[PATCH u-boot-marvell 0/5] kwboot fix for AXP + some others

[PATCH u-boot-marvell 0/5] kwboot fix for AXP + some others

[Marek Behún]

From: Marek Behún <marek.behun@nic.cz>

Hi Stefan,

here are some more Pali's changes for kwboot, reviewed by me. The last
one should fix boot on AXP.

Marek

Pali Rohár (5):
  tools: kwboot: Fix sending retry of last header packet
  tools: kwboot: Do not call tcdrain() after each sent packet
  tools: kwboot: Increase delay after changing baudrate in ARM code
  tools: kwboot: Replace ARM mov + movt instruction pair by mov + orr
  tools: kwboot: Do not use stack when setting baudrate back to default
    value

 tools/kwboot.c | 148 +++++++++++++++++++++++++++----------------------
 1 file changed, 83 insertions(+), 65 deletions(-)

-- 
2.32.0

[PATCH u-boot-marvell 1/5] tools: kwboot: Fix sending retry of last header packet

[Marek Behún]

From: Pali Rohár <pali@kernel.org>

After the trasfer of last header packet, it is possible that baudrate
change pattern is received, and also that NAK byte is received so that
the packet should be sent again.

Thus we should not clear the baudrate change state when sending retry
of that packet.

Move code for initializing state variables from kwboot_xm_recv_reply()
to kwboot_xm_sendblock().

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
---
 tools/kwboot.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/tools/kwboot.c b/tools/kwboot.c
index bb7cae9f05..b2c48812c3 100644
--- a/tools/kwboot.c
+++ b/tools/kwboot.c
@@ -859,11 +859,6 @@ kwboot_xm_recv_reply(int fd, char *c, int nak_on_non_xm,
 	uint64_t recv_until = _now() + timeout;
 	int rc;
 
-	if (non_xm_print)
-		*non_xm_print = 0;
-	if (baud_changed)
-		*baud_changed = 0;
-
 	while (1) {
 		rc = kwboot_tty_recv(fd, c, 1, timeout);
 		if (rc) {
@@ -929,6 +924,8 @@ kwboot_xm_sendblock(int fd, struct kwboot_block *block, int allow_non_xm,
 	char c;
 
 	*done_print = 0;
+	non_xm_print = 0;
+	baud_changed = 0;
 
 	retries = 0;
 	do {
-- 
2.32.0

[PATCH u-boot-marvell 2/5] tools: kwboot: Do not call tcdrain() after each sent packet

[Marek Behún]

From: Pali Rohár <pali@kernel.org>

Kwboot puts each xmodem packet to kernel queue, then waits until all bytes
of that packet are transmitted over UART and then waits for xmodem reply
until it is received into kernel queue.

If some reply is received during the time we are waiting until all bytes
are transmitted, then kernel puts them into the queue and returns it to
kwboot in next read() call.

So there is no need to wait (with tcdrain() function) until all bytes from
xmodem packet are transmitted over UART, since any reply received either
during that time or after is returned to kwboot with the next read().

Therefore do not call tcdrain() after each xmodem packet sent. Instead
directly wait for any reply after putting xmodem packet into write kernel
queue.

This change could speed up xmodem transfer in case tcdrain() function waits
for a longer time.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
---
 tools/kwboot.c | 13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)

diff --git a/tools/kwboot.c b/tools/kwboot.c
index b2c48812c3..a6bfd3d4ce 100644
--- a/tools/kwboot.c
+++ b/tools/kwboot.c
@@ -404,7 +404,7 @@ out:
 }
 
 static int
-kwboot_tty_send(int fd, const void *buf, size_t len)
+kwboot_tty_send(int fd, const void *buf, size_t len, int nodrain)
 {
 	if (!buf)
 		return 0;
@@ -412,13 +412,16 @@ kwboot_tty_send(int fd, const void *buf, size_t len)
 	if (kwboot_write(fd, buf, len) < 0)
 		return -1;
 
+	if (nodrain)
+		return 0;
+
 	return tcdrain(fd);
 }
 
 static int
 kwboot_tty_send_char(int fd, unsigned char c)
 {
-	return kwboot_tty_send(fd, &c, 1);
+	return kwboot_tty_send(fd, &c, 1, 0);
 }
 
 static speed_t
@@ -705,7 +708,7 @@ kwboot_bootmsg(int tty, void *msg)
 			break;
 
 		for (count = 0; count < 128; count++) {
-			rc = kwboot_tty_send(tty, msg, 8);
+			rc = kwboot_tty_send(tty, msg, 8, 0);
 			if (rc) {
 				usleep(msg_req_delay * 1000);
 				continue;
@@ -737,7 +740,7 @@ kwboot_debugmsg(int tty, void *msg)
 		if (rc)
 			break;
 
-		rc = kwboot_tty_send(tty, msg, 8);
+		rc = kwboot_tty_send(tty, msg, 8, 0);
 		if (rc) {
 			usleep(msg_req_delay * 1000);
 			continue;
@@ -929,7 +932,7 @@ kwboot_xm_sendblock(int fd, struct kwboot_block *block, int allow_non_xm,
 
 	retries = 0;
 	do {
-		rc = kwboot_tty_send(fd, block, sizeof(*block));
+		rc = kwboot_tty_send(fd, block, sizeof(*block), 1);
 		if (rc)
 			return rc;
 
-- 
2.32.0

[PATCH u-boot-marvell 3/5] tools: kwboot: Increase delay after changing baudrate in ARM code

[Marek Behún]

From: Pali Rohár <pali@kernel.org>

Increase loop cycles from 600000 to 2998272, which should increase delay
from 1ms to about 5ms on 1200 MHz CPU.

