flashrom December 2012

flashrom@flashrom.org

33 participants
119 discussions

Re: [flashrom] MX25L1605DM2I-12G with Buspirate
by Andrew Goodbody 06 Dec '12

06 Dec '12

On 01/12/12 21:11, Daniel . wrote: > So at last I just need to know now how I can merge the needed > decompressed files back into a flashable ROM file. No, this is the wrong way. You do not need to decompress the .wph file. The .wph file in this case is 1.5MB long with some extra data and code stuck on the end. So the first thing to do is to cut the excess after 1.5MB off the end of that file. This 1.5MB is the top 1.5MB of the 2MB flash image**. Read the first 0.5MB from the flash device* and prefix it to the 1.5MB image to create the 2MB image. You can then program that 2MB image to the flash device. Andrew * If the failed update has corrupted this first 0.5MB on the flash device then you have big trouble. ** If you want to confirm that, disassemble the last 16 bytes of the 1.5MB image, there should be a real mode far jump.

1 0

Fwd: BPv3.6 w flashrom
by Idwer Vollering 05 Dec '12

05 Dec '12

Forwarding. Please "reply to all". ---------- Forwarded message ---------- From: Nabs <ch12011-7(a)yahoo.com.au> Date: 2012/12/5 Subject: Re: [flashrom] BPv3.6 w flashrom To: "vidwer(a)gmail.com" <vidwer(a)gmail.com> Just had a Eureka moment. I think what confused me was the fact that the cables were soldered onto the actual soic clip! So all I need to do is desolder and resolder them into their correct layouts, or desolder and use alligator clips instead? Is that right? Cheers, Nabs. ------------------------------ * From: * Idwer Vollering <vidwer(a)gmail.com>; * To: * Nabs <ch12011-7(a)yahoo.com.au>; * Cc: * flashrom(a)flashrom.org <flashrom(a)flashrom.org>; * Subject: * Re: [flashrom] BPv3.6 w flashrom * Sent: * Wed, Dec 5, 2012 12:05:04 PM 2012/12/5 Nabs <ch12011-7(a)yahoo.com.au>: > Hi, sorry for the trouble. I'm running bus piratev3.6, bootloaderv4.4, > firmwarev6.1 on flashrom-0.9.6.1, i plan to run the BP at 1M. > > When i run the command "flashrom -p buspirate_spi:dev=/dev/ttyUSB0 -Vr > oldbios.bin" > I get: > " > Found Generic flash chip "unknown SPI chip (RDID)" (0 kB, SPI) on > buspirate_spi. > Probing for Generic unknown SPI chip (REMS), 0 kB: probe_spi_rems: id1 0x0, > id2 0x0 > Found Generic flash chip "unknown SPI chip (RDID)" (0 kB, SPI). > === > This flash part has status NOT WORKING for operations: PROBE READ ERASE > WRITE > " etc... > > > I think i have the wiring wrong which is why i get that error, i wonder if > you could help me as its my first time? > I'm trying to use In-circuit programming with a soldered bios chip using an > 'SOIC 8 clip' and it has 8 wires. But the bus pirate has 10 pins. > > From another thread, these are the connections i need: > Bus Pirate "MISO" (Pin 1) <----- SPI Device "DO" (Wire/Pin 2) > Bus Pirate "CS" (Pin 2) -----> SPI Device "CS" (Wire/Pin 1) > Bus Pirate "MOSI" (Pin 3) -----> SPI Device "DI" (Wire/Pin 5) > Bus Pirate "CLK" (Pin 4) -----> SPI Device "CLK" (Wire/Pin 6) > Bus Pirate "GND" (Pin 10) <----> SPI Device "GND" (Wire/Pin 4) > > It appears to me that GND is pin 10 and that the wires of the SOIC8 clip > only contact pins 1-8, so how can this work? > > Also, could i just check the layout of the pins on the bus pirate if i'm > reading it correctly on BPv3.6: > 10 9 > 8 7 > 6 5 > 4 3 > 2 1 > > Corresponding to: > GND 3V3 > +5V ADC > VPU AUX > CLK MOSI > CS MISO Forget the existence of +5V, VPU, ADC and AUX for now ( http://dangerousprototypes.com/docs/Bus_Pirate_I/O_Pin_Descriptions ), connect the clip to the buspirate so that the clip passes the buspirate pins like mentioned in this post: http://www.coreboot.org/pipermail/flashrom/2012-November/010178.html This works for me, with a MX25L1605A soic8 chip: CS# (CS) [1 chip 8] VCC (bridge with buspirate pin 3+7) SO (MISO) [2 chip 7] HOLD# (+3V3) WP# (bridge with buspirate pin 7+8) [3 chip 6] SCLK (CLK) GND (GND) [4 chip 5] SI (MOSI) > > Or are the pin numbers read in the opposite direction? > ie > 9 10 > 7 8 > 5 6 > 3 4 > 1 2 > > Here's an image of my bus pirate if it helps: > http://dangerousprototypes.com/wp-content/media/2012/11/Bus-Pirate-v3.6inte… > > Your help is much appreciated and i appreciate your patience, Nabs. > > _______________________________________________ > flashrom mailing list > flashrom(a)flashrom.org > http://www.flashrom.org/mailman/listinfo/flashrom

