public inbox for gdb-patches@sourceware.org
 help / color / mirror / Atom feed
From: Andrew Burgess <aburgess@redhat.com>
To: gdb-patches@sourceware.org
Cc: Andrew Burgess <aburgess@redhat.com>
Subject: [PATCH 2/2] gdb: fix reopen_exec_file for files with target: prefix
Date: Wed, 25 Oct 2023 16:08:18 +0100	[thread overview]
Message-ID: <8f5ded0b6bac82dcc59fc028aad239012f08bf1a.1698246342.git.aburgess@redhat.com> (raw)
In-Reply-To: <cover.1698246342.git.aburgess@redhat.com>

Following on from this commit:

  commit f2c4f78c813a9cef38b7e9c9ad18822fb9e19345
  Date:   Thu Sep 21 16:35:30 2023 +0100

      gdb: fix reread_symbols when an objfile has target: prefix

In this commit I update reopen_exec_file to correctly handle
executables with a target: prefix.  Before this commit we used the
system 'stat' call, which obviously isn't going to work for files with
a target: prefix (files located on a possibly remote target machine).

By switching to bfd_stat we will use remote fileio to stat the remote
files, which means we should now correctly detect changes in a remote
executable.

The program_space::ebfd_mtime variable, with which we compare the
result of bfd_stat is set with a call to bfd_get_mtime, which in turn
calls bfd_stat, so comparing to the result of calling bfd_stat makes
sense (I think).

As I discussed in the commit f2c4f78c813a, if a BFD is an in-memory
BFD, then calling bfd_stat will always return 0, while bfd_get_mtime
will always return the time at which the BFD was created.  As a result
comparing the results will always show the file having changed.

I don't believe that GDB can set the main executable to an in-memory
BFD object, so, in this commit, I simply assert that the executable is
not in-memory.  If this ever changes then we would need to decide how
to handle this case -- always reload, or never reload.  The assert
doesn't appear to trigger for our current test suite.
---
 gdb/corefile.c                                | 17 +++--
 gdb/testsuite/gdb.server/target-exec-file.exp | 70 +++++++++++++++++++
 2 files changed, 80 insertions(+), 7 deletions(-)

diff --git a/gdb/corefile.c b/gdb/corefile.c
index 19a96bc6f86..cfbb852b3c2 100644
--- a/gdb/corefile.c
+++ b/gdb/corefile.c
@@ -105,21 +105,24 @@ specify_exec_file_hook (void (*hook) (const char *))
 void
 reopen_exec_file (void)
 {
-  int res;
-  struct stat st;
-
   /* Don't do anything if there isn't an exec file.  */
   if (current_program_space->exec_bfd () == NULL)
     return;
 
+  bfd *exec_bfd = current_program_space->exec_bfd ();
+
+  /* The main executable can't be an in-memory BFD object.  If it was then
+     the use of bfd_stat below would not work as expected.  */
+  gdb_assert ((exec_bfd->flags & BFD_IN_MEMORY) == 0);
+
   /* If the timestamp of the exec file has changed, reopen it.  */
-  std::string filename = bfd_get_filename (current_program_space->exec_bfd ());
-  res = stat (filename.c_str (), &st);
+  struct stat st;
+  int res = bfd_stat (exec_bfd, &st);
 
   if (res == 0
-      && current_program_space->ebfd_mtime
+      && current_program_space->ebfd_mtime != 0
       && current_program_space->ebfd_mtime != st.st_mtime)
-    exec_file_attach (filename.c_str (), 0);
+    exec_file_attach (bfd_get_filename (exec_bfd), 0);
 }
 \f
 /* If we have both a core file and an exec file,
diff --git a/gdb/testsuite/gdb.server/target-exec-file.exp b/gdb/testsuite/gdb.server/target-exec-file.exp
index 9260df8b88d..40863538785 100644
--- a/gdb/testsuite/gdb.server/target-exec-file.exp
+++ b/gdb/testsuite/gdb.server/target-exec-file.exp
@@ -52,6 +52,19 @@ set target_exec [gdb_remote_download target $binfile]
 # prompt us if this is the right thing to do.
 gdb_test_no_output "set confirm off"
 
+if { [allow_python_tests] } {
+    # Register an event handler for the executable changed event.
+    # This handler just copies the event into a global Python object.
+    gdb_test_multiline "Add connection_removed event" \
+	"python" "" \
+	"global_exec_changed_event = None" "" \
+	"def executable_changed(event):" "" \
+	"   global global_exec_changed_event" "" \
+	"   global_exec_changed_event = event" "" \
+	"gdb.events.executable_changed.connect (executable_changed)" "" \
+	"end" ""
+}
+
 # Start gdbserver, but always in extended-remote mode, and then
 # connect to it from GDB.
 set res [gdbserver_start "--multi" $target_exec]
@@ -59,6 +72,22 @@ set gdbserver_protocol "extended-remote"
 set gdbserver_gdbport [lindex $res 1]
 gdb_target_cmd $gdbserver_protocol $gdbserver_gdbport
 
