sf: Add documentation and an 'sf mmap' command
v4 cover sf: Add documentation and an 'sf mmap' command
This little series adds documentation and a few other tidy-ups to the 'sf' command.
It also provides a way to access memory-mapped SPI via the command line.
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examples
Changes in v3:
- Add configuration and return value also
Changes in v2:
- Explain why 'usage' cannot be const
- Many fixes as suggested by Heinrich
Simon Glass (5):
command: Use a constant pointer for the help
sf: Use const for the stage name
sf: Tidy up code to avoid #ifdef
sf: doc: Add documentation for the 'sf' command
sf: Provide a command to access memory-mapped SPI
arch/Kconfig | 2 + cmd/Kconfig | 8 ++ cmd/sf.c | 67 +++++++--- doc/usage/index.rst | 1 + doc/usage/sf.rst | 308 ++++++++++++++++++++++++++++++++++++++++++++ include/command.h | 2 +- 6 files changed, 369 insertions(+), 19 deletions(-) create mode 100644 doc/usage/sf.rst
Signature · 1 line
2.33.0.464.g1972c5931b-goog
PATCH v4 1/5 command: Use a constant pointer for the help
This text should never change during execution, so it makes sense to use a const char * so that it can be declared as const in the code. Update struct cmd_tbl with a const char * pointer for 'help'.
We cannot make usage const because of the bmode command, used on mx53ppd for example.
---
(no changes since v2)
Changes in v2:
- Explain why 'usage' cannot be const
include/command.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
include/command.h +1 -1
@@ -45,7 +45,7 @@ struct cmd_tbl { char *const argv[]);char *usage;/* Usage message (short) */ #ifdef CONFIG_SYS_LONGHELP -char *help;/* Help message (long) */ +const char *help;/* Help message (long) */ #endif #ifdef CONFIG_AUTO_COMPLETE /* do auto completion on the arguments */
Signature · 1 line
2.33.0.464.g1972c5931b-goog
Re: [PATCH v4 1/5] command: Use a constant pointer for the help
Quoted 8 lines · Simon Glass
This text should never change during execution, so it makes sense to use a const char * so that it can be declared as const in the code. Update struct
cmd_tblwith a const char * pointer for 'help'.We cannot make usage const because of the bmode command, used on mx53ppd for example.
Quoted 7 lines
Signed-off-by Simon Glass <[email protected]>---
(no changes since v2)
Changes in v2:
- Explain why 'usage' cannot be const
Re: [PATCH v4 1/5] command: Use a constant pointer for the help
In reply to Simon GlassQuoted 8 lines · Simon Glass
This text should never change during execution, so it makes sense to use a const char * so that it can be declared as const in the code. Update struct
cmd_tblwith a const char * pointer for 'help'.We cannot make usage const because of the bmode command, used on mx53ppd for example.
Signed-off-by Simon Glass <[email protected]>
Quoted 24 lines
---
(no changes since v2)
Changes in v2:
- Explain why 'usage' cannot be constinclude/command.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)include/command.h +1
-1@@ -45,7 +45,7 @@ struct cmd_tbl { char *const argv[]);char *usage;/* Usage message (short) */ #ifdef CONFIG_SYS_LONGHELP -char *help;/* Help message (long) */ +const char *help;/* Help message (long) */ #endif #ifdef CONFIG_AUTO_COMPLETE /* do auto completion on the arguments */
Simon Glass
Reviewed-by Heinrich SchuchardtRe: [PATCH v4 1/5] command: Use a constant pointer for the help
Quoted 8 lines · Simon Glass
This text should never change during execution, so it makes sense to use a const char * so that it can be declared as const in the code. Update struct
cmd_tblwith a const char * pointer for 'help'.We cannot make usage const because of the bmode command, used on mx53ppd for example.
Signed-off-by Simon Glass <[email protected]>
Quoted 10 lines
---
(no changes since v2)
Changes in v2:
- Explain why 'usage' cannot be constinclude/command.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Applied to u-boot-dm, thanks!
Simon Glass
Reviewed-by Heinrich SchuchardtRe: [PATCH v4 1/5] command: Use a constant pointer for the help
Quoted 8 lines · Simon Glass
This text should never change during execution, so it makes sense to use a const char * so that it can be declared as const in the code. Update struct
cmd_tblwith a const char * pointer for 'help'.We cannot make usage const because of the bmode command, used on mx53ppd for example.
Signed-off-by Simon Glass <[email protected]>
Quoted 10 lines
---
(no changes since v2)
Changes in v2:
- Explain why 'usage' cannot be constinclude/command.h | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Applied to u-boot-dm, thanks!
Applied to u-boot-dm/next, thanks!
PATCH v4 2/5 sf: Use const for the stage name
In reply to Simon GlassThis is not updated at runtime so should be marked const. Update the code accordingly.
---
(no changes since v1)
cmd/sf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
cmd/sf.c +1 -1
@@ -394,7 +394,7 @@ enum { STAGE_COUNT, }; -static char *stage_name[STAGE_COUNT ] = { +static const char *stage_name[STAGE_COUNT ] = {"erase" ,"check" ,"write" ,
Signature · 1 line
2.33.0.464.g1972c5931b-goog
Re: [PATCH v4 2/5] sf: Use const for the stage name
Quoted 7 lines · Simon Glass
This is not updated at runtime so should be marked const. Update the code accordingly.
Signed-off-by Simon Glass <[email protected]>---
Re: [PATCH v4 2/5] sf: Use const for the stage name
In reply to Simon GlassQuoted 4 lines · Simon Glass
This is not updated at runtime so should be marked const. Update the code accordingly.
Signed-off-by Simon Glass <[email protected]>
Quoted 21 lines
---
(no changes since v1)
cmd/sf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)cmd/sf.c +1
-1@@ -394,7 +394,7 @@ enum { STAGE_COUNT, }; -static char *stage_name[STAGE_COUNT ] = { +static const char *stage_name[STAGE_COUNT ] = {"erase" ,"check" ,"write" ,
Simon Glass
Reviewed-by Heinrich SchuchardtRe: [PATCH v4 2/5] sf: Use const for the stage name
Quoted 4 lines · Simon Glass
This is not updated at runtime so should be marked const. Update the code accordingly.
Signed-off-by Simon Glass <[email protected]>
Quoted 7 lines
---
(no changes since v1)
cmd/sf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Applied to u-boot-dm, thanks!
