| @c OBSOLETE |
| @c OBSOLETE @node Convex,,, Top |
| @c OBSOLETE @appendix Convex-specific info |
| @c OBSOLETE @cindex Convex notes |
| @c OBSOLETE |
| @c OBSOLETE Scalar registers are 64 bits long, which is a pain since |
| @c OBSOLETE left half of an S register frequently contains noise. |
| @c OBSOLETE Therefore there are two ways to obtain the value of an S register. |
| @c OBSOLETE |
| @c OBSOLETE @table @kbd |
| @c OBSOLETE @item $s0 |
| @c OBSOLETE returns the low half of the register as an int |
| @c OBSOLETE |
| @c OBSOLETE @item $S0 |
| @c OBSOLETE returns the whole register as a long long |
| @c OBSOLETE @end table |
| @c OBSOLETE |
| @c OBSOLETE You can print the value in floating point by using @samp{p/f $s0} or @samp{p/f $S0} |
| @c OBSOLETE to print a single or double precision value. |
| @c OBSOLETE |
| @c OBSOLETE @cindex vector registers |
| @c OBSOLETE Vector registers are handled similarly, with @samp{$V0} denoting the whole |
| @c OBSOLETE 64-bit register and @kbd{$v0} denoting the 32-bit low half; @samp{p/f $v0} |
| @c OBSOLETE or @samp{p/f $V0} can be used to examine the register in floating point. |
| @c OBSOLETE The length of the vector registers is taken from @samp{$vl}. |
| @c OBSOLETE |
| @c OBSOLETE Individual elements of a vector register are denoted in the obvious way; |
| @c OBSOLETE @samp{print $v3[9]} prints the tenth element of register @kbd{v3}, and |
| @c OBSOLETE @samp{set $v3[9] = 1234} alters it. |
| @c OBSOLETE |
| @c OBSOLETE @kbd{$vl} and @kbd{$vs} are int, and @kbd{$vm} is an int vector. |
| @c OBSOLETE Elements of @kbd{$vm} can't be assigned to. |
| @c OBSOLETE |
| @c OBSOLETE @cindex communication registers |
| @c OBSOLETE @kindex info comm-registers |
| @c OBSOLETE Communication registers have names @kbd{$C0 .. $C63}, with @kbd{$c0 .. $c63} |
| @c OBSOLETE denoting the low-order halves. @samp{info comm-registers} will print them |
| @c OBSOLETE all out, and tell which are locked. (A communication register is |
| @c OBSOLETE locked when a value is sent to it, and unlocked when the value is |
| @c OBSOLETE received.) Communication registers are, of course, global to all |
| @c OBSOLETE threads, so it does not matter what the currently selected thread is. |
| @c OBSOLETE @samp{info comm-reg @var{name}} prints just that one communication |
| @c OBSOLETE register; @samp{name} may also be a communication register number |
| @c OBSOLETE @samp{nn} or @samp{0xnn}. |
| @c OBSOLETE @samp{info comm-reg @var{address}} prints the contents of the resource |
| @c OBSOLETE structure at that address. |
| @c OBSOLETE |
| @c OBSOLETE @kindex info psw |
| @c OBSOLETE The command @samp{info psw} prints the processor status word @kbd{$ps} |
| @c OBSOLETE bit by bit. |
| @c OBSOLETE |
| @c OBSOLETE @kindex set base |
| @c OBSOLETE GDB normally prints all integers in base 10, but the leading |
| @c OBSOLETE @kbd{0x80000000} of pointers is intolerable in decimal, so the default |
| @c OBSOLETE output radix has been changed to try to print addresses appropriately. |
| @c OBSOLETE The @samp{set base} command can be used to change this. |
| @c OBSOLETE |
| @c OBSOLETE @table @code |
| @c OBSOLETE @item set base 10 |
| @c OBSOLETE Integer values always print in decimal. |
| @c OBSOLETE |
| @c OBSOLETE @item set base 16 |
| @c OBSOLETE Integer values always print in hex. |
| @c OBSOLETE |
| @c OBSOLETE @item set base |
| @c OBSOLETE Go back to the initial state, which prints integer values in hex if they |
| @c OBSOLETE look like pointers (specifically, if they start with 0x8 or 0xf in the |
| @c OBSOLETE stack), otherwise in decimal. |
| @c OBSOLETE @end table |
| @c OBSOLETE |
| @c OBSOLETE @kindex set pipeline |
| @c OBSOLETE When an exception such as a bus error or overflow happens, usually the PC |
| @c OBSOLETE is several instructions ahead by the time the exception is detected. |
| @c OBSOLETE The @samp{set pipe} command will disable this. |
| @c OBSOLETE |
| @c OBSOLETE @table @code |
| @c OBSOLETE @item set pipeline off |
| @c OBSOLETE Forces serial execution of instructions; no vector chaining and no |
| @c OBSOLETE scalar instruction overlap. With this, exceptions are detected with |
| @c OBSOLETE the PC pointing to the instruction after the one in error. |
| @c OBSOLETE |
| @c OBSOLETE @item set pipeline on |
| @c OBSOLETE Returns to normal, fast, execution. This is the default. |
| @c OBSOLETE @end table |
| @c OBSOLETE |
| @c OBSOLETE @cindex parallel |
| @c OBSOLETE In a parallel program, multiple threads may be executing, each |
| @c OBSOLETE with its own registers, stack, and local memory. When one of them |
| @c OBSOLETE hits a breakpoint, that thread is selected. Other threads do |
| @c OBSOLETE not run while the thread is in the breakpoint. |
