[flashrom] [PATCH] Added support to control the GPIOs in the ft222_spi driver from the command line

Reggie McMurtrey reggie.mcmurtrey at gmail.com
Sat Feb 21 16:34:23 CET 2015


Thanks for the feedback Stefan, Trying again :)

Signed-off-by: Reggie McMurtrey <reggie.mcmurtrey at gmail.com>

---
 flashrom.8.tmpl | 17 ++++++++++++
 ft2232_spi.c    | 81
++++++++++++++++++++++++++++++++++++++++++++-------------
 2 files changed, 80 insertions(+), 18 deletions(-)

diff --git a/flashrom.8.tmpl b/flashrom.8.tmpl
index 8b5f98c..40d8d84 100644
--- a/flashrom.8.tmpl
+++ b/flashrom.8.tmpl
@@ -681,6 +681,23 @@ parameter with the
 .B "  flashrom \-p ft2232_spi:divisor=div"
 .sp
 syntax.
+.sp
+If the ft2232_spi device has been integrated into a custom board, or you
are using a breakout cable
+like the C232HM cable from FTDI you can control the upper bits of the port
( bits 7-4  ) via the
+.sp
+.B "  flashrom \-p ft2232_spi:gpio_dir=dir,gpio_value=value"
+.sp
+syntax where
+.B dir
+is a mask defining the directions of the port's upper bits.
+.B value
+is the value of the bits for the port. Only bits defined as outputs with
the
+.B dir
+argument are affected. For example
+.B gpio_dir=0xf0,gpio_value=0x80
+would set the four upper bits of the port as outputs with bit 7 as high,
and bits 6 through 4 as
+low. This could be useful for controlling write protects, muxing options,
or other in
+circuit programming related features.
 .SS
 .BR "serprog " programmer
 A mandatory parameter specifies either a serial
diff --git a/ft2232_spi.c b/ft2232_spi.c
index 4e23797..75cb772 100644
--- a/ft2232_spi.c
+++ b/ft2232_spi.c
@@ -93,6 +93,9 @@ const struct dev_entry devs_ft2232spi[] = {
  */
 static uint8_t cs_bits = 0x08;
 static uint8_t pindir = 0x0b;
+static uint8_t gpio_dir = 0x00;
+static uint8_t gpio_value = 0x00;
+static uint8_t adapter_mask = 0x00;
 static struct ftdi_context ftdic_context;