Simon Glass
Reviewed-by Heinrich SchuchardtRe: [PATCH v4 2/5] sf: Use const for the stage name
Quoted 4 lines · Simon Glass
This is not updated at runtime so should be marked const. Update the code accordingly.
Signed-off-by Simon Glass <[email protected]>
Quoted 7 lines
---
(no changes since v1)
cmd/sf.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)
Applied to u-boot-dm, thanks!
Applied to u-boot-dm/next, thanks!
PATCH v4 3/5 sf: Tidy up code to avoid #ifdef
In reply to Simon GlassUpdate this code to use IS_ENABLED() instead.
---
(no changes since v1)
cmd/sf.c | 30 +++++++++++++----------------- 1 file changed, 13 insertions(+), 17 deletions(-)
cmd/sf.c +13 -17
@@ -384,7 +384,6 @@ static int do_spi_protect(int argc, char *const argv[])return ret ==0 ?0 :1 ; } -#ifdef CONFIG_CMD_SF_TEST enum {STAGE_ERASE ,STAGE_CHECK ,
@@ -548,7 +547,6 @@ static int do_spi_flash_test(int argc, char *const argv[])return 0 ; } -#endif /* CONFIG_CMD_SF_TEST */ static int do_spi_flash (struct cmd_tbl *cmdtp,int flag,int argc,char *const argv[])
@@ -582,10 +580,8 @@ static int do_spi_flash(struct cmd_tbl *cmdtp, int flag, int argc, ret =do_spi_flash_erase (argc, argv); else if (strcmp(cmd, "protect") == 0) ret =do_spi_protect (argc, argv); -#ifdef CONFIG_CMD_SF_TEST - else if (!strcmp(cmd, "test")) + else if (IS_ENABLED(CONFIG_CMD_SF_TEST) && !strcmp(cmd, "test")) ret =do_spi_flash_test (argc, argv); -#endif else ret =-1 ;
@@ -597,16 +593,8 @@ usage:return CMD_RET_USAGE ; } -#ifdef CONFIG_CMD_SF_TEST -#define SF_TEST_HELP "\nsf test offset len " \ - "- run a very basic destructive test" -#else -#define SF_TEST_HELP -#endif - -U_BOOT_CMD ( - sf,5 ,1 , do_spi_flash, -"SPI flash sub-system" , +#ifdef CONFIG_SYS_LONGHELP +static const char long_help[] ="probe [[bus:]cs] [hz] [mode] - init flash device on given SPI bus\n" " and chip select\n" "sf read addr offset|partition len - read `len' bytes starting at\n"
@@ -622,6 +610,14 @@ U_BOOT_CMD(" at `addr' to flash at `offset'\n" " or to start of mtd `partition'\n" "sf protect lock/unlock sector len - protect/unprotect 'len' bytes starting\n" -" at address 'sector'\n" -SF_TEST_HELP +" at address 'sector'" +#ifdef CONFIG_CMD_SF_TEST + "\nsf test offset len - run a very basic destructive test" +#endif +#endif /* CONFIG_SYS_LONGHELP */ + ; + +U_BOOT_CMD ( + sf,5 ,1 , do_spi_flash, +"SPI flash sub-system" , long_help );
Signature · 1 line
2.33.0.464.g1972c5931b-goog
Re: [PATCH v4 3/5] sf: Tidy up code to avoid #ifdef
Hi Simon,
Quoted 5 lines · Simon Glass
Update this code to use
IS_ENABLED()instead.Signed-off-by Simon Glass <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>
Nitpick: Trailers shouldn't have a blank line between them. I see it for this patch and 4/5 as well. It probably doesn't matter, but I wonder if it will trip up some tools that work on commit trailers like git-interpret-trailers. Something you might want to fix in your workflow...
Signature · 3 lines
Regards, Pratyush Yadav Texas Instruments Inc.
Re: [PATCH v4 3/5] sf: Tidy up code to avoid #ifdef
Hi Pratyush,
Quoted 8 lines · Pratyush Yadav
Hi Simon,
Quoted 4 lines · Simon Glass
Update this code to use
IS_ENABLED()instead.Signed-off-by Simon Glass <[email protected]>
Quoted 7 lines
Quoted 1 line
Reviewed-by Pratyush Yadav <[email protected]>Nitpick: Trailers shouldn't have a blank line between them. I see it for this patch and 4/5 as well. It probably doesn't matter, but I wonder if it will trip up some tools that work on commit trailers like git-interpret-trailers. Something you might want to fix in your workflow...
The fix to 'patman status -d <new_branch>' is in mainline but was not in the tree where I ran this tool, unfortunately.
So hopefully this won't happen again.
Regards,
Simon
Re: [PATCH v4 3/5] sf: Tidy up code to avoid #ifdef
In reply to Simon GlassQuoted 8 lines · Simon Glass
Update this code to use
IS_ENABLED()instead.Signed-off-by Simon Glass <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>---
Simon Glass
Reviewed-by Jagan TekiRe: [PATCH v4 3/5] sf: Tidy up code to avoid #ifdef
Quoted 8 lines · Simon Glass
Update this code to use
IS_ENABLED()instead.Signed-off-by Simon Glass <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>---
Applied to u-boot-dm, thanks!
Simon Glass
Reviewed-by Jagan TekiRe: [PATCH v4 3/5] sf: Tidy up code to avoid #ifdef
Quoted 8 lines · Simon Glass
Update this code to use
IS_ENABLED()instead.Signed-off-by Simon Glass <[email protected]>Reviewed-by Pratyush Yadav <[email protected]>---
Applied to u-boot-dm, thanks!
Applied to u-boot-dm/next, thanks!
PATCH v4 4/5 sf: doc: Add documentation for the 'sf' command
In reply to Simon GlassThis command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
---
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examples
Changes in v2:
- Many fixes as suggested by Heinrich
doc/usage/index.rst | 1 + doc/usage/sf.rst | 245 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+) create mode 100644 doc/usage/sf.rst
doc/usage/index.rst +1 -0
@@ -43,6 +43,7 @@ Shell commands qfw reset sbi + sf scp03 setexpr size
doc/usage/sf.rst new file +245 -0
245 lines not on this page · raw message
Signature · 1 line
2.33.0.464.g1972c5931b-goog
Re: [PATCH v4 4/5] sf: doc: Add documentation for the 'sf' command
Quoted 9 lines · Simon Glass
This command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
Signed-off-by Simon Glass <[email protected]>Acked-by Pratyush Yadav <[email protected]>---
Re: [PATCH v4 4/5] sf: doc: Add documentation for the 'sf' command
In reply to Simon GlassQuoted 26 lines · Simon Glass
This command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
Signed-off-by Simon Glass <[email protected]>Acked-by Pratyush Yadav <[email protected]>---
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examplesChanges in v2:
- Many fixes as suggested by Heinrichdoc/usage/index.rst | 1 + doc/usage/sf.rst | 245 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+)create mode 100644
doc/usage/sf.rstdoc/usage/index.rst +0
-0
doc/usage/index.rst Shell commands · line 46
Quoted 4 lines
doc/usage/index.rst +1
-0@@ -43,6 +43,7 @@ Shell commands qfw reset sbi + sf
Please, keep this list in alphabetical order.
