I followed the instructions given here to debug simulated code.
I run into the following issue with gdb-7.8.1.Does anyone know how to resolve this?
(gdb) set remote Z-packet on
(gdb) set tdesc filename features/arm-with-neon.xml
(gdb) symbol-file linux-linaro-gem5/vmlinux
Reading symbols from linux-linaro-gem5/vmlinux...done.
(gdb) target remote localhost:7000
Remote debugging using localhost:7000
Remote 'g' packet reply is too long: 7d3cc510e0080000000000884833608000c0648000c064800039698010ef6d806a400080f0c00f410000000000000000f8b60180ec3260807480008010336080d30100000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000
Which architecture?
I have been playing in this area for a while and among the ISAs I tried, this only seems to work on MIPS.
Hopefully my implementation of remote debugging on POWER will be accepted.
I suspect the current code may generate bad-formed g-packet replies on ARM and SPARC. The problem is the selection between 32- and 64-bit mode. The constructor allocates enough space for the biggest possible case, and stores that size in the 'size' variable, which is later used to determine how many bytes to actually transmit on the wire; the unused space is padded with zeros, and these zeros are also transmitted. The GDB client has the following check in remote.c, function process_g_packet():
if (buf_len > 2 * rsa->sizeof_g_packet)
error (_("Remote 'g' packet reply is too long: %s"), rs->buf);
This check will object to the extra padding.
I have been using a temporary workaround for this which isn't a real fix, but if there is interest, I can spend some time towards an actual patch which would consult the current bit-width (e.g., the AM flag in PSR on SPARC) and allocate the reply size accordingly.
Boris
_______________________________________________ gem5-users mailing list [email protected] http://m5sim.org/cgi-bin/mailman/listinfo/gem5-users