The Number 2998272 was chosen as the nearest value around 3000000 which can
be encoded into one ARM mov instruction. It avoids usage of movt instruction
which is not supported by ARMv5te cores.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
---
 tools/kwboot.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/tools/kwboot.c b/tools/kwboot.c
index a6bfd3d4ce..84294cadfe 100644
--- a/tools/kwboot.c
+++ b/tools/kwboot.c
@@ -119,7 +119,7 @@ static unsigned char kwboot_baud_code[] = {
 				/* ;   writel(UART_BASE + DLL, new_dll);      */
 				/* ;   writel(UART_BASE + DLH, new_dlh);      */
 				/* ;   writel(UART_BASE + LCR, lcr & ~DLAB);  */
-				/* ;   msleep(1);                             */
+				/* ;   msleep(5);                             */
 				/* ;   return 0;                              */
 				/* ; }                                        */
 
@@ -130,7 +130,7 @@ static unsigned char kwboot_baud_code[] = {
 	0x01, 0x00, 0x4d, 0xe3, /* movt  r0, #0xd001                          */
 
 				/*  ; r2 = address of preamble string         */
-	0xd0, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
+	0xcc, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
 
 				/*  ; Send preamble string over UART          */
 				/* .Lloop_preamble:                           */
@@ -177,15 +177,15 @@ static unsigned char kwboot_baud_code[] = {
 
 				/*  ; Read old baudrate value                 */
 				/*  ; r2 = old_baudrate                       */
-	0x8c, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
+	0x88, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
 
 				/*  ; Calculate base clock                    */
 				/*  ; r1 = r2 * r1                            */
 	0x92, 0x01, 0x01, 0xe0, /* mul   r1, r2, r1                           */
 
 				/*  ; Read new baudrate value                 */
-				/*  ; r2 = baudrate                           */
-	0x88, 0x20, 0x9f, 0xe5, /* ldr   r2, baudrate                         */
+				/*  ; r2 = new_baudrate                       */
+	0x84, 0x20, 0x9f, 0xe5, /* ldr   r2, new_baudrate                     */
 
 				/*  ; Calculate new Divisor Latch             */
 				/*  ; r1 = DIV_ROUND(r1, r2) =                */
@@ -225,10 +225,10 @@ static unsigned char kwboot_baud_code[] = {
 	0x80, 0x10, 0xc1, 0xe3, /* bic   r1, r1, #0x80                        */
 	0x0c, 0x10, 0x80, 0xe5, /* str   r1, [r0, #0x0c]                      */
 
-				/*  ; Sleep 1ms ~~ 600000 cycles at 1200 MHz  */
-				/*  ; r1 = 600000                             */
-	0x9f, 0x1d, 0xa0, 0xe3, /* mov   r1, #0x27c0                          */
-	0x09, 0x10, 0x40, 0xe3, /* movt  r1, #0x0009                          */
+				/*  ; Loop 0x2dc000 (2998272) cycles          */
+				/*  ; which is about 5ms on 1200 MHz CPU      */
+				/*  ; r1 = 0x2dc000                           */
+	0xb7, 0x19, 0xa0, 0xe3, /* mov   r1, #0x2dc000                        */
 				/* .Lloop_sleep:                              */
 	0x01, 0x10, 0x41, 0xe2, /* sub   r1, r1, #1                           */
 	0x00, 0x00, 0x51, 0xe3, /* cmp   r1, #0                               */
-- 
2.32.0

[PATCH u-boot-marvell 4/5] tools: kwboot: Replace ARM mov + movt instruction pair by mov + orr

[Marek Behún]

From: Pali Rohár <pali@kernel.org>

Older Armada SoCs have custom ARMv5te compatible core which does not
support movt instruction. So replace mov + movt instruction pair used for
immediate move construction by mov + orr instructions which are supported
also by ARMv5te.

After this change kwboot ARM code should be compatible with any 32-bit ARM
core compatible by ARMv2 or new. At least GNU AS does not throw any error
or warning.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
---
 tools/kwboot.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/kwboot.c b/tools/kwboot.c
index 84294cadfe..62c218ef64 100644
--- a/tools/kwboot.c
+++ b/tools/kwboot.c
@@ -126,8 +126,8 @@ static unsigned char kwboot_baud_code[] = {
 	0xfe, 0x5f, 0x2d, 0xe9, /* push  { r1 - r12, lr }                     */
 
 				/*  ; r0 = UART_BASE                          */
-	0x02, 0x0a, 0xa0, 0xe3, /* mov   r0, #0x2000                          */
-	0x01, 0x00, 0x4d, 0xe3, /* movt  r0, #0xd001                          */
+	0x0d, 0x02, 0xa0, 0xe3, /* mov   r0, #0xd0000000                      */
+	0x12, 0x0a, 0x80, 0xe3, /* orr   r0, r0, #0x12000                     */
 
 				/*  ; r2 = address of preamble string         */
 	0xcc, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
-- 
2.32.0

[PATCH u-boot-marvell 5/5] tools: kwboot: Do not use stack when setting baudrate back to default value

[Marek Behún]

From: Pali Rohár <pali@kernel.org>

The ARM code we inject into the image to change baudrate back to the
default value of 115200 Baud, which is run after successful UART transfer
of the whole image, cannot use stack as at this stage stack pointer is not
initialized yet.

Stack can only be used when BootROM is executing binary header, to
preserve state of registers, since BootROM expects that.

Change the ARM baudrate code to not use stack at all and put binary
header specific pre + post code (which stores and restores registers) into
separate arrays.

The baudrate change code now jumps at it's end and expects that there is
either code which returns to the BootROM or jumps to the original exec
address.

