From: Hui Zhu <hui_zhu@mentor.com>
To: Jan Kratochvil <jan.kratochvil@redhat.com>
Cc: <gdb-patches@sourceware.org>
Subject: Re: [PATCH] breakpoint remove fail handle bug fix
Date: Thu, 12 Apr 2012 04:04:00 -0000 [thread overview]
Message-ID: <4F863B12.2040304@mentor.com> (raw)
In-Reply-To: <20120411203549.GA4715@host2.jankratochvil.net>
On 04/12/12 04:35, Jan Kratochvil wrote:
> On Wed, 11 Apr 2012 11:07:08 +0200, Hui Zhu wrote:
>> (gdb) d
>> Delete all breakpoints? (y or n) y
>> warning: Error removing breakpoint 2
>
> I would propose the attached patch instead.
>
> It needs a testcase, would you write one?
>
> Not sure if gdbserver also needs a fix or not.
>
> No regressions on {x86_64,x86_64-m32,i686}-fedora17-linux-gnu.
Hi Jan,
Thanks for your patch. And it handle the local debug very well.
But it didn't handle the issue when the target is remote. I am sorry
that I didn't talk clear in my mail. I use the gdb that patch your
patch and do a test with remote:
(gdb) set target-async on
(reverse-i-search)`target': set ^CQuit-async on
(gdb) disassemble
No frame selected.
(gdb) target remote :1234
Remote debugging using :1234
Reading symbols from /lib64/ld-linux-x86-64.so.2...(no debugging symbols
found)...done.
Loaded symbols for /lib64/ld-linux-x86-64.so.2
0x00007ffff7ddcaf0 in ?? () from /lib64/ld-linux-x86-64.so.2
(gdb) tb main
Temporary breakpoint 1 at 0x4004c8: file tmp.c, line 4.
(gdb) disassemble
No function contains program counter for selected frame.
(gdb) c
Continuing.
Program received signal SIGINT, Interrupt.
0x00007ffff7ddcaf0 in ?? () from /lib64/ld-linux-x86-64.so.2
(gdb) c
Continuing.
Temporary breakpoint 1, main () at tmp.c:4
4 sleep (20);
(gdb) disas
Dump of assembler code for function main:
0x00000000004004c4 <+0>: push %rbp
0x00000000004004c5 <+1>: mov %rsp,%rbp
=> 0x00000000004004c8 <+4>: mov $0x14,%edi
0x00000000004004cd <+9>: mov $0x0,%eax
0x00000000004004d2 <+14>: callq 0x4003d0 <sleep@plt>
0x00000000004004d7 <+19>: mov $0x0,%eax
0x00000000004004dc <+24>: pop %rbp
0x00000000004004dd <+25>: retq
End of assembler dump.
(gdb) b 6
Breakpoint 2 at 0x4004d7: file tmp.c, line 6.
(gdb) c &
Continuing.
(gdb) d
Delete all breakpoints? (y or n) y
Cannot execute this command while the target is running.
(gdb)
Program received signal SIGTRAP, Trace/breakpoint trap.
0x00000000004004d8 in main () at tmp.c:6
6 return 0;
c
Continuing.
Program received signal SIGSEGV, Segmentation fault.
0x00000000004004d8 in main () at tmp.c:6
6 return 0;
(gdb) c
Continuing.
Program terminated with signal SIGSEGV, Segmentation fault.
The program no longer exists.
(gdb)
Do you have some good idea on this part?
Thanks,
Hui
>
>
> Thanks,
> Jan
>
>
> gdb/
> 2012-04-11 Jan Kratochvil<jan.kratochvil@redhat.com>
>
> * linux-nat.c (linux_proc_xfer_partial): Do not check for LEN size and
> support also WRITEBUF.
> (linux_xfer_partial): Move here the LEN check from
> linux_proc_xfer_partial but also call linux_proc_xfer_partial as a last
> resort if super_xfer_partial fails.