Quoted 33 lines
doc/usage/index.rst +0
-0scp03 setexpr sizedoc/usage/sf.rst new file +24
-0@@ -0,0 +1,245 @@ +.. SPDX-License-Identifier: GPL-2.0+: + +sf command +========== + +Synopis +------- + +:: + + sf probe [[[<bus>:]<cs>] [<hz> [<mode>]]] + sf read <addr> <offset>|<partition> <len> + sf write <addr> <offset>|<partition> <len> + sf erase <offset>|<partition> <len> + sf update <addr> <offset>|<partition> <len> + sf protect lock|unlock <sector> <len> + sf test <offset>|<partition> <len> + +Description +----------- + +The *sf* command is used to access SPI flash, supporting read/write/erase and +a few other functions. +
doc/usage/sf.rst line 34
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+Probe +----- + +The flash must first be probed with *sf probe* before any of the other +subcommands can be used. All of the parameters are optional: + +bus + SPI bus number containing the SPI-flash chip, e.g. 0. If you don't know + the number, you can use 'dm uclass' to see all the spi devices, + and check the value for 'seq' for each one (here 0 and 2)::
I would have expected the 'spi' command to have an info sub-command like the other subsystems. But that is missing.
Quoted 28 lines
doc/usage/sf.rst new file +28
-0+ + uclass 89: spi + 0 spi@0 @ 05484960, seq 0 + 1 spi@1 @ 05484b40, seq 2 + +cs + SPI chip-select to use for the chip. This is often 0 and can be omitted, + but in some cases multiple slaves are attached to a SPI controller, + selected by a chip-select line for each one. + +hz + Speed of the SPI bus in hertz. This normally defaults to 100000, i.e. + 100KHz, which is very slow. Note that if the device exists in the + device tree, there might be a speed provided there, in which case this + setting is ignored. + +mode + SPI mode to use: + + ===== ================ + Mode Meaning + ===== ================ + 0 CPOL=0, CPHA=0 + 1 CPOL=0, CPHA=1 + 2 CPOL=1, CPHA=0 + 3 CPOL=1, CPHA=1 + ===== ================ +
doc/usage/sf.rst line 72
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+ Clock phase (CPHA) 0 means that data is transferred (sampled) on the + first clock edge; 1 means the second. + + Clock polarity (CPOL) controls the idle state of the clock, 0 for low, + 1 for high. + The active state is the opposite of idle. + + You may find this `SPI documentation`_ useful. + +Parameters for other subcommands (described below) are as follows:
I would not expect parameters for other sub-commands in chapter "Probe".
Please put all parameters into a separate section "Parameters". This makes navigation easier.
Quoted 84 lines
doc/usage/sf.rst new file +84
-0+ +addr + Memory address to start transfer + +offset + Flash offset to start transfer + +partition + If the parameter is not numeric, it is assumed to be a partition + description in the format <dev_type><dev_num>,<part_num> which is not + covered here. This requires CONFIG_CMD_MTDPARTS. + +len + Number of bytes to transfer + +Read +~~~~ + +Use *sf read* to read from SPI flash to memory. The read will fail if an +attempt is made to read past the end of the flash. + + +Write +~~~~~ + +Use *sf write* to write from memory to SPI flash. The SPI flash should be +erased first, since otherwise the result is undefined. + +The write will fail if an attempt is made to read past the end of the flash. + + +Erase +~~~~~ + +Use *sf erase* to erase a region of SPI flash. The erase will fail if any part +of the region to be erased is protected or lies past the end of the flash. It +may also fail if the start offset or length are not aligned to an erase region +(e.g. 256 bytes). + + +Update +~~~~~~ + +Use *sf update* to automatically erase and update a region of SPI flash from +memory. This works a sector at a time (typical 4KB or 64KB). For each +sector it first checks if the sector already has the right data. If so it is +skipped. If not, the sector is erased and the new data written. Note that if +the length is not a multiple of the erase size, the space after the data in +the last sector will be erased. If the offset does not start at the beginning +of an erase block, the operation will fail. + +Speed statistics are shown including the number of bytes that were already +correct. + + +Protect +~~~~~~~ + +SPI-flash chips often have a protection feature where the chip is split up into +regions which can be locked or unlocked. With *sf protect* it is possible to +change these settings, if supported by the driver. + +lock|unlock + Selects whether to lock or unlock the sectors + +<sector> + Start sector number to lock/unlock. This may be the byte offset or some + other value, depending on the chip. + +<len> + Number of bytes to lock/unlock + + +Test +~~~~ + +A convenient and fast *sf test* subcommand provides a way to check that SPI +flash is working as expected. This works in four stages: + + * erase - erases the entire region + * check - checks that the region is erased + * write - writes a test pattern to the region, consisting of the U-Boot code + * read - reads back the test pattern to check that it was written correctly +
doc/usage/sf.rst line 166
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+Memory is allocated for two buffers, each <len> bytes in size. At typical +size is 64KB to 1MB. The offset and size must be aligned to an erase boundary. + +Note that this test will fail if any part of the SPI flash is write-protected. + + +Examples +-------- + +This first example uses sandbox::
"the sandbox"
Quoted 37 lines
doc/usage/sf.rst new file +37
-0+ + => sf probe + SF: Detected m25p16 with page size 256 Bytes, erase size 64 KiB, total 2 MiB + => sf read 1000 1100 80000 + device 0 offset 0x1100, size 0x80000 + SF: 524288 bytes @ 0x1100 Read: OK + => md 1000 + 00001000: edfe0dd0 f33a0000 78000000 84250000 ......:....x..%. + 00001010: 28000000 11000000 10000000 00000000 ...(............ + 00001020: 6f050000 0c250000 00000000 00000000 ...o..%......... + 00001030: 00000000 00000000 00000000 00000000 ................ + 00001040: 00000000 00000000 00000000 00000000 ................ + 00001050: 00000000 00000000 00000000 00000000 ................ + 00001060: 00000000 00000000 00000000 00000000 ................ + 00001070: 00000000 00000000 01000000 00000000 ................ + 00001080: 03000000 04000000 00000000 01000000 ................ + 00001090: 03000000 04000000 0f000000 01000000 ................ + 000010a0: 03000000 08000000 1b000000 646e6173 ............sand + 000010b0: 00786f62 03000000 08000000 21000000 box............! + 000010c0: 646e6173 00786f62 01000000 61696c61 sandbox.....alia + 000010d0: 00736573 03000000 07000000 2c000000 ses............, + 000010e0: 6332692f 00003040 03000000 07000000 /i2c@0.......... + 000010f0: 31000000 6963702f 00003040 03000000 ...1/pci@0...... + => sf erase 0 80000 + SF: 524288 bytes @ 0x0 Erased: OK + => sf read 1000 1100 80000 + device 0 offset 0x1100, size 0x80000 + SF: 524288 bytes @ 0x1100 Read: OK + => md 1000 + 00001000: ffffffff ffffffff ffffffff ffffffff ................ + 00001010: ffffffff ffffffff ffffffff ffffffff ................ + 00001020: ffffffff ffffffff ffffffff ffffffff ................ + 00001030: ffffffff ffffffff ffffffff ffffffff ................ + 00001040: ffffffff ffffffff ffffffff ffffffff ................ + 00001050: ffffffff ffffffff ffffffff ffffffff ................ + 00001060: ffffffff ffffffff ffffffff ffffffff ................ + 00001070: ffffffff ffffffff ffffffff ffffffff ................