1 0

Re: [flashrom] BPv3.6 w flashrom
by Idwer Vollering 05 Dec '12

05 Dec '12

Please "reply to all". 2012/12/5 Nabil <nabil.gergis(a)yahoo.com.au>: > Sorry there's one thing I'm not sure I understand... > Is it possible to adjust the locations of the wires of the soic clip so they match up with the correct pinouts? > I just dont get how gnd (pin 10 of bus pirate) can be connected to the soic clip (only 8 cables which connect to pins 1-8) > there's something I'm missing here i think, could you explain to me how to use an soic clip in detail, sorry for the trouble, much obliged. Nabs Which soic8 clip do you have? I have a Pomona 5250; it is possible to adjust the wire locations (pinouts) if you have the probe cable (that fits on the open end of the wires - the other end is a box header that connects to the buspirate): http://dangerousprototypes.com/2011/08/03/new-bus-pirate-probe-cable/ > > Idwer Vollering <vidwer(a)gmail.com> wrote: > >>2012/12/5 Nabs <ch12011-7(a)yahoo.com.au>: >>> Hi, sorry for the trouble. I'm running bus piratev3.6, bootloaderv4.4, >>> firmwarev6.1 on flashrom-0.9.6.1, i plan to run the BP at 1M. >>> >>> When i run the command "flashrom -p buspirate_spi:dev=/dev/ttyUSB0 -Vr >>> oldbios.bin" >>> I get: >>> " >>> Found Generic flash chip "unknown SPI chip (RDID)" (0 kB, SPI) on >>> buspirate_spi. >>> Probing for Generic unknown SPI chip (REMS), 0 kB: probe_spi_rems: id1 0x0, >>> id2 0x0 >>> Found Generic flash chip "unknown SPI chip (RDID)" (0 kB, SPI). >>> === >>> This flash part has status NOT WORKING for operations: PROBE READ ERASE >>> WRITE >>> " etc... >>> >>> >>> I think i have the wiring wrong which is why i get that error, i wonder if >>> you could help me as its my first time? >>> I'm trying to use In-circuit programming with a soldered bios chip using an >>> 'SOIC 8 clip' and it has 8 wires. But the bus pirate has 10 pins. >>> >>> From another thread, these are the connections i need: >>> Bus Pirate "MISO" (Pin 1) <----- SPI Device "DO" (Wire/Pin 2) >>> Bus Pirate "CS" (Pin 2) -----> SPI Device "CS" (Wire/Pin 1) >>> Bus Pirate "MOSI" (Pin 3) -----> SPI Device "DI" (Wire/Pin 5) >>> Bus Pirate "CLK" (Pin 4) -----> SPI Device "CLK" (Wire/Pin 6) >>> Bus Pirate "GND" (Pin 10) <----> SPI Device "GND" (Wire/Pin 4) >>> >>> It appears to me that GND is pin 10 and that the wires of the SOIC8 clip >>> only contact pins 1-8, so how can this work? >>> >>> Also, could i just check the layout of the pins on the bus pirate if i'm >>> reading it correctly on BPv3.6: >>> 10 9 >>> 8 7 >>> 6 5 >>> 4 3 >>> 2 1 >>> >>> Corresponding to: >>> GND 3V3 >>> +5V ADC >>> VPU AUX >>> CLK MOSI >>> CS MISO >> >>Forget the existence of +5V, VPU, ADC and AUX for now ( >>http://dangerousprototypes.com/docs/Bus_Pirate_I/O_Pin_Descriptions ), >>connect the clip to the buspirate so that the clip passes the >>buspirate pins like mentioned in this post: >>http://www.coreboot.org/pipermail/flashrom/2012-November/010178.html >> >>This works for me, with a MX25L1605A soic8 chip: >> >>CS# (CS) [1 chip 8] VCC >>(bridge with buspirate pin 3+7) >>SO (MISO) [2 chip 7] HOLD# (+3V3) >>WP# (bridge with buspirate pin 7+8) [3 chip 6] SCLK (CLK) >>GND (GND) [4 chip 5] SI (MOSI) >> >>> >>> Or are the pin numbers read in the opposite direction? >>> ie >>> 9 10 >>> 7 8 >>> 5 6 >>> 3 4 >>> 1 2 >>> >>> Here's an image of my bus pirate if it helps: >>> http://dangerousprototypes.com/wp-content/media/2012/11/Bus-Pirate-v3.6inte… >>> >>> Your help is much appreciated and i appreciate your patience, Nabs. >>> >>> _______________________________________________ >>> flashrom mailing list >>> flashrom(a)flashrom.org >>> http://www.flashrom.org/mailman/listinfo/flashrom