Signed-off-by: Pali Rohár <pali@kernel.org>
Reviewed-by: Marek Behún <marek.behun@nic.cz>
---
 tools/kwboot.c | 112 ++++++++++++++++++++++++++++---------------------
 1 file changed, 65 insertions(+), 47 deletions(-)

diff --git a/tools/kwboot.c b/tools/kwboot.c
index 62c218ef64..359b43c0d8 100644
--- a/tools/kwboot.c
+++ b/tools/kwboot.c
@@ -78,14 +78,7 @@ struct kwboot_block {
 #define KWBOOT_BLK_RSP_TIMEO 1000 /* ms */
 #define KWBOOT_HDR_RSP_TIMEO 10000 /* ms */
 
-/* ARM code making baudrate changing function return to original exec address */
-static unsigned char kwboot_pre_baud_code[] = {
-				/* exec_addr:                                 */
-	0x00, 0x00, 0x00, 0x00, /* .word 0                                    */
-	0x0c, 0xe0, 0x1f, 0xe5, /* ldr lr, exec_addr                          */
-};
-
-/* ARM code for binary header injection to change baudrate */
+/* ARM code to change baudrate */
 static unsigned char kwboot_baud_code[] = {
 				/* ; #define UART_BASE 0xd0012000             */
 				/* ; #define THR       0x00                   */
@@ -123,14 +116,12 @@ static unsigned char kwboot_baud_code[] = {
 				/* ;   return 0;                              */
 				/* ; }                                        */
 
-	0xfe, 0x5f, 0x2d, 0xe9, /* push  { r1 - r12, lr }                     */
-
 				/*  ; r0 = UART_BASE                          */
 	0x0d, 0x02, 0xa0, 0xe3, /* mov   r0, #0xd0000000                      */
 	0x12, 0x0a, 0x80, 0xe3, /* orr   r0, r0, #0x12000                     */
 
 				/*  ; r2 = address of preamble string         */
-	0xcc, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
+	0xc8, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
 
 				/*  ; Send preamble string over UART          */
 				/* .Lloop_preamble:                           */
@@ -177,7 +168,7 @@ static unsigned char kwboot_baud_code[] = {
 
 				/*  ; Read old baudrate value                 */
 				/*  ; r2 = old_baudrate                       */
-	0x88, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
+	0x84, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
 
 				/*  ; Calculate base clock                    */
 				/*  ; r1 = r2 * r1                            */
@@ -185,7 +176,7 @@ static unsigned char kwboot_baud_code[] = {
 
 				/*  ; Read new baudrate value                 */
 				/*  ; r2 = new_baudrate                       */
-	0x84, 0x20, 0x9f, 0xe5, /* ldr   r2, new_baudrate                     */
+	0x80, 0x20, 0x9f, 0xe5, /* ldr   r2, new_baudrate                     */
 
 				/*  ; Calculate new Divisor Latch             */
 				/*  ; r1 = DIV_ROUND(r1, r2) =                */
@@ -234,9 +225,7 @@ static unsigned char kwboot_baud_code[] = {
 	0x00, 0x00, 0x51, 0xe3, /* cmp   r1, #0                               */
 	0xfc, 0xff, 0xff, 0x1a, /* bne   .Lloop_sleep                         */
 
-				/*  ; Return 0 - no error                     */
-	0x00, 0x00, 0xa0, 0xe3, /* mov   r0, #0                               */
-	0xfe, 0x9f, 0xbd, 0xe8, /* pop   { r1 - r12, pc }                     */
+	0x05, 0x00, 0x00, 0xea, /* b     end                                  */
 
 				/*  ; Preamble string                         */
 				/* preamble:                                  */
@@ -252,10 +241,29 @@ static unsigned char kwboot_baud_code[] = {
 				/*  ; Placeholder for new baudrate value      */
 				/* new_baudrate:                              */
 	0x00, 0x00, 0x00, 0x00, /* .word 0                                    */
+
+				/* end:                                       */
+};
+
+/* ARM code for storing registers for future returning back to the bootrom */
+static unsigned char kwboot_baud_code_binhdr_pre[] = {
+	0xfe, 0x5f, 0x2d, 0xe9, /* push  { r1 - r12, lr }                     */
 };
 
-#define KWBOOT_BAUDRATE_BIN_HEADER_SZ (sizeof(kwboot_baud_code) + \
-				       sizeof(struct opt_hdr_v1) + 8 + 16)
+/* ARM code for returning back to the bootrom */
+static unsigned char kwboot_baud_code_binhdr_post[] = {
+				/*  ; Return 0 - no error                     */
+	0x00, 0x00, 0xa0, 0xe3, /* mov   r0, #0                               */
+	0xfe, 0x9f, 0xbd, 0xe8, /* pop   { r1 - r12, pc }                     */
+};
+
+/* ARM code for jumping to the original image exec_addr */
+static unsigned char kwboot_baud_code_data_jump[] = {
+	0x04, 0xf0, 0x1f, 0xe5, /* ldr   pc, exec_addr                        */
+				/*  ; Placeholder for exec_addr               */
+				/* exec_addr:                                 */
+	0x00, 0x00, 0x00, 0x00, /* .word 0                                    */
+};
 
 static const char kwb_baud_magic[16] = "$baudratechange";
 
@@ -1409,44 +1417,51 @@ kwboot_add_bin_ohdr_v1(void *img, size_t *size, uint32_t binsz)
 }
 
 static void
-_inject_baudrate_change_code(void *img, size_t *size, int pre,
+_inject_baudrate_change_code(void *img, size_t *size, int for_data,
 			     int old_baud, int new_baud)
 {
-	uint32_t codesz = sizeof(kwboot_baud_code);
 	struct main_hdr_v1 *hdr = img;
+	uint32_t orig_datasz;
+	uint32_t codesz;
 	uint8_t *code;
 
