Mirror of the gdb-patches mailing list
 help / color / mirror / Atom feed
From: "Pierre Muller" <pierre.muller@ics-cnrs.unistra.fr>
To: "'Pedro Alves'" <pedro@codesourcery.com>
Cc: <gdb-patches@sourceware.org>
Subject: RE: [RFC] Add watchpoint hit address function to procfs.c
Date: Mon, 10 May 2010 07:16:00 -0000	[thread overview]
Message-ID: <004e01caf010$ac780450$05680cf0$@muller@ics-cnrs.unistra.fr> (raw)
In-Reply-To: <201005081823.56505.pedro@codesourcery.com>


  Hi Pedro,

  I tried to take all your remarks into account and
decreased further the size of the patch.

  I also checked the testsuite by 
running with RUNTESTFLAGS=gdb*/watch*.exp
with both CVS GDB and with patch applied.
(I used my modified gdb.exp to avoid problems
with the 'DOS-like' newlines..)
 Here is the output:
pierre@fpcOpenSolaris:~/gdbcvs/build64/gdb/testsuite$ diff gdb.sum gdbv4.sum
1c1
< Test Run By pierre on Mon May 10 09:02:27 2010
---
> Test Run By pierre on Mon May 10 09:01:18 2010
80,81c80,81
< FAIL: gdb.base/watchpoint-hw-hit-once.exp: continue
< FAIL: gdb.base/watchpoint-hw-hit-once.exp: continue to break-at-exit (the
prog
ram exited)
---
> PASS: gdb.base/watchpoint-hw-hit-once.exp: continue
> PASS: gdb.base/watchpoint-hw-hit-once.exp: continue to break-at-exit
129,135c129,135
< FAIL: gdb.base/watch-read.exp: read watchpoint triggers on first read
(timeout
)
< FAIL: gdb.base/watch-read.exp: read watchpoint triggers on read after
value ch
anged (timeout)
< FAIL: gdb.base/watch-read.exp: set write watchpoint on global variable
(timeou
t)
< FAIL: gdb.base/watch-read.exp: write watchpoint triggers (timeout)
< FAIL: gdb.base/watch-read.exp: only write watchpoint triggers when value
chang
es (timeout)
< FAIL: gdb.base/watch-read.exp: read watchpoint triggers when value doesn't
cha
nge, trapping reads and writes (timeout)
< FAIL: gdb.base/watch-read.exp: only read watchpoint triggers when value
doesn'
t change (timeout)
---
> PASS: gdb.base/watch-read.exp: read watchpoint triggers on first read
> PASS: gdb.base/watch-read.exp: read watchpoint triggers on read after
value ch
anged
> PASS: gdb.base/watch-read.exp: set write watchpoint on global variable
> PASS: gdb.base/watch-read.exp: write watchpoint triggers
> PASS: gdb.base/watch-read.exp: only write watchpoint triggers when value
chang
es
> FAIL: gdb.base/watch-read.exp: read watchpoint triggers when value doesn't
cha
nge, trapping reads and writes
> FAIL: gdb.base/watch-read.exp: only read watchpoint triggers when value
doesn'
t change
192,193c192,193
< # of expected passes          145
< # of unexpected failures      15
---
> # of expected passes          152
> # of unexpected failures      8
196c196
< /export/home/pierre/bin/gdbcvs version  7.1.50.20100510-cvs -nw -nx
---
> /usr/local/src/gdb/build64/gdb/testsuite/../../gdb/gdb version
7.1.50.2010051
0-cvs -nw -nx

  So there are no regressions.
Remaining failures are:
pierre@fpcOpenSolaris:~/gdbcvs/build64/gdb/testsuite$ grep FAIL gdbv4.sum
FAIL: gdb.base/watch-read.exp: read watchpoint triggers when value doesn't
chang
e, trapping reads and writes
FAIL: gdb.base/watch-read.exp: only read watchpoint triggers when value
doesn't
change
KFAIL: gdb.threads/watchthreads2.exp: gdb can drop watchpoints in
multithreaded
app (PRMS: gdb/10116)
FAIL: gdb.threads/watchthreads.exp: threaded watch loop
FAIL: gdb.threads/watchthreads.exp: first watchpoint on args[0] hit
FAIL: gdb.threads/watchthreads.exp: first watchpoint on args[1] hit
FAIL: gdb.threads/watchthreads.exp: watchpoint on args[0] hit in thread
FAIL: gdb.threads/watchthreads.exp: watchpoint on args[1] hit in thread
FAIL: gdb.threads/watchthreads.exp: combination of threaded watchpoints = 30

  I also checked without adding TIDGET(inferior_ptid), leaving 0 instead,
for both find_procinfo_or_die calls.
This had no effect on the testsuite output (same 8 failures).

Pierre





2010-05-10  Pierre Muller  <muller@ics.u-strasbg.fr>

	* procfs.c (proc_watchpoint_address): New function.
	(procfs_stopped_by_watchpoint): Change second parameter of
	find_procinfo_or_die call to TIDGET (INFERIOR_PTID).
	Remove useless check after find_procinfo_or_die call.
	(procfs_stopped_data_address): New function.
	(procfs_use_watchpoints): Register new watchpoint related
	function.