1 0

BPv3.6 w flashrom
by Nabs 05 Dec '12

05 Dec '12

Hi, sorry for the trouble. I'm running bus piratev3.6, bootloaderv4.4, firmwarev6.1 on flashrom-0.9.6.1, i plan to run the BP at 1M. When i run the command "flashrom -p buspirate_spi:dev=/dev/ttyUSB0 -Vr oldbios.bin" I get: " Found Generic flash chip "unknown SPI chip (RDID)" (0 kB, SPI) on buspirate_spi. Probing for Generic unknown SPI chip (REMS), 0 kB: probe_spi_rems: id1 0x0, id2 0x0 Found Generic flash chip "unknown SPI chip (RDID)" (0 kB, SPI). === This flash part has status NOT WORKING for operations: PROBE READ ERASE WRITE " etc... I think i have the wiring wrong which is why i get that error, i wonder if you could help me as its my first time? I'm trying to use In-circuit programming with a soldered bios chip using an 'SOIC 8 clip' and it has 8 wires. But the bus pirate has 10 pins. From another thread, these are the connections i need: Bus Pirate "MISO" (Pin 1) <----- SPI Device "DO" (Wire/Pin 2) Bus Pirate "CS" (Pin 2) -----> SPI Device "CS" (Wire/Pin 1) Bus Pirate "MOSI" (Pin 3) -----> SPI Device "DI" (Wire/Pin 5) Bus Pirate "CLK" (Pin 4) -----> SPI Device "CLK" (Wire/Pin 6) Bus Pirate "GND" (Pin 10) <----> SPI Device "GND" (Wire/Pin 4) It appears to me that GND is pin 10 and that the wires of the SOIC8 clip only contact pins 1-8, so how can this work? Also, could i just check the layout of the pins on the bus pirate if i'm reading it correctly on BPv3.6: 10 9 8 7 6 5 4 3 2 1 Corresponding to: GND 3V3 +5V ADC VPU AUX CLK MOSI CS MISO Or are the pin numbers read in the opposite direction? ie 9 10 7 8 5 6 3 4 1 2 Here's an image of my bus pirate if it helps: http://dangerousprototypes.com/wp-content/media/2012/11/Bus-Pirate-v3.6inte… Your help is much appreciated and i appreciate your patience, Nabs.

2 1

Bus Pirate + Flashrom
by Nabs 05 Dec '12

05 Dec '12

Hi, i get the following error below when trying to read a bios chip, what options do i have for troubleshooting and what's causing the problem? Thanks Nabs "flashrom -f -Vr bios.img.bin -p buspirate_spi:dev=/dev/tty8,spispeed=1M" flashrom v0.9.5.2-r1517 on Linux 3.2.0-23-generic (x86_64), built with libpci 3.1.8, GCC 4.6.3, little endian flashrom is free software, get the source code at http://www.flashrom.org Calibrating delay loop... OS timer resolution is 1 usecs, 1372M loops per second, 10 myus = 10 us, 100 myus = 113 us, 1000 myus = 1036 us, 10000 myus = 10346 us, 4 myus = 6 us, OK. Initializing buspirate_spi programmer SPI speed is 1MHz ^[[BEntering raw bitbang mode failed! Got 1c9ce050e0 Error: Programmer initialization failed. ^[[AEntering raw bitbang mode failed!

