[PATCH v2 00/60] Fix mtd parent device not showing up in sysfs

[Brian Norris]

Hi Frans,

On Wed, Jun 10, 2015 at 10:38:14PM +0200, Frans Klaver wrote:
> Here's the second iteration of a series stemming from a previous discussion
> about mtd devices not showing up in sysfs properly[1]. While the original
> series 27 patches, this one should cover all drivers that suffer from this
> issue, and shamelessly hits 60. Again, if it's preferred that the patches are
> grouped differently, like by mtd subdir, I'd be happy to resend. Until then,
> we'll make do with this.
> 
> This series teaches add_mtd_device() to set name and owner based on the set
> parent device, if dev.parent is filled in, and name resp. owner isn't already
> filled in. All drivers that I could find that don't set the parent are touched
> to set the parent device, and remove the name and owner assignment in cases
> where I'm sure the resulting name will not change. Only the latter is done in
> cases where I would find the driver explicitly filling in the owner.
> 
> Thanks for having a go at your favorite MTD,
> Frans
> 
> v1..v2
>   - Overall: make it more clear that it's about a parent device symlink gone
>     missing (thanks, Brian)
>   - Fix compile error in 30: "fsmc_nand: show parent device in sysfs"
>   - Fix possible null-pointer dereference in 02: "mtd: core: set some
>     defaults..."
> 
> [1] https://lkml.org/lkml/2014/5/26/586

Thanks for the patches. I've tested a few drivers and compile tested as
much as I can, and everything looks good. Patches 5 and 59 (m25p80.c and
fsl-quadspi.c) didn't apply cleanly yet, as those drivers have had some
concurrent changes, so I didn't take them yet. If it's trivial to fix
up, I'll apply themself. Stay tuned.

Regards,
Brian