-	if (pre) {
-		uint32_t presz = sizeof(kwboot_pre_baud_code);
-		uint32_t orig_datasz;
-
+	if (for_data) {
 		orig_datasz = le32_to_cpu(hdr->blocksize) - sizeof(uint32_t);
 
-		code = kwboot_img_grow_data_right(img, size, presz + codesz);
-
-		/*
-		 * We need to prepend code that loads lr register with original
-		 * value of hdr->execaddr. We do this by putting the original
-		 * exec address before the code that loads it relatively from
-		 * it's beginning.
-		 * Afterwards we change the exec address to this code (which is
-		 * at offset 4, because the first 4 bytes contain the original
-		 * exec address).
-		 */
-		memcpy(code, kwboot_pre_baud_code, presz);
-		*(uint32_t *)code = hdr->execaddr;
-
-		hdr->execaddr = cpu_to_le32(le32_to_cpu(hdr->destaddr) +
-					    orig_datasz + 4);
-
-		code += presz;
+		codesz = sizeof(kwboot_baud_code) +
+			 sizeof(kwboot_baud_code_data_jump);
+		code = kwboot_img_grow_data_right(img, size, codesz);
 	} else {
+		codesz = sizeof(kwboot_baud_code_binhdr_pre) +
+			 sizeof(kwboot_baud_code) +
+			 sizeof(kwboot_baud_code_binhdr_post);
 		code = kwboot_add_bin_ohdr_v1(img, size, codesz);
+
+		codesz = sizeof(kwboot_baud_code_binhdr_pre);
+		memcpy(code, kwboot_baud_code_binhdr_pre, codesz);
+		code += codesz;
 	}
 
-	memcpy(code, kwboot_baud_code, codesz - 8);
-	*(uint32_t *)(code + codesz - 8) = cpu_to_le32(old_baud);
-	*(uint32_t *)(code + codesz - 4) = cpu_to_le32(new_baud);
+	codesz = sizeof(kwboot_baud_code) - 2 * sizeof(uint32_t);
+	memcpy(code, kwboot_baud_code, codesz);
+	code += codesz;
+	*(uint32_t *)code = cpu_to_le32(old_baud);
+	code += sizeof(uint32_t);
+	*(uint32_t *)code = cpu_to_le32(new_baud);
+	code += sizeof(uint32_t);
+
+	if (for_data) {
+		codesz = sizeof(kwboot_baud_code_data_jump) - sizeof(uint32_t);
+		memcpy(code, kwboot_baud_code_data_jump, codesz);
+		code += codesz;
+		*(uint32_t *)code = hdr->execaddr;
+		code += sizeof(uint32_t);
+		hdr->execaddr = cpu_to_le32(le32_to_cpu(hdr->destaddr) + orig_datasz);
+	} else {
+		codesz = sizeof(kwboot_baud_code_binhdr_post);
+		memcpy(code, kwboot_baud_code_binhdr_post, codesz);
+		code += codesz;
+	}
 }
 
 static int
@@ -1729,10 +1744,13 @@ main(int argc, char **argv)
 		baudrate = 0;
 	else
 		/* ensure we have enough space for baudrate change code */
-		after_img_rsv += KWBOOT_BAUDRATE_BIN_HEADER_SZ +
+		after_img_rsv += sizeof(struct opt_hdr_v1) + 8 + 16 +
+				 sizeof(kwboot_baud_code_binhdr_pre) +
+				 sizeof(kwboot_baud_code) +
+				 sizeof(kwboot_baud_code_binhdr_post) +
 				 KWBOOT_XM_BLKSZ +
-				 sizeof(kwboot_pre_baud_code) +
 				 sizeof(kwboot_baud_code) +
+				 sizeof(kwboot_baud_code_data_jump) +
 				 KWBOOT_XM_BLKSZ;
 
 	if (imgpath) {
-- 
2.32.0

Re: [PATCH u-boot-marvell 1/5] tools: kwboot: Fix sending retry of last header packet

[Stefan Roese]

On 27.10.21 20:56, Marek Behún wrote:
> From: Pali Rohár <pali@kernel.org>
> 
> After the trasfer of last header packet, it is possible that baudrate
> change pattern is received, and also that NAK byte is received so that
> the packet should be sent again.
> 
> Thus we should not clear the baudrate change state when sending retry
> of that packet.
> 
> Move code for initializing state variables from kwboot_xm_recv_reply()
> to kwboot_xm_sendblock().
> 
> Signed-off-by: Pali Rohár <pali@kernel.org>
> Reviewed-by: Marek Behún <marek.behun@nic.cz>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   tools/kwboot.c | 7 ++-----
>   1 file changed, 2 insertions(+), 5 deletions(-)
> 
> diff --git a/tools/kwboot.c b/tools/kwboot.c
> index bb7cae9f05..b2c48812c3 100644
> --- a/tools/kwboot.c
> +++ b/tools/kwboot.c
> @@ -859,11 +859,6 @@ kwboot_xm_recv_reply(int fd, char *c, int nak_on_non_xm,
>   	uint64_t recv_until = _now() + timeout;
>   	int rc;
>   
> -	if (non_xm_print)
> -		*non_xm_print = 0;
> -	if (baud_changed)
> -		*baud_changed = 0;
> -
>   	while (1) {
>   		rc = kwboot_tty_recv(fd, c, 1, timeout);
>   		if (rc) {
> @@ -929,6 +924,8 @@ kwboot_xm_sendblock(int fd, struct kwboot_block *block, int allow_non_xm,
>   	char c;
>   
>   	*done_print = 0;
> +	non_xm_print = 0;
> +	baud_changed = 0;
>   
>   	retries = 0;
>   	do {
> 


