flashrom November 2009

flashrom@flashrom.org

47 participants
148 discussions

Verbose output of lspci, superiotool and flashrom - Intel DP55WB (unsupported)
by Adrian Glaubitz 26 Nov '09

26 Nov '09

Hi, tested an Intel DP55WB, however the chipset was not detected and therefore flashing was not possible. Please see the attached logs of lspci -nnvvvxxx, superiotool -dV and flashrom -V. Website: <http://www.intel.com/products/desktop/motherboards/DP55WB/DP55WB-overview.h…> Adrian

2 3

[RFC] Abort early or abort late on failure?
by Carl-Daniel Hailfinger 26 Nov '09

26 Nov '09

Hi, flashrom currently doesn't have a coherent policy about failed writes. Some chip drivers abort on the first failed write, others keep going until the end. Either variant makes sense, but mixing them is suboptimal. The following policies are available: - hardcode abort early (on first error) - hardcode abort never (ignore all errors until the end) - switch behaviour based on commandline parameter. Opinions? Regards, Carl-Daniel -- Developer quote of the month: "We are juggling too many chainsaws and flaming arrows and tigers."

2 1

Success with DFI Blood-Iron P35 T2RL
by Erno Vaurio 26 Nov '09

26 Nov '09

Hello, I was able to successfully flash my board with a new bios. There was one warning about spi_chip_erase_60 failing, this didn't affect the outcome and flash was verified. Board model: DFI Infinity Blood-Iron P35-T2RL Manual: http://img.lanparty.tw/Upload/Manual/CM/021107321.pdf Bios download link: http://img.lanparty.tw/Upload/BIOS/CM/I35RD924.zip $ sudo flashrom -w I35RD924.BIN flashrom v0.9.1-r706 No coreboot table found. Found chipset "Intel ICH9R", enabling flash write... OK. This chipset supports the following protocols: LPC,FWH,SPI. Calibrating delay loop... OK. Found chip "SST SST25VF080B" (1024 KB, SPI) at physical address 0xfff00000. Flash image seems to be a legacy BIOS. Disabling checks. Writing flash chip... Erasing flash before programming... spi_chip_erase_60 failed during command execution done. COMPLETE. Verifying flash... VERIFIED. Requested output attached as files.

1 0

[PATCH] winbondw25x32/64 support
by Bao, Zheng 26 Nov '09

26 Nov '09

Add entries of W25x32 and W25x64. The modle_ids are already in the header. W25x32 has been successfully probed. W25x64 is not available. I just get information from datasheet. Signed-off-by: Zheng Bao <zheng.bao(a)amd.com> Index: flashchips.c =================================================================== --- flashchips.c (revision 781) +++ flashchips.c (working copy) @@ -3038,6 +3038,38 @@ { .vendor = "Winbond", + .name = "W25x32", + .bustype = CHIP_BUSTYPE_SPI, + .manufacture_id = WINBOND_NEX_ID, + .model_id = W_25X32, + .total_size = 4096, + .page_size = 256, + .tested = TEST_OK_PROBE, + .probe = probe_spi_rdid, + .probe_timing = TIMING_ZERO, + .erase = spi_chip_erase_c7, + .write = spi_chip_write_256, + .read = spi_chip_read, + }, + + { + .vendor = "Winbond", + .name = "W25x64", + .bustype = CHIP_BUSTYPE_SPI, + .manufacture_id = WINBOND_NEX_ID, + .model_id = W_25X64, + .total_size = 8192, + .page_size = 256, + .tested = TEST_UNTESTED, + .probe = probe_spi_rdid, + .probe_timing = TIMING_ZERO, + .erase = spi_chip_erase_c7, + .write = spi_chip_write_256, + .read = spi_chip_read, + }, + + { + .vendor = "Winbond", .name = "W29C011", .bustype = CHIP_BUSTYPE_PARALLEL, .manufacture_id = WINBOND_ID,

