remote protocol: ambiguity of 'E*' response packets

remote protocol: ambiguity of 'E*' response packets

[Frank Ch. Eigler]

Hi -

For the 'g' and 'm' query packets, it seems like a response packet
starting with "$E..."  is entirely valid, if the register contents or
memory region starts with a 0xE0-0xEF byte.  But remote.c:remote_send()
seems to interpret $E... anything as an error, and send_g_packet() uses
this one.

Am I misreading the code or the specs?

- FChE

Re: remote protocol: ambiguity of 'E*' response packets

[Pedro Alves]

On Tuesday 07 July 2009 04:49:46, Frank Ch. Eigler wrote:
> Hi -
> 
> For the 'g' and 'm' query packets, it seems like a response packet
> starting with "$E..."  is entirely valid, if the register contents or
> memory region starts with a 0xE0-0xEF byte.  But remote.c:remote_send()
> seems to interpret $E... anything as an error, and send_g_packet() uses
> this one.
> 
> Am I misreading the code or the specs?

Have the stub return lowercase hex digits [0-9a-z], like 0xe0-0xef
to work around it.

This is http://sourceware.org/bugzilla/show_bug.cgi?id=9665

-- 
Pedro Alves