doc/usage/sf.rst line 213
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+ 00001080: ffffffff ffffffff ffffffff ffffffff ................ + 00001090: ffffffff ffffffff ffffffff ffffffff ................ + 000010a0: ffffffff ffffffff ffffffff ffffffff ................ + 000010b0: ffffffff ffffffff ffffffff ffffffff ................ + 000010c0: ffffffff ffffffff ffffffff ffffffff ................ + 000010d0: ffffffff ffffffff ffffffff ffffffff ................ + 000010e0: ffffffff ffffffff ffffffff ffffffff ................ + 000010f0: ffffffff ffffffff ffffffff ffffffff ................ + +This second example is running on coral, an x86 Chromebook::
Proper names should be capitalized: "Google Coral".
Quoted 23 lines
doc/usage/sf.rst new file +22
-0+ + => sf probe + SF: Detected w25q128fw with page size 256 Bytes, erase size 4 KiB, total 16 MiB + => sf erase 300000 80000 + SF: 524288 bytes @ 0x300000 Erased: OK + => sf update 1110000 300000 80000 + device 0 offset 0x300000, size 0x80000 + 524288 bytes written, 0 bytes skipped in 0.457s, speed 1164578 B/s + + # This does nothing as the flash is already updated + => sf update 1110000 300000 80000 + device 0 offset 0x300000, size 0x80000 + 0 bytes written, 524288 bytes skipped in 0.196s, speed 2684354 B/s + => sf test 00000 80000 # try a protected region + SPI flash test: + Erase failed (err = -5) + Test failed + => sf test 800000 80000 + SPI flash test: + 0 erase: 18 ticks, 28444 KiB/s 227.552 Mbps + 1 check: 192 ticks, 2666 KiB/s 21.328 Mbps + 2 write: 227 ticks, 2255 KiB/s 18.040 Mbps
doc/usage/sf.rst line 245
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+ 3 read: 189 ticks, 2708 KiB/s 21.664 Mbps + Test passed + 0 erase: 18 ticks, 28444 KiB/s 227.552 Mbps + 1 check: 192 ticks, 2666 KiB/s 21.328 Mbps + 2 write: 227 ticks, 2255 KiB/s 18.040 Mbps + 3 read: 189 ticks, 2708 KiB/s 21.664 Mbps + + +.. _SPI documentation: + https://en.wikipedia.org/wiki/Serial_Peripheral_Interface
Please, add a section "Configuration" (cf. loady command). For other commands we also describe the return value of $? to expect.
Best regards
Heinrich
Re: [PATCH v4 4/5] sf: doc: Add documentation for the 'sf' command
Hi Heinrich,
Quoted 123 lines · Heinrich Schuchardt
Quoted 26 lines · Simon Glass
This command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
Signed-off-by Simon Glass <[email protected]>Acked-by Pratyush Yadav <[email protected]>---
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examplesChanges in v2:
- Many fixes as suggested by Heinrichdoc/usage/index.rst | 1 + doc/usage/sf.rst | 245 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+)create mode 100644
doc/usage/sf.rstdoc/usage/index.rst +0
-0doc/usage/index.rst Shell commands · line 46
Quoted 4 lines
doc/usage/index.rst +1
-0@@ -43,6 +43,7 @@ Shell commands qfw reset sbi + sfPlease, keep this list in alphabetical order.
Quoted 33 lines
doc/usage/index.rst +0
-0scp03 setexpr sizedoc/usage/sf.rst new file +24
-0@@ -0,0 +1,245 @@ +.. SPDX-License-Identifier: GPL-2.0+: + +sf command +========== + +Synopis +------- + +:: + + sf probe [[[<bus>:]<cs>] [<hz> [<mode>]]] + sf read <addr> <offset>|<partition> <len> + sf write <addr> <offset>|<partition> <len> + sf erase <offset>|<partition> <len> + sf update <addr> <offset>|<partition> <len> + sf protect lock|unlock <sector> <len> + sf test <offset>|<partition> <len> + +Description +----------- + +The *sf* command is used to access SPI flash, supporting read/write/erase and +a few other functions. +doc/usage/sf.rst line 34
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+Probe +----- + +The flash must first be probed with *sf probe* before any of the other +subcommands can be used. All of the parameters are optional: + +bus + SPI bus number containing the SPI-flash chip, e.g. 0. If you don't know + the number, you can use 'dm uclass' to see all the spi devices, + and check the value for 'seq' for each one (here 0 and 2)::I would have expected the 'spi' command to have an info sub-command like the other subsystems. But that is missing.