Viele Grüße,
Stefan

-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH u-boot-marvell 2/5] tools: kwboot: Do not call tcdrain() after each sent packet

[Stefan Roese]

On 27.10.21 20:56, Marek Behún wrote:
> From: Pali Rohár <pali@kernel.org>
> 
> Kwboot puts each xmodem packet to kernel queue, then waits until all bytes
> of that packet are transmitted over UART and then waits for xmodem reply
> until it is received into kernel queue.
> 
> If some reply is received during the time we are waiting until all bytes
> are transmitted, then kernel puts them into the queue and returns it to
> kwboot in next read() call.
> 
> So there is no need to wait (with tcdrain() function) until all bytes from
> xmodem packet are transmitted over UART, since any reply received either
> during that time or after is returned to kwboot with the next read().
> 
> Therefore do not call tcdrain() after each xmodem packet sent. Instead
> directly wait for any reply after putting xmodem packet into write kernel
> queue.
> 
> This change could speed up xmodem transfer in case tcdrain() function waits
> for a longer time.
> 
> Signed-off-by: Pali Rohár <pali@kernel.org>
> Reviewed-by: Marek Behún <marek.behun@nic.cz>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   tools/kwboot.c | 13 ++++++++-----
>   1 file changed, 8 insertions(+), 5 deletions(-)
> 
> diff --git a/tools/kwboot.c b/tools/kwboot.c
> index b2c48812c3..a6bfd3d4ce 100644
> --- a/tools/kwboot.c
> +++ b/tools/kwboot.c
> @@ -404,7 +404,7 @@ out:
>   }
>   
>   static int
> -kwboot_tty_send(int fd, const void *buf, size_t len)
> +kwboot_tty_send(int fd, const void *buf, size_t len, int nodrain)
>   {
>   	if (!buf)
>   		return 0;
> @@ -412,13 +412,16 @@ kwboot_tty_send(int fd, const void *buf, size_t len)
>   	if (kwboot_write(fd, buf, len) < 0)
>   		return -1;
>   
> +	if (nodrain)
> +		return 0;
> +
>   	return tcdrain(fd);
>   }
>   
>   static int
>   kwboot_tty_send_char(int fd, unsigned char c)
>   {
> -	return kwboot_tty_send(fd, &c, 1);
> +	return kwboot_tty_send(fd, &c, 1, 0);
>   }
>   
>   static speed_t
> @@ -705,7 +708,7 @@ kwboot_bootmsg(int tty, void *msg)
>   			break;
>   
>   		for (count = 0; count < 128; count++) {
> -			rc = kwboot_tty_send(tty, msg, 8);
> +			rc = kwboot_tty_send(tty, msg, 8, 0);
>   			if (rc) {
>   				usleep(msg_req_delay * 1000);
>   				continue;
> @@ -737,7 +740,7 @@ kwboot_debugmsg(int tty, void *msg)
>   		if (rc)
>   			break;
>   
> -		rc = kwboot_tty_send(tty, msg, 8);
> +		rc = kwboot_tty_send(tty, msg, 8, 0);
>   		if (rc) {
>   			usleep(msg_req_delay * 1000);
>   			continue;
> @@ -929,7 +932,7 @@ kwboot_xm_sendblock(int fd, struct kwboot_block *block, int allow_non_xm,
>   
>   	retries = 0;
>   	do {
> -		rc = kwboot_tty_send(fd, block, sizeof(*block));
> +		rc = kwboot_tty_send(fd, block, sizeof(*block), 1);
>   		if (rc)
>   			return rc;
>   
> 


Viele Grüße,
Stefan

-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH u-boot-marvell 3/5] tools: kwboot: Increase delay after changing baudrate in ARM code

[Stefan Roese]