 static const char *get_ft2232_devicename(int ft2232_vid, int ft2232_type)
@@ -204,10 +207,11 @@ int ft2232_spi_init(void)
  ft2232_type = AMONTEC_JTAGKEY_PID;
  channel_count = 2;
  /* JTAGkey(2) needs to enable its output via Bit4 / GPIOL0
- *  value: 0x18  OE=high, CS=high, DI=low, DO=low, SK=low
- *    dir: 0x1b  OE=output, CS=output, DI=input, DO=output, SK=output */
- cs_bits = 0x18;
- pindir = 0x1b;
+ *  value: 0x10  OE=high, Reserved (CS), Reserved (DI), Reserved (DO),
Reserved (SK)
+ *    dir: 0x10  OE=output, Reserved (CS), Reserved (DI), Reserved (DO),
Reserved (SK) */
+ gpio_dir = 0x10;
+ gpio_value = 0x10;
+ adapter_mask = 0x10;
  } else if (!strcasecmp(arg, "picotap")) {
  ft2232_vid = GOEPEL_VID;
  ft2232_type = GOEPEL_PICOTAP_PID;
@@ -224,8 +228,9 @@ int ft2232_spi_init(void)
  /* In its default configuration it is a jtagkey clone */
  ft2232_type = FTDI_FT2232H_PID;
  channel_count = 2;
- cs_bits = 0x18;
- pindir = 0x1b;
+ gpio_dir = 0x10;
+ gpio_value = 0x10;
+ adapter_mask = 0x10;
  } else if (!strcasecmp(arg, "openmoko")) {
  ft2232_vid = FIC_VID;
  ft2232_type = OPENMOKO_DBGBOARD_PID;
@@ -235,10 +240,11 @@ int ft2232_spi_init(void)
  ft2232_type = OLIMEX_ARM_OCD_PID;
  channel_count = 2;
  /* arm-usb-ocd(-h) has an output buffer that needs to be enabled by
pulling ADBUS4 low.
- *  value: 0x08  #OE=low, CS=high, DI=low, DO=low, SK=low
- *    dir: 0x1b  #OE=output, CS=output, DI=input, DO=output, SK=output */
- cs_bits = 0x08;
- pindir = 0x1b;
+ *  value: 0x00  #OE=low, Reserved (CS), Reserved (DI), Reserved (DO),
Reserved (SK)
+ *    dir: 0x10  #OE=output, Reserved (CS), Reserved (DI), Reserved (DO),
Reserved (SK) */
+ gpio_dir = 0x10;
+ gpio_value = 0x00;
+ adapter_mask = 0x10;
  } else if (!strcasecmp(arg, "arm-usb-tiny")) {
  ft2232_vid = OLIMEX_VID;
  ft2232_type = OLIMEX_ARM_TINY_PID;
@@ -248,8 +254,9 @@ int ft2232_spi_init(void)
  ft2232_type = OLIMEX_ARM_OCD_H_PID;
  channel_count = 2;
  /* See arm-usb-ocd */
- cs_bits = 0x08;
- pindir = 0x1b;
+ gpio_dir = 0x10;
+ gpio_value = 0x00;
+ adapter_mask = 0x10;
  } else if (!strcasecmp(arg, "arm-usb-tiny-h")) {
  ft2232_vid = OLIMEX_VID;
  ft2232_type = OLIMEX_ARM_TINY_H_PID;
@@ -262,6 +269,38 @@ int ft2232_spi_init(void)
  }
  free(arg);

+ arg = extract_programmer_param("gpio_dir");
+ if (arg && strlen(arg)) {
+ unsigned int temp = 0;
+ char *endptr;
+ temp = strtoul(arg, &endptr, 16);
+ if (*endptr || temp > 0xFF) {
+ msg_perr("Error: Invalid GPIO direction mask \"%s\".\n"
+ "Valid values are between 0x0 and 0xff (the lower four bits are
ignored).\n", arg);
+ free(arg);
+ return -2;
+ } else {
+ gpio_dir = ((uint32_t) temp & (0xf0 & ~adapter_mask)) | gpio_dir;
+ }
+ }
+ free(arg);
+
+ arg = extract_programmer_param("gpio_value");
+ if (arg && strlen(arg)) {
+ unsigned int temp = 0;
+ char *endptr;
+ temp = strtoul(arg, &endptr, 16);
+ if (*endptr || temp > 0xFF) {
+ msg_perr("Error: Invalid GPIO value mask \"%s\".\n"
+ "Valid values are between 0x0 and 0xff (the lower four bits are
ignored).\n", arg);
+ free(arg);
+ return -2;
+ } else {
+ gpio_value = ((uint32_t) temp & (0xf0 & ~adapter_mask)) | gpio_value;
+ }
+ }
+ free(arg);
+
  arg = extract_programmer_param("port");
  if (arg) {
  switch (toupper((unsigned char)*arg)) {
@@ -391,9 +430,11 @@ int ft2232_spi_init(void)
  }