| @c OBSOLETE |
| @c OBSOLETE @kindex 1cont |
| @c OBSOLETE The selected thread can be single-stepped, given signals, and so |
| @c OBSOLETE on. Any other threads remain stopped. When a @samp{cont} command is given, |
| @c OBSOLETE all threads are resumed. To resume just the selected thread, use |
| @c OBSOLETE the command @samp{1cont}. |
| @c OBSOLETE |
| @c OBSOLETE @kindex thread |
| @c OBSOLETE The @samp{thread} command will show the active threads and the |
| @c OBSOLETE instruction they are about to execute. The selected thread is marked |
| @c OBSOLETE with an asterisk. The command @samp{thread @var{n}} will select thread @var{n}, |
| @c OBSOLETE shifting the debugger's attention to it for single-stepping, |
| @c OBSOLETE registers, local memory, and so on. |
| @c OBSOLETE |
| @c OBSOLETE @kindex info threads |
| @c OBSOLETE The @samp{info threads} command will show what threads, if any, have |
| @c OBSOLETE invisibly hit breakpoints or signals and are waiting to be noticed. |
| @c OBSOLETE |
| @c OBSOLETE @kindex set parallel |
| @c OBSOLETE The @samp{set parallel} command controls how many threads can be active. |
| @c OBSOLETE |
| @c OBSOLETE @table @code |
| @c OBSOLETE @item set parallel off |
| @c OBSOLETE One thread. Requests by the program that other threads join in |
| @c OBSOLETE (spawn and pfork instructions) do not cause other threads to start up. |
| @c OBSOLETE This does the same thing as the @samp{limit concurrency 1} command. |
| @c OBSOLETE |
| @c OBSOLETE @item set parallel fixed |
| @c OBSOLETE All CPUs are assigned to your program whenever it runs. When it |
| @c OBSOLETE executes a pfork or spawn instruction, it begins parallel execution |
| @c OBSOLETE immediately. This does the same thing as the @samp{mpa -f} command. |
| @c OBSOLETE |
| @c OBSOLETE @item set parallel on |
| @c OBSOLETE One or more threads. Spawn and pfork cause CPUs to join in when and if |
| @c OBSOLETE they are free. This is the default. It is very good for system |
| @c OBSOLETE throughput, but not very good for finding bugs in parallel code. If you |
| @c OBSOLETE suspect a bug in parallel code, you probably want @samp{set parallel fixed.} |
| @c OBSOLETE @end table |
| @c OBSOLETE |
| @c OBSOLETE @subsection Limitations |
| @c OBSOLETE |
| @c OBSOLETE WARNING: Convex GDB evaluates expressions in long long, because S |
| @c OBSOLETE registers are 64 bits long. However, GDB expression semantics are not |
| @c OBSOLETE exactly C semantics. This is a bug, strictly speaking, but it's not one I |
| @c OBSOLETE know how to fix. If @samp{x} is a program variable of type int, then it |
| @c OBSOLETE is also type int to GDB, but @samp{x + 1} is long long, as is @samp{x + y} |
| @c OBSOLETE or any other expression requiring computation. So is the expression |
| @c OBSOLETE @samp{1}, or any other constant. You only really have to watch out for |
| @c OBSOLETE calls. The innocuous expression @samp{list_node (0x80001234)} has an |
| @c OBSOLETE argument of type long long. You must explicitly cast it to int. |
| @c OBSOLETE |
| @c OBSOLETE It is not possible to continue after an uncaught fatal signal by using |
| @c OBSOLETE @samp{signal 0}, @samp{return}, @samp{jump}, or anything else. The difficulty is with |
| @c OBSOLETE Unix, not GDB. |
| @c OBSOLETE |
| @c OBSOLETE I have made no big effort to make such things as single-stepping a |
| @c OBSOLETE @kbd{join} instruction do something reasonable. If the program seems to |
| @c OBSOLETE hang when doing this, type @kbd{ctrl-c} and @samp{cont}, or use |
| @c OBSOLETE @samp{thread} to shift to a live thread. Single-stepping a @kbd{spawn} |
| @c OBSOLETE instruction apparently causes new threads to be born with their T bit set; |
| @c OBSOLETE this is not handled gracefully. When a thread has hit a breakpoint, other |
| @c OBSOLETE threads may have invisibly hit the breakpoint in the background; if you |
| @c OBSOLETE clear the breakpoint gdb will be surprised when threads seem to continue |
| @c OBSOLETE to stop at it. All of these situations produce spurious signal 5 traps; |
| @c OBSOLETE if this happens, just type @samp{cont}. If it becomes a nuisance, use |
| @c OBSOLETE @samp{handle 5 nostop}. (It will ask if you are sure. You are.) |
| @c OBSOLETE |
| @c OBSOLETE There is no way in GDB to store a float in a register, as with |
| @c OBSOLETE @kbd{set $s0 = 3.1416}. The identifier @kbd{$s0} denotes an integer, |
| @c OBSOLETE and like any C expression which assigns to an integer variable, the |
| @c OBSOLETE right-hand side is casted to type int. If you should need to do |
| @c OBSOLETE something like this, you can assign the value to @kbd{@{float@} ($sp-4)} |
| @c OBSOLETE and then do @kbd{set $s0 = $sp[-4]}. Same deal with @kbd{set $v0[69] = 6.9}. |
| |