On 27.10.21 20:57, Marek Behún wrote:
> From: Pali Rohár <pali@kernel.org>
> 
> Increase loop cycles from 600000 to 2998272, which should increase delay
> from 1ms to about 5ms on 1200 MHz CPU.
> 
> The Number 2998272 was chosen as the nearest value around 3000000 which can
> be encoded into one ARM mov instruction. It avoids usage of movt instruction
> which is not supported by ARMv5te cores.
> 
> Signed-off-by: Pali Rohár <pali@kernel.org>
> Reviewed-by: Marek Behún <marek.behun@nic.cz>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   tools/kwboot.c | 18 +++++++++---------
>   1 file changed, 9 insertions(+), 9 deletions(-)
> 
> diff --git a/tools/kwboot.c b/tools/kwboot.c
> index a6bfd3d4ce..84294cadfe 100644
> --- a/tools/kwboot.c
> +++ b/tools/kwboot.c
> @@ -119,7 +119,7 @@ static unsigned char kwboot_baud_code[] = {
>   				/* ;   writel(UART_BASE + DLL, new_dll);      */
>   				/* ;   writel(UART_BASE + DLH, new_dlh);      */
>   				/* ;   writel(UART_BASE + LCR, lcr & ~DLAB);  */
> -				/* ;   msleep(1);                             */
> +				/* ;   msleep(5);                             */
>   				/* ;   return 0;                              */
>   				/* ; }                                        */
>   
> @@ -130,7 +130,7 @@ static unsigned char kwboot_baud_code[] = {
>   	0x01, 0x00, 0x4d, 0xe3, /* movt  r0, #0xd001                          */
>   
>   				/*  ; r2 = address of preamble string         */
> -	0xd0, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
> +	0xcc, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
>   
>   				/*  ; Send preamble string over UART          */
>   				/* .Lloop_preamble:                           */
> @@ -177,15 +177,15 @@ static unsigned char kwboot_baud_code[] = {
>   
>   				/*  ; Read old baudrate value                 */
>   				/*  ; r2 = old_baudrate                       */
> -	0x8c, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
> +	0x88, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
>   
>   				/*  ; Calculate base clock                    */
>   				/*  ; r1 = r2 * r1                            */
>   	0x92, 0x01, 0x01, 0xe0, /* mul   r1, r2, r1                           */
>   
>   				/*  ; Read new baudrate value                 */
> -				/*  ; r2 = baudrate                           */
> -	0x88, 0x20, 0x9f, 0xe5, /* ldr   r2, baudrate                         */
> +				/*  ; r2 = new_baudrate                       */
> +	0x84, 0x20, 0x9f, 0xe5, /* ldr   r2, new_baudrate                     */
>   
>   				/*  ; Calculate new Divisor Latch             */
>   				/*  ; r1 = DIV_ROUND(r1, r2) =                */
> @@ -225,10 +225,10 @@ static unsigned char kwboot_baud_code[] = {
>   	0x80, 0x10, 0xc1, 0xe3, /* bic   r1, r1, #0x80                        */
>   	0x0c, 0x10, 0x80, 0xe5, /* str   r1, [r0, #0x0c]                      */
>   
> -				/*  ; Sleep 1ms ~~ 600000 cycles at 1200 MHz  */
> -				/*  ; r1 = 600000                             */
> -	0x9f, 0x1d, 0xa0, 0xe3, /* mov   r1, #0x27c0                          */
> -	0x09, 0x10, 0x40, 0xe3, /* movt  r1, #0x0009                          */
> +				/*  ; Loop 0x2dc000 (2998272) cycles          */
> +				/*  ; which is about 5ms on 1200 MHz CPU      */
> +				/*  ; r1 = 0x2dc000                           */
> +	0xb7, 0x19, 0xa0, 0xe3, /* mov   r1, #0x2dc000                        */
>   				/* .Lloop_sleep:                              */
>   	0x01, 0x10, 0x41, 0xe2, /* sub   r1, r1, #1                           */
>   	0x00, 0x00, 0x51, 0xe3, /* cmp   r1, #0                               */
> 


Viele Grüße,
Stefan

-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH u-boot-marvell 4/5] tools: kwboot: Replace ARM mov + movt instruction pair by mov + orr

[Stefan Roese]

On 27.10.21 20:57, Marek Behún wrote:
> From: Pali Rohár <pali@kernel.org>
> 
> Older Armada SoCs have custom ARMv5te compatible core which does not
> support movt instruction. So replace mov + movt instruction pair used for
> immediate move construction by mov + orr instructions which are supported
> also by ARMv5te.
> 
> After this change kwboot ARM code should be compatible with any 32-bit ARM
> core compatible by ARMv2 or new. At least GNU AS does not throw any error
> or warning.
> 
> Signed-off-by: Pali Rohár <pali@kernel.org>
> Reviewed-by: Marek Behún <marek.behun@nic.cz>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   tools/kwboot.c | 4 ++--
>   1 file changed, 2 insertions(+), 2 deletions(-)
> 
> diff --git a/tools/kwboot.c b/tools/kwboot.c
> index 84294cadfe..62c218ef64 100644
> --- a/tools/kwboot.c
> +++ b/tools/kwboot.c
> @@ -126,8 +126,8 @@ static unsigned char kwboot_baud_code[] = {
>   	0xfe, 0x5f, 0x2d, 0xe9, /* push  { r1 - r12, lr }                     */
>   
>   				/*  ; r0 = UART_BASE                          */
> -	0x02, 0x0a, 0xa0, 0xe3, /* mov   r0, #0x2000                          */
> -	0x01, 0x00, 0x4d, 0xe3, /* movt  r0, #0xd001                          */
> +	0x0d, 0x02, 0xa0, 0xe3, /* mov   r0, #0xd0000000                      */
> +	0x12, 0x0a, 0x80, 0xe3, /* orr   r0, r0, #0x12000                     */
>   
>   				/*  ; r2 = address of preamble string         */
>   	0xcc, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
> 


Viele Grüße,
Stefan

-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH u-boot-marvell 5/5] tools: kwboot: Do not use stack when setting baudrate back to default value

[Stefan Roese]

On 27.10.21 20:57, Marek Behún wrote:
> From: Pali Rohár <pali@kernel.org>
> 
> The ARM code we inject into the image to change baudrate back to the
> default value of 115200 Baud, which is run after successful UART transfer
> of the whole image, cannot use stack as at this stage stack pointer is not
> initialized yet.
> 
> Stack can only be used when BootROM is executing binary header, to
> preserve state of registers, since BootROM expects that.
> 
> Change the ARM baudrate code to not use stack at all and put binary
> header specific pre + post code (which stores and restores registers) into
> separate arrays.
> 
> The baudrate change code now jumps at it's end and expects that there is
> either code which returns to the BootROM or jumps to the original exec
> address.
> 
> Signed-off-by: Pali Rohár <pali@kernel.org>
> Reviewed-by: Marek Behún <marek.behun@nic.cz>