2 1

[commit] r782 - trunk
by svn＠flashrom.org 26 Nov '09

26 Nov '09

Author: hailfinger Date: 2009-11-26 12:05:01 +0100 (Thu, 26 Nov 2009) New Revision: 782 Modified: trunk/flashchips.c Log: Add entries of W25x32 and W25x64. The model_ids are already in the header. W25x32 has been successfully probed. W25x64 is not available, the entry is based on the datasheet. Signed-off-by: Zheng Bao <zheng.bao(a)amd.com> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006(a)gmx.net> Modified: trunk/flashchips.c =================================================================== --- trunk/flashchips.c 2009-11-25 17:05:52 UTC (rev 781) +++ trunk/flashchips.c 2009-11-26 11:05:01 UTC (rev 782) @@ -3038,6 +3038,38 @@ { .vendor = "Winbond", + .name = "W25x32", + .bustype = CHIP_BUSTYPE_SPI, + .manufacture_id = WINBOND_NEX_ID, + .model_id = W_25X32, + .total_size = 4096, + .page_size = 256, + .tested = TEST_OK_PROBE, + .probe = probe_spi_rdid, + .probe_timing = TIMING_ZERO, + .erase = spi_chip_erase_c7, + .write = spi_chip_write_256, + .read = spi_chip_read, + }, + + { + .vendor = "Winbond", + .name = "W25x64", + .bustype = CHIP_BUSTYPE_SPI, + .manufacture_id = WINBOND_NEX_ID, + .model_id = W_25X64, + .total_size = 8192, + .page_size = 256, + .tested = TEST_UNTESTED, + .probe = probe_spi_rdid, + .probe_timing = TIMING_ZERO, + .erase = spi_chip_erase_c7, + .write = spi_chip_write_256, + .read = spi_chip_read, + }, + + { + .vendor = "Winbond", .name = "W29C011", .bustype = CHIP_BUSTYPE_PARALLEL, .manufacture_id = WINBOND_ID,

1 0

Generated zeros and reading firmware after flashing them
by Xangel2Net 25 Nov '09

25 Nov '09

1 0

[commit] r781 - trunk
by svn＠flashrom.org 25 Nov '09

25 Nov '09

Author: hailfinger Date: 2009-11-25 18:05:52 +0100 (Wed, 25 Nov 2009) New Revision: 781 Modified: trunk/flashrom.c Log: Clarify a comment about verification routine usage. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006(a)gmx.net> Acked-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006(a)gmx.net> Modified: trunk/flashrom.c =================================================================== --- trunk/flashrom.c 2009-11-25 16:58:17 UTC (rev 780) +++ trunk/flashrom.c 2009-11-25 17:05:52 UTC (rev 781) @@ -407,7 +407,8 @@ } /** - * @cmpbuf buffer to compare against + * @cmpbuf buffer to compare against, cmpbuf[0] is expected to match the + flash content at location start * @start offset to the base address of the flash chip * @len length of the verified area * @message string to print in the "FAILED" message

1 0

[PATCH] Reduce realloc syscall overhead for ft2232 and bitbang
by Carl-Daniel Hailfinger 25 Nov '09

