[RFA] Implement darwin remote TARGET_OBJECT_DARWIN_DYLD_INFO.
Commit Message
From: Xavier Roirand <xavier.roirand@adacore.com>
gdb/ChangeLog (Tristan Gingold <gingold@adacore.com>):
* remote.c (remote_xfer_partial): Handle DARWIN_DYLD_INFO.
---
gdb/ChangeLog | 6 ++++++
gdb/remote.c | 33 +++++++++++++++++++++++++++++++++
2 files changed, 39 insertions(+)
Comments
Hi,
A bit of context.
I have never submitted this code before because it doesn't use the official gdb remote packet (‘qXfer:libraries:read’) but the one used by Apple (aka lldb) debugserver.
So with this patch, gdb is able to use debugserver.
Tristan.
----- Mail original -----
> From: Xavier Roirand <xavier.roirand@adacore.com>
>
> gdb/ChangeLog (Tristan Gingold <gingold@adacore.com>):
>
> * remote.c (remote_xfer_partial): Handle DARWIN_DYLD_INFO.
> ---
> gdb/ChangeLog | 6 ++++++
> gdb/remote.c | 33 +++++++++++++++++++++++++++++++++
> 2 files changed, 39 insertions(+)
>
> diff --git a/gdb/ChangeLog b/gdb/ChangeLog
> index b9a9bcd..098ebda 100644
> --- a/gdb/ChangeLog
> +++ b/gdb/ChangeLog
> @@ -1,3 +1,9 @@
> +
> +2018-02-27 Tristan Gingold <gingold@adacore.com>
> +
> + Pushed by Xavier Roirand <roirand@adacore.com>.
> + * remote.c (remote_xfer_partial): Handle DARWIN_DYLD_INFO.
> +
> 2018-02-26 Maciej W. Rozycki <macro@mips.com>
>
> * mips-tdep.c (mips_gdbarch_init): Don't use a 32-bit BFD
> diff --git a/gdb/remote.c b/gdb/remote.c
> index 15d6c5b..95fbfc1 100644
> --- a/gdb/remote.c
> +++ b/gdb/remote.c
> @@ -10495,6 +10495,39 @@ remote_xfer_partial (struct target_ops *ops,
> enum target_object object,
> return TARGET_XFER_E_IO;
> }
>
> + if (object == TARGET_OBJECT_DARWIN_DYLD_INFO)
> + {
> + /* Can only read. */
> + if (readbuf == NULL)
> + return TARGET_XFER_E_IO;
> +
> + if (offset == 0 && len <= 8)
> + {
> + ULONGEST addr;
> + int i;
> +
> + putpkt ("qShlibInfoAddr");
> + getpkt (&rs->buf, &rs->buf_size, 0);
> + if (packet_check_result (rs->buf) != PACKET_OK)
> + return TARGET_XFER_E_IO;
> +
> + /* The packet contains the address in hex (but not suitable for
> + hex2bin as the number of digits is arbitrary. */
> + addr = strtoulst (rs->buf, NULL, 16);
> +
> + /* Store it in big-endian. */
> + for (i = len - 1; i >= 0; i--)
> + {
> + readbuf[i] = addr & 0xff;
> + addr >>= 8;
> + }
> +
> + *xfered_len = len;
> + return TARGET_XFER_OK;
> + }
> + else
> + return TARGET_XFER_EOF;
> + }
> /* Only handle flash writes. */
> if (writebuf != NULL)
> {
> --
> 2.7.4
>
>
tgingold@free.fr writes:
Hi Tristan,
Thanks for the input,
> I have never submitted this code before because it doesn't use the
> official gdb remote packet (‘qXfer:libraries:read’) but the one used
> by Apple (aka lldb) debugserver.
With this patch applied, GDB starts to send packet qShlibInfoAddr.
> So with this patch, gdb is able to use debugserver.
This is a good justification to add this packet, but we still need to
document this packet.
@@ -1,3 +1,9 @@
+
+2018-02-27 Tristan Gingold <gingold@adacore.com>
+
+ Pushed by Xavier Roirand <roirand@adacore.com>.
+ * remote.c (remote_xfer_partial): Handle DARWIN_DYLD_INFO.
+
2018-02-26 Maciej W. Rozycki <macro@mips.com>
* mips-tdep.c (mips_gdbarch_init): Don't use a 32-bit BFD
@@ -10495,6 +10495,39 @@ remote_xfer_partial (struct target_ops *ops, enum target_object object,
return TARGET_XFER_E_IO;
}
+ if (object == TARGET_OBJECT_DARWIN_DYLD_INFO)
+ {
+ /* Can only read. */
+ if (readbuf == NULL)
+ return TARGET_XFER_E_IO;
+
+ if (offset == 0 && len <= 8)
+ {
+ ULONGEST addr;
+ int i;
+
+ putpkt ("qShlibInfoAddr");
+ getpkt (&rs->buf, &rs->buf_size, 0);
+ if (packet_check_result (rs->buf) != PACKET_OK)
+ return TARGET_XFER_E_IO;
+
+ /* The packet contains the address in hex (but not suitable for
+ hex2bin as the number of digits is arbitrary. */
+ addr = strtoulst (rs->buf, NULL, 16);
+
+ /* Store it in big-endian. */
+ for (i = len - 1; i >= 0; i--)
+ {
+ readbuf[i] = addr & 0xff;
+ addr >>= 8;
+ }
+
+ *xfered_len = len;
+ return TARGET_XFER_OK;
+ }
+ else
+ return TARGET_XFER_EOF;
+ }
/* Only handle flash writes. */
if (writebuf != NULL)
{