Change common unhandled packet error message to clarify origin

[sjmudd]

conn.go has 3 identical error messages for different situations so it's harder to identify where an error has been triggered. This change simply identifies the part of the code affected for easier analysis.

I basically saw this and without protocol knowledge it's not clear which part of the code is triggering the error:

E0124 12:51:40.010442     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 68 ...]
E0124 12:51:40.012055     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 72 ...]
E0124 12:51:40.012139     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 72 ...]
E0124 12:51:40.012210     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 82 ...]
E0124 12:51:40.012270     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 82 ...]
E0124 12:51:40.012329     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 85 ...]
E0124 12:51:40.012386     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 85 ...]
E0124 12:51:40.012450     838 conn.go:801] Got unhandled packet from client 478 (a.b.c.d:58121), returning error: [4 100 ...]

[sjmudd]

Oddly enough this came I think from a 8.0.13 mysql client. Which is odd as https://dev.mysql.com/doc/dev/mysql-server/8.0.11/page_protocol_com_field_list.html indicates this is COM_FIELD_LIST and is a deprecated command as of 5.7.X. So odd the client is still using it.

[dweitzman]

Was this a vtgate version older than December 22nd? #4486 fixed a bug in com_set_option which wasn't mentioned in the PR description. There was a missing call to recycleReadPacket

[sjmudd]

I think the code I'm using is newer. The code I'm running is 6cf46db so not sure.
Either way it's good to make the debug logging a bit tidier, and maybe if byte 0 is the packet type try to label that better than we do now so it's more obvious but I'm not 100% sure if this is always the case as not looked much at the protocol.

[sjmudd]

I messed up something trying to fix the sign-off so have done a new PR. Bit stupid, sorry. so closing this one.