Reviewed-by: Stefan Roese <sr@denx.de>

Thanks,
Stefan

> ---
>   tools/kwboot.c | 112 ++++++++++++++++++++++++++++---------------------
>   1 file changed, 65 insertions(+), 47 deletions(-)
> 
> diff --git a/tools/kwboot.c b/tools/kwboot.c
> index 62c218ef64..359b43c0d8 100644
> --- a/tools/kwboot.c
> +++ b/tools/kwboot.c
> @@ -78,14 +78,7 @@ struct kwboot_block {
>   #define KWBOOT_BLK_RSP_TIMEO 1000 /* ms */
>   #define KWBOOT_HDR_RSP_TIMEO 10000 /* ms */
>   
> -/* ARM code making baudrate changing function return to original exec address */
> -static unsigned char kwboot_pre_baud_code[] = {
> -				/* exec_addr:                                 */
> -	0x00, 0x00, 0x00, 0x00, /* .word 0                                    */
> -	0x0c, 0xe0, 0x1f, 0xe5, /* ldr lr, exec_addr                          */
> -};
> -
> -/* ARM code for binary header injection to change baudrate */
> +/* ARM code to change baudrate */
>   static unsigned char kwboot_baud_code[] = {
>   				/* ; #define UART_BASE 0xd0012000             */
>   				/* ; #define THR       0x00                   */
> @@ -123,14 +116,12 @@ static unsigned char kwboot_baud_code[] = {
>   				/* ;   return 0;                              */
>   				/* ; }                                        */
>   
> -	0xfe, 0x5f, 0x2d, 0xe9, /* push  { r1 - r12, lr }                     */
> -
>   				/*  ; r0 = UART_BASE                          */
>   	0x0d, 0x02, 0xa0, 0xe3, /* mov   r0, #0xd0000000                      */
>   	0x12, 0x0a, 0x80, 0xe3, /* orr   r0, r0, #0x12000                     */
>   
>   				/*  ; r2 = address of preamble string         */
> -	0xcc, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
> +	0xc8, 0x20, 0x8f, 0xe2, /* adr   r2, preamble                         */
>   
>   				/*  ; Send preamble string over UART          */
>   				/* .Lloop_preamble:                           */
> @@ -177,7 +168,7 @@ static unsigned char kwboot_baud_code[] = {
>   
>   				/*  ; Read old baudrate value                 */
>   				/*  ; r2 = old_baudrate                       */
> -	0x88, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
> +	0x84, 0x20, 0x9f, 0xe5, /* ldr   r2, old_baudrate                     */
>   
>   				/*  ; Calculate base clock                    */
>   				/*  ; r1 = r2 * r1                            */
> @@ -185,7 +176,7 @@ static unsigned char kwboot_baud_code[] = {
>   
>   				/*  ; Read new baudrate value                 */
>   				/*  ; r2 = new_baudrate                       */
> -	0x84, 0x20, 0x9f, 0xe5, /* ldr   r2, new_baudrate                     */
> +	0x80, 0x20, 0x9f, 0xe5, /* ldr   r2, new_baudrate                     */
>   
>   				/*  ; Calculate new Divisor Latch             */
>   				/*  ; r1 = DIV_ROUND(r1, r2) =                */
> @@ -234,9 +225,7 @@ static unsigned char kwboot_baud_code[] = {
>   	0x00, 0x00, 0x51, 0xe3, /* cmp   r1, #0                               */
>   	0xfc, 0xff, 0xff, 0x1a, /* bne   .Lloop_sleep                         */
>   
> -				/*  ; Return 0 - no error                     */
> -	0x00, 0x00, 0xa0, 0xe3, /* mov   r0, #0                               */
> -	0xfe, 0x9f, 0xbd, 0xe8, /* pop   { r1 - r12, pc }                     */
> +	0x05, 0x00, 0x00, 0xea, /* b     end                                  */
>   
>   				/*  ; Preamble string                         */
>   				/* preamble:                                  */
> @@ -252,10 +241,29 @@ static unsigned char kwboot_baud_code[] = {
>   				/*  ; Placeholder for new baudrate value      */
>   				/* new_baudrate:                              */
>   	0x00, 0x00, 0x00, 0x00, /* .word 0                                    */
> +
> +				/* end:                                       */
> +};
> +
> +/* ARM code for storing registers for future returning back to the bootrom */
> +static unsigned char kwboot_baud_code_binhdr_pre[] = {
> +	0xfe, 0x5f, 0x2d, 0xe9, /* push  { r1 - r12, lr }                     */
>   };
>   
> -#define KWBOOT_BAUDRATE_BIN_HEADER_SZ (sizeof(kwboot_baud_code) + \
> -				       sizeof(struct opt_hdr_v1) + 8 + 16)
> +/* ARM code for returning back to the bootrom */
> +static unsigned char kwboot_baud_code_binhdr_post[] = {
> +				/*  ; Return 0 - no error                     */
> +	0x00, 0x00, 0xa0, 0xe3, /* mov   r0, #0                               */
> +	0xfe, 0x9f, 0xbd, 0xe8, /* pop   { r1 - r12, pc }                     */
> +};
> +
> +/* ARM code for jumping to the original image exec_addr */
> +static unsigned char kwboot_baud_code_data_jump[] = {
> +	0x04, 0xf0, 0x1f, 0xe5, /* ldr   pc, exec_addr                        */
> +				/*  ; Placeholder for exec_addr               */
> +				/* exec_addr:                                 */
> +	0x00, 0x00, 0x00, 0x00, /* .word 0                                    */
> +};
>   
>   static const char kwb_baud_magic[16] = "$baudratechange";
>   
> @@ -1409,44 +1417,51 @@ kwboot_add_bin_ohdr_v1(void *img, size_t *size, uint32_t binsz)
>   }
>   
>   static void
> -_inject_baudrate_change_code(void *img, size_t *size, int pre,
> +_inject_baudrate_change_code(void *img, size_t *size, int for_data,
>   			     int old_baud, int new_baud)
>   {
> -	uint32_t codesz = sizeof(kwboot_baud_code);
>   	struct main_hdr_v1 *hdr = img;
> +	uint32_t orig_datasz;
> +	uint32_t codesz;
>   	uint8_t *code;
>   
> -	if (pre) {
> -		uint32_t presz = sizeof(kwboot_pre_baud_code);
> -		uint32_t orig_datasz;
> -
> +	if (for_data) {
>   		orig_datasz = le32_to_cpu(hdr->blocksize) - sizeof(uint32_t);
>   
> -		code = kwboot_img_grow_data_right(img, size, presz + codesz);
> -
> -		/*
> -		 * We need to prepend code that loads lr register with original
> -		 * value of hdr->execaddr. We do this by putting the original
> -		 * exec address before the code that loads it relatively from
> -		 * it's beginning.
> -		 * Afterwards we change the exec address to this code (which is
> -		 * at offset 4, because the first 4 bytes contain the original
> -		 * exec address).
> -		 */
> -		memcpy(code, kwboot_pre_baud_code, presz);
> -		*(uint32_t *)code = hdr->execaddr;
> -
> -		hdr->execaddr = cpu_to_le32(le32_to_cpu(hdr->destaddr) +
> -					    orig_datasz + 4);
> -
> -		code += presz;
> +		codesz = sizeof(kwboot_baud_code) +
> +			 sizeof(kwboot_baud_code_data_jump);
> +		code = kwboot_img_grow_data_right(img, size, codesz);
>   	} else {
> +		codesz = sizeof(kwboot_baud_code_binhdr_pre) +
> +			 sizeof(kwboot_baud_code) +
> +			 sizeof(kwboot_baud_code_binhdr_post);
>   		code = kwboot_add_bin_ohdr_v1(img, size, codesz);
> +
> +		codesz = sizeof(kwboot_baud_code_binhdr_pre);
> +		memcpy(code, kwboot_baud_code_binhdr_pre, codesz);
> +		code += codesz;
>   	}
>   
> -	memcpy(code, kwboot_baud_code, codesz - 8);
> -	*(uint32_t *)(code + codesz - 8) = cpu_to_le32(old_baud);
> -	*(uint32_t *)(code + codesz - 4) = cpu_to_le32(new_baud);
> +	codesz = sizeof(kwboot_baud_code) - 2 * sizeof(uint32_t);
> +	memcpy(code, kwboot_baud_code, codesz);
> +	code += codesz;
> +	*(uint32_t *)code = cpu_to_le32(old_baud);
> +	code += sizeof(uint32_t);
> +	*(uint32_t *)code = cpu_to_le32(new_baud);
> +	code += sizeof(uint32_t);
> +
> +	if (for_data) {
> +		codesz = sizeof(kwboot_baud_code_data_jump) - sizeof(uint32_t);
> +		memcpy(code, kwboot_baud_code_data_jump, codesz);
> +		code += codesz;
> +		*(uint32_t *)code = hdr->execaddr;
> +		code += sizeof(uint32_t);
> +		hdr->execaddr = cpu_to_le32(le32_to_cpu(hdr->destaddr) + orig_datasz);
> +	} else {
> +		codesz = sizeof(kwboot_baud_code_binhdr_post);
> +		memcpy(code, kwboot_baud_code_binhdr_post, codesz);
> +		code += codesz;
> +	}
>   }
>   
>   static int
> @@ -1729,10 +1744,13 @@ main(int argc, char **argv)
>   		baudrate = 0;
>   	else
>   		/* ensure we have enough space for baudrate change code */
> -		after_img_rsv += KWBOOT_BAUDRATE_BIN_HEADER_SZ +
> +		after_img_rsv += sizeof(struct opt_hdr_v1) + 8 + 16 +
> +				 sizeof(kwboot_baud_code_binhdr_pre) +
> +				 sizeof(kwboot_baud_code) +
> +				 sizeof(kwboot_baud_code_binhdr_post) +
>   				 KWBOOT_XM_BLKSZ +
> -				 sizeof(kwboot_pre_baud_code) +
>   				 sizeof(kwboot_baud_code) +
> +				 sizeof(kwboot_baud_code_data_jump) +
>   				 KWBOOT_XM_BLKSZ;
>   
>   	if (imgpath) {
> 


