diff options
author | Pierre Muller <muller@sourceware.org> | 2009-09-26 12:08:31 +0000 |
---|---|---|
committer | Pierre Muller <muller@sourceware.org> | 2009-09-26 12:08:31 +0000 |
commit | 087b74b25b24a901bc78eb5a2d436dd1fd5fda21 (patch) | |
tree | 0ae1db239ea2876e43bcb4dc799d025fb73d7bd2 /gdb/i386-nat.c | |
parent | 39236edf0673953bea1f931718c68b966817225e (diff) | |
download | gdb-087b74b25b24a901bc78eb5a2d436dd1fd5fda21.zip gdb-087b74b25b24a901bc78eb5a2d436dd1fd5fda21.tar.gz gdb-087b74b25b24a901bc78eb5a2d436dd1fd5fda21.tar.bz2 |
gdb ChangeLog entry:
* i386-nat.c (i386_stopped_data_address): Also check that
hitted watch register is not vacant.
gdb/testsuite ChangeLog entry:
New test for two watchpoints, with disabling of
the first inserted.
* testsuite/gdb.base/watchpoints.c: New file.
* testsuite/gdb.base/watchpoints.exp: New file.
Diffstat (limited to 'gdb/i386-nat.c')
-rw-r--r-- | gdb/i386-nat.c | 5 |
1 files changed, 4 insertions, 1 deletions
diff --git a/gdb/i386-nat.c b/gdb/i386-nat.c index ab1bd8a..325b4df 100644 --- a/gdb/i386-nat.c +++ b/gdb/i386-nat.c @@ -555,7 +555,10 @@ i386_stopped_data_address (struct target_ops *ops, CORE_ADDR *addr_p) that GDB doesn't call the target_stopped_data_address method except for data watchpoints. In other words, I'm being paranoiac. */ - && I386_DR_GET_RW_LEN (i) != 0) + && I386_DR_GET_RW_LEN (i) != 0 + /* This third condition makes sure DRi is not vacant, this + avoids false positives in windows-nat.c. */ + && !I386_DR_VACANT (i)) { addr = dr_mirror[i]; rc = 1; |