From: Tom de Vries <tdevries@suse.de>
To: Pedro Alves <palves@redhat.com>,
"Aktemur, Tankut Baris" <tankut.baris.aktemur@intel.com>,
"gdb-patches@sourceware.org" <gdb-patches@sourceware.org>
Subject: [PATCH][gdb/testsuite] Warn about leaked global array
Date: Mon, 18 May 2020 08:18:28 +0200 [thread overview]
Message-ID: <5c3d667e-7a88-0de6-6a19-6b3c4a9130e4@suse.de> (raw)
In-Reply-To: <eef8d619-8a5f-046f-012a-88bf36accf6b@suse.de>
[-- Attachment #1: Type: text/plain, Size: 697 bytes --]
[ was: Re: [PATCH v8 6/6] gdb/infrun: handle already-exited threads when
attempting to stop ]
On 15-05-2020 19:17, Tom de Vries wrote:
> On 15-05-2020 17:46, Pedro Alves wrote:
>> Curious -- is there a reason for the underscores?
>>
>
> Yeah, I'm trying to avoid this:
> ...
> $ cat test.tcl
> #!/usr/bin/tclsh
>
> proc foo { var } {
> global $var
> }
>
> foo var
> $ ./test.tcl
> variable "var" already exists
> while executing
> "global $var"
> (procedure "foo" line 2)
> invoked from within
> "foo var"
> (file "./test.tcl" line 7)
> ...
I managed to find a better way of dealing with this, using "upvar #0".
I've also added more comments to the patch.
Thanks,
- Tom
[-- Attachment #2: 0001-gdb-testsuite-Warn-about-leaked-global-array.patch --]
[-- Type: text/x-patch, Size: 4585 bytes --]
[gdb/testsuite] Warn about leaked global array
A variable name cannot be used both as scalar and array without an
intermediate unset. Trying to do so will result in tcl errors, for
example, for:
...
set var "bla"
set var(1) "bla"
...
we get:
...
can't set "var(1)": variable isn't array
...
and for the reverse statement order we get:
...
can't set "var": variable is array
...
So, since a global name in one test-case can leak to another
test-case, setting a global name in one test-case can result in
a tcl error in another test-case that reuses the name in a different
way.
Warn about leaking a global array from a test-case.
Tested on x86_64-linux.
gdb/testsuite/ChangeLog:
2020-05-18 Tom de Vries <tdevries@suse.de>
* lib/gdb.exp (global_array_exists, global_unset, save_global_vars)
(check_global_vars): New proc.
(gdb_init): Call save_global_vars.
(gdb_finish): Call check_global_vars.
---
gdb/testsuite/lib/gdb.exp | 91 +++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 91 insertions(+)
diff --git a/gdb/testsuite/lib/gdb.exp b/gdb/testsuite/lib/gdb.exp
index f7d20bd94f..285736ee92 100644
--- a/gdb/testsuite/lib/gdb.exp
+++ b/gdb/testsuite/lib/gdb.exp
@@ -5048,6 +5048,95 @@ proc standard_testfile {args} {
}
}
+# Returns 1 if GLOBALVARNAME is a global array.
+
+proc global_array_exists { globalvarname } {
+ # Introduce local alias of global variable $globalvarname. Using
+ # "global $globalvarname" instead is simpler, but this may result in a
+ # clash with local name "globalvarname".
+ upvar #0 $globalvarname globalvar
+ return [array exists globalvar]
+}
+
+# Unset global variable GLOBALVARNAME.
+
+proc global_unset { globalvarname } {
+ # Introduce local alias of global variable $globalvarname.
+ upvar #0 $globalvarname globalvar
+ unset globalvar
+}
+
+# Save global vars to variable gdb_global_vars.
+
+proc save_global_vars { test_file_name } {
+ # Save for the warning.
+ global gdb_test_file_name
+ set gdb_test_file_name $test_file_name
+
+ # Sample state before running test.
+ global gdb_global_vars
+ set gdb_global_vars [info globals]
+}
+
+# Check global variables not in gdb_global_vars.
+
+proc check_global_vars { } {
+ # Sample state after running test.
+ global gdb_global_vars
+ set vars [info globals]
+
+ # I'm not sure these two should actually be global, but at least there
+ # seems to be no harm in having these as globals, given that we don't
+ # expect to reuse these names as scalars.
+ set skip [list "expect_out" "spawn_out"]
+
+ foreach var $vars {
+ if { ![global_array_exists $var] } {
+ continue
+ }
+
+ set found [lsearch -exact $gdb_global_vars $var]
+ if { $found != -1 } {
+ # Already present before running test.
