Re: [PATCH v7 5/7] nvmem: core: Rework layouts to become platform devices

[Dan Carpenter]

On Fri, Aug 04, 2023 at 05:39:03PM +0200, Miquel Raynal wrote:
> Hi Dan,
> 
> dan.carpenter@xxxxxxxxxx wrote on Thu, 3 Aug 2023 13:13:04 +0300:
> 
> > Hi Miquel,
> > 
> > kernel test robot noticed the following build warnings:
> > 
> > https://git-scm.com/docs/git-format-patch#_base_tree_information]
> > 
> > url:    https://github.com/intel-lab-lkp/linux/commits/Miquel-Raynal/nvmem-core-Create-all-cells-before-adding-the-nvmem-device/20230802-022331
> > base:   char-misc/char-misc-testing
> > patch link:    https://lore.kernel.org/r/20230801182132.1058707-6-miquel.raynal%40bootlin.com
> > patch subject: [PATCH v7 5/7] nvmem: core: Rework layouts to become platform devices
> > config: x86_64-randconfig-m001-20230730 (https://download.01.org/0day-ci/archive/20230803/202308030002.DnSFOrMB-lkp@xxxxxxxxx/config)
> > compiler: gcc-12 (Debian 12.2.0-14) 12.2.0
> > reproduce: (https://download.01.org/0day-ci/archive/20230803/202308030002.DnSFOrMB-lkp@xxxxxxxxx/reproduce)
> > 
> > If you fix the issue in a separate patch/commit
> > (i.e. not just a new version of the same patch/commit),
> 
> (Nice addition, a lot of newcomers would always add these tags
> otherwise.)

The Intel kbuild devs add this stuff, I just look it over and hit
forward.

> 
> > kindly add following tags
> > | Reported-by: kernel test robot <lkp@xxxxxxxxx>
> > | Reported-by: Dan Carpenter <dan.carpenter@xxxxxxxxxx>
> > | Closes: https://lore.kernel.org/r/202308030002.DnSFOrMB-lkp@xxxxxxxxx/
> > 
> > New smatch warnings:
> > drivers/nvmem/core.c:1003 nvmem_register() warn: 'layout_np' is an error pointer or valid
> > drivers/nvmem/core.c:2130 nvmem_try_loading_layout_driver() warn: 'layout_np' is an error pointer or valid
> > 
> > Old smatch warnings:
> > drivers/nvmem/core.c:761 nvmem_add_cells_from_fixed_layout() warn: 'layout_np' is an error pointer or valid
> > drivers/nvmem/core.c:802 nvmem_layout_get() warn: 'layout_np' is an error pointer or valid
> > 
> > vim +/layout_np +1003 drivers/nvmem/core.c
> > 
> > 266570f496b90d Michael Walle         2023-04-04  1000  
> > 00d059fd6702f0 Miquel Raynal         2023-08-01  1001  	/* Populate layouts as devices */
> > 00d059fd6702f0 Miquel Raynal         2023-08-01  1002  	layout_np = of_nvmem_layout_get_container(nvmem);
> > 00d059fd6702f0 Miquel Raynal         2023-08-01 @1003  	if (layout_np) {
> > 
> > So, ugh, of_nvmem_layout_get_container() return NULL on error or error
> > pointer if either CONFIG_NVMEM or CONFIG_OF is turned off.  I feel like
> > that's a mistake.  Normally when a function returns both error pointers
> > and NULL then the NULL means the feature is disabled and the error
> > pointers mean there was an error.  Here it is the opposite.
> > 
> > I have written a blog about this:
> > https://staticthinking.wordpress.com/2022/08/01/mixing-error-pointers-and-null/
> 
> Nice (besides the huge spider which stared at me unexpectedly :-) )

Those are a species of jumping spiders.  They are shiny and golden in
real life, but it never shows up properly in photos.  :)

> 
> > At first I thought that this was to do with CONFIG_COMPILE_TEST but
> > actually that is disabled.  The issue here is that CONFIG_OF is turned
> > off.  So this is a genuine bug, we're compiling a module which will
> > always crash.
> > 
> > So I guess the fix is easy that this should return NULL if either
> > CONFIG_NVMEM or CONFIG_OF is turned off.  That was a long explanation
> > which is no longer required now that it's not a COMPILE_TEST issue.  :P
> 
> I wanted to disable CONFIG_OF to make the test, I totally forget, I'll
> handle this case and return NULL when this happens.
> 
> However I don't understand why you mention CONFIG_NVMEM, because if it
> is not defined, this file will not compile at all?

Yeah.  You're right.  I wrote this email thinking it was a
CONFIG_COMPILE_TEST issue and didn't edit it properly in the end.

regards,
dan carpenter