[PATCHv6 4/8] zswap: add to mm/

Dan Magenheimer dan.magenheimer at oracle.com
Thu Feb 28 18:13:28 UTC 2013


> From: Seth Jennings [mailto:sjenning at linux.vnet.ibm.com]
> Subject: [PATCHv6 4/8] zswap: add to mm/
> 
> +/*
> + * Maximum compression ratio, as as percentage, for an acceptable
> + * compressed page. Any pages that do not compress by at least
> + * this ratio will be rejected.
> +*/
> +static unsigned int zswap_max_compression_ratio = 80;
> +module_param_named(max_compression_ratio,
> +			zswap_max_compression_ratio, uint, 0644);

Unless this is a complete coincidence, I believe that
the default value "80" is actually:

(100 * (1L >> ZS_MAX_ZSPAGE_ORDER)) /
        ((1L >> ZS_MAX_ZSPAGE_ORDER)) + 1)

(though the constant ZS_MAX_ZSPAGE_ORDER is not currently
defined outside of zsmalloc.c) because pages that compress
less efficiently than this always require a full pageframe
in zsmalloc.  True?

If this change were made, is there any real reason for this
to be a user-selectable parameter, i.e. given the compression
internals knowledge necessary to understand what value should
be selected, would any mortal sysadmin ever want to change it
or know what would be a reasonable value to change it to?



More information about the devel mailing list