Fix fastboot handling of partitions when no slots are, supported
PATCH 1/1 Fix fastboot handling of partitions when no slots are, supported
The fastboot module has a bug that prevents some command to work properly on devices that haven't an Android-like partition scheme, that is, just one spl and one kernel partition, instead of the redundant scheme with _a and _b slots.
This is the schema of our NAND storage (board is based on an AllWinner A33 sunxi chip):
=> mtdparts
device nand0 <1c03000.nand>, # parts = 4
#: name size net size offset mask_flags 0: spl 0x00020000 0x00020000 0x00000000 0 1: uboot 0x00100000 0x00100000 0x00020000 0 2: kernel 0x00400000 0x00400000 0x00120000 0 3: ubi 0x07ae0000 0x079e0000 (!) 0x00520000 0
active partition: nand0,0 - (spl) 0x00020000 @ 0x00000000
This happens when we try to erase the spl partition using fastboot:
$ fastboot erase spl Erasing 'spl_a' FAILED (remote: 'invalid NAND device') fastboot: error: Command failed
The error occurs because getvars fails to handle the error returned by nand layer when a partition cannot be found.
Indeed, getvar_get_part_info returns what is returned by fastboot_nand_get_part_info (0 on success, 1 on failure) but it should return -ENODEV or -EINVAL instead. Since the cause of failure is not returned by the nand function, I decided to return -EINVAL to make it simple.
---
drivers/fastboot/fb_getvar.c | 5 ++++-
drivers/fastboot/fb_nand.c | 2 +-
2 files changed, 5 insertions(+), 2 deletions(-)
drivers/fastboot/fb_getvar.c +0 -0
@@ -121,8 +121,11 @@ static int getvar_get_part_info(const char *part_name, char *response,
*size = disk_part.size * disk_part.blksz;
} else if (IS_ENABLED(CONFIG_FASTBOOT_FLASH_NAND)) {
r = fastboot_nand_get_part_info(part_name, &part_info, response); - if (r >= 0 && size)
+ if (r == 0 && size) {
*size = part_info->size;
+ } else {
+ r = -EINVAL;
+ }
} else {
fastboot_fail("this storage is not supported in bootloader", response); r = -ENODEV;
drivers/fastboot/fb_nand.c +0 -0
@@ -141,7 +141,7 @@ static lbaint_t fb_nand_sparse_reserve(struct sparse_storage *info,
*
* @part_name: Named device to lookup
* @part_info: Pointer to returned part_info pointer
- * @response: Pointer to fastboot response buffer
+ * @response: 0 on success, 1 otherwise
*/
int fastboot_nand_get_part_info(const char *part_name, struct part_info **part_info, char *response)
Signature · 1 line
2.48.1
PATCH 1/1 Fix fastboot handling of partitions when no slots are, supported
The fastboot module has a bug that prevents some command to work properly on devices that haven't an Android-like partition scheme, that is, just one spl and one kernel partition, instead of the redundant scheme with _a and _b slots.
This is the schema of our NAND storage (board is based on an AllWinner A33 sunxi chip):
=> mtdparts
device nand0 <1c03000.nand>, # parts = 4
#: name size net size offset mask_flags 0: spl 0x00020000 0x00020000 0x00000000 0 1: uboot 0x00100000 0x00100000 0x00020000 0 2: kernel 0x00400000 0x00400000 0x00120000 0 3: ubi 0x07ae0000 0x079e0000 (!) 0x00520000 0
active partition: nand0,0 - (spl) 0x00020000 @ 0x00000000
This happens when we try to erase the spl partition using fastboot:
$ fastboot erase spl Erasing 'spl_a' FAILED (remote: 'invalid NAND device') fastboot: error: Command failed
The error occurs because getvars fails to handle the error returned by nand layer when a partition cannot be found.
Indeed, getvar_get_part_info returns what is returned by fastboot_nand_get_part_info (0 on success, 1 on failure) but it should return -ENODEV or -EINVAL instead. Since the cause of failure is not returned by the nand function, I decided to return -EINVAL to make it simple.
--- drivers/fastboot/fb_getvar.c | 5 ++++- drivers/fastboot/fb_nand.c | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-)
drivers/fastboot/fb_getvar.c +4 -1
@@ -121,8 +121,11 @@ static int getvar_get_part_info(const char *part_name, char *response, *size = disk_part.size * disk_part.blksz; } else if (IS_ENABLED(CONFIG_FASTBOOT_FLASH_NAND)) { r = fastboot_nand_get_part_info(part_name, &part_info, response); - if (r >= 0 && size) + if (r == 0 && size) { *size = part_info->size; + } else { + r = -EINVAL; + } } else { fastboot_fail("this storage is not supported in bootloader",
response);
r = -ENODEV;
drivers/fastboot/fb_nand.c +1 -1
@@ -141,7 +141,7 @@ static lbaint_t fb_nand_sparse_reserve(struct sparse_storage *info,* * @part_name: Named device to lookup * @part_info: Pointer to returned part_info pointer -* @response: Pointer to fastboot response buffer +* @response: 0 on success, 1 otherwise */ int fastboot_nand_get_part_info(const char *part_name, struct part_info **part_info, char *response)
Signature · 1 line
2.48.1
Hi Federico,
Thank you for the patch.
