've written but never finished 
for various reasons through the years.




The version I've been using actually has two flavors: FreeBSD 8 and
below and FreeBSD 9 and above.  However, the last of the FreeBSD 8
releases reaches EOL at the end of June so the unwind support I've
attached here drops the FreeBSD 8 variation for simplicity's sake.
Seems reasonable.  We're not nearly as aggressive at dropping dead code 
as we perhaps could/should be.



Note that I provided a similar unwind support for DragonFly a few months
ago.  Please consider applying the attached patch to gcc trunk.   (copy
of patch found here:
http://leaf.dragonflybsd.org/~marino/freebsd/freebsd-unwind-support.diff )
Thanks for pointing that out.  It's a shame that the BSDs can't share 
this code, but such is life.




Suggested text for libgcc/ChangeLog:

2015-05-XX  John Marino 

* config.host (i[34567]86-*-freebsd*, x86_64-*-freebsd*):
Set md_unwind_header
* config/i386/freebsd-unwind.h: New.


Also please recall that my copyright assignment to FSF is in order!
Thanks for pointing that out.  Otherwise I'd probably have asked the 
redundant question :-)


Installed on the trunk.

Thanks,
jeff

Re: [PATCH] Contribute FreeBSD unwind support (x86_64 and x86)

[Andreas Tobler]

On 20.05.15 22:21, Andreas Tobler wrote:

On 20.05.15 21:49, John Marino wrote:

I have maintained unwind support for FreeBSD i386 and x86_64 in my
"gnat-aux" repository for many years (I created it).  I've always
intended on contributing it back to GCC, but I never got around to
"proving" it worked until now.

The version I've been using actually has two flavors: FreeBSD 8 and
below and FreeBSD 9 and above.  However, the last of the FreeBSD 8
releases reaches EOL at the end of June so the unwind support I've
attached here drops the FreeBSD 8 variation for simplicity's sake.

I was under the impression that MD unwinding was used for more than just
GNAT but it looks like that impression was wrong.  When I ran the
testsuite, the only tests affected were Ada tests.


It is, libjava uses it.



Note that I provided a similar unwind support for DragonFly a few months
ago.  Please consider applying the attached patch to gcc trunk.   (copy
of patch found here:
http://leaf.dragonflybsd.org/~marino/freebsd/freebsd-unwind-support.diff )

Suggested text for libgcc/ChangeLog:

2015-05-XX  John Marino 

* config.host (i[34567]86-*-freebsd*, x86_64-*-freebsd*):
Set md_unwind_header
* config/i386/freebsd-unwind.h: New.


Also please recall that my copyright assignment to FSF is in order!


Testing patch locally now.


So, finally my findings. With some additional work on libjava for 
FreeBSD (pending patch on my side), I can confirm that this patch is a 
working prerequisite for proper signal handling in libjava. E.g, the 
magic Throw_2 test case succeeds.


I have not found any regressions on the FreeBSD side. And I really would 
like to see this patch go into trunk.


My tests were done on amd64 and i386 FreeBSD -CURRENT.

Thanks a lot!

Andreas

Re: [PATCH] Contribute FreeBSD unwind support (x86_64 and x86)

[Andreas Tobler]

On 20.05.15 21:49, John Marino wrote:

I have maintained unwind support for FreeBSD i386 and x86_64 in my
"gnat-aux" repository for many years (I created it).  I've always
intended on contributing it back to GCC, but I never got around to
"proving" it worked until now.

The version I've been using actually has two flavors: FreeBSD 8 and
below and FreeBSD 9 and above.  However, the last of the FreeBSD 8
releases reaches EOL at the end of June so the unwind support I've
attached here drops the FreeBSD 8 variation for simplicity's sake.

I was under the impression that MD unwinding was used for more than just
GNAT but it looks like that impression was wrong.  When I ran the
testsuite, the only tests affected were Ada tests.


It is, libjava uses it.



Note that I provided a similar unwind support for DragonFly a few months
ago.  Please consider applying the attached patch to gcc trunk.   (copy
of patch found here:
http://leaf.dragonflybsd.org/~marino/freebsd/freebsd-unwind-support.diff )

Suggested text for libgcc/ChangeLog:

2015-05-XX  John Marino 

* config.host (i[34567]86-*-freebsd*, x86_64-*-freebsd*):
Set md_unwind_header
* config/i386/freebsd-unwind.h: New.


Also please recall that my copyright assignment to FSF is in order!


Testing patch locally now.

Thanks,
Andreas

[PATCH] Contribute FreeBSD unwind support (x86_64 and x86)

[John Marino]

I have maintained unwind support for FreeBSD i386 and x86_64 in my
"gnat-aux" repository for many years (I created it).  I've always
intended on contributing it back to GCC, but I never got around to
"proving" it worked until now.

The version I've been using actually has two flavors: FreeBSD 8 and
below and FreeBSD 9 and above.  However, the last of the FreeBSD 8
releases reaches EOL at the end of June so the unwind support I've
attached here drops the FreeBSD 8 variation for simplicity's sake.

I was under the impression that MD unwinding was used for more than just
GNAT but it looks like that impression was wrong.  When I ran the
testsuite, the only tests affected were Ada tests.