1 0

Dell Alienware Aurora R2
by Gilles Antonetti / Gmail 03 Dec '12

03 Dec '12

Apparently my machine has/had been hijacked somehow. The lshw was showing that it was running on 15 logicalcpu. Let me know if you need more. Thanks for help, Regards, Gilles

1 0

[PATCH 4/4] Enable serprog on Windows.
by Stefan Tauner 03 Dec '12

03 Dec '12

Sockets are not ported (yet). Signed-off-by: Stefan Tauner <stefan.tauner(a)student.tuwien.ac.at> Signed-off-by: Idwer Vollering <vidwer(a)gmail.com> --- Makefile | 8 +------- serprog.c | 18 ++++++++++++------ 2 files changed, 13 insertions(+), 13 deletions(-) diff --git a/Makefile b/Makefile index f3cb07a..b8cb7b7 100644 --- a/Makefile +++ b/Makefile @@ -138,12 +138,6 @@ FLASHROM_CFLAGS += -Dffs=__builtin_ffs # libusb-win32/libftdi stuff is usually installed in /usr/local. CPPFLAGS += -I/usr/local/include LDFLAGS += -L/usr/local/lib -# Serprog is not supported under Windows/MinGW (missing sockets support). -ifeq ($(CONFIG_SERPROG), yes) -UNSUPPORTED_FEATURES += CONFIG_SERPROG=yes -else -override CONFIG_SERPROG = no -endif # For now we disable all PCI-based programmers on Windows/MinGW (no libpci). ifeq ($(CONFIG_INTERNAL), yes) UNSUPPORTED_FEATURES += CONFIG_INTERNAL=yes @@ -324,7 +318,7 @@ SVNDEF := -D'FLASHROM_VERSION="$(VERSION)"' # Always enable internal/onboard support for now. CONFIG_INTERNAL ?= yes -# Always enable serprog for now. Needs to be disabled on Windows. +# Always enable serprog for now. CONFIG_SERPROG ?= yes # RayeR SPIPGM hardware support diff --git a/serprog.c b/serprog.c index 7f432ee..630039c 100644 --- a/serprog.c +++ b/serprog.c @@ -20,21 +20,23 @@ */ #include <stdio.h> +#ifndef _WIN32 /* stuff (presumably) needed for sockets only */ #include <stdlib.h> #include <unistd.h> -#include <strings.h> -#include <string.h> -#include <ctype.h> #include <fcntl.h> #include <sys/socket.h> #include <arpa/inet.h> #include <netinet/in.h> #include <netinet/tcp.h> #include <netdb.h> -#include <sys/stat.h> -#include <errno.h> -#include <inttypes.h> +#endif +#ifdef _WIN32 +#include <conio.h> +#else #include <termios.h> +#endif +#include <string.h> +#include <errno.h> #include "flash.h" #include "programmer.h" #include "chipdrivers.h" @@ -81,6 +83,7 @@ static int sp_opbuf_usage = 0; whether the command is supported before doing it */ static int sp_check_avail_automatic = 0; +#ifndef WIN32 static int sp_opensocket(char *ip, unsigned int port) { int flag = 1; @@ -114,6 +117,7 @@ static int sp_opensocket(char *ip, unsigned int port) setsockopt(sock, IPPROTO_TCP, TCP_NODELAY, &flag, sizeof(int)); return sock; } +#endif /* Synchronize: a bit tricky algorithm that tries to (and in my tests has * * always succeeded in) bring the serial protocol to known waiting-for- * @@ -351,6 +355,7 @@ int serprog_init(void) } free(device); +#ifndef _WIN32 device = extract_programmer_param("ip"); if (have_device && device) { msg_perr("Error: Both host and device specified.\n" @@ -394,6 +399,7 @@ int serprog_init(void) "flashrom -p serprog:ip=ipaddr:port\n"); return 1; } +#endif if (register_shutdown(serprog_shutdown, NULL)) return 1; -- Kind regards, Stefan Tauner

1 0

[PATCH 3/4] Replace native calls in serprog with wrapper calls.
by Stefan Tauner 03 Dec '12