Index: src/gdb/procfs.c
===================================================================
RCS file: /cvs/src/src/gdb/procfs.c,v
retrieving revision 1.131
diff -u -p -r1.131 procfs.c
--- src/gdb/procfs.c	20 Apr 2010 22:36:35 -0000	1.131
+++ src/gdb/procfs.c	10 May 2010 06:24:55 -0000
@@ -1313,6 +1313,34 @@ proc_what (procinfo *pi)
 #endif
 }
 
+/*
+ * Function: proc_watchpoint_address
+ *
+ *   This function is only called when PI is stopped by a watchpoint.
+ *   Assuming OS supports it, write to *ADDR the data address which
+ * triggered it and return 1.
+ * Return 0 if it is not possible to know the address.
+ */
+
+static int
+proc_watchpoint_address (procinfo *pi, CORE_ADDR *addr)
+{
+  if (!pi->status_valid)
+    if (!proc_get_status (pi))
+      return 0;
+
+#ifdef NEW_PROC_API
+  *addr = (CORE_ADDR) unsigned_pointer_to_address (target_gdbarch,
+	    builtin_type (target_gdbarch)->builtin_data_ptr,
+	    (gdb_byte *) &pi->prstatus.pr_lwp.pr_info.si_addr);
+#else
+  *addr = (CORE_ADDR) unsigned_pointer_to_address (target_gdbarch,
+	    builtin_type (target_gdbarch)->builtin_data_ptr,
+	    (gdb_byte *) &pi->prstatus.pr_info.si_addr);
+#endif
+  return 1;
+}
+
 #ifndef PIOCSSPCACT	/* The following is not supported on OSF.  */
 /*
  * Function: proc_nsysarg
@@ -5539,10 +5567,7 @@ procfs_stopped_by_watchpoint (void)
 {
   procinfo *pi;
 
-  pi = find_procinfo_or_die (PIDGET (inferior_ptid), 0);
-
-  if (!pi)	/* If no process, then not stopped by watchpoint!  */
-    return 0;
+  pi = find_procinfo_or_die (PIDGET (inferior_ptid), TIDGET
(inferior_ptid));
 
   if (proc_flags (pi) & (PR_STOPPED | PR_ISTOP))
     {
@@ -5560,6 +5585,26 @@ procfs_stopped_by_watchpoint (void)
     }
   return 0;
 }
+/*
+ * Function procfs_stopped_data_address
+ * 
+ * Returns 1 if we the OS knows the position of the triggered
+ * watchpoint.  Sets *ADDR to that address.
+ * Returns 0 if OS cannot report that address.
+ * This function is only called if procfs_stopped_by_watchpoint
+ * returned 1, thus no further checks are done.
+ * The function also assumes that ADDR is not NULL.
+ */
+
+static int
+procfs_stopped_data_address (struct target_ops *targ, CORE_ADDR *addr)
+{
+  procinfo *pi;
+
+  pi = find_procinfo_or_die (PIDGET (inferior_ptid), TIDGET
(inferior_ptid));
+
+  return proc_watchpoint_address (pi, addr);
+}
 
 static int
 procfs_insert_watchpoint (CORE_ADDR addr, int len, int type)
@@ -5607,6 +5652,7 @@ procfs_use_watchpoints (struct target_op
   t->to_remove_watchpoint = procfs_remove_watchpoint;
   t->to_region_ok_for_hw_watchpoint = procfs_region_ok_for_hw_watchpoint;
   t->to_can_use_hw_breakpoint = procfs_can_use_hw_breakpoint;
+  t->to_stopped_data_address = procfs_stopped_data_address;
 }
 
 /*


  reply	other threads:[~2010-05-10  7:16 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2010-04-26 14:17 Pierre Muller
2010-05-07 15:25 ` Pedro Alves
2010-05-08 16:46   ` Pierre Muller
2010-05-08 17:24     ` Pedro Alves
2010-05-10  7:16       ` Pierre Muller [this message]
2010-05-10 19:27         ` Pedro Alves
2010-05-10 21:23           ` Joel Brobecker
2010-05-12  2:31             ` Reformat procfs.c (was: Re: [RFC] Add watchpoint hit address function to procfs.c) Pedro Alves
2010-05-17 10:50               ` Pedro Alves
2010-05-10 21:33           ` [RFC] Add watchpoint hit address function to procfs.c Pierre Muller
2010-05-10 21:49             ` Pedro Alves
2010-05-10 22:19               ` [RFA] " Pierre Muller
2010-05-11 11:49                 ` Pedro Alves

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='004e01caf010$ac780450$05680cf0$@muller@ics-cnrs.unistra.fr' \
    --to=pierre.muller@ics-cnrs.unistra.fr \
    --cc=gdb-patches@sourceware.org \
    --cc=pedro@codesourcery.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