[PATCH] nvme: Fixes u64 division which breaks i386 builds

[Christoph Hellwig]

On Tue, Jul 21, 2015 at 03:08:13PM -0600, Jon Derrick wrote:
> Uses div_u64 for u64 division and round_down, a bitwise operation,
> instead of rounddown, which uses a modulus.
> 
> Signed-off-by: Jon Derrick <jonathan.derrick at intel.com>
> ---
> Hi folks. I accidentally broke i386 builds with my u64 math. This is the
> attempt to use safe operators instead of direct operator on u64 variables
> 
> I will, of course, be mindful of i386 builds from now on. :)

Yes, it's pain.  Might be time to troll Linus to finally link against
libgcc for 64-bit division helpers to avoid these issues..

Reviewed-by: Christoph Hellwig <hch at lst.de>

> 
>  drivers/block/nvme-core.c | 5 +++--
>  1 file changed, 3 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/block/nvme-core.c b/drivers/block/nvme-core.c
> index 82b4ffb..666e994 100644
> --- a/drivers/block/nvme-core.c
> +++ b/drivers/block/nvme-core.c
> @@ -1454,8 +1454,9 @@ static int nvme_cmb_qdepth(struct nvme_dev *dev, int nr_io_queues,
>  	unsigned q_size_aligned = roundup(q_depth * entry_size, dev->page_size);
>  
>  	if (q_size_aligned * nr_io_queues > dev->cmb_size) {
> -		q_depth = rounddown(dev->cmb_size / nr_io_queues,
> -					dev->page_size) / entry_size;
> +		u64 mem_per_q = div_u64(dev->cmb_size, nr_io_queues);
> +		mem_per_q = round_down(mem_per_q, dev->page_size);
> +		q_depth = div_u64(mem_per_q, entry_size);
>  
>  		/*
>  		 * Ensure the reduced q_depth is above some threshold where it
> -- 
> 2.1.4
> 
> 
> _______________________________________________
> Linux-nvme mailing list
> Linux-nvme at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-nvme
---end quoted text---