Add work size argument (zoogie#6)

* reduced work size to make nvidia runtime happy
give nvidia GPU more weight

* Minor fixes to Makefile comments

* Add arguments that specify type of work size
Applicable to LFCS and MSKY mining

* Allow work size argument to be optional
I still highly advise one is provided

Makefile

@@ -6,12 +6,12 @@ ifdef SYSTEMROOT
 	LDFLAGS += -L$(INTELOCLSDKROOT)\lib\x64
 else
 	ifeq ($(shell uname), Linux)
-		# Intel's OpenCL SDK installer doesn't set an environmenr variable on Linux, so we'll have to specify its default installation location instead.
+		# Intel's OpenCL SDK installer doesn't set an environment variable on Linux, so we'll have to specify its default installation location instead.
 		CFLAGS += -std=c11 -Wall -Werror -O2 -mrdrnd -I/opt/intel/opencl-sdk/include
 		LDFLAGS += -L/opt/intel/opencl-sdk/lib64
 	endif
 	ifeq ($(shell uname), Darwin)
-		# macOS's "ld" likes to warn you about library dirs not being found. That being said, macOS includes its own implementation of OpenCL.
+		# macOS's linker likes to warn you about library dirs not being found. That being said, macOS includes its own implementation of OpenCL, so CFLAGS and LDFLAGS are not needed.
 		CFLAGS += -std=c11 -Wall -Werror -O2 -mrdrnd
 	endif
 endif
@@ -21,12 +21,12 @@ all: $(PNAME)
 $(PNAME): $(OBJS)
 ifeq ($(shell uname), Darwin)
 	$(CC) -o $@ $^ -framework OpenCL -lmbedcrypto
-# If you want to use the mbedcrypto static library instead (on macOS), change "-lmbedcrypto" to "/usr/local/lib/libmbedcrypto.a" (or wherever else it may be) with the quotes.
+# If you want to use the mbedcrypto static library instead (on macOS), change "-lmbedcrypto" to "/usr/local/lib/libmbedcrypto.a" (if you downloaded mbedtls through Homebrew) with the quotes.
 else
 	$(CC) $(LDFLAGS) -o $@ $^ -lOpenCL -lmbedcrypto
 # If you want to use the mbedcrypto static library instead (whether you're using MSYS2 or are on Linux), change "-lmbedcrypto" to "-l:libmbedcrypto.a" without the quotes.
 # Note: Ubuntu (probably Debian as well) doesn't install "libmbedcrypto.a" through apt-get, thus you would have to compile mbedtls yourself.
 endif
 
 clean:
-	rm -f $(PNAME) *.o
+	rm -f $(PNAME) *.o

bfcl.c

@@ -49,7 +49,7 @@ int main(int argc, const char *argv[]) {
 			puts(invalid_parameters);
 			ret = -1;
 		}
-	} else if (argc == 7) {
+	} else if (argc == 7 && strcmp(argv[1], "lfcs")) {
 		unsigned char console_id[8], emmc_cid[16], offset[2], src[16], ver[16];
 		hex2bytes(console_id, 8, argv[2], 1);
 		hex2bytes(emmc_cid, 16, argv[3], 1);
@@ -69,19 +69,53 @@ int main(int argc, const char *argv[]) {
 			puts(invalid_parameters);
 			ret = -1;
 		}
-	} else if(argc == 5 && !strcmp(argv[1], "msky")){
+	} else if(argc == 6 && !strcmp(argv[1], "msky") && !strcmp(argv[5], "sws")){ // "sws" is standard work size
 		uint32_t msky[4], ver[4], msky_offset;
 		hex2bytes((unsigned char*)msky, 16, argv[2], 1);
 		hex2bytes((unsigned char*)ver, 16, argv[3], 1);
 		hex2bytes((unsigned char*)&msky_offset, 4, argv[4], 1);
+		group_bits = 28;
 		ret = ocl_brute_msky(msky, ver, msky_offset);
-	} else if(argc == 6 && !strcmp(argv[1], "lfcs")){
+	} else if(argc == 6 && !strcmp(argv[1], "msky") && !strcmp(argv[5], "rws")){ // "rws" is reduced work size
+		uint32_t msky[4], ver[4], msky_offset;
+		hex2bytes((unsigned char*)msky, 16, argv[2], 1);
+		hex2bytes((unsigned char*)ver, 16, argv[3], 1);
+		hex2bytes((unsigned char*)&msky_offset, 4, argv[4], 1);
+		group_bits = 20;
+		ret = ocl_brute_msky(msky, ver, msky_offset);
+	} else if(argc == 5 && !strcmp(argv[1], "msky")){ // In the event a work size argument isn't provided, the standard work size shall be used.
+		uint32_t msky[4], ver[4], msky_offset;
+		hex2bytes((unsigned char*)msky, 16, argv[2], 1);
+		hex2bytes((unsigned char*)ver, 16, argv[3], 1);
+		hex2bytes((unsigned char*)&msky_offset, 4, argv[4], 1);
+		group_bits = 28;
+		ret = ocl_brute_msky(msky, ver, msky_offset);
+	} else if(argc == 7 && !strcmp(argv[1], "lfcs") && !strcmp(argv[6], "sws")){ // "sws" is standard work size
+		uint32_t lfcs, ver[2], lfcs_offset;
+		uint16_t newflag;
+		hex2bytes((unsigned char*)&lfcs, 4, argv[2], 1);
+		hex2bytes((unsigned char*)&newflag, 2, argv[3], 1);
+		hex2bytes((unsigned char*)ver, 8, argv[4], 1);
+		hex2bytes((unsigned char*)&lfcs_offset, 4, argv[5], 1);
+		group_bits = 28;
+		ret = ocl_brute_lfcs(lfcs, newflag, ver, lfcs_offset);
+	} else if(argc == 7 && !strcmp(argv[1], "lfcs") && !strcmp(argv[6], "rws")){ // "rws" is reduced work size
+		uint32_t lfcs, ver[2], lfcs_offset;
+		uint16_t newflag;
+		hex2bytes((unsigned char*)&lfcs, 4, argv[2], 1);
+		hex2bytes((unsigned char*)&newflag, 2, argv[3], 1);
+		hex2bytes((unsigned char*)ver, 8, argv[4], 1);
+		hex2bytes((unsigned char*)&lfcs_offset, 4, argv[5], 1);
+		group_bits = 20;
+		ret = ocl_brute_lfcs(lfcs, newflag, ver, lfcs_offset);
+	} else if(argc == 6 && !strcmp(argv[1], "lfcs")){ // In the event a work size argument isn't provided, the standard work size shall be used.
 		uint32_t lfcs, ver[2], lfcs_offset;
 		uint16_t newflag;
 		hex2bytes((unsigned char*)&lfcs, 4, argv[2], 1);
 		hex2bytes((unsigned char*)&newflag, 2, argv[3], 1);
 		hex2bytes((unsigned char*)ver, 8, argv[4], 1);
 		hex2bytes((unsigned char*)&lfcs_offset, 4, argv[5], 1);
+		group_bits = 28;
 		ret = ocl_brute_lfcs(lfcs, newflag, ver, lfcs_offset);
 	} else {
 		printf(invalid_parameters);