Viele Grüße,
Stefan

-- 
DENX Software Engineering GmbH,      Managing Director: Wolfgang Denk
HRB 165235 Munich, Office: Kirchenstr.5, D-82194 Groebenzell, Germany
Phone: (+49)-8142-66989-51 Fax: (+49)-8142-66989-80 Email: sr@denx.de

Re: [PATCH u-boot-marvell 0/5] kwboot fix for AXP + some others

[Stefan Roese]

On 27.10.21 20:56, Marek Behún wrote:
> From: Marek Behún <marek.behun@nic.cz>
> 
> Hi Stefan,
> 
> here are some more Pali's changes for kwboot, reviewed by me. The last
> one should fix boot on AXP.
> 
> Marek
> 
> Pali Rohár (5):
>    tools: kwboot: Fix sending retry of last header packet
>    tools: kwboot: Do not call tcdrain() after each sent packet
>    tools: kwboot: Increase delay after changing baudrate in ARM code
>    tools: kwboot: Replace ARM mov + movt instruction pair by mov + orr
>    tools: kwboot: Do not use stack when setting baudrate back to default
>      value
> 
>   tools/kwboot.c | 148 +++++++++++++++++++++++++++----------------------
>   1 file changed, 83 insertions(+), 65 deletions(-)
> 

Applied to u-boot-marvell/master

Thanks,
Stefan