03 Dec '12

read(), write(), usleep() are replaced. Signed-off-by: Stefan Tauner <stefan.tauner(a)student.tuwien.ac.at> Signed-off-by: Idwer Vollering <vidwer(a)gmail.com> --- serprog.c | 43 +++++++++++++++---------------------------- 1 file changed, 15 insertions(+), 28 deletions(-) diff --git a/serprog.c b/serprog.c index cf329a3..7f432ee 100644 --- a/serprog.c +++ b/serprog.c @@ -133,7 +133,7 @@ static int sp_synchronize(void) goto err_out; } /* A second should be enough to get all the answers to the buffer */ - usleep(1000 * 1000); + internal_delay(1000 * 1000); sp_flush_incoming(); /* Then try up to 8 times to send syncnop and get the correct special * @@ -206,15 +206,15 @@ static int sp_docommand(uint8_t command, uint32_t parmlen, unsigned char c; if (sp_automatic_cmdcheck(command)) return 1; - if (write(sp_fd, &command, 1) != 1) { + if (serialport_write(&command, 1) != 0) { msg_perr("Error: cannot write op code: %s\n", strerror(errno)); return 1; } - if (write(sp_fd, params, parmlen) != (parmlen)) { + if (serialport_write(params, parmlen) != 0) { msg_perr("Error: cannot write parameters: %s\n", strerror(errno)); return 1; } - if (read(sp_fd, &c, 1) != 1) { + if (serialport_read(&c, 1) != 0) { msg_perr("Error: cannot read from device: %s\n", strerror(errno)); return 1; } @@ -225,17 +225,10 @@ static int sp_docommand(uint8_t command, uint32_t parmlen, return 1; } if (retlen) { - int rd_bytes = 0; - do { - int r; - r = read(sp_fd, retparms + rd_bytes, - retlen - rd_bytes); - if (r <= 0) { - msg_perr("Error: cannot read return parameters: %s\n", strerror(errno)); - return 1; - } - rd_bytes += r; - } while (rd_bytes != retlen); + if (serialport_read(retparms, retlen) != 0) { + msg_perr("Error: cannot read return parameters: %s\n", strerror(errno)); + return 1; + } } return 0; } @@ -245,7 +238,7 @@ static void sp_flush_stream(void) if (sp_streamed_transmit_ops) do { unsigned char c; - if (read(sp_fd, &c, 1) != 1) { + if (serialport_read(&c, 1) != 0) { sp_die("Error: cannot read from device (flushing stream)"); } if (c == S_NAK) { @@ -272,7 +265,7 @@ static int sp_stream_buffer_op(uint8_t cmd, uint32_t parmlen, uint8_t * parms) memcpy(&(sp[1]), parms, parmlen); if (sp_streamed_transmit_bytes >= (1 + parmlen + sp_device_serbuf_size)) sp_flush_stream(); - if (write(sp_fd, sp, 1 + parmlen) != (1 + parmlen)) + if (serialport_write(sp, 1 + parmlen) != 0) sp_die("Error: cannot write command"); free(sp); sp_streamed_transmit_ops += 1; @@ -684,10 +677,9 @@ static void sp_pass_writen(void) header[4] = (sp_write_n_addr >> 0) & 0xFF; header[5] = (sp_write_n_addr >> 8) & 0xFF; header[6] = (sp_write_n_addr >> 16) & 0xFF; - if (write(sp_fd, header, 7) != 7) + if (serialport_write(header, 7) != 0) sp_die("Error: cannot write write-n command\n"); - if (write(sp_fd, sp_write_n_buf, sp_write_n_bytes) != - sp_write_n_bytes) + if (serialport_write(sp_write_n_buf, sp_write_n_bytes) != 0) sp_die("Error: cannot write write-n data"); sp_streamed_transmit_bytes += 7 + sp_write_n_bytes; sp_streamed_transmit_ops += 1; @@ -788,7 +780,7 @@ static uint8_t serprog_chip_readb(const struct flashctx *flash, buf[2] = ((addr >> 16) & 0xFF); sp_stream_buffer_op(S_CMD_R_BYTE, 3, buf); sp_flush_stream(); - if (read(sp_fd, &c, 1) != 1) + if (serialport_read(&c, 1) != 0) sp_die("readb byteread"); msg_pspew("%s addr=0x%lx returning 0x%02X\n", __func__, addr, c); return c; @@ -797,7 +789,6 @@ static uint8_t serprog_chip_readb(const struct flashctx *flash, /* Local version that really does the job, doesn't care of max_read_n. */ static void sp_do_read_n(uint8_t * buf, const chipaddr addr, size_t len) { - int rd_bytes = 0; unsigned char sbuf[6]; msg_pspew("%s: addr=0x%lx len=%lu\n", __func__, addr, (unsigned long)len); /* Stream the read-n -- as above. */ @@ -811,12 +802,8 @@ static void sp_do_read_n(uint8_t * buf, const chipaddr addr, size_t len) sbuf[5] = ((len >> 16) & 0xFF); sp_stream_buffer_op(S_CMD_R_NBYTES, 6, sbuf); sp_flush_stream(); - do { - int r = read(sp_fd, buf + rd_bytes, len - rd_bytes); - if (r <= 0) - sp_die("Error: cannot read read-n data"); - rd_bytes += r; - } while (rd_bytes != len); + if (serialport_read(buf, len) != 0) + sp_die("Error: cannot read read-n data"); return; } -- Kind regards, Stefan Tauner