Quoted 28 lines
doc/usage/sf.rst new file +28
-0+ + uclass 89: spi + 0 spi@0 @ 05484960, seq 0 + 1 spi@1 @ 05484b40, seq 2 + +cs + SPI chip-select to use for the chip. This is often 0 and can be omitted, + but in some cases multiple slaves are attached to a SPI controller, + selected by a chip-select line for each one. + +hz + Speed of the SPI bus in hertz. This normally defaults to 100000, i.e. + 100KHz, which is very slow. Note that if the device exists in the + device tree, there might be a speed provided there, in which case this + setting is ignored. + +mode + SPI mode to use: + + ===== ================ + Mode Meaning + ===== ================ + 0 CPOL=0, CPHA=0 + 1 CPOL=0, CPHA=1 + 2 CPOL=1, CPHA=0 + 3 CPOL=1, CPHA=1 + ===== ================ +doc/usage/sf.rst line 72
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+ Clock phase (CPHA) 0 means that data is transferred (sampled) on the + first clock edge; 1 means the second. + + Clock polarity (CPOL) controls the idle state of the clock, 0 for low, + 1 for high. + The active state is the opposite of idle. + + You may find this `SPI documentation`_ useful. + +Parameters for other subcommands (described below) are as follows:I would not expect parameters for other sub-commands in chapter "Probe".
Quoted 2 lines
Please put all parameters into a separate section "Parameters". This makes navigation easier.
This series was sent back in April and is now at version 5, after multiple rounds of changes. This version alone sat here for nearly two months. Who will want to write documentation in U-Boot if this is the process?
I don't disagree with most of your comments, just the timing, although the 'spi info' thing is highly debatable, as you cannot memory-map SPI itself, only SPI flash.
The common.h header removal suffered a similar fate and we have never resolved that, now 18 months later.
So please, let's get something in and move forward.
Regards,
Simon
Heinrich Schuchardt
Acked-byRe: [PATCH v4 4/5] sf: doc: Add documentation for the 'sf' command
Am 13. November 2021 15:21:13 MEZ schrieb Simon Glass <[email protected]>:
Quoted 134 lines
Hi Heinrich,
Quoted 123 lines · Heinrich Schuchardt
Quoted 26 lines · Simon Glass
This command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
Signed-off-by Simon Glass <[email protected]>Acked-by Pratyush Yadav <[email protected]>---
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examplesChanges in v2:
- Many fixes as suggested by Heinrichdoc/usage/index.rst | 1 + doc/usage/sf.rst | 245 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+)create mode 100644
doc/usage/sf.rstdoc/usage/index.rst +0
-0doc/usage/index.rst Shell commands · line 46
Quoted 4 lines
doc/usage/index.rst +1
-0@@ -43,6 +43,7 @@ Shell commands qfw reset sbi + sfPlease, keep this list in alphabetical order.
Quoted 33 lines
doc/usage/index.rst +0
-0scp03 setexpr sizedoc/usage/sf.rst new file +24
-0@@ -0,0 +1,245 @@ +.. SPDX-License-Identifier: GPL-2.0+: + +sf command +========== + +Synopis +------- + +:: + + sf probe [[[<bus>:]<cs>] [<hz> [<mode>]]] + sf read <addr> <offset>|<partition> <len> + sf write <addr> <offset>|<partition> <len> + sf erase <offset>|<partition> <len> + sf update <addr> <offset>|<partition> <len> + sf protect lock|unlock <sector> <len> + sf test <offset>|<partition> <len> + +Description +----------- + +The *sf* command is used to access SPI flash, supporting read/write/erase and +a few other functions. +doc/usage/sf.rst line 34
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+Probe +----- + +The flash must first be probed with *sf probe* before any of the other +subcommands can be used. All of the parameters are optional: + +bus + SPI bus number containing the SPI-flash chip, e.g. 0. If you don't know + the number, you can use 'dm uclass' to see all the spi devices, + and check the value for 'seq' for each one (here 0 and 2)::I would have expected the 'spi' command to have an info sub-command like the other subsystems. But that is missing.
Quoted 28 lines
doc/usage/sf.rst new file +28
-0+ + uclass 89: spi + 0 spi@0 @ 05484960, seq 0 + 1 spi@1 @ 05484b40, seq 2 + +cs + SPI chip-select to use for the chip. This is often 0 and can be omitted, + but in some cases multiple slaves are attached to a SPI controller, + selected by a chip-select line for each one. + +hz + Speed of the SPI bus in hertz. This normally defaults to 100000, i.e. + 100KHz, which is very slow. Note that if the device exists in the + device tree, there might be a speed provided there, in which case this + setting is ignored. + +mode + SPI mode to use: + + ===== ================ + Mode Meaning + ===== ================ + 0 CPOL=0, CPHA=0 + 1 CPOL=0, CPHA=1 + 2 CPOL=1, CPHA=0 + 3 CPOL=1, CPHA=1 + ===== ================ +doc/usage/sf.rst line 72
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+ Clock phase (CPHA) 0 means that data is transferred (sampled) on the + first clock edge; 1 means the second. + + Clock polarity (CPOL) controls the idle state of the clock, 0 for low, + 1 for high. + The active state is the opposite of idle. + + You may find this `SPI documentation`_ useful. + +Parameters for other subcommands (described below) are as follows:I would not expect parameters for other sub-commands in chapter "Probe".
Quoted 2 lines
Please put all parameters into a separate section "Parameters". This makes navigation easier.
This series was sent back in April and is now at version 5, after multiple rounds of changes. This version alone sat here for nearly two months. Who will want to write documentation in U-Boot if this is the process?
Quoted 8 lines
I don't disagree with most of your comments, just the timing, although the 'spi info' thing is highly debatable, as you cannot memory-map SPI itself, only SPI flash.
The
common.hheader removal suffered a similar fate and we have never resolved that, now 18 months later.So please, let's get something in and move forward.
We can still in improve the documentation in a follow up patch.
Quoted 3 lines
Regards,
Simon
Re: [PATCH v4 4/5] sf: doc: Add documentation for the 'sf' command
Hi Heinrich,
Quoted 140 lines · Heinrich Schuchardt
Am 13. November 2021 15:21:13 MEZ schrieb Simon Glass <[email protected]>:
Quoted 138 lines
Hi Heinrich,
Quoted 123 lines · Heinrich Schuchardt
Quoted 26 lines · Simon Glass
This command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
Signed-off-by Simon Glass <[email protected]>Acked-by Pratyush Yadav <[email protected]>---
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examplesChanges in v2:
- Many fixes as suggested by Heinrichdoc/usage/index.rst | 1 + doc/usage/sf.rst | 245 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+)create mode 100644
doc/usage/sf.rstdoc/usage/index.rst +0
-0doc/usage/index.rst Shell commands · line 46
Quoted 4 lines
doc/usage/index.rst +1
-0@@ -43,6 +43,7 @@ Shell commands qfw reset sbi + sfPlease, keep this list in alphabetical order.
