From: David Mosberger <davidm@hpl.hp.com>
To: linux-ia64@vger.kernel.org
Subject: [Linux-ia64] gas fix for .restore directive
Date: Sat, 17 Feb 2001 20:39:22 +0000 [thread overview]
Message-ID: <marc-linux-ia64-105590693005170@msgid-missing> (raw)
The attached patch fixes a typo in the ".restore" directive and brings
it in line with the Assembly Language Reference Manual. According to
the manual, the directive:
.restore sp
should pop all open regions (i.e., ecount = # of prologues - 1) and
.restore sp, N
sets the ecount field to N. Unfortunately, I couldn't cross-check
this against the Intel assembler because, no matter what I tried, it
always complained about "illegal operand combination for .RESTORE".
The patch is relative to the CVS tree, but applies cleanly on top
of the Nov 17 toolchain as well.
Jim or Rich, could you look over it and apply it if it looks OK?
Thanks,
--david
2001-02-17 David Mosberger <davidm@hpl.hp.com>
* config/tc-ia64.c (struct unwind): Add member "prologue_count".
(dot_proc): Clear unwind.prologue_count to zero.
(dot_prologue): Increment unwind.prologue_count.
(dot_restore): If second operand is omitted, use
unwind.prologue_count -1 for "ecount" (# of additional regions to
pop). Decrement unwind.prologue_count by number of regions
popped.
Index: config/tc-ia64.c
=================================RCS file: /cvs/src/src/gas/config/tc-ia64.c,v
retrieving revision 1.41
diff -u -r1.41 tc-ia64.c
--- tc-ia64.c 2001/02/10 01:42:04 1.41
+++ tc-ia64.c 2001/02/17 20:16:18
@@ -624,6 +624,7 @@
/* TRUE if processing unwind directives in a prologue region. */
int prologue;
int prologue_mask;
+ unsigned int prologue_count; /* number of .prologues seen so far */
} unwind;
typedef void (*vbyte_func) PARAMS ((int, char *, char *));
@@ -3047,7 +3048,7 @@
int dummy ATTRIBUTE_UNUSED;
{
expressionS e1, e2;
- unsigned long ecount = 0;
+ unsigned long ecount; /* # of _additional_ regions to pop */
int sep;
sep = parse_operand (&e1);
@@ -3060,14 +3061,21 @@
if (sep = ',')
{
parse_operand (&e2);
- if (e1.X_op != O_constant)
+ if (e2.X_op != O_constant || e2.X_add_number < 0)
{
- as_bad ("Second operand to .restore must be constant");
+ as_bad ("Second operand to .restore must be a constant >= 0");
return;
}
- ecount = e1.X_op;
+ ecount = e2.X_add_number;
}
+ else
+ ecount = unwind.prologue_count - 1;
add_unwind_entry (output_epilogue (ecount));
+
+ if (ecount < unwind.prologue_count)
+ unwind.prologue_count -= ecount + 1;
+ else
+ unwind.prologue_count = 0;
}
static void
@@ -3710,6 +3718,7 @@
demand_empty_rest_of_line ();
ia64_do_align (16);
+ unwind.prologue_count = 0;
unwind.list = unwind.tail = unwind.current_entry = NULL;
unwind.personality_routine = 0;
}
@@ -3764,6 +3773,7 @@
unwind.prologue = 1;
unwind.prologue_mask = mask;
+ ++unwind.prologue_count;
}
static void
next reply other threads:[~2001-02-17 20:39 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2001-02-17 20:39 David Mosberger [this message]
2001-02-21 22:48 ` [Linux-ia64] gas fix for .restore directive Jim Wilson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=marc-linux-ia64-105590693005170@msgid-missing \
--to=davidm@hpl.hp.com \
--cc=linux-ia64@vger.kernel.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox