linux-kernel - Re: [PATCH 2/2] xen/blkback: turn the cache purge percent into a parameter

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Thu, 15 Oct 2020 16:44:01 +0200
From:   Roger Pau Monné <roger.pau@...rix.com>
To:     Jürgen Groß <jgross@...e.com>
CC:     <linux-kernel@...r.kernel.org>,
        Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>,
        Jens Axboe <axboe@...nel.dk>,
        Boris Ostrovsky <boris.ostrovsky@...cle.com>,
        SeongJae Park <sjpark@...zon.de>,
        <xen-devel@...ts.xenproject.org>, <linux-block@...r.kernel.org>,
        "J . Roeleveld" <joost@...arean.org>
Subject: Re: [PATCH 2/2] xen/blkback: turn the cache purge percent into a
 parameter

On Thu, Oct 15, 2020 at 04:37:52PM +0200, Jürgen Groß wrote:
> On 15.10.20 16:24, Roger Pau Monne wrote:
> > Assume that reads and writes to the variable will be atomic. The worse
> > that could happen is that one of the purges removes a partially
> > written percentage of grants, but the cache itself will recover.
> > 
> > Signed-off-by: Roger Pau Monné <roger.pau@...rix.com>
> > ---
> > Cc: Konrad Rzeszutek Wilk <konrad.wilk@...cle.com>
> > Cc: Jens Axboe <axboe@...nel.dk>
> > Cc: Boris Ostrovsky <boris.ostrovsky@...cle.com>
> > Cc: SeongJae Park <sjpark@...zon.de>
> > Cc: xen-devel@...ts.xenproject.org
> > Cc: linux-block@...r.kernel.org
> > Cc: J. Roeleveld <joost@...arean.org>
> > Cc: Jürgen Groß <jgross@...e.com>
> > ---
> >   Documentation/ABI/testing/sysfs-driver-xen-blkback | 9 +++++++++
> >   drivers/block/xen-blkback/blkback.c                | 7 +++++--
> >   2 files changed, 14 insertions(+), 2 deletions(-)
> > 
> > diff --git a/Documentation/ABI/testing/sysfs-driver-xen-blkback b/Documentation/ABI/testing/sysfs-driver-xen-blkback
> > index 776f25d335ca..7de791ad61f9 100644
> > --- a/Documentation/ABI/testing/sysfs-driver-xen-blkback
> > +++ b/Documentation/ABI/testing/sysfs-driver-xen-blkback
> > @@ -45,3 +45,12 @@ Description:
> >                   to be executed periodically. This parameter controls the time
> >                   interval between consecutive executions of the purge mechanism
> >                   is set in ms.
> > +
> > +What:           /sys/module/xen_blkback/parameters/lru_percent_clean
> > +Date:           October 2020
> > +KernelVersion:  5.10
> > +Contact:        Roger Pau Monné <roger.pau@...rix.com>
> > +Description:
> > +                When the persistent grants list is full we will remove unused
> > +                grants from the list. The percent number of grants to be
> > +                removed at each LRU execution.
> > diff --git a/drivers/block/xen-blkback/blkback.c b/drivers/block/xen-blkback/blkback.c
> > index 6ad9b76fdb2b..772852d45a5a 100644
> > --- a/drivers/block/xen-blkback/blkback.c
> > +++ b/drivers/block/xen-blkback/blkback.c
> > @@ -127,7 +127,10 @@ MODULE_PARM_DESC(lru_internval,
> >    * from the list. The percent number of grants to be removed at each LRU
> >    * execution.
> >    */
> > -#define LRU_PERCENT_CLEAN 5
> > +static unsigned int lru_percent_clean = 5;
> > +module_param_named(lru_percent_clean, lru_percent_clean, uint, 0644);
> > +MODULE_PARM_DESC(lru_percent_clean,
> > +		 "Percentage of persistent grants to remove from the cache when full");
> >   /* Run-time switchable: /sys/module/blkback/parameters/ */
> >   static unsigned int log_stats;
> > @@ -404,7 +407,7 @@ static void purge_persistent_gnt(struct xen_blkif_ring *ring)
> >   	    !ring->blkif->vbd.overflow_max_grants)) {
> >   		num_clean = 0;
> >   	} else {
> > -		num_clean = (max_pgrants / 100) * LRU_PERCENT_CLEAN;
> > +		num_clean = (max_pgrants / 100) * lru_percent_clean;
> 
> Hmm, wouldn't it be better to use (max_grants * lru_percent_clean) / 100
> here in order to support max_grants values less than 100?

Yes, we should have done that when moving max_pgrants to a parameter,
it used to be fixed first.

Will do in next version since I'm already changing the line.

Thanks, Roger.