Re: RFC: remove set_fs for m68k

[Michael Schmitz <schmitzmic@gmail.com>]

Hi Christoph,

Am 25.07.2021 um 19:36 schrieb Michael Schmitz:
> Hi Christoph,
>
> Am 23.07.2021 um 17:11 schrieb Christoph Hellwig:
>> On Fri, Jul 23, 2021 at 04:00:26PM +1200, Michael Schmitz wrote:
>>>> FYI, I've pushed out a new version - this contains the fix from Andreas
>>>> and drops all these uninlining tweaks as they very obviously cause more
>>>> harm than good for now.
>>>>
>>>> And sorry for the delay, life has been a little busy lately.
>>>
>>> The new version passed two rounds of tests OK - looking good so far.
>>
>> Thanks a lot!  I'll wait a bit and will send it out as a formal series.
>
> Just saw this:
>
> [256262.110000] *** FORMAT ERROR ***   FORMAT=0
> [256262.140000] Current process id is 2969
> [256262.170000] BAD KERNEL TRAP: 00000000
> [256262.190000] Modules linked in: atari_scsi ne 8390p
> [256262.220000] PC: [<00002a8c>] resume_userspace+0x14/0x16
> [256262.250000] SR: 2200  SP: 4b4338bf  a2: 00000000
> [256262.290000] d0: 00000008    d1: 00000003    d2: 00000ba1    d3:
> 00000000
> [256262.300000] d4: ffffffff    d5: 00000000    a0: 00000ba1    a1:
> 00000080
> [256262.320000] Process savelog (pid: 2969, task=d81f6132)
> [256262.340000] Frame format=0
> [256262.360000] Stack from 00c1bfa4:
> [256262.360000]         02108005 0d06b008 00000009 00000000 00000001
> 000347ba 0080ab00 00778b6c
> [256262.360000]         00c1be0c 0003e68e 00d11970 00000003 00000000
> 002c7bca 00035a66 00d11970
> [256262.360000]         00778b6c 00809014 00000000 00002700 0000006d
> 00002600 00049ef6
> [256262.470000] Call Trace: [<000347ba>] get_work_pool+0x0/0x38
> [256262.480000]  [<0003e68e>] wake_up_process+0x12/0x16
> [256262.500000]  [<002c7bca>] printk+0x0/0x18
> [256262.520000]  [<00035a66>] __queue_work+0x1a8/0x1be
> [256262.540000]  [<00002700>] wait_for_initramfs+0x50/0x58
> [256262.570000]  [<00002600>] name_to_dev_t+0x2b0/0x360
> [256262.580000]  [<00049ef6>] __irq_wake_thread+0x0/0x40
> [256262.600000]
> [256262.610000] Code: 1029 0007 660c 4cdf 073e 201f 588f dfdf <4e73>
> 254f 040c e308 660a 487a ffe0 60ff 002d 336e 598f 48e7 031e 486f 001c 61ff
>
> Same PC as the one I reported on the 18th. Kernel locked up hard and
> needing a hardware reset this time.
>
> I'll back out the top commit of your series (the one removing set_fs())
> and retest. Again, this may take a few days to show any results.

No further format errors, kernel panics or lockups seen after almost a 
week of tests.

That suggests something is wrong with your 'm68k: remove set_fs()' commit.

I'll try adding a get_fc() and warn whenever FC does not match what your 
patch expects, maybe that can help to get a clearer picture.

Cheers,

	Michael