Quoted 33 lines
doc/usage/index.rst +0
-0scp03 setexpr sizedoc/usage/sf.rst new file +24
-0@@ -0,0 +1,245 @@ +.. SPDX-License-Identifier: GPL-2.0+: + +sf command +========== + +Synopis +------- + +:: + + sf probe [[[<bus>:]<cs>] [<hz> [<mode>]]] + sf read <addr> <offset>|<partition> <len> + sf write <addr> <offset>|<partition> <len> + sf erase <offset>|<partition> <len> + sf update <addr> <offset>|<partition> <len> + sf protect lock|unlock <sector> <len> + sf test <offset>|<partition> <len> + +Description +----------- + +The *sf* command is used to access SPI flash, supporting read/write/erase and +a few other functions. +doc/usage/sf.rst line 34
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+Probe +----- + +The flash must first be probed with *sf probe* before any of the other +subcommands can be used. All of the parameters are optional: + +bus + SPI bus number containing the SPI-flash chip, e.g. 0. If you don't know + the number, you can use 'dm uclass' to see all the spi devices, + and check the value for 'seq' for each one (here 0 and 2)::I would have expected the 'spi' command to have an info sub-command like the other subsystems. But that is missing.
Quoted 28 lines
doc/usage/sf.rst new file +28
-0+ + uclass 89: spi + 0 spi@0 @ 05484960, seq 0 + 1 spi@1 @ 05484b40, seq 2 + +cs + SPI chip-select to use for the chip. This is often 0 and can be omitted, + but in some cases multiple slaves are attached to a SPI controller, + selected by a chip-select line for each one. + +hz + Speed of the SPI bus in hertz. This normally defaults to 100000, i.e. + 100KHz, which is very slow. Note that if the device exists in the + device tree, there might be a speed provided there, in which case this + setting is ignored. + +mode + SPI mode to use: + + ===== ================ + Mode Meaning + ===== ================ + 0 CPOL=0, CPHA=0 + 1 CPOL=0, CPHA=1 + 2 CPOL=1, CPHA=0 + 3 CPOL=1, CPHA=1 + ===== ================ +doc/usage/sf.rst line 72
Quoted 10 lines
doc/usage/sf.rst new file +10
-0+ Clock phase (CPHA) 0 means that data is transferred (sampled) on the + first clock edge; 1 means the second. + + Clock polarity (CPOL) controls the idle state of the clock, 0 for low, + 1 for high. + The active state is the opposite of idle. + + You may find this `SPI documentation`_ useful. + +Parameters for other subcommands (described below) are as follows:I would not expect parameters for other sub-commands in chapter "Probe".
Quoted 2 lines
Please put all parameters into a separate section "Parameters". This makes navigation easier.
This series was sent back in April and is now at version 5, after multiple rounds of changes. This version alone sat here for nearly two months. Who will want to write documentation in U-Boot if this is the process?
I don't disagree with most of your comments, just the timing, although the 'spi info' thing is highly debatable, as you cannot memory-map SPI itself, only SPI flash.
Quoted 8 lines
Quoted 4 lines
The
common.hheader removal suffered a similar fate and we have never resolved that, now 18 months later.So please, let's get something in and move forward.
We can still in improve the documentation in a follow up patch.
Thank you.
Regards,
Simon
Re: [PATCH v4 4/5] sf: doc: Add documentation for the 'sf' command
Hi Heinrich,
Quoted 28 lines · Heinrich Schuchardt
Am 13. November 2021 15:21:13 MEZ schrieb Simon Glass <[email protected]>:
Quoted 26 lines
Hi Heinrich,
Quoted 23 lines · Heinrich Schuchardt
Quoted 21 lines · Simon Glass
This command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
Signed-off-by Simon Glass <[email protected]>Acked-by Pratyush Yadav <[email protected]>---
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examplesChanges in v2:
- Many fixes as suggested by Heinrichdoc/usage/index.rst | 1 + doc/usage/sf.rst | 245 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+)create mode 100644
doc/usage/sf.rst
Applied to u-boot-dm, thanks!
Re: [PATCH v4 4/5] sf: doc: Add documentation for the 'sf' command
Hi Heinrich,
Quoted 28 lines · Heinrich Schuchardt
Am 13. November 2021 15:21:13 MEZ schrieb Simon Glass <[email protected]>:
Quoted 26 lines
Hi Heinrich,
Quoted 23 lines · Heinrich Schuchardt
Quoted 21 lines · Simon Glass
This command is fairly complicated so documentation is useful. Unfortunately I an not sure how the MTD side of things works and cannot find information about that.
Signed-off-by Simon Glass <[email protected]>Acked-by Pratyush Yadav <[email protected]>---
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examplesChanges in v2:
- Many fixes as suggested by Heinrichdoc/usage/index.rst | 1 + doc/usage/sf.rst | 245 ++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 246 insertions(+)create mode 100644
doc/usage/sf.rst
Applied to u-boot-dm, thanks!
Applied to u-boot-dm/next, thanks!
PATCH v4 5/5 sf: Provide a command to access memory-mapped SPI
In reply to Simon GlassAdd a new 'sf mmap' function to show the address of a SPI offset, if the hardware supports it. This is useful on x86 systems.
