On 3/30/23 18:09, Claudio Imbrenda wrote:
On Thu, 30 Mar 2023 11:42:40 +0000
Janosch Frank <frankja@xxxxxxxxxxxxx> wrote:
Add functions and definitions needed to test the Adjunct
Processor (AP) crypto interface.
Signed-off-by: Janosch Frank <frankja@xxxxxxxxxxxxx>
[...]
+bool ap_check(void)
+{
+ struct ap_queue_status r1 = {};
+ struct pqap_r2 r2 = {};
+
+ /* Base AP support has no STFLE or SCLP feature bit */
this is true, but you are also indiscriminately using a feature for
which there is a STFLE feature. since it seems you depend on that, you
might as well just check bit for STFLE.12 and assume the base support
is there if it's set
Fair enough.
+ expect_pgm_int();
+ ap_pqap_tapq(0, 0, &r1, &r2);
+
+ if (clear_pgm_int() == PGM_INT_CODE_OPERATION)
+ return false;
+
+ return true;
+}
[...]
+struct ap_config_info {
+ uint8_t apsc : 1; /* S bit */
+ uint8_t apxa : 1; /* N bit */
+ uint8_t qact : 1; /* C bit */
+ uint8_t rc8a : 1; /* R bit */
+ uint8_t l : 1; /* L bit */
+ uint8_t lext : 3; /* Lext bits */
+ uint8_t reserved2[3];
+ uint8_t Na; /* max # of APs - 1 */
+ uint8_t Nd; /* max # of Domains - 1 */
+ uint8_t reserved6[10];
+ uint32_t apm[8]; /* AP ID mask */
is there a specific reason why these are uint32_t?
uint64_t would maybe make your life easier in subsequent patches (see my
comments there)
That's how the architecture specifies it.
That part of the IO architecture works with words, it seems to be quite old.
