Am seeing a build error in next-0514. -0420 built OK.
building a 'make allmodconfig' on a RPi4 in 32-bit mode.
MODPOST 7575 modules
ERROR: modpost: "__aeabi_uldivmod" [drivers/md/dm-zoned.ko] undefined!
objdump and 'make drivers/md/dm-zoned-target.s' tells
me that the problem is in function dmz_fixup_devices(), near here:
@ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
ldr r0, [r6, #56] @ reg_dev_166->capacity, reg_dev_166->capacity
adds r1, r3, r1 @ tmp316, _227, reg_dev_166->capacity
adc r0, r2, r0 @ tmp315, _227, reg_dev_166->capacity
subs r1, r1, #1 @, tmp316,
@ drivers/md/dm-zoned-target.c:805: reg_dev->zone_nr_sectors = zoned_dev->zone_nr_sectors;
strd r2, [r6, #80] @, reg_dev,
@ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
sbc r0, r0, #0 @, tmp315,
bl __aeabi_uldivmod @
@ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
str r1, [r6, #64] @ tmp306, reg_dev_166->nr_zones
git blame points at this commit:
commit 70978208ec91d798066f4c291bc98ff914bea222
Author: Hannes Reinecke <[email protected]>
Date: Mon May 11 10:24:30 2020 +0200
dm zoned: metadata version 2
Reverting that commit lets the build complete.
On Sat, 2020-05-16 at 07:19 -0400, Valdis Klētnieks wrote:
> Am seeing a build error in next-0514. -0420 built OK.
> building a 'make allmodconfig' on a RPi4 in 32-bit mode.
>
> MODPOST 7575 modules
> ERROR: modpost: "__aeabi_uldivmod" [drivers/md/dm-zoned.ko]
> undefined!
>
> objdump and 'make drivers/md/dm-zoned-target.s' tells
> me that the problem is in function dmz_fixup_devices(), near here:
>
> @ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones =
> DIV_ROUND_UP(reg_dev->capacity,
> ldr r0, [r6, #56] @ reg_dev_166->capacity, reg_dev_166-
> >capacity
> adds r1, r3, r1 @ tmp316, _227, reg_dev_166->capacity
> adc r0, r2, r0 @ tmp315, _227, reg_dev_166->capacity
> subs r1, r1, #1 @, tmp316,
> @ drivers/md/dm-zoned-target.c:805: reg_dev-
> >zone_nr_sectors = zoned_dev->zone_nr_sectors;
> strd r2, [r6, #80] @, reg_dev,
> @ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones =
> DIV_ROUND_UP(reg_dev->capacity,
> sbc r0, r0, #0 @, tmp315,
> bl __aeabi_uldivmod @
> @ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones =
> DIV_ROUND_UP(reg_dev->capacity,
> str r1, [r6, #64] @ tmp306, reg_dev_166->nr_zones
>
> git blame points at this commit:
>
> commit 70978208ec91d798066f4c291bc98ff914bea222
> Author: Hannes Reinecke <[email protected]>
> Date: Mon May 11 10:24:30 2020 +0200
>
> dm zoned: metadata version 2
>
> Reverting that commit lets the build complete.
That's because the DIV_ROUND_UP in the patch should actually be
DIV_ROUND_UP_SECTOR_T I think.
James
On 5/16/20 1:19 PM, Valdis Klētnieks wrote:
> Am seeing a build error in next-0514. -0420 built OK.
> building a 'make allmodconfig' on a RPi4 in 32-bit mode.
>
> MODPOST 7575 modules
> ERROR: modpost: "__aeabi_uldivmod" [drivers/md/dm-zoned.ko] undefined!