---
(no changes since v3)
Changes in v3:
- Add configuration and return value also
arch/Kconfig | 2 ++ cmd/Kconfig | 8 ++++++ cmd/sf.c | 35 +++++++++++++++++++++++++++ doc/usage/sf.rst | 63 ++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 108 insertions(+)
arch/Kconfig +2 -0
@@ -136,6 +136,7 @@ config SANDBOX imply CMD_LZMADEC imply CMD_SATA imply CMD_SF + imply CMD_SF_MMAP imply CMD_SF_TEST imply CRC32_VERIFY imply FAT_WRITE
@@ -200,6 +201,7 @@ config X86 imply CMD_IRQ imply CMD_PCI imply CMD_SF + imply CMD_SF_MMAP imply CMD_SF_TEST imply CMD_ZBOOT imply DM_ETH
cmd/Kconfig +8 -0
@@ -1273,6 +1273,14 @@ config CMD_SF help SPI Flash support +config CMD_SF_MMAP + bool "sf mmap - Access memory-mapped SPI flash" + depends on CMD_SF + help + On some systems part of the SPI flash is mapped into mmemory. This + command provides a way to map a SPI-flash offset to a memory address, + so that the contents can be browsed using 'md', for example. + config CMD_SF_TEST bool "sf test - Allow testing of SPI flash" depends on CMD_SF
cmd/sf.c +35 -0
@@ -384,6 +384,36 @@ static int do_spi_protect(int argc, char *const argv[])return ret ==0 ?0 :1 ; } +static int do_spi_flash_mmap (int argc,char *const argv[]) +{ +loff_t offset, len, maxsize; +uint map_offset, map_size; +ulong map_base; +int dev =0 ; +int ret; + +if (argc <2 ) +return CMD_RET_USAGE ; + + ret =mtd_arg_off_size (argc -1 , &argv[1 ], &dev, &offset, &len, + &maxsize,MTD_DEV_TYPE_NOR , flash->size); +if (ret) +return ret; + + ret =dm_spi_get_mmap (flash->dev, &map_base, &map_size, &map_offset); +if (ret) { +printf ("Mapping not available (err=%d)\n" , ret); +return CMD_RET_FAILURE ; + } +if (offset <0 || offset + len > map_size) { +printf ("Offset out of range (map size %x)\n" , map_size); +return CMD_RET_FAILURE ; + } +printf ("%lx\n" , map_base + (ulong )offset); + +return 0 ; +} +enum {STAGE_ERASE ,STAGE_CHECK ,
@@ -580,6 +610,8 @@ static int do_spi_flash(struct cmd_tbl *cmdtp, int flag, int argc, ret =do_spi_flash_erase (argc, argv); else if (strcmp(cmd, "protect") == 0) ret =do_spi_protect (argc, argv); + else if (IS_ENABLED(CONFIG_CMD_SF_MMAP) && !strcmp(cmd, "mmap")) + ret =do_spi_flash_mmap (argc, argv); else if (IS_ENABLED(CONFIG_CMD_SF_TEST) && !strcmp(cmd, "test")) ret =do_spi_flash_test (argc, argv); else
@@ -611,6 +643,9 @@ static const char long_help[] = " or to start of mtd `partition'\n" "sf protect lock/unlock sector len - protect/unprotect 'len' bytes starting\n" " at address 'sector'" +#ifdef CONFIG_CMD_SF_MMAP + "\nsf mmap offset len - get memory address of SPI-flash offset\n" +#endif#ifdef CONFIG_CMD_SF_TEST "\nsf test offset len - run a very basic destructive test"#endif
doc/usage/sf.rst +63 -0
81 lines not on this page · raw message
Signature · 1 line
2.33.0.464.g1972c5931b-goog
Re: [PATCH v4 5/5] sf: Provide a command to access memory-mapped SPI
Quoted 3 lines · Simon Glass
Add a new 'sf mmap' function to show the address of a SPI offset, if the hardware supports it. This is useful on x86 systems.
I'm not quite sure about growing sf for limited use cases, maybe support it in existing arguments might be a good option.
Jagan.
Re: [PATCH v4 5/5] sf: Provide a command to access memory-mapped SPI
Hi Jagan,
Quoted 8 lines · Jagan Teki
Quoted 3 lines · Simon Glass
Add a new 'sf mmap' function to show the address of a SPI offset, if the hardware supports it. This is useful on x86 systems.
I'm not quite sure about growing sf for limited use cases, maybe support it in existing arguments might be a good option.
At least on x86 this is a common feature.
What do you suggest for this? Can you give an example of what arguments might be used?
Regards,
Simon
Re: [PATCH v4 5/5] sf: Provide a command to access memory-mapped SPI
In reply to Simon GlassQuoted 1 line · Simon Glass
Add a new 'sf mmap' function to show the address of a SPI offset, if the
I would expect a 'spi info' command to provide this information.
Best regards
Heinrich
Quoted 204 lines
hardware supports it. This is useful on x86 systems.
Signed-off-by Simon Glass <[email protected]>---
(no changes since v3)
Changes in v3:
- Add configuration and return value alsoarch/Kconfig | 2 ++ cmd/Kconfig | 8 ++++++ cmd/sf.c | 35 +++++++++++++++++++++++++++ doc/usage/sf.rst | 63 ++++++++++++++++++++++++++++++++++++++++++++++++ 4 files changed, 108 insertions(+)arch/Kconfig +2
-0@@ -136,6 +136,7 @@ config SANDBOX imply CMD_LZMADEC imply CMD_SATA imply CMD_SF + imply CMD_SF_MMAP imply CMD_SF_TEST imply CRC32_VERIFY imply FAT_WRITE@@ -200,6 +201,7 @@ config X86 imply CMD_IRQ imply CMD_PCI imply CMD_SF + imply CMD_SF_MMAP imply CMD_SF_TEST imply CMD_ZBOOT imply DM_ETHcmd/Kconfig +8
-0@@ -1273,6 +1273,14 @@ config CMD_SF help SPI Flash support +config CMD_SF_MMAP + bool "sf mmap - Access memory-mapped SPI flash" + depends on CMD_SF + help + On some systems part of the SPI flash is mapped into mmemory. This + command provides a way to map a SPI-flash offset to a memory address, + so that the contents can be browsed using 'md', for example. + config CMD_SF_TEST bool "sf test - Allow testing of SPI flash" depends on CMD_SFcmd/sf.c +35