+ continue
+ }
+
+ set found [lsearch -exact $skip $var]
+ if { $found != -1 } {
+ continue
+ }
+
+ # A variable name cannot be used both as scalar and array without an
+ # intermediate unset. Trying to do so will result in tcl errors, for
+ # example, for:
+ # set var "bla"
+ # set var(1) "bla"
+ # we get:
+ # can't set "var(1)": variable isn't array
+ # and for the reverse statement order we get:
+ # can't set "var": variable is array
+ #
+ # So, since a global name in one test-case can leak to another
+ # test-case, setting a global name in one test-case can result in
+ # a tcl error in another test-case that reuses the name in a different
+ # way.
+ #
+ # Warn about leaking a global array from the test-case.
+ # The way to fix this is to wrap the test-case in a namespace and to
+ # change the global variable into a namespace variable:
+ # namespace eval $testfile {
+ # variable var
+ # ...
+ # }
+ global gdb_test_file_name
+ warning "$gdb_test_file_name.exp defined global array $var"
+
+ # If the variable remains set, we won't warn for the next test where
+ # it's leaked, so unset.
+ global_unset $var
+ }
+}
+
# The default timeout used when testing GDB commands. We want to use
# the same timeout as the default dejagnu timeout, unless the user has
# already provided a specific value (probably through a site.exp file).
@@ -5177,10 +5266,12 @@ proc gdb_init { test_file_name } {
global gdb_instances
set gdb_instances 0
+ save_global_vars $test_file_name
return [default_gdb_init $test_file_name]
}
proc gdb_finish { } {
+ check_global_vars
global gdbserver_reconnect_p
global gdb_prompt
global cleanfiles
next prev parent reply other threads:[~2020-05-18 6:18 UTC|newest]
Thread overview: 31+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-05-13 20:53 [PATCH v8 0/6] Handle already-exited threads in 'stop_all_threads' Pedro Alves
2020-05-13 20:53 ` [PATCH v8 1/6] gdb: protect some 'regcache_read_pc' calls Pedro Alves
2020-05-13 20:53 ` [PATCH v8 2/6] gdb/infrun: move a 'regcache_read_pc' call down to first use Pedro Alves
2020-05-13 20:53 ` [PATCH v8 3/6] gdb/infrun: extract out a code piece into 'mark_non_executing_threads' function Pedro Alves
2020-05-13 20:53 ` [PATCH v8 4/6] gdb: introduce 'all_non_exited_process_targets' and 'switch_to_target_no_thread' Pedro Alves
2020-05-14 8:44 ` Aktemur, Tankut Baris
2020-05-14 11:12 ` Pedro Alves
2020-05-14 11:23 ` Aktemur, Tankut Baris
2020-05-13 20:53 ` [PATCH v8 5/6] gdb/infrun: enable/disable thread events of all targets in stop_all_threads Pedro Alves
2020-05-14 8:44 ` Aktemur, Tankut Baris
2020-05-14 11:16 ` Pedro Alves
2020-05-14 11:30 ` Aktemur, Tankut Baris
2020-05-13 20:53 ` [PATCH v8 6/6] gdb/infrun: handle already-exited threads when attempting to stop Pedro Alves
2020-05-14 8:47 ` Aktemur, Tankut Baris
2020-05-14 11:16 ` Pedro Alves
2020-05-14 11:40 ` Aktemur, Tankut Baris
2020-05-14 18:00 ` Tom de Vries
2020-05-14 18:54 ` Aktemur, Tankut Baris
2020-05-14 18:58 ` Pedro Alves
2020-05-15 7:53 ` Aktemur, Tankut Baris
2020-05-15 10:14 ` Pedro Alves
2020-05-15 10:17 ` Tom de Vries
2020-05-15 10:35 ` Pedro Alves
2020-05-15 11:53 ` Tom de Vries
2020-05-15 12:02 ` Pedro Alves
2020-05-15 14:16 ` Tom de Vries
2020-05-15 15:46 ` Pedro Alves
2020-05-15 17:17 ` Tom de Vries
2020-05-18 6:18 ` Tom de Vries [this message]
2020-05-18 10:41 ` [PATCH][gdb/testsuite] Warn about leaked global array Pedro Alves
2020-05-19 16:34 ` Tom de Vries
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=5c3d667e-7a88-0de6-6a19-6b3c4a9130e4@suse.de \
--to=tdevries@suse.de \
--cc=gdb-patches@sourceware.org \
--cc=palves@redhat.com \
--cc=tankut.baris.aktemur@intel.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).