25 Nov '09

Reduce realloc syscall overhead for FT2232 and bitbang. FT2232 ran realloc() for every executed command. Start with a big enough buffer and don't touch buffer size unless it needs to grow. Bitbang was slightly better: It only ran realloc() if buffer size changed. Still, the solution above improves performance and reliability. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006(a)gmx.net> Index: flashrom-realloc_overhead/bitbang_spi.c =================================================================== --- flashrom-realloc_overhead/bitbang_spi.c (Revision 777) +++ flashrom-realloc_overhead/bitbang_spi.c (Arbeitskopie) @@ -87,14 +87,16 @@ static unsigned char *bufout = NULL; static unsigned char *bufin = NULL; static int oldbufsize = 0; - int bufsize = max(writecnt + readcnt, 260); + int bufsize; int i; /* Arbitrary size limitation here. We're only constrained by memory. */ if (writecnt > 65536 || readcnt > 65536) return SPI_INVALID_LENGTH; - if (bufsize != oldbufsize) { + bufsize = max(writecnt + readcnt, 260); + /* Never shrink. realloc() calls are expensive. */ + if (bufsize > oldbufsize) { bufout = realloc(bufout, bufsize); if (!bufout) { fprintf(stderr, "Out of memory!\n"); @@ -109,6 +111,7 @@ free(bufout); exit(1); } + oldbufsize = bufsize; } memcpy(bufout, writearr, writecnt); Index: flashrom-realloc_overhead/ft2232_spi.c =================================================================== --- flashrom-realloc_overhead/ft2232_spi.c (Revision 777) +++ flashrom-realloc_overhead/ft2232_spi.c (Arbeitskopie) @@ -200,14 +200,22 @@ static unsigned char *buf = NULL; /* failed is special. We use bitwise ops, but it is essentially bool. */ int i = 0, ret = 0, failed = 0; + int bufsize; + static int oldbufsize = 0; if (writecnt > 65536 || readcnt > 65536) return SPI_INVALID_LENGTH; - buf = realloc(buf, writecnt + readcnt + 100); - if (!buf) { - fprintf(stderr, "Out of memory!\n"); - exit(1); // -1 + /* buf is not used for the response from the chip. */ + bufsize = max(writecnt + 9, 260); + /* Never shrink. realloc() calls are expensive. */ + if (bufsize > oldbufsize) { + buf = realloc(buf, bufsize); + if (!buf) { + fprintf(stderr, "Out of memory!\n"); + exit(1); + } + oldbufsize = bufsize; } /* -- Developer quote of the month: "We are juggling too many chainsaws and flaming arrows and tigers."

3 4

[commit] r780 - trunk
by svn＠flashrom.org 25 Nov '09

25 Nov '09

Author: hailfinger Date: 2009-11-25 17:58:17 +0100 (Wed, 25 Nov 2009) New Revision: 780 Modified: trunk/bitbang_spi.c trunk/ft2232_spi.c Log: Reduce realloc syscall overhead for FT2232 and bitbang. FT2232 ran realloc() for every executed command. Start with a big enough buffer and don't touch buffer size unless it needs to grow. Bitbang was slightly better: It only ran realloc() if buffer size changed. Still, the solution above improves performance and reliability. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006(a)gmx.net> Acked-by: Sean Nelson <audiohacked(a)gmail.com> Modified: trunk/bitbang_spi.c =================================================================== --- trunk/bitbang_spi.c 2009-11-25 16:41:50 UTC (rev 779) +++ trunk/bitbang_spi.c 2009-11-25 16:58:17 UTC (rev 780) @@ -87,14 +87,16 @@ static unsigned char *bufout = NULL; static unsigned char *bufin = NULL; static int oldbufsize = 0; - int bufsize = max(writecnt + readcnt, 260); + int bufsize; int i; /* Arbitrary size limitation here. We're only constrained by memory. */ if (writecnt > 65536 || readcnt > 65536) return SPI_INVALID_LENGTH; - if (bufsize != oldbufsize) { + bufsize = max(writecnt + readcnt, 260); + /* Never shrink. realloc() calls are expensive. */ + if (bufsize > oldbufsize) { bufout = realloc(bufout, bufsize); if (!bufout) { fprintf(stderr, "Out of memory!\n"); @@ -109,6 +111,7 @@ free(bufout); exit(1); } + oldbufsize = bufsize; } memcpy(bufout, writearr, writecnt); Modified: trunk/ft2232_spi.c =================================================================== --- trunk/ft2232_spi.c 2009-11-25 16:41:50 UTC (rev 779) +++ trunk/ft2232_spi.c 2009-11-25 16:58:17 UTC (rev 780) @@ -200,14 +200,22 @@ static unsigned char *buf = NULL; /* failed is special. We use bitwise ops, but it is essentially bool. */ int i = 0, ret = 0, failed = 0; + int bufsize; + static int oldbufsize = 0; if (writecnt > 65536 || readcnt > 65536) return SPI_INVALID_LENGTH; - buf = realloc(buf, writecnt + readcnt + 100); - if (!buf) { - fprintf(stderr, "Out of memory!\n"); - exit(1); // -1 + /* buf is not used for the response from the chip. */ + bufsize = max(writecnt + 9, 260 + 9); + /* Never shrink. realloc() calls are expensive. */ + if (bufsize > oldbufsize) { + buf = realloc(buf, bufsize); + if (!buf) { + fprintf(stderr, "Out of memory!\n"); + exit(1); + } + oldbufsize = bufsize; } /*

