[PATCH] Revert "drivers: scsi: fix inaccurate block count reporting in scsi operations"

[Balaji Selvanathan]

Hi Tom,

Apologies for late reply. Somehow missed the mail.

Upon analysis, it seems the most likely root cause is Tony's 238GB disk 
triggers the CONFIG_SYS_64BIT_LBA code path in scsi_read function. Any 
read beyond block 268,435,455 blocks enters this path.

'blks' is decremented both inside the 64-bit LBA block AND after 
successful scsi_exec(), causing incorrect block count tracking and (most 
likely) the data abort observed.

The fix is to remove 'blks -= blocks;' from inside the 
CONFIG_SYS_64BIT_LBA block in scsi_read() function.

Hi Tony, can you please check with this change?

Tom, shall I raise a new patch for this now?

Regards,

Balaji

On 12/16/2025 10:36 PM, Tom Rini wrote:
> On Fri, Dec 12, 2025 at 08:50:56AM -0600, Tom Rini wrote:
>> On Mon, Dec 08, 2025 at 06:53:52PM -0800, Tony Dinh wrote:
>>
>>> This reverts commit b32dda34506b4f486bc803d0c7251f987edd2455.
>>>
>>> Currently, scsi is broken and causes data abort. I did a bisect and found
>>> that this commit was the cause. Reverting it and scsi is working again.
>>>
>>> === Error Log with data abort
>>>
>>> U-Boot 2026.01-rc4-tld-1-g0e0a198a68be (Dec 08 2025 - 17:42:05 -0800)
>>> Synology DS116
>>>
>>> DS116> scsi reset
>>>
>>> Reset SCSI
>>> scanning bus for devices...
>>> Target spinup took 0 ms.
>>> SATA link 1 timeout.
>>> AHCI 0001.0000 32 slots 2 ports 6 Gbps 0x3 impl SATA mode
>>> flags: 64bit ncq led only pmp fbss pio slum part sxs
>>>    Device 0: (0:0) Vendor: ATA Prod.: FUJITSU MHY2250B Rev: 890B
>>>              Type: Hard Disk
>>>              Capacity: 238475.1 MB = 232.8 GB (488397169 x 512)
>>>
>>> DS116> ext2ls scsi 0:1 /boot
>>> data abort
>>> pc : [<3ffa5d3e>]	   lr : [<3ffc1515>]
>>> sp : 3fb65ad8  ip : 0000000c	 fp : 3fb7f220
>>> r10: 00002710  r9 : 3fb6aed0	 r8 : 3fb85e40
>>> r7 : f10a8100  r6 : 000003e8	 r5 : 000003e8  r4 : 3fb65ae0
>>> r3 : e38e39c1  r2 : 00000000	 r1 : 3fb65ae0  r0 : 00931767
>>> Flags: Nzcv  IRQs off  FIQs off  Mode SVC_32 (T)
>>> Code: 3ffd b510 6803 460c (6bdb) 681b
>>>
>>> Signed-off-by: Tony Dinh <mibodhi at gmail.com>
>> Any comments Balaji ?
> Again, any comments Balaji? I'm otherwise inclined to take this revert
> this week and sort things out again in the next branch / post v2026.01.
>