>
> objdump and 'make drivers/md/dm-zoned-target.s' tells
> me that the problem is in function dmz_fixup_devices(), near here:
>
> @ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
> ldr r0, [r6, #56] @ reg_dev_166->capacity, reg_dev_166->capacity
> adds r1, r3, r1 @ tmp316, _227, reg_dev_166->capacity
> adc r0, r2, r0 @ tmp315, _227, reg_dev_166->capacity
> subs r1, r1, #1 @, tmp316,
> @ drivers/md/dm-zoned-target.c:805: reg_dev->zone_nr_sectors = zoned_dev->zone_nr_sectors;
> strd r2, [r6, #80] @, reg_dev,
> @ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
> sbc r0, r0, #0 @, tmp315,
> bl __aeabi_uldivmod @
> @ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
> str r1, [r6, #64] @ tmp306, reg_dev_166->nr_zones
>
> git blame points at this commit:
>
> commit 70978208ec91d798066f4c291bc98ff914bea222
> Author: Hannes Reinecke <[email protected]>
> Date: Mon May 11 10:24:30 2020 +0200
>
> dm zoned: metadata version 2
>
> Reverting that commit lets the build complete.
>
>
I thought I've send a patch to fix that up; DIV_ROUND_UP() needs to be
changed to DIV_ROUND_UP_ULL().
I'll be checking and will be sending a patch if necessary.
Cheers,
Hannes
--
Dr. Hannes Reinecke Teamlead Storage & Networking
[email protected] +49 911 74053 688
SUSE Software Solutions GmbH, Maxfeldstr. 5, 90409 Nürnberg
HRB 36809 (AG Nürnberg), Geschäftsführer: Felix Imendörffer
On Mon, May 18 2020 at 2:25am -0400,
Hannes Reinecke <[email protected]> wrote:
> On 5/16/20 1:19 PM, Valdis Klētnieks wrote:
> >Am seeing a build error in next-0514. -0420 built OK.
> >building a 'make allmodconfig' on a RPi4 in 32-bit mode.
> >
> > MODPOST 7575 modules
> >ERROR: modpost: "__aeabi_uldivmod" [drivers/md/dm-zoned.ko] undefined!
> >
> >objdump and 'make drivers/md/dm-zoned-target.s' tells
> >me that the problem is in function dmz_fixup_devices(), near here:
> >
> >@ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
> > ldr r0, [r6, #56] @ reg_dev_166->capacity, reg_dev_166->capacity
> > adds r1, r3, r1 @ tmp316, _227, reg_dev_166->capacity
> > adc r0, r2, r0 @ tmp315, _227, reg_dev_166->capacity
> > subs r1, r1, #1 @, tmp316,
> >@ drivers/md/dm-zoned-target.c:805: reg_dev->zone_nr_sectors = zoned_dev->zone_nr_sectors;
> > strd r2, [r6, #80] @, reg_dev,
> >@ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
> > sbc r0, r0, #0 @, tmp315,
> > bl __aeabi_uldivmod @
> >@ drivers/md/dm-zoned-target.c:806: reg_dev->nr_zones = DIV_ROUND_UP(reg_dev->capacity,
> > str r1, [r6, #64] @ tmp306, reg_dev_166->nr_zones
> >
> >git blame points at this commit:
> >
> >commit 70978208ec91d798066f4c291bc98ff914bea222
> >Author: Hannes Reinecke <[email protected]>
> >Date: Mon May 11 10:24:30 2020 +0200
> >
> > dm zoned: metadata version 2
> >
> >Reverting that commit lets the build complete.
> >
> >
> I thought I've send a patch to fix that up; DIV_ROUND_UP() needs to
> be changed to DIV_ROUND_UP_ULL().
> I'll be checking and will be sending a patch if necessary.
Unless I'm missing something it was fixed up with this commit last
wednesday (13th):
https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=dm-5.8&id=81a3a1453ec4e5da081e1395732801a600feb352
On Mon, 18 May 2020 12:44:49 -0400, Mike Snitzer said:
> Unless I'm missing something it was fixed up with this commit last
> wednesday (13th):
>
> https://git.kernel.org/pub/scm/linux/kernel/git/device-mapper/linux-dm.git/commit/?h=dm-5.8&id=81a3a1453ec4e5da081e1395732801a600feb352
That says:
author Nathan Chancellor <[email protected]> 2020-05-13 01:45:22 -0700
committer Mike Snitzer <[email protected]> 2020-05-15 10:29:39 -0400
So it didn't make it into next-0514, which is why I got bit by it. It's in today's linux-next
and life is good. :)