>
> --- a/gdb/linux-nat.c
> +++ b/gdb/linux-nat.c
> @@ -4494,9 +4494,9 @@ linux_nat_make_corefile_notes (bfd *obfd, int *note_size)
> }
>
> /* Implement the to_xfer_partial interface for memory reads using the /proc
> - filesystem. Because we can use a single read() call for /proc, this
> - can be much more efficient than banging away at PTRACE_PEEKTEXT,
> - but it doesn't support writes. */
> + filesystem. Because we can use a single read or write call for /proc, this
> + can be much more efficient than banging away at PTRACE_PEEKTEXT or
> + PTRACE_POKETEXT. */
>
> static LONGEST
> linux_proc_xfer_partial (struct target_ops *ops, enum target_object object,
> @@ -4508,29 +4508,35 @@ linux_proc_xfer_partial (struct target_ops *ops, enum target_object object,
> int fd;
> char filename[64];
>
> - if (object != TARGET_OBJECT_MEMORY || !readbuf)
> - return 0;
> -
> - /* Don't bother for one word. */
> - if (len< 3 * sizeof (long))
> + if (object != TARGET_OBJECT_MEMORY)
> return 0;
>
> /* We could keep this file open and cache it - possibly one per
> thread. That requires some juggling, but is even faster. */
> sprintf (filename, "/proc/%d/mem", PIDGET (inferior_ptid));
> - fd = open (filename, O_RDONLY | O_LARGEFILE);
> + fd = open (filename, (readbuf ? O_RDONLY : O_WRONLY) | O_LARGEFILE);
> if (fd == -1)
> return 0;
>
> - /* If pread64 is available, use it. It's faster if the kernel
> + /* If pread64 or pwrite64 is available, use it. It's faster if the kernel
> supports it (only one syscall), and it's 64-bit safe even on
> 32-bit platforms (for instance, SPARC debugging a SPARC64
> application). */
> + if ((readbuf != NULL
> +#ifdef HAVE_PREAD64
> +&& (pread64 (fd, readbuf, len, offset) != len)
> +#else
> +&& (lseek (fd, offset, SEEK_SET) == -1 || read (fd, readbuf, len) != len)
> +#endif
> + )
> + || (writebuf != NULL
> #ifdef HAVE_PREAD64
> - if (pread64 (fd, readbuf, len, offset) != len)
> + && (pwrite64 (fd, writebuf, len, offset) != len)
> #else
> - if (lseek (fd, offset, SEEK_SET) == -1 || read (fd, readbuf, len) != len)
> + && (lseek (fd, offset, SEEK_SET) == -1
> + || write (fd, writebuf, len) != len)
> #endif
> + ))
> ret = 0;
> else
> ret = len;
> @@ -4759,13 +4765,24 @@ linux_xfer_partial (struct target_ops *ops, enum target_object object,
> offset&= ((ULONGEST) 1<< addr_bit) - 1;
> }
>
> - xfer = linux_proc_xfer_partial (ops, object, annex, readbuf, writebuf,
> - offset, len);
> + /* Use more expensive linux_proc_xfer_partial only for larger transfers. */
> + if (len>= 3 * sizeof (long))
> + {
> + xfer = linux_proc_xfer_partial (ops, object, annex, readbuf, writebuf,
> + offset, len);
> + if (xfer != 0)
> + return xfer;
> + }
> +
> + xfer = super_xfer_partial (ops, object, annex, readbuf, writebuf,
> + offset, len);
> if (xfer != 0)
> return xfer;
>
> - return super_xfer_partial (ops, object, annex, readbuf, writebuf,
> - offset, len);
> + /* PTRACE_* of super_xfer_partial may not work if the inferior is running.
> + linux_proc_xfer_partial still may work in such case. */
> + return linux_proc_xfer_partial (ops, object, annex, readbuf, writebuf,
> + offset, len);
> }
>
> static void
next prev parent reply other threads:[~2012-04-12 2:17 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-04-11 9:32 Hui Zhu
2012-04-11 10:53 ` Yao Qi
2012-04-11 21:33 ` Jan Kratochvil
2012-04-12 4:04 ` Hui Zhu [this message]
2012-04-16 10:01 ` Jan Kratochvil
2012-04-16 16:42 ` Hui Zhu
2012-04-17 21:15 ` Jan Kratochvil
2012-04-12 4:45 ` Doug Evans
2012-04-12 4:35 ` Doug Evans
2012-04-16 9:50 ` Hui Zhu
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=4F863B12.2040304@mentor.com \
--to=hui_zhu@mentor.com \
--cc=gdb-patches@sourceware.org \
--cc=jan.kratochvil@redhat.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