  msg_pdbg("Set data bits\n");
+ msg_pspew("Value: 0x%02x\n", cs_bits | gpio_value);
+ msg_pspew("Dir: 0x%02x\n", pindir | gpio_dir);
  buf[0] = SET_BITS_LOW;
- buf[1] = cs_bits;
- buf[2] = pindir;
+ buf[1] = cs_bits | gpio_value;
+ buf[2] = pindir | gpio_dir;
  if (send_buf(ftdic, buf, 3)) {
  ret = -8;
  goto ftdi_err;
@@ -446,9 +487,11 @@ static int ft2232_spi_send_command(struct flashctx
*flash,
  * operations.
  */
  msg_pspew("Assert CS#\n");
+ msg_pspew("Value: 0x%02x\n", ~cs_bits & gpio_value);
+ msg_pspew("Dir: 0x%02x\n", pindir | gpio_dir);
  buf[i++] = SET_BITS_LOW;
- buf[i++] = 0 & ~cs_bits; /* assertive */
- buf[i++] = pindir;
+ buf[i++] = ~cs_bits & gpio_value; /* assertive */
+ buf[i++] = pindir | gpio_dir;

  if (writecnt) {
  buf[i++] = MPSSE_DO_WRITE | MPSSE_WRITE_NEG;
@@ -487,9 +530,11 @@ static int ft2232_spi_send_command(struct flashctx
*flash,
  }

  msg_pspew("De-assert CS#\n");
+ msg_pspew("Value: 0x%02x\n", cs_bits | gpio_value);
+ msg_pspew("Dir: 0x%02x\n", pindir | gpio_dir);
  buf[i++] = SET_BITS_LOW;
- buf[i++] = cs_bits;
- buf[i++] = pindir;
+ buf[i++] = cs_bits | gpio_value;
+ buf[i++] = pindir | gpio_dir;
  ret = send_buf(ftdic, buf, i);
  failed |= ret;
  if (ret)
-- 
1.9.1


On Sat, Feb 21, 2015 at 4:20 AM, Stefan Tauner <
stefan.tauner at alumni.tuwien.ac.at> wrote:

> On Fri, 20 Feb 2015 18:21:09 -0600
> Reggie McMurtrey <reggie.mcmurtrey at gmail.com> wrote:
>
> > This is my first time submitting a patch to a project, so let me know
> > if I have done something wrong here.
> >
> >
> > I commonly use the C232HM cable from FTDI to program various parts of
> > boards I design.  This one cable allows me to program SPI, I2C, and
> > JTAG parts.  Simply adding some muxes to the design I have a single
> > header to attach this cable to.  The patch below allows control of the
> > muxes from the GPIOL3-GPIOL0 signals on the cable and disable write
> > protection.  I figured other people may like this capability if they
> > are using a similar breakout cable of have designed the ft2232 chip
> > into their design.
> >
> > Signed-off-by: Reggie McMurtrey <reggie.mcmurtrey at gmail.com>
> >
> > @@ -262,6 +269,18 @@ int ft2232_spi_init(void)
> >       }
> >       free(arg);
> >
> > +     arg = extract_programmer_param("gpio_dir");
> > +     if (arg) {
> > +             gpio_dir = (strtoul(arg, 0, 16) & (0xf0 & ~adapter_mask))
> | gpio_dir;
> > +     }
> > +     free(arg);
> > +
> > +     arg = extract_programmer_param("gpio_value");
> > +     if (arg) {
> > +             gpio_value = (strtoul(arg, 0, 16) & (0xf0 &
> ~adapter_mask)) | gpio_value;
> > +     }
> > +     free(arg);
> > +
> >       arg = extract_programmer_param("port");
> >       if (arg) {
> >               switch (toupper((unsigned char)*arg)) {
>
> Hello Reggie,
>
> thanks for your patch! I had no time to review it in detail but we
> definitely want some error/wrong input handling for the parameter
> parsing above. There are various similar checks throughout the code
> base that you may wanna look at... not all are perfect yet though. If
> you have any doubts then please refer to the strtol(3) manpage.
>
> A more detailed review will follow when that's fixed.
> --
> Kind regards/Mit freundlichen Grüßen, Stefan Tauner
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.flashrom.org/pipermail/flashrom/attachments/20150221/701650c9/attachment.html>


More information about the flashrom mailing list