Hi Bob,
It was 7 years ago, but today I saw a similar message with a USB Thumb
drive:nsh> ehci_waiter: connected
nsh> ls /dev
/dev:
console
null
sda
ttyS0
nsh> mount -t vfat /dev/sda /mnt
fat_checkbootrecord: WARNING: Signature: aa55 FS sectorsize: 20487 HW
sectorsize: 512
nsh>
Maybe it sector size is wrong, this is a USB Stick that came with my 3D
Printer and it was support to work, at least it mounts correctly and I can
see the content:
nsh> ls /mnt
/mnt:
System Volume Information/
Delta filament barrel base_Hyper
K1C air filter-K1C_0.4_PLA-CF_37
K1C_supplementary files_EN_V1.4.
Logo filament barrel base_Hyper
Side fan cowl_Generic-ABS_47m.gc
3DBenchy.gcode
600S-TEST_Hyper PLA_7m.gcode
CR4CU220812S11_ota_img_V1.3.3.30
usb-sniffer-v6-Case_PLA_37m52s.g
nsh>
If you can share more info about your debug process for that similar issue,
it could help.
BR,
Alan
On Sun, Jul 22, 2018 at 2:15 PM <bob.feret...@rafresearch.com> wrote:
> OK. Works for me. Let's consider this issue "Solved".
>
> On Friday, July 20, 2018 at 6:21:38 PM UTC-7, bob.fe...@rafresearch.com
> wrote:
>>
>> I have been observing the below error messages every time I automount an
>> SD-card.
>>
>> fat_checkbootrecord: ERROR: Signature: aa55 FS: 0 HW sectorsize: 512
>> (That is with a new, manufacturer formatted SD-card. After Linux
>> re-formats the card it's FS: 36352.)
>>
>> It did not seem to cause any failures, so I made it a lower priority for
>> investigation. Now that I had a bit of time to look further, these are my
>> initial findings.
>>
>> First let's refresh our memories with the FAT32 file system structure.
>>
>> Master Boot Record (MBR) located in the first sector of a SD-card (sector
>> 0x0).
>> It contains executable boot code, four partition entries, and a 0x55AA
>> signature in bytes 510 & 511.
>>
>> FAT32 Boot Record located at sector 0x800.
>> This sector contains a jump instruction, BytesPerSector,
>> SectorsPerCluster, and all of the other things that fs/fat/fs_fat32.h
>> claims are in the MBR. Also there is a 0x55AA signature in bytes 510 & 511.
>>
>> The printf() function that generates this error message is...
>> if (MBR_GETSIGNATURE(fs->fs_buffer) != BOOT_SIGNATURE16 ||
>> MBR_GETBYTESPERSEC(fs->fs_buffer) != fs->fs_hwsectorsize)
>> {
>> ferr("ERROR: Signature: %04x FS sectorsize: %d HW sectorsize: %d\n",
>> MBR_GETSIGNATURE(fs->fs_buffer),
>> MBR_GETBYTESPERSEC(fs->fs_buffer),
>> fs->fs_hwsectorsize);
>>
>> return -EINVAL;
>> }
>>
>> The problem is that is that fat is really reading these fields from the
>> MBR (sector 0) when they are in the FAT32 Boot Record located at sector
>> 0x800.
>>
>> I am posting this to the Group, because I don't know how far this
>> confusion has spread in the fatfs code. This aspect of it seems to have
>> only cosmetic consequences, but I am still seeing occasional automount bind
>> failures. (Not debounce related). And I am wondering if this MBR vs FAT32
>> Boot Record confusion has been more widespread than this cosmetic issue.
>>
>> Has anyone else been noticing anomalies during mount or automount?
>>
>> Does anyone know the history of this code and can shed some light on
>> other sections that can have been infected by the MBR vs FAT32 Boot Record
>> confusion?
>>
>> Regards,
>> Bob Feretich
>>
>> --
> You received this message because you are subscribed to the Google Groups
> "NuttX" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to nuttx+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>
