[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <87fry2t0zw.fsf@linux.intel.com>
Date: Fri, 09 Feb 2024 01:56:03 -0800
From: Andi Kleen <ak@...ux.intel.com>
To: Chunsheng Luo <luochunsheng@...c.edu>
Cc: gregkh@...uxfoundation.org, rafael@...nel.org,
akpm@...ux-foundation.org, linux-kernel@...r.kernel.org,
linux-mm@...ck.org
Subject: Re: [PATCH] meminfo: provide estimated per-node's available memory
Chunsheng Luo <luochunsheng@...c.edu> writes:
> + available = sum_zone_node_page_state(nid, NR_FREE_PAGES) - pgdat->totalreserve_pages;
> +
> + /*
> + * Not all the page cache can be freed, otherwise the system will
> + * start swapping or thrashing. Assume at least half of the page
> + * cache, or the low watermark worth of cache, needs to stay.
> + */
> + pagecache = node_page_state(pgdat, NR_ACTIVE_FILE) +
> + node_page_state(pgdat, NR_INACTIVE_FILE);
> + pagecache -= min(pagecache / 2, wmark_low);
The magic number 2 should be a define (or maybe even a tunable). Similar
below. It seems quite arbitrary, but I don't have a better solution
either. Maybe could handle dirty differently, but nothing stands out here
> + node_page_state(pgdat, NR_KERNEL_MISC_RECLAIMABLE);
> + reclaimable -= min(reclaimable / 2, wmark_low);
> + available += reclaimable;
> +
> + if (available < 0)
> + available = 0;
That would be a bug? Perhaps add a WARN_ON
With those changes:
Reviewed-by: Andi Kleen <ak@...ux.intel.com>
-Andi
Powered by blists - more mailing lists