+if { [allow_python_tests] } {
+    # When connecting to a remote target, if the user has not told GDB
+    # which executable to use, then GDB will figure out an executable
+    # from the remote target.
+    #
+    # As a result we expect to have seen an executable changed event.
+    with_test_prefix "after connecting" {
+	gdb_test "python print(global_exec_changed_event)" \
+	    "<gdb.ExecutableChangedEvent object at $hex>"
+	gdb_test "python print(global_exec_changed_event.progspace.executable_filename)" \
+	    [string_to_regexp target:$target_exec]
+	gdb_test "python print(global_exec_changed_event.reload)" "False"
+	gdb_test_no_output "python global_exec_changed_event = None"
+    }
+}
+
 # Issue a 'file' command and parse the output.  We look for a couple
 # of specific things to ensure that we are correctly reading the exec
 # from the remote target.
@@ -104,6 +133,20 @@ gdb_assert { $saw_read_of_remote_exec } \
 gdb_assert { $saw_read_of_syms_from_exec } \
     "symbols were read from remote exec file"
 
+if { [allow_python_tests] } {
+    # The 'file' command forces GDB to always load the executable,
+    # even if the same filename is used.  In this case, as the
+    # filename is the same, this will show as a reload event.
+    with_test_prefix "after 'file' command" {
+	gdb_test "python print(global_exec_changed_event)" \
+	    "<gdb.ExecutableChangedEvent object at $hex>"
+	gdb_test "python print(global_exec_changed_event.progspace.executable_filename)" \
+	    [string_to_regexp target:$target_exec]
+	gdb_test "python print(global_exec_changed_event.reload)" "True"
+	gdb_test_no_output "python global_exec_changed_event = None"
+    }
+}
+
 # Start the inferior (with the 'start' command), use TESTNAME for any
 # pass/fail calls.  EXPECT_REREAD should be true or false and
 # indicates if we expect to too a line like:
@@ -155,10 +198,24 @@ proc start_inferior { testname expect_reread } {
 # see the symbols re-read now.
 start_inferior "start inferior the first time" false
 
+if { [allow_python_tests] } {
+    # The executable hasn't changed.
+    with_test_prefix "after starting inferior for the first time" {
+	gdb_test "python print(global_exec_changed_event)" "None"
+    }
+}
+
 # Re-start the inferior.  The executable is unchanged so we should not
 # see the symbol file being re-read.
 start_inferior "start inferior a second time" false
 
+if { [allow_python_tests] } {
+    # The executable still hasn't changed.
+    with_test_prefix "after starting inferior for the second time" {
+	gdb_test "python print(global_exec_changed_event)" "None"
+    }
+}
+
 # Delay for a short while so, when we touch the exec, we know the
 # timestamp will change.
 sleep 1
@@ -172,3 +229,16 @@ if { $status != 0 } {
 # Start the inferior again, we expect to see the symbols being re-read
 # from the remote file.
 start_inferior "start inferior a third time" true
+
+if { [allow_python_tests] } {
+    # The executable has now changed on disk.  This will be a reload
+    # event.
+    with_test_prefix "after starting inferior for the third time" {
+	gdb_test "python print(global_exec_changed_event)" \
+	    "<gdb.ExecutableChangedEvent object at $hex>"
+	gdb_test "python print(global_exec_changed_event.progspace.executable_filename)" \
+	    [string_to_regexp target:$target_exec]
+	gdb_test "python print(global_exec_changed_event.reload)" "True"
+	gdb_test_no_output "python global_exec_changed_event = None"
+    }
+}
-- 
2.25.4


  parent reply	other threads:[~2023-10-25 15:08 UTC|newest]

Thread overview: 15+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-10-25 15:08 [PATCH 0/2] BFD cache management And Exec file " Andrew Burgess
2023-10-25 15:08 ` [PATCH 1/2] gdb: move all bfd_cache_close_all calls in gdb_bfd.c Andrew Burgess
2023-10-26  9:06   ` Lancelot SIX
2023-10-27  8:49     ` Andrew Burgess
2023-10-27 19:30   ` Tom Tromey
2023-10-30 10:20     ` Andrew Burgess
2023-10-31 18:28       ` Tom Tromey
2023-11-01 10:46         ` Andrew Burgess
2023-11-12 23:38           ` Tom Tromey
2023-10-25 15:08 ` Andrew Burgess [this message]
2023-10-27 18:39   ` [PATCH 2/2] gdb: fix reopen_exec_file for files with target: prefix Tom Tromey
2023-10-30 13:41 ` [PATCH 0/2] BFD cache management And Exec file " Andrew Burgess
2023-10-30 13:41   ` [PATCH 1/2] gdb: move all bfd_cache_close_all calls in gdb_bfd.c Andrew Burgess
2023-10-30 13:41   ` [PATCH 2/2] gdb: fix reopen_exec_file for files with target: prefix Andrew Burgess
2023-11-12 23:40   ` [PATCH 0/2] BFD cache management And Exec file " Tom Tromey

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=8f5ded0b6bac82dcc59fc028aad239012f08bf1a.1698246342.git.aburgess@redhat.com \
    --to=aburgess@redhat.com \
    --cc=gdb-patches@sourceware.org \
    /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).