FreeBSD 10.1 / gcc-6.0.0 before patch applied:
http://leaf.dragonflybsd.org/~marino/freebsd/summary-100F64.txt

FreeBSD 10.1 / gcc-6.0.0 after patch applied:
http://leaf.dragonflybsd.org/~marino/freebsd/summary-unwind-100F64.txt

Difference between runs:
http://leaf.dragonflybsd.org/~marino/freebsd/unwind-diff-100F64.txt

FreeBSD 11/amd64 with patch (Ada only):
http://leaf.dragonflybsd.org/~marino/freebsd/summary-110F64.txt

FreeBSD 11/i386 with patch (Ada only)
http://leaf.dragonflybsd.org/~marino/freebsd/summary-110F32.txt


Note that I provided a similar unwind support for DragonFly a few months
ago.  Please consider applying the attached patch to gcc trunk.   (copy
of patch found here:
http://leaf.dragonflybsd.org/~marino/freebsd/freebsd-unwind-support.diff )

Suggested text for libgcc/ChangeLog:

2015-05-XX  John Marino 

* config.host (i[34567]86-*-freebsd*, x86_64-*-freebsd*):
Set md_unwind_header
* config/i386/freebsd-unwind.h: New.


Also please recall that my copyright assignment to FSF is in order!
Thanks,
John Marino

--- /dev/null
+++ libgcc/config/i386/freebsd-unwind.h
@@ -0,0 +1,173 @@
+/* DWARF2 EH unwinding support for FreeBSD: AMD x86-64 and x86.
+   Copyright (C) 2015 Free Software Foundation, Inc.
+   Contributed by John Marino 
+
+This file is part of GCC.
+
+GCC is free software; you can redistribute it and/or modify
+it under the terms of the GNU General Public License as published by
+the Free Software Foundation; either version 3, or (at your option)
+any later version.
+
+GCC is distributed in the hope that it will be useful,
+but WITHOUT ANY WARRANTY; without even the implied warranty of
+MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+GNU General Public License for more details.
+
+Under Section 7 of GPL version 3, you are granted additional
+permissions described in the GCC Runtime Library Exception, version
+3.1, as published by the Free Software Foundation.
+
+You should have received a copy of the GNU General Public License and
+a copy of the GCC Runtime Library Exception along with this program;
+see the files COPYING3 and COPYING.RUNTIME respectively.  If not, see
+.  */
+
+/* Do code reading to identify a signal frame, and set the frame
+   state data appropriately.  See unwind-dw2.c for the structs. */
+
+#include 
+#include 
+#include 
+#include 
+
+#define REG_NAME(reg)  sf_uc.uc_mcontext.mc_## reg
+
+#ifdef __x86_64__
+#define MD_FALLBACK_FRAME_STATE_FOR x86_64_freebsd_fallback_frame_state
+
+static _Unwind_Reason_Code
+x86_64_freebsd_fallback_frame_state
+(struct _Unwind_Context *context, _Unwind_FrameState *fs)
+{
+  struct sigframe *sf;
+  long new_cfa;
+
+  /* Prior to FreeBSD 9, the signal trampoline was located immediately
+ before the ps_strings.  To support non-executable stacks on AMD64,
+ the sigtramp was moved to a shared page for FreeBSD 9.  Unfortunately
+ this means looking frame patterns again (sys/amd64/amd64/sigtramp.S)
+ rather than using the robust and convenient KERN_PS_STRINGS trick.
+
+ :  lea 0x10(%rsp),%rdi
+ :  pushq   $0x0
+ :  mov $0x1a1,%rax
+ :  syscall
+
+ If we can't find this pattern, we're at the end of the stack.
+  */
+
+  if (!(   *(unsigned int *)(context->ra)  == 0x247c8d48
+&& *(unsigned int *)(context->ra +  4) == 0x48006a10
+&& *(unsigned int *)(context->ra +  8) == 0x01a1c0c7
+&& *(unsigned int *)(context->ra + 12) == 0x050f ))
+return _URC_END_OF_STACK;
+
+  sf = (struct sigframe *) context->cfa;
+  new_cfa = sf->REG_NAME(rsp);
+  fs->regs.cfa_how = CFA_REG_OFFSET;
+  /* Register 7 is rsp  */
+  fs->regs.cfa_reg = 7;
+  fs->regs.cfa_offset = new_cfa - (long) context->cfa;
+
+  /* The SVR4 register numbering macros aren't usable in libgcc.  */
+  fs->regs.reg[0].how = REG_SAVED_OFFSET;
+  fs->regs.reg[0].loc.offset = (long)&sf->REG_NAME(rax) - new_cfa;
+  fs->regs.reg[1].how = REG_SAVED_OFFSET;
+  fs->regs.reg[1].loc.offset = (long)&sf->REG_NAME(rdx) - new_cfa;
+  fs->regs.reg[2].how = REG_SAVED_OFFSET;
+  fs->regs.reg[2].loc.offset = (long)&sf->REG_NAME(rcx) - new_cfa;
+  fs->regs.reg[3].how = REG_SAVED_OFFSET;
+  fs->regs.reg[3].loc.offset = (long)&sf->REG_NAME(rbx) - new_cfa;
+  fs->regs.reg[4].how