This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [commit] gdbserver async I/O simplifications
- From: Markus Deuling <deuling at de dot ibm dot com>
- To: Daniel Jacobowitz <drow at false dot org>
- Cc: gdb-patches at sourceware dot org
- Date: Fri, 07 Dec 2007 08:00:55 +0100
- Subject: Re: [commit] gdbserver async I/O simplifications
- References: <20071207014108.GA6743@caradoc.them.org>
Hi Daniel,
Daniel Jacobowitz schrieb:
For the patch I'm working on, I needed asynchronous I/O to be disabled
during start_inferior and it was ending up enabled. This version is
considerably simpler, and does not have that problem. Async I/O can
and should be enabled once we have received a packet that will take a
long time, and must be disabled before we reply to GDB about it.
Tested on x86_64-linux and committed.
it seems own_buf is missing in gdbserver/server.c (myresume). This fixes the build
but I havent tested it. Its just what I applied to my dev tree.
ChangeLog:
* gdbserver/server.c (myresume): Add own_buf variable.
--
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com
diff -urpN src/gdb/gdbserver/server.c dev/gdb/gdbserver/server.c
--- src/gdb/gdbserver/server.c 2007-12-07 06:32:33.000000000 +0100
+++ dev/gdb/gdbserver/server.c 2007-12-07 07:56:54.000000000 +0100
@@ -803,9 +803,11 @@ void
myresume (int step, int *signalp, char *statusp)
{
struct thread_resume resume_info[2];
+ char *own_buf;
int n = 0;
int sig = *signalp;
+ own_buf = malloc (PBUFSIZ + 1);
set_desired_inferior (0);
if (step || sig || (cont_thread != 0 && cont_thread != -1))