Quoted 28 lines · Federico Fuga
The fastboot module has a bug that prevents some command to work properly on devices that haven't an Android-like partition scheme, that is, just one spl and one kernel partition, instead of the redundant scheme with _a and _b slots.
This is the schema of our NAND storage (board is based on an AllWinner A33 sunxi chip):
=> mtdparts
device nand0 <1c03000.nand>, # parts = 4
#: name size net size offsetmask_flags0: spl 0x00020000 0x00020000 0x00000000 0 1: uboot 0x00100000 0x00100000 0x00020000 0 2: kernel 0x00400000 0x00400000 0x00120000 0 3: ubi 0x07ae00000x079e0000(!) 0x00520000 0active partition: nand0,0 - (spl)
0x00020000@0x00000000This happens when we try to erase the spl partition using fastboot:
$ fastboot erase spl Erasing 'spl_a' FAILED (remote: 'invalid NAND device') fastboot: error: Command failedThe error occurs because getvars fails to handle the error returned by nand layer when a partition cannot be found.
Quoted 7 lines
Indeed,
getvar_get_part_inforeturns what is returned byfastboot_nand_get_part_info(0 on success, 1 on failure) but it should return-ENODEVor-EINVALinstead. Since the cause of failure is not returned by the nand function, I decided to return-EINVALto make it simple.Signed-off-by Federico Fuga <[email protected]>
Unfortunately, I'm unable to apply this. I've tried using b4 and via git am.
$ b4 shazam -s -l --check [email protected]
Grabbing thread from lore.kernel.org/all/[email protected]/t.mbox.gz Checking for newer revisions
Grabbing search results from lore.kernel.org
Analyzing 2 messages in the thread
Assuming new revision: v2 ([PATCH] Fix fastboot handling of partitions when no slots are, supported) Analyzing 0 code-review messages
Will use the latest revision: v2
You can pick other revisions using the -vN flag
Checking attestation on all messages, may take a moment...
---
[PATCH] Fix fastboot handling of partitions when no slots are, supported
+ Link: https://lore.kernel.org/r/[email protected]
+ Signed-off-by: Mattijs Korpershoek <[email protected]>
Traceback (most recent call last):
File "/mnt/work/upstream/b4/src/b4/command.py", line 435, in <module>
cmd()
~~~^^ File "/mnt/work/upstream/b4/src/b4/command.py", line 417, in cmd
cmdargs.func(cmdargs)
Downloading the thread from this patchwork link (clicking on "mbox"):
https://patchwork.ozlabs.org/project/uboot/patch/[email protected]/
$ git am Fix-fastboot-handling-of-partitions-when-no-slots-are-supported.patch Applying: Fix fastboot handling of partitions when no slots are, supported error: corrupt patch at line 13 error: could not build fake ancestor Patch failed at 0001 Fix fastboot handling of partitions when no slots are, supported hint: Use 'git am --show-current-patch=diff' to see the failed patch hint: When you have resolved this problem, run "git am --continue". hint: If you prefer to skip this patch, run "git am --skip" instead. hint: To restore the original branch and stop patching, run "git am --abort". hint: Disable this message with "git config advice.mergeConflict false"
Can you please confirm that you are able to apply it yourself? Maybe the email client is misconfigured somehow?
I've also tried to apply the one that has been send yesterday
(https://patchwork.ozlabs.org/project/uboot/patch/[email protected]/)
but I'm encountering the same problem.
Quoted 40 lines
--- drivers/fastboot/fb_getvar.c | 5 ++++- drivers/fastboot/fb_nand.c | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-)drivers/fastboot/fb_getvar.c +4
-1@@ -121,8 +121,11 @@ static int getvar_get_part_info(const char *part_name, char *response, *size = disk_part.size * disk_part.blksz; } else if (IS_ENABLED(CONFIG_FASTBOOT_FLASH_NAND)) { r = fastboot_nand_get_part_info(part_name, &part_info, response); - if (r >= 0 && size) + if (r == 0 && size) { *size = part_info->size; + } else { + r = -EINVAL; + } } else { fastboot_fail("this storage is not supported in bootloader",response);
r =-ENODEV;drivers/fastboot/fb_nand.c +1
-1@@ -141,7 +141,7 @@ static lbaint_t fb_nand_sparse_reserve(structsparse_storage *info, * * @part_name: Named device to lookup * @part_info: Pointer to returned part_info pointer -* @response: Pointer to fastboot response buffer +* @response: 0 on success, 1 otherwise */ int fastboot_nand_get_part_info(const char *part_name,struct
part_info**part_info, char *response)Signature · 1 line
2.48.1
Hi Mattijs,
sorry, my bad: I exported from the wrong branch, as I was working on the latest release, not from main.
