summaryrefslogtreecommitdiff
path: root/gdb
diff options
context:
space:
mode:
authorMichael Snyder <msnyder@specifix.com>2000-11-17 01:30:53 +0000
committerMichael Snyder <msnyder@specifix.com>2000-11-17 01:30:53 +0000
commit3d14fc8af771fa9ba337c9bc2ac71110279459fa (patch)
treebf174223f80c1e114f92c1e3f98c108ea1a13836 /gdb
parentef53c28cb62963cb4faf81ddd006db0590f0790e (diff)
downloadgdb-3d14fc8af771fa9ba337c9bc2ac71110279459fa.tar.gz
2000-11-16 Michael Snyder <msnyder@cleaver.cygnus.com>
* regcache.c (read_register_bytes): Failing to set register_valid is not necessarily an error, if the register is a pseudo-register. Some pseudo-registers are never marked as valid, so that they will be read anew every time. Determining if a pseudo-register is valid (or should be marked invalid) may be difficult, whereas just recomputing it may be cheap.
Diffstat (limited to 'gdb')
-rw-r--r--gdb/ChangeLog9
-rw-r--r--gdb/regcache.c9
2 files changed, 17 insertions, 1 deletions
diff --git a/gdb/ChangeLog b/gdb/ChangeLog
index af10432003c..8507debe8c8 100644
--- a/gdb/ChangeLog
+++ b/gdb/ChangeLog
@@ -1,3 +1,12 @@
+2000-11-16 Michael Snyder <msnyder@cleaver.cygnus.com>
+
+ * regcache.c (read_register_bytes): Failing to set register_valid
+ is not necessarily an error, if the register is a pseudo-register.
+ Some pseudo-registers are never marked as valid, so that they will
+ be read anew every time. Determining if a pseudo-register is valid
+ (or should be marked invalid) may be difficult, whereas just
+ recomputing it may be cheap.
+
Thu Nov 16 09:47:57 2000 David Taylor <taylor@redhat.com>
* tracepoint.c (trace_find_tracepoint_command): Replace call to
diff --git a/gdb/regcache.c b/gdb/regcache.c
index 2326b3485b9..104bff7b179 100644
--- a/gdb/regcache.c
+++ b/gdb/regcache.c
@@ -391,7 +391,14 @@ read_register_bytes (int inregbyte, char *myaddr, int inlen)
FETCH_PSEUDO_REGISTER (regno);
if (!register_valid[regno])
- error ("read_register_bytes: Couldn't update register %d.", regno);
+ {
+ /* Sometimes pseudoregs are never marked valid, so that they
+ will be fetched every time (it can be complicated to know
+ if a pseudoreg is valid, while "fetching" them can be cheap).
+ */
+ if (regno < NUM_REGS)
+ error ("read_register_bytes: Couldn't update register %d.", regno);
+ }
}
if (myaddr != NULL)