From: Richard Earnshaw <rearnsha@armltd.co.uk>
To: raeburn@cygnus.com
Cc: gas2@cygnus.com
Subject: Inconsistent behaviour (in resolve_symbol_value?)
Date: Mon, 05 Aug 1996 09:31:00 -0000 [thread overview]
Message-ID: <9608051502.AA02481@sun52.armltd> (raw)
The following ARM assembler file should generate errors for both lines, but
only generates an error for the second:
foo.s:
mov r0, #1 | bar
mov r1, # bar | 1
as.new foo.s
foo.s: Assembler messages:
foo.s:2: Error: L0 set to illegal operation on non-absolute symbols
The problem seems to be the code in resolve_symbol_value that handles
various binary operations. I'm not sure what the code is trying to check
for so I'm not sure what the fix should be. Both 'if' clauses involve
checks for undefined sections, but the cases are very specific. Further,
after the logical operation has been performed the type of the node is
changed to O_constant, so it is no-longer possible to recover the
operation in the back end.
...
case O_bit_inclusive_or:
...
case O_logical_or:
resolve_symbol_value (symp->sy_value.X_add_symbol);
resolve_symbol_value (symp->sy_value.X_op_symbol);
seg_left = S_GET_SEGMENT (symp->sy_value.X_add_symbol);
seg_right = S_GET_SEGMENT (symp->sy_value.X_op_symbol);
if (seg_left != seg_right
&& seg_left != undefined_section
&& seg_right != undefined_section)
{
char *file;
unsigned int line;
if (expr_symbol_where (symp, &file, &line))
as_bad_where
(file, line,
"illegal operation on symbols in different sections");
else
as_bad
("%s set to illegal operation on symbols in different section
S_GET_NAME (symp));
}
if ((S_GET_SEGMENT (symp->sy_value.X_add_symbol)
!= absolute_section)
&& symp->sy_value.X_op != O_subtract)
{
char *file;
unsigned int line;
if (expr_symbol_where (symp, &file, &line))
as_bad_where (file, line,
"illegal operation on non-absolute symbols");
else
as_bad ("%s set to illegal operation on non-absolute symbols",
S_GET_NAME (symp));
}
left = S_GET_VALUE (symp->sy_value.X_add_symbol);
right = S_GET_VALUE (symp->sy_value.X_op_symbol);
...
This is a serious problem for a user, since it quietly generating code that
is incorrect.
next reply other threads:[~1996-08-05 9:31 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
1996-08-05 9:31 Richard Earnshaw [this message]
1996-08-07 12:34 ` Ian Lance Taylor
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=9608051502.AA02481@sun52.armltd \
--to=rearnsha@armltd.co.uk \
--cc=gas2@cygnus.com \
--cc=raeburn@cygnus.com \
/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;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).