1 0

[PATCH 2/4] Introduce serialport_write_nonblock().
by Stefan Tauner 03 Dec '12

03 Dec '12

It seems useful to have a generic and platform-independent method to read and write to a serial port without blocking. This is the write part. This allows to get rid of the explicit temporary disabling of blocking I/O in serprog's sp_synchronize(). Signed-off-by: Stefan Tauner <stefan.tauner(a)student.tuwien.ac.at> --- programmer.h | 1 + serial.c | 72 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ serprog.c | 19 ++++------------ 3 files changed, 78 insertions(+), 14 deletions(-) diff --git a/programmer.h b/programmer.h index e385314..722fa54 100644 --- a/programmer.h +++ b/programmer.h @@ -649,6 +649,7 @@ extern fdtype sp_fd; /* expose serialport_shutdown as it's currently used by buspirate */ int serialport_shutdown(void *data); int serialport_write(unsigned char *buf, unsigned int writecnt); +int serialport_write_nonblock(unsigned char *buf, unsigned int writecnt, unsigned int timeout, unsigned int *really_wrote); int serialport_read(unsigned char *buf, unsigned int readcnt); int serialport_read_nonblock(unsigned char *c, unsigned int readcnt, unsigned int timeout, unsigned int *really_read); diff --git a/serial.c b/serial.c index 15fafbb..7c136c9 100644 --- a/serial.c +++ b/serial.c @@ -387,3 +387,75 @@ int serialport_read_nonblock(unsigned char *c, unsigned int readcnt, unsigned in #endif return ret; } + +/* Tries up to timeout ms to write writecnt characters from the array starting at buf. Returns + * 0 on success, positive values on temporary errors (e.g. timeouts) and negative ones on permanent errors. + * If really_wrote is not NULL, this function sets its contents to the number of bytes written successfully. */ +int serialport_write_nonblock(unsigned char *buf, unsigned int writecnt, unsigned int timeout, unsigned int *really_wrote) +{ + int ret = 1; + /* disable blocked i/o and declare platform-specific variables */ +#ifdef _WIN32 + DWORD rv; + COMMTIMEOUTS oldTimeout; + COMMTIMEOUTS newTimeout = { + .ReadIntervalTimeout = MAXDWORD, + .ReadTotalTimeoutMultiplier = 0, + .ReadTotalTimeoutConstant = 0, + .WriteTotalTimeoutMultiplier = 0, + .WriteTotalTimeoutConstant = 0 + }; + if(GetCommTimeouts(sp_fd, &oldTimeout) != 0) { + msg_perr("Could not get serial port timeout settings!\n"); + return -1; + } + if(SetCommTimeouts(sp_fd, &newTimeout) != 0) { + msg_perr("Could not set serial port timeout settings!