Resending it.
Sorry for the mistake.
Federico
Quoted 127 lines · Mattijs Korpershoek
Hi Federico,
Thank you for the patch.
Quoted 28 lines · Federico Fuga
The fastboot module has a bug that prevents some command to work properly on devices that haven't an Android-like partition scheme, that is, just one spl and one kernel partition, instead of the redundant scheme with _a and _b slots.
This is the schema of our NAND storage (board is based on an AllWinner A33 sunxi chip):
=> mtdparts
device nand0 <1c03000.nand>, # parts = 4
#: name size net size offsetmask_flags0: spl 0x00020000 0x00020000 0x00000000 0 1: uboot 0x00100000 0x00100000 0x00020000 0 2: kernel 0x00400000 0x00400000 0x00120000 0 3: ubi 0x07ae00000x079e0000(!) 0x00520000 0active partition: nand0,0 - (spl)
0x00020000@0x00000000This happens when we try to erase the spl partition using fastboot:
$ fastboot erase spl Erasing 'spl_a' FAILED (remote: 'invalid NAND device') fastboot: error: Command failedThe error occurs because getvars fails to handle the error returned by nand layer when a partition cannot be found.
Quoted 7 lines
Indeed,
getvar_get_part_inforeturns what is returned byfastboot_nand_get_part_info(0 on success, 1 on failure) but it should return-ENODEVor-EINVALinstead. Since the cause of failure is not returned by the nand function, I decided to return-EINVALto make it simple.Signed-off-by Federico Fuga <[email protected]>Unfortunately, I'm unable to apply this. I've tried using b4 and via git am.
$ b4 shazam -s -l --check [email protected]Grabbing thread from lore.kernel.org/all/[email protected]/t.mbox.gz Checking for newer revisions
Grabbing search results from lore.kernel.org
Analyzing 2 messages in the thread
Assuming new revision: v2 ([PATCH] Fix fastboot handling of partitions when no slots are, supported) Analyzing 0 code-review messages
Will use the latest revision: v2
You can pick other revisions using the -vN flag
Checking attestation on all messages, may take a moment...
---
[PATCH] Fix fastboot handling of partitions when no slots are, supported
+ Link: https://lore.kernel.org/r/[email protected]
+ Signed-off-by: Mattijs Korpershoek <[email protected]>
Traceback (most recent call last):File "/mnt/work/upstream/b4/src/b4/command.py", line 435, in <module> cmd() ~~~^^ File "/mnt/work/upstream/b4/src/b4/command.py", line 417, in cmd cmdargs.func(cmdargs)Downloading the thread from this patchwork link (clicking on "mbox"):
https://patchwork.ozlabs.org/project/uboot/patch/[email protected]/$ git am Fix-fastboot-handling-of-partitions-when-no-slots-are-supported.patch Applying: Fix fastboot handling of partitions when no slots are, supported error: corrupt patch at line 13 error: could not build fake ancestor Patch failed at 0001 Fix fastboot handling of partitions when no slots are, supported hint: Use 'git am --show-current-patch=diff' to see the failed patch hint: When you have resolved this problem, run "git am --continue". hint: If you prefer to skip this patch, run "git am --skip" instead. hint: To restore the original branch and stop patching, run "git am --abort". hint: Disable this message with "git config advice.mergeConflict false"Can you please confirm that you are able to apply it yourself? Maybe the email client is misconfigured somehow?
I've also tried to apply the one that has been send yesterday
(https://patchwork.ozlabs.org/project/uboot/patch/[email protected]/)
but I'm encountering the same problem.Quoted 40 lines
--- drivers/fastboot/fb_getvar.c | 5 ++++- drivers/fastboot/fb_nand.c | 2 +- 2 files changed, 5 insertions(+), 2 deletions(-)drivers/fastboot/fb_getvar.c +4
-1@@ -121,8 +121,11 @@ static int getvar_get_part_info(const char *part_name, char *response, *size = disk_part.size * disk_part.blksz; } else if (IS_ENABLED(CONFIG_FASTBOOT_FLASH_NAND)) { r = fastboot_nand_get_part_info(part_name, &part_info, response); - if (r >= 0 && size) + if (r == 0 && size) { *size = part_info->size; + } else { + r = -EINVAL; + } } else { fastboot_fail("this storage is not supported in bootloader",response);
r =-ENODEV;drivers/fastboot/fb_nand.c +1
-1@@ -141,7 +141,7 @@ static lbaint_t fb_nand_sparse_reserve(structsparse_storage *info, * * @part_name: Named device to lookup * @part_info: Pointer to returned part_info pointer -* @response: Pointer to fastboot response buffer +* @response: 0 on success, 1 otherwise */ int fastboot_nand_get_part_info(const char *part_name,struct
part_info**part_info, char *response)Signature · 1 line
2.48.1