cl/dsi.h

@@ -1,10 +1,10 @@
 // more about this: https://github.com/Jimmy-Z/TWLbf/blob/master/dsi.c
 
 __constant static const u64 DSi_KEY_Y[2] =
-	{0xbd4dc4d30ab9dc76ull, 0xe1a00005202ddd1dull};
+	{0xbd4dc4d30ab9dc76ul, 0xe1a00005202ddd1dul};
 
 __constant static const u64 DSi_KEY_MAGIC[2] =
-	{0x2a680f5f1a4f3e79ull, 0xfffefb4e29590258ull};
+	{0x2a680f5f1a4f3e79ul, 0xfffefb4e29590258ul};
 
 // CAUTION this one doesn't work in-place
 inline void byte_reverse_16(u8 *out, const u8 *in){

ocl_brute.c

@@ -130,7 +130,8 @@ int ocl_brute_console_id(const cl_uchar *console_id, const cl_uchar *emmc_cid,
 		// I wish we could use 1e10 in C, counting 0 is not good to your eye
 		total = from_bcd(1ull << 40);
 		// work items variations on lower bits per enqueue, 8 + 1 digits, including the known digit
-		group_bits = 36;
+		// reduced from 36 to 28 to make nvidia runtime happy
+		group_bits = 28;
 		// work items per enqueue, don't count the known digit here
 		num_items = from_bcd(1ull << (group_bits - 4));
 		// between the template bits and group bits, it's the loop bits
@@ -350,7 +351,6 @@ int ocl_brute_msky(const cl_uint *msky, const cl_uint *ver, cl_uint msky_offset)
 	OCL_ASSERT(clEnqueueWriteBuffer(command_queue, mem_out, CL_TRUE, 0, sizeof(cl_uint), &out, 0, NULL, NULL));
 
 	unsigned brute_bits = 32;
-	unsigned group_bits = 28;
 	unsigned loop_bits = brute_bits - group_bits;
 	unsigned loops = 1ull << loop_bits;
 	size_t num_items = 1ull << group_bits;
@@ -471,7 +471,6 @@ int ocl_brute_lfcs(cl_uint lfcs_template, cl_ushort newflag, const cl_uint *ver,
 	OCL_ASSERT(clEnqueueWriteBuffer(command_queue, mem_out, CL_TRUE, 0, sizeof(cl_uint), &out, 0, NULL, NULL));
 
 	unsigned brute_bits = 32;
-	unsigned group_bits = 28;
 	unsigned loop_bits = brute_bits - group_bits;
 	unsigned loops = 1ull << loop_bits;
 	size_t num_items = 1ull << group_bits;

ocl_brute.h

@@ -18,3 +18,4 @@ int ocl_brute_msky(const cl_uint *msky, const cl_uint *ver, cl_uint msky_offset)
 
 int ocl_brute_lfcs(cl_uint lfcs_template, cl_ushort newflag, const cl_uint *ver, cl_uint lfcs_offset);
 
+unsigned group_bits;

ocl_util.c

@@ -189,8 +189,10 @@ void ocl_get_device(cl_platform_id *p_platform_id, cl_device_id *p_device_id) {
 				&& devices[j].c_avail == CL_TRUE){
 				cl_ulong cap = 1ull * devices[j].max_compute_units * devices[j].freq;
 				// unfortunately that metric is not comparable between different vendors
-				if (strstr((const char*)devices[j].name, "Intel") == 0) {
+				if (strstr((const char*)devices[j].vendor, "Advanced Micro Devices") != 0) {
 					cap *= 64;
+				} else if(strstr((const char*)devices[j].vendor, "NVIDIA") != 0) {
+					cap *= 128;
 				}
 				if (cap > maximum) {
 					maximum = cap;