\n"); + return -1; + } +#else + ssize_t rv; + const int flags = fcntl(sp_fd, F_GETFL); + fcntl(sp_fd, F_SETFL, flags | O_NONBLOCK); +#endif + + int i; + int wr_bytes = 0; + for (i = 0; i < timeout; i++) { + msg_pspew("writecnt %d wr_bytes %d\n", writecnt, wr_bytes); +#ifdef _WIN32 + WriteFile(sp_fd, buf + wr_bytes, writecnt - wr_bytes, &rv, NULL); + msg_pspew("wrote %lu bytes\n", rv); +#else + rv = write(sp_fd, buf + wr_bytes, writecnt - wr_bytes); + msg_pspew("wrote %zd bytes\n", rv); +#endif + if ((rv == -1) && (errno != EAGAIN)) { + msg_perr("Serial port write error: %s\n", strerror(errno)); + ret = -1; + break; + } + if (rv > 0) { + wr_bytes += rv; + if (wr_bytes == writecnt) { + msg_pspew("write successful\n"); + ret = 0; + break; + } + } + internal_delay(1000); /* 1ms units */ + } + if (really_wrote != NULL) + *really_wrote = wr_bytes; + + /* restore original blocking behavior */ +#ifdef _WIN32 + if(SetCommTimeouts(sp_fd, &oldTimeout) != 0) { + msg_perr("Could not restore serial port timeout settings!\n"); + return -1; + } +#else + fcntl(sp_fd, F_SETFL, flags); +#endif + return ret; +} diff --git a/serprog.c b/serprog.c index cc06798..cf329a3 100644 --- a/serprog.c +++ b/serprog.c @@ -117,23 +117,19 @@ static int sp_opensocket(char *ip, unsigned int port) /* Synchronize: a bit tricky algorithm that tries to (and in my tests has * * always succeeded in) bring the serial protocol to known waiting-for- * - * command state - uses nonblocking read - rest of the driver uses * + * command state - uses nonblocking I/O - rest of the driver uses * * blocking read - TODO: add an alarm() timer for the rest of the app on * * serial operations, though not such a big issue as the first thing to * * do is synchronize (eg. check that device is alive). */ static int sp_synchronize(void) { int i; - int flags = fcntl(sp_fd, F_GETFL); unsigned char buf[8]; - flags |= O_NONBLOCK; - fcntl(sp_fd, F_SETFL, flags); /* First sends 8 NOPs, then flushes the return data - should cause * * the device serial parser to get to a sane state, unless if it * * is waiting for a real long write-n. */ memset(buf, S_CMD_NOP, 8); - if (write(sp_fd, buf, 8) != 8) { - msg_perr("flush write: %s\n", strerror(errno)); + if (serialport_write_nonblock(buf, 8, 1, NULL) != 0) { goto err_out; } /* A second should be enough to get all the answers to the buffer */ @@ -147,8 +143,7 @@ static int sp_synchronize(void) for (i = 0; i < 8; i++) { int n; unsigned char c = S_CMD_SYNCNOP; - if (write(sp_fd, &c, 1) != 1) { - msg_perr("sync write: %s\n", strerror(errno)); + if (serialport_write_nonblock(&c, 1, 1, NULL) != 0) { goto err_out; } msg_pdbg("."); @@ -165,9 +160,8 @@ static int sp_synchronize(void) if (ret > 0 || c != S_ACK) continue; c = S_CMD_SYNCNOP; - if (write(sp_fd, &c, 1) != 1) { - msg_perr("sync write: %s\n", strerror(errno)); - return 1; + if (serialport_write_nonblock(&c, 1, 1, NULL) != 0) { + goto err_out; } ret = serialport_read_nonblock(&c, 1, 500, NULL); if (ret < 0) @@ -179,9 +173,6 @@ static int sp_synchronize(void) goto err_out; if (c != S_ACK) break; /* fail */ - /* Ok, synchronized; back to blocking reads and return. */ - flags &= ~O_NONBLOCK; - fcntl(sp_fd, F_SETFL, flags); msg_pdbg("\n"); return 0; } -- Kind regards, Stefan Tauner

