| 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
| ||
|
Message-ID: <20150125143834.GA951@swordfish>
Date: Sun, 25 Jan 2015 23:38:34 +0900
From: Sergey Senozhatsky <sergey.senozhatsky@...il.com>
To: Ganesh Mahendran <opensource.ganesh@...il.com>
Cc: Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
Minchan Kim <minchan@...nel.org>,
Andrew Morton <akpm@...ux-foundation.org>,
linux-kernel <linux-kernel@...r.kernel.org>,
Linux-MM <linux-mm@...ck.org>, Nitin Gupta <ngupta@...are.org>,
Jerome Marchand <jmarchan@...hat.com>
Subject: Re: [PATCH 2/2] zram: protect zram->stat race with init_lock
Hello,
On (01/24/15 21:17), Ganesh Mahendran wrote:
> Hello Sergey
>
> 2015-01-23 22:38 GMT+08:00 Sergey Senozhatsky <sergey.senozhatsky@...il.com>:
> > On (01/23/15 14:58), Minchan Kim wrote:
> >> The zram->stat handling should be procted by init_lock.
> >> Otherwise, user could see stale value from the stat.
> >>
> >> Signed-off-by: Minchan Kim <minchan@...nel.org>
> >> ---
> >>
> >> I don't think it's stable material. The race is rare in real practice
> >> and this stale stat value read is not a critical.
> >>
> >> drivers/block/zram/zram_drv.c | 37 ++++++++++++++++++++++++++++---------
> >> 1 file changed, 28 insertions(+), 9 deletions(-)
> >>
> >> diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
> >> index 0299d82275e7..53f176f590b0 100644
> >> --- a/drivers/block/zram/zram_drv.c
> >> +++ b/drivers/block/zram/zram_drv.c
> >> @@ -48,8 +48,13 @@ static ssize_t name##_show(struct device *d, \
> >> struct device_attribute *attr, char *b) \
> >> { \
> >
> > a side note: I wasn't Cc'd in that patchset and found out it only when it's
> > been merged. I'm not sure I understand, why it has been renamed from specific
> > zram_X_show to X_show. what gives?
>
> I changed from zram_attr_##name##_show to name##_show in commit:
> fcf1bce zram: use DEVICE_ATTR_[RW|RO|WO] to define zram sys device attribute
>
> I just want to keep the name consistent with others, like
> disksize_show(), initstate_show().
aha, I see.
-ss
> Thanks.
>
> >
> >
> > can't help, catches my eye every time, that rename has broken the original
> > formatting:
> >
> >
> > diff --git a/drivers/block/zram/zram_drv.c b/drivers/block/zram/zram_drv.c
> > index 9250b3f..c567af5 100644
> > --- a/drivers/block/zram/zram_drv.c
> > +++ b/drivers/block/zram/zram_drv.c
> > @@ -44,7 +44,7 @@ static const char *default_compressor = "lzo";
> > static unsigned int num_devices = 1;
> >
> > #define ZRAM_ATTR_RO(name) \
> > -static ssize_t name##_show(struct device *d, \
> > +static ssize_t name##_show(struct device *d, \
> > struct device_attribute *attr, char *b) \
> > { \
> > struct zram *zram = dev_to_zram(d); \
> >
> >
> >
> > I don't have any objections. but do we really want to wrap atomic ops in
> > semaphore? it is really such serious race?
> >
> >
> > -ss
> >
> >> struct zram *zram = dev_to_zram(d); \
> >> - return scnprintf(b, PAGE_SIZE, "%llu\n", \
> >> - (u64)atomic64_read(&zram->stats.name)); \
> >> + u64 val = 0; \
> >> + \
> >> + down_read(&zram->init_lock); \
> >> + if (init_done(zram)) \
> >> + val = atomic64_read(&zram->stats.name); \
> >> + up_read(&zram->init_lock); \
> >> + return scnprintf(b, PAGE_SIZE, "%llu\n", val); \
> >> } \
> >> static DEVICE_ATTR_RO(name);
> >>
> >> @@ -67,8 +72,14 @@ static ssize_t disksize_show(struct device *dev,
> >> struct device_attribute *attr, char *buf)
> >> {
> >> struct zram *zram = dev_to_zram(dev);
> >> + u64 val = 0;
> >> +
> >> + down_read(&zram->init_lock);
> >> + if (init_done(zram))
> >> + val = zram->disksize;
> >> + up_read(&zram->init_lock);
> >>
> >> - return scnprintf(buf, PAGE_SIZE, "%llu\n", zram->disksize);
> >> + return scnprintf(buf, PAGE_SIZE, "%llu\n", val);
> >> }
> >>
> >> static ssize_t initstate_show(struct device *dev,
> >> @@ -88,9 +99,14 @@ static ssize_t orig_data_size_show(struct device *dev,
> >> struct device_attribute *attr, char *buf)
> >> {
> >> struct zram *zram = dev_to_zram(dev);
> >> + u64 val = 0;
> >> +
> >> + down_read(&zram->init_lock);
> >> + if (init_done(zram))
> >> + val = atomic64_read(&zram->stats.pages_stored) << PAGE_SHIFT;
> >> + up_read(&zram->init_lock);
> >>
> >> - return scnprintf(buf, PAGE_SIZE, "%llu\n",
> >> - (u64)(atomic64_read(&zram->stats.pages_stored)) << PAGE_SHIFT);
> >> + return scnprintf(buf, PAGE_SIZE, "%llu\n", val);
> >> }
> >>
> >> static ssize_t mem_used_total_show(struct device *dev,
> >> @@ -957,10 +973,6 @@ static int zram_rw_page(struct block_device *bdev, sector_t sector,
> >> struct bio_vec bv;
> >>
> >> zram = bdev->bd_disk->private_data;
> >> - if (!valid_io_request(zram, sector, PAGE_SIZE)) {
> >> - atomic64_inc(&zram->stats.invalid_io);
> >> - return -EINVAL;
> >> - }
> >>
> >> down_read(&zram->init_lock);
> >> if (unlikely(!init_done(zram))) {
> >> @@ -968,6 +980,13 @@ static int zram_rw_page(struct block_device *bdev, sector_t sector,
> >> goto out_unlock;
> >> }
> >>
> >> + if (!valid_io_request(zram, sector, PAGE_SIZE)) {
> >> + atomic64_inc(&zram->stats.invalid_io);
> >> + err = -EINVAL;
> >> + goto out_unlock;
> >> + }
> >> +
> >> +
> >> index = sector >> SECTORS_PER_PAGE_SHIFT;