new sim module firmware, shorter delays for T=0 transfers

This commit is contained in:
iceman1001 committed 2026-08-27 01:38:46 +02:00
1 parent a76399c030
commit 84ad2fb8d4
9 files changed
+116 -220

No files matched your search

+12 -40
View File
@@ -327,18 +327,9 @@ static int smart_wait(uint8_t *out, int maxoutlen, bool verbose) {
return -1;
}
// Which class byte a GET RESPONSE following a 61xx / 9Fxx should carry.
//
// There is no rule that can be derived from the command's own class, because
// two cards that both use CLA 'A0' disagree:
//
// - a GSM 11.11 / TS 51.011 SIM answers 6D00 to '00 C0 ...' and needs 'A0 C0'
// - a HID iCLASS SE SAM answers 6D00 to 'A0 C0 ...' and needs '00 C0'
//
// and EMV (Book 1, 9.3.1) fixes GET RESPONSE at '00' even after a proprietary
// '80' command. So '00' is tried first - the historical behaviour, right for
// EMV and for the SAM - and a card that rejects the class gets one retry with
// the class of the command that produced the status word.
// Class byte for a GET RESPONSE after 61xx / 9Fxx. No single rule works: a GSM
// SIM needs 'A0 C0' and rejects '00 C0', a HID iCLASS SE SAM is the other way
// round, and EMV fixes it at '00'. Try '00' first, then the command's own.
#define GETRESP_TRY_FIRST 0
#define GETRESP_TRY_RETRY 1
@@ -400,8 +391,7 @@ static int smart_responseEx(uint8_t *out, int maxoutlen, bool verbose, uint8_t c
if (verbose) PrintAndLogEx(INFO, "Requesting " _YELLOW_("0x%02X") " bytes response", len);
// '00' first, then one retry with the command's own class if the card
// rejects it - see get_response_cla() for why neither works everywhere.
// '00' first, then the command's own class if the card rejects it
for (int attempt = GETRESP_TRY_FIRST; attempt <= GETRESP_TRY_RETRY; attempt++) {
uint8_t cmd_getresp[] = {
@@ -442,20 +432,9 @@ static int smart_responseEx(uint8_t *out, int maxoutlen, bool verbose, uint8_t c
goto out;
}
/*
* Two shapes are valid here, depending on how the GET RESPONSE went out:
*
* len + 2 the data and its status word. This is what SEND_T0
* gives, because the module runs the procedure byte
* exchange itself and strips the echoed INS.
* len + 2 + 1 the same with that procedure byte still in front,
* which is what the raw pass through leaves behind.
*
* Only the second needs unwrapping - but the first used to fall through
* both branches without ever adding to totallen, so the caller got zero
* bytes and reported "result length = 0" while the card had answered
* perfectly. It went unnoticed while GET RESPONSE was always sent raw.
*/
// Two valid shapes: len+2 is data plus SW, which SEND_T0 gives since
// the module strips the procedure byte; len+2+1 still has it in front,
// as the raw pass through leaves it.
if (datalen == len + 2) {
totallen += datalen;
} else {
@@ -823,15 +802,14 @@ static int CmdSmartInfo(const char *Cmd) {
SendCommandNG(CMD_SMART_ATR, NULL, 0);
PacketResponseNG resp;
if (WaitForResponseTimeout(CMD_SMART_ATR, &resp, 2500) == false) {
if (verbose) {
PrintAndLogEx(WARNING, "smart card timeout");
}
PrintAndLogEx(WARNING, "smart card timeout");
return PM3_ETIMEOUT;
}
if (resp.status != PM3_SUCCESS) {
PrintAndLogEx(WARNING, "no ATR - check the card is present and seated");
if (verbose) {
PrintAndLogEx(WARNING, "smart card select failed");
PrintAndLogEx(INFO, "module returned status %d", resp.status);
}
return PM3_ESOFT;
}
@@ -1653,14 +1631,8 @@ int CmdSmartcard(const char *Cmd) {
return CmdsParse(CommandTable, Cmd);
}
/*
* Which protocol the contact exchanges below ask for.
*
* T=0 by default, which is what this has always sent. The ARM side already
* redirects a T=0 request to T=1 when the card's ATR offers no T=0 at all - a
* request that could not have worked as asked. This is for the other case: a
* card that offers both, where T=0 would work but you want T=1 anyway.
*/
// Protocol the contact exchanges ask for. T=0 by default; the ARM redirects a
// card offering no T=0 by itself, so this is for one that offers both.
static smartcard_command_t s_sc_protocol = SC_RAW_T0;
void SetSmartcardProtocolT1(bool use_t1) {