1 0

[PATCH] Clean up and fix jedec.c error handling
by Carl-Daniel Hailfinger 25 Nov '09

25 Nov '09

jedec.c was missing error handling in a few cases. Fix. jedec.c error handling used double negation in too many places for no good reason. Clean up. Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006(a)gmx.net> Index: flashrom-jedec_errorhandling/jedec.c =================================================================== --- flashrom-jedec_errorhandling/jedec.c (Revision 777) +++ flashrom-jedec_errorhandling/jedec.c (Arbeitskopie) @@ -262,7 +262,7 @@ int write_page_write_jedec(struct flashchip *flash, uint8_t *src, int start, int page_size) { - int i, tried = 0, ok; + int i, tried = 0, failed; uint8_t *s = src; chipaddr bios = flash->virtual_memory; chipaddr dst = bios + start; @@ -287,27 +287,27 @@ dst = d; src = s; - ok = !verify_range(flash, src, start, page_size, NULL); + failed = verify_range(flash, src, start, page_size, NULL); - if (!ok && tried++ < MAX_REFLASH_TRIES) { + if (failed && tried++ < MAX_REFLASH_TRIES) { fprintf(stderr, "retrying.\n"); goto retry; } - if (!ok) { + if (failed) { fprintf(stderr, " page 0x%lx failed!\n", (d - bios) / page_size); } - return !ok; + return failed; } int write_byte_program_jedec(chipaddr bios, uint8_t *src, chipaddr dst) { - int tried = 0, ok = 1; + int tried = 0, failed = 0; - /* If the data is 0xFF, don't program it */ + /* If the data is 0xFF, don't program it and don't complain. */ if (*src == 0xFF) { - return -1; + return 0; } retry: @@ -325,27 +325,30 @@ } if (tried >= MAX_REFLASH_TRIES) - ok = 0; + failed = 1; - return !ok; + return failed; } int write_sector_jedec(chipaddr bios, uint8_t *src, chipaddr dst, unsigned int page_size) { - int i; + int i, failed = 0; for (i = 0; i < page_size; i++) { - write_byte_program_jedec(bios, src, dst); + if (write_byte_program_jedec(bios, src, dst)) + failed = 1; dst++, src++; } + if (failed) + fprintf(stderr, " writing sector at 0x%lx failed!\n", dst); - return 0; + return failed; } int write_jedec(struct flashchip *flash, uint8_t *buf) { - int i; + int i, failed = 0; int total_size = flash->total_size * 1024; int page_size = flash->page_size; chipaddr bios = flash->virtual_memory; @@ -358,12 +361,13 @@ printf("Programming page: "); for (i = 0; i < total_size / page_size; i++) { printf("%04d at address: 0x%08x", i, i * page_size); - write_page_write_jedec(flash, buf + i * page_size, - i * page_size, page_size); + if (write_page_write_jedec(flash, buf + i * page_size, + i * page_size, page_size)) + failed = 1; printf("\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b"); } printf("\n"); protect_jedec(bios); - return 0; + return failed; } -- Developer quote of the month: "We are juggling too many chainsaws and flaming arrows and tigers."

2 3

← Newer
1
...
4
5
6
7
8
9
10
...
15
Older →

Jump to page:

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

flashrom November 2009