-0@@ -384,6 +384,36 @@ static int do_spi_protect(int argc, char *const argv[])return ret ==0 ?0 :1 ; } +static int do_spi_flash_mmap (int argc,char *const argv[]) +{ +loff_t offset, len, maxsize; +uint map_offset, map_size; +ulong map_base; +int dev =0 ; +int ret; + +if (argc <2 ) +return CMD_RET_USAGE ; + + ret =mtd_arg_off_size (argc -1 , &argv[1 ], &dev, &offset, &len, + &maxsize,MTD_DEV_TYPE_NOR , flash->size); +if (ret) +return ret; + + ret =dm_spi_get_mmap (flash->dev, &map_base, &map_size, &map_offset); +if (ret) { +printf ("Mapping not available (err=%d)\n" , ret); +return CMD_RET_FAILURE ; + } +if (offset <0 || offset + len > map_size) { +printf ("Offset out of range (map size %x)\n" , map_size); +return CMD_RET_FAILURE ; + } +printf ("%lx\n" , map_base + (ulong )offset); + +return 0 ; +} +enum {STAGE_ERASE ,STAGE_CHECK ,@@ -580,6 +610,8 @@ static int do_spi_flash(struct cmd_tbl *cmdtp, int flag, int argc, ret =do_spi_flash_erase (argc, argv); else if (strcmp(cmd, "protect") == 0) ret =do_spi_protect (argc, argv); + else if (IS_ENABLED(CONFIG_CMD_SF_MMAP) && !strcmp(cmd, "mmap")) + ret =do_spi_flash_mmap (argc, argv); else if (IS_ENABLED(CONFIG_CMD_SF_TEST) && !strcmp(cmd, "test")) ret =do_spi_flash_test (argc, argv); else@@ -611,6 +643,9 @@ static const char long_help[] = " or to start of mtd `partition'\n" "sf protect lock/unlock sector len - protect/unprotect 'len' bytes starting\n" " at address 'sector'" +#ifdef CONFIG_CMD_SF_MMAP + "\nsf mmap offset len - get memory address of SPI-flash offset\n" +#endif#ifdef CONFIG_CMD_SF_TEST "\nsf test offset len - run a very basic destructive test"#endif doc/usage/sf.rst +63
-0@@ -14,6 +14,7 @@ Synopis sf erase <offset>|<partition> <len> sf update <addr> <offset>|<partition> <len> sf protect lock|unlock <sector> <len> + sf mmap <offset>|<partition> <len> sf test <offset>|<partition> <len> Description@@ -143,6 +144,16 @@ lock|unlock Number of bytes to lock/unlock +Memory-mapped flash +------------------- + +On some systems part of the SPI flash is mapped into mmemory. With *sf mmap* +you can map a SPI-flash offset to a memory address, so that the contents can be +browsed using 'md', for example. + +The command will fail if this is not supported by the hardware or driver. + + Test ~~~~@@ -240,6 +251,58 @@ This second example is running on coral, an x86 Chromebook:: 2 write: 227 ticks, 2255 KiB/s 18.040 Mbps 3 read: 189 ticks, 2708 KiB/s 21.664 Mbps + # On coral, SPI flash offset 0 corresponds to address ff081000 in memory + => sf mmap 0 1000 + device 0 offset 0x0, size 0x1000 + ff081000 + + # See below for how this address was obtained + => sf mmap e80000 11e18 + device 0 offset 0xe80000, size 0x11e18 + fff01000 + => md fff01000 + fff01000: b2e8e089 89000030 30b4e8c4 c0310000 ....0......0..1. + fff01010: 002c95e8 2ce8e800 feeb0000 dfe8c489 ..,....,........ + fff01020: f400002c 83f4fdeb d4e80cec 3100001e ,..............1 + fff01030: 0cc483c0 f883c3c3 8b0b770a df408504 .........w....@. + fff01040: c085fef1 c8b80575 c3fef1e5 53565755 ....u.......UWVS + fff01050: 89c38951 2404c7c5 80000002 8924048b Q......$......$. + fff01060: 89a20fdf 89fb89de 75890045 084d8904 ........E..u..M. + fff01070: ff0c5589 c5832404 243c8110 80000005 .U...$....<$.... + fff01080: 43c6da75 3b800030 43037520 d889f8eb u..C0..; u.C.... + fff01090: 5f5e5b5a 80e6c35d 535657c3 e7e8c689 Z[^_]....WVS.... + fff010a0: 89000069 00a164c3 8b000000 408b4c56 i....d......VL.@ + fff010b0: 0cec837c ddb9ff6a e8fef1e5 00004e01 |...j........N.. + fff010c0: a1640389 00000000 1c80b60f 66000001 ..d............f + fff010d0: b80c4389 00000001 a20fdf89 fb89de89 .C.............. + fff010e0: 89104389 c4831453 5bc03110 56c35f5e .C..S....1.[^_.V + fff010f0: 14ec8353 ce89d389 0000a164 b60f0000 S.......d....... + + +The offset e80000 was obtained using the following binman command, to find the +location of U-Boot SPL:: + + $ binman ls -i /tmp/b/chromebook_coral/u-boot.rom + Name Image-pos Size Entry-type Offset Uncomp-size + ------------------------------------------------------------------------------------------------------ + main-section 0 1000000 section 0 + spl e80000 11e18 u-boot-spl ffe80000 + u-boot d00000 9106e u-boot ffd00000 + ... + .. _SPI documentation: https://en.wikipedia.org/wiki/Serial_Peripheral_Interface + + +Configuration +------------- + +The *sf* command is only available if `CONFIG_CMD_SF=y`. Note that it depends on +`CONFIG_DM_SPI_FLASH`. + +Return value +------------ + +The return value $? is set to 0 (true) if the command succeded and to 1 (false) +otherwise.
Re: [PATCH v4 5/5] sf: Provide a command to access memory-mapped SPI
Hi Heinrich,
Quoted 7 lines · Heinrich Schuchardt
Quoted 1 line · Simon Glass
Add a new 'sf mmap' function to show the address of a SPI offset, if the
I would expect a 'spi info' command to provide this information.
Why is that? It depends on the size of actual SPI-flash device. It is not a property of the SPI bus, so far as I understand it.
Regards,
Simon
Hi Jagan,
Quoted 24 lines · Simon Glass
This little series adds documentation and a few other tidy-ups to the 'sf' command.
It also provides a way to access memory-mapped SPI via the command line.
Changes in v4:
- Split out the 'const' change into a separate patch
- Show the 'sf probe' output in the examplesChanges in v3:
- Add configuration and return value alsoChanges in v2:
- Explain why 'usage' cannot be const
- Many fixes as suggested by HeinrichSimon Glass (5):
command: Use a constant pointer for the help
sf: Use const for the stage name
sf: Tidy up code to avoid #ifdef
sf: doc: Add documentation for the 'sf' command
sf: Provide a command to access memory-mapped SPI
Quoted 8 lines
arch/Kconfig | 2 + cmd/Kconfig | 8 ++ cmd/sf.c | 67 +++++++--- doc/usage/index.rst | 1 + doc/usage/sf.rst | 308 ++++++++++++++++++++++++++++++++++++++++++++ include/command.h | 2 +- 6 files changed, 369 insertions(+), 19 deletions(-) create mode 100644 doc/usage/sf.rst
Any word on this series please?
Regards,
Simon