target/loongarch: Support QMP dump-guest-memory

[Richard Henderson]

On 9/13/24 11:44, scan-ad...@coverity.com wrote:

** CID 1561133:  Integer handling issues  (BAD_SHIFT)
/builds/qemu-project/qemu/target/loongarch/arch_dump.c: 102 in 
loongarch_write_elf64_fprpreg()


________________________________________________________________________________________________________
*** CID 1561133:  Integer handling issues  (BAD_SHIFT)
/builds/qemu-project/qemu/target/loongarch/arch_dump.c: 102 in 
loongarch_write_elf64_fprpreg()
96         int ret, i;
97
98         loongarch_note_init(&note, s, "CORE", 5, NT_PRFPREG, 
sizeof(note.fpu));
99         note.fpu.fcsr = cpu_to_dump64(s, env->fcsr0);
100
101         for (i = 0; i < 8; i++) {

     CID 1561133:  Integer handling issues  (BAD_SHIFT)
     In expression "env->cf[i] << 8 * i", left shifting by more than 31 bits has 
undefined behavior.  The shift amount, "8 * i", is as much as 56.

102             note.fpu.fcc |= env->cf[i] << (8 * i);

Coverity is correct that 'bool' can't be shifted by 56.
However there's an existing read_fcc() that gets this correct.


r~