On Tue, Jun 01, 2021 at 04:23:44PM +0000, Diego Santa Cruz wrote:
> > -----Original Message-----
> > From: Alexander Fomichev <fomichev.ru@xxxxxxxxx>
> > Sent: 01 June 2021 14:16
> > To: Diego Santa Cruz <Diego.SantaCruz@xxxxxxxxxxxx>; linux-
> > i2c@xxxxxxxxxxxxxxx
> > Cc: Bartosz Golaszewski <bgolaszewski@xxxxxxxxxxxx>; linux@xxxxxxxxx
> > Subject: Re: [PATCH] misc: eeprom: at24: fix NVMEM name with custom
> > AT24 device name
> >
> > On Thu, Dec 03, 2020 at 09:47:03PM +0000, Diego Santa Cruz wrote:
> > > When the "label" property is set on the AT24 EEPROM the NVMEM devid is
> > > set to NVMEM_DEVID_NONE, but it is not effective since there is a
> > > leftover line setting it back to NVMEM_DEVID_AUTO a few lines after.
> > >
> > > Fixes: 61f764c307f6 ("eeprom: at24: Support custom device names for AT24
> > EEPROMs")
> > > Signed-off-by: Diego Santa Cruz <Diego.SantaCruz@xxxxxxxxxxxx>
> > > ---
> > > drivers/misc/eeprom/at24.c | 1 -
> > > 1 file changed, 1 deletion(-)
> > >
> > > diff --git a/drivers/misc/eeprom/at24.c b/drivers/misc/eeprom/at24.c
> > > index 35fabaf539b7..fbf69148b5ad 100644
> > > --- a/drivers/misc/eeprom/at24.c
> > > +++ b/drivers/misc/eeprom/at24.c
> > > @@ -704,7 +704,6 @@ static int at24_probe(struct i2c_client *client)
> > >
> > > nvmem_config.type = NVMEM_TYPE_EEPROM;
> > > nvmem_config.dev = dev;
> > > - nvmem_config.id = NVMEM_DEVID_AUTO;
> > > nvmem_config.read_only = !writable;
> > > nvmem_config.root_only = !(flags & AT24_FLAG_IRUGO);
> > > nvmem_config.owner = THIS_MODULE;
> >
> > This commit (4e302c3b568eaf2aeebba804c07aba5d921a8c9e) makes a
> > regress
> > on powerpc64 systems. Tested on YADRO VESNIN (POWER8) and
> > SuperMicro/IBM
> > 9006-22P (POWER9) servers.
> >
>
> That commit is only making the previous commit 61f764c307f6 ("eeprom: at24: Support custom device names for AT24 EEPROMs") take its intended effect in full.
> And if I get 61f764c307f6 correctly it makes the name from the "label" property and sets the id to NVMEM_DEVID_NONE, and commit 4e302c3b568e just removes the leftover override to NVMEM_DEVID_AUTO so that 61f764c307f6 does what was intended.
>
> The -17 is actually EEXIST, so probably the name is already being used by another device.
> Although reverting 4e302c3b568e fixes the issue, it is probably not the correct solution. At a minimum the line that sets the id to NVMEM_DEVID_NONE would have to be removed and the comments from 61f764c307f6 adjusted to reflect reality.
>
> I have no powerpc64 systems nor any knowledge to find the real issue, although I expect a better solution can be found.
>
> > The error is logged multiple times in dmesg as following fragment:
> >
> > [ 5.164711] CPU: 26 PID: 1532 Comm: systemd-udevd Tainted: G W
> > 5.12.0-at24-catch+ #10
> > [ 5.164714] Call Trace:
> > [ 5.164715] [c000000052e2f010] [c000000000913610]
> > dump_stack+0xc4/0x114 (unreliable)
> > [ 5.164721] [c000000052e2f060] [c00000000061bb88]
> > sysfs_warn_dup+0x88/0xc0
> > [ 5.164726] [c000000052e2f0e0] [c00000000061c2ac]
> > sysfs_do_create_link_sd+0x17c/0x190
> > [ 5.164730] [c000000052e2f130] [c000000000ac1854]
> > bus_add_device+0x94/0x1d0
> > [ 5.164735] [c000000052e2f1b0] [c000000000abcff8]
> > device_add+0x428/0xb90
> > [ 5.164738] [c000000052e2f2a0] [c000000000dea270]
> > nvmem_register+0x220/0xe00
> > [ 5.164743] [c000000052e2f390] [c000000000deaeac]
> > devm_nvmem_register+0x5c/0xc0
> > [ 5.164747] [c000000052e2f3d0] [c008000016ca0c1c]
> > at24_probe+0x664/0x8d0 [at24]
> > [ 5.164753] [c000000052e2f650] [c000000000cfd404]
> > i2c_device_probe+0x194/0x650
> > [ 5.164757] [c000000052e2f6f0] [c000000000ac357c]
> > really_probe+0x1cc/0x790
> > [ 5.164761] [c000000052e2f790] [c000000000ac3c9c]
> > driver_probe_device+0x15c/0x200
> > [ 5.164766] [c000000052e2f810] [c000000000ac470c]
> > device_driver_attach+0x11c/0x130
> > [ 5.164771] [c000000052e2f850] [c000000000ac4810]
> > __driver_attach+0xf0/0x200
> > [ 5.164775] [c000000052e2f8d0] [c000000000abf998]
> > bus_for_each_dev+0xa8/0x130
> > [ 5.164779] [c000000052e2f930] [c000000000ac2944]
> > driver_attach+0x34/0x50
> > [ 5.164783] [c000000052e2f950] [c000000000ac1e30]
> > bus_add_driver+0x1b0/0x2f0
> > [ 5.164788] [c000000052e2f9e0] [c000000000ac58f4]
> > driver_register+0xb4/0x1c0
> > [ 5.164792] [c000000052e2fa50] [c000000000cfcbc8]
> > i2c_register_driver+0x78/0x120
> > [ 5.164796] [c000000052e2fad0] [c008000016ca11f0] at24_init+0x6c/0x88
> > [at24]
> > [ 5.164801] [c000000052e2fb30] [c0000000000122c0]
> > do_one_initcall+0x60/0x2c0
> > [ 5.164805] [c000000052e2fc00] [c000000000253bdc]
> > do_init_module+0x7c/0x350
> > [ 5.164809] [c000000052e2fc90] [c000000000257d24]
> > __do_sys_finit_module+0xd4/0x160
> > [ 5.164813] [c000000052e2fdb0] [c00000000002bfb4]
> > system_call_exception+0xf4/0x200
> > [ 5.164817] [c000000052e2fe10] [c00000000000cf70]
> > system_call_vectored_common+0xf0/0x268
> > [ 5.164821] --- interrupt: 3000 at 0x7f35ee3fb4c4
> > [ 5.164823] NIP: 00007f35ee3fb4c4 LR: 0000000000000000 CTR:
> > 0000000000000000
> > [ 5.164825] REGS: c000000052e2fe80 TRAP: 3000 Tainted: G W
> > (5.12.0-at24-catch+)
> > [ 5.164828] MSR: 900000000280f033
> > <SF,HV,VEC,VSX,EE,PR,FP,ME,IR,DR,RI,LE> CR: 44222884 XER: 00000000
> > [ 5.164842] IRQMASK: 0
> > GPR00: 0000000000000161 00007ffffbf36390 00007f35ee4e7000
> > 0000000000000006
> > GPR04: 00007f35ee6ab510 0000000000000000 0000000000000006
> > 0000000000000000
> > GPR08: 0000000000000000 0000000000000000 0000000000000000
> > 0000000000000000
> > GPR12: 0000000000000000 00007f35ee78e680 0000000020000000
> > 0000000000000000
> > GPR16: 0000000000000000 00000c0344781980 00000c0344781a18
> > 00000c0344781948
> > GPR20: 0000000000000000 00000c0350142a20 00007ffffbf365f0
> > 00000c03501b4d90
> > GPR24: 0000000000000000 00000c0350142a20 0000000000020000
> > 00000c0350139830
> > GPR28: 00007f35ee6ab510 0000000000020000 0000000000000000
> > 00000c0350142a20
> > [ 5.164883] NIP [00007f35ee3fb4c4] 0x7f35ee3fb4c4
> > [ 5.164885] LR [0000000000000000] 0x0
> > [ 5.164887] --- interrupt: 3000
> > [ 5.165102] at24: probe of 8-0057 failed with error -17
> >
> >
> > Reverting the patch solves the issue.
> >
> >
> > Reported-by: Alexander Fomichev <fomichev.ru@xxxxxxxxx>
> > CC: Bartosz Golaszewski <bgolaszewski@xxxxxxxxxxxx>
> > CC: linux@xxxxxxxxx
> >
>
That's right. More digging showed that commit 61f764c307f6 causes the
described problem, not 4e302c3b568e.
I'll report it to appropriate thread.
--
Regards,
Alexander