1 0

[PATCH 1/4] Replace sp_sync_read_timeout() with serialport_read_nonblock().
by Stefan Tauner 03 Dec '12

03 Dec '12

It seems useful to have a generic and platform-independent method to read and write to a serial port without blocking. This is the read part. It stores the current blocking properties before disabling blocking and restores them after reading. The timeout is implemented as previously by retrying every millisecond until the timeout is reached or enough characters are available. TODO: testing on windows and then removing the spew prints. Signed-off-by: Stefan Tauner <stefan.tauner(a)student.tuwien.ac.at> --- programmer.h | 1 + serial.c | 71 ++++++++++++++++++++++++++++++++++++++++++++++++++++++++++ serprog.c | 28 ++++------------------- 3 files changed, 76 insertions(+), 24 deletions(-) diff --git a/programmer.h b/programmer.h index dedec67..e385314 100644 --- a/programmer.h +++ b/programmer.h @@ -650,6 +650,7 @@ extern fdtype sp_fd; int serialport_shutdown(void *data); int serialport_write(unsigned char *buf, unsigned int writecnt); int serialport_read(unsigned char *buf, unsigned int readcnt); +int serialport_read_nonblock(unsigned char *c, unsigned int readcnt, unsigned int timeout, unsigned int *really_read); /* Serial port/pin mapping: diff --git a/serial.c b/serial.c index 9446ce3..15fafbb 100644 --- a/serial.c +++ b/serial.c @@ -316,3 +316,74 @@ int serialport_read(unsigned char *buf, unsigned int readcnt) return 0; } + +/* Tries up to timeout ms to read readcnt characters and places them into the array starting at c. Returns + * 0 on success, positive values on temporary errors (e.g. timeouts) and negative ones on permanent errors. + * If really_read is not NULL, this function sets its contents to the number of bytes read successfully. */ +int serialport_read_nonblock(unsigned char *c, unsigned int readcnt, unsigned int timeout, unsigned int *really_read) +{ + int ret = 1; + /* disable blocked i/o and declare platform-specific variables */ +#ifdef _WIN32 + DWORD rv; + COMMTIMEOUTS oldTimeout; + COMMTIMEOUTS newTimeout = { + .ReadIntervalTimeout = MAXDWORD, + .ReadTotalTimeoutMultiplier = 0, + .ReadTotalTimeoutConstant = 0, + .WriteTotalTimeoutMultiplier = 0, + .WriteTotalTimeoutConstant = 0 + }; + if(GetCommTimeouts(sp_fd, &oldTimeout) != 0) { + msg_perr("Could not get serial port timeout settings!\n"); + return -1; + } + if(SetCommTimeouts(sp_fd, &newTimeout) != 0) { + msg_perr("Could not set serial port timeout settings!\n"); + return -1; + } +#else + ssize_t rv; + const int flags = fcntl(sp_fd, F_GETFL); + fcntl(sp_fd, F_SETFL, flags | O_NONBLOCK); +#endif + + int i; + int rd_bytes = 0; + for (i = 0; i < timeout; i++) { + msg_pspew("readcnt %d rd_bytes %d\n", readcnt, rd_bytes); +#ifdef _WIN32 + ReadFile(sp_fd, c + rd_bytes, readcnt - rd_bytes, &rv, NULL); + msg_pspew("read %lu bytes\n", rv); +#else + rv = read(sp_fd, c + rd_bytes, readcnt - rd_bytes); + msg_pspew("read %zd bytes\n", rv); +#endif + if ((rv == -1) && (errno != EAGAIN)) { + msg_perr("Serial port read error: %s\n", strerror(errno)); + ret = -1; + break; + } + if (rv > 0) + rd_bytes += rv; + if (rd_bytes == readcnt) { + msg_perr("read successful\n"); + ret = 0; + break; + } + internal_delay(1000); /* 1ms units */ + } + if (really_read != NULL) + *really_read = rd_bytes; + + /* restore original blocking behavior */ +#ifdef _WIN32 + if(SetCommTimeouts(sp_fd, &oldTimeout) != 0) { + msg_perr("Could not restore serial port timeout settings!\n"); + return -1; + } +#else + fcntl(sp_fd, F_SETFL, flags); +#endif + return ret; +} diff --git a/serprog.c b/serprog.c index 98c839e..cc06798 100644 --- a/serprog.c +++ b/serprog.c @@ -115,26 +115,6 @@ static int sp_opensocket(char *ip, unsigned int port) return sock; } -/* Returns 0 on success and places the character read into the location pointed to by c. - * Returns positive values on temporary errors and negative ones on permanent errors. - * An iteration of one loop takes up to 1ms. */ -static int sp_sync_read_timeout(unsigned int loops, unsigned char *c) -{ - int i; - for (i = 0; i < loops; i++) { - ssize_t rv; - rv = read(sp_fd, c, 1); - if (rv == 1) - return 0; - if ((rv == -1) && (errno != EAGAIN)) { - msg_perr("read: %s\n", strerror(errno)); - return -1; - } - usleep(1000); /* 1ms units */ - } - return 1; -} - /* Synchronize: a bit tricky algorithm that tries to (and in my tests has * * always succeeded in) bring the serial protocol to known waiting-for- * * command state - uses nonblocking read - rest of the driver uses * @@ -174,12 +154,12 @@ static int sp_synchronize(void) msg_pdbg("."); fflush(stdout); for (n = 0; n < 10; n++) { - int ret = sp_sync_read_timeout(50, &c); + int ret = serialport_read_nonblock(&c, 1, 50, NULL); if (ret < 0) goto err_out; if (ret > 0 || c != S_NAK) continue; - ret = sp_sync_read_timeout(20, &c); + ret = serialport_read_nonblock(&c, 1, 20, NULL); if (ret < 0) goto err_out; if (ret > 0 || c != S_ACK) @@ -189,12 +169,12 @@ static int sp_synchronize(void) msg_perr("sync write: %s\n", strerror(errno)); return 1; } - ret = sp_sync_read_timeout(500, &c); + ret = serialport_read_nonblock(&c, 1, 500, NULL); if (ret < 0) goto err_out; if (ret > 0 || c != S_NAK) break; /* fail */ - ret = sp_sync_read_timeout(100, &c); + ret = serialport_read_nonblock(&c, 1, 100, NULL); if (ret > 0 || ret < 0) goto err_out; if (c != S_ACK) -- Kind regards, Stefan Tauner

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

flashrom December 2012