Re: [linux-next:master] BUILD REGRESSION 8cb8311e95e3bb58bd84d6350365f14a718faa6d

Dan Carpenter <[email protected]> Thu, 26 May 2022 11:48:32 +0300

Newsgroups	gmane.comp.freedesktop.amd-gfx,gmane.linux.alsa.devel,gmane.comp.emulators.kvm.devel,gmane.linux.drivers.rdma,gmane.linux.network,gmane.linux.kernel.bpf,gmane.comp.video.dri.devel,gmane.linux.kernel.virtualization,gmane.linux.kernel.mm,gmane.linux.kernel.pci,gmane.linux.ports.arm.omap,gmane.linux.ports.riscv,gmane.linux.ports.arm.kernel,gmane.linux.parport
Message-ID	<20220526084832.GC2146@kadam>

On Thu, May 26, 2022 at 02:16:34AM +0100, Matthew Wilcox wrote:
> Bizarre this started showing up now.  The recent patch was:
> 
> -       info->alloced += compound_nr(page);
> -       inode->i_blocks += BLOCKS_PER_PAGE << compound_order(page);
> +       info->alloced += folio_nr_pages(folio);
> +       inode->i_blocks += BLOCKS_PER_PAGE << folio_order(folio);
> 
> so it could tell that compound_order() was small, but folio_order()
> might be large?

The old code also generates a warning on my test system.  Smatch thinks
both compound_order() and folio_order() are 0-255.  I guess because of
the "unsigned char compound_order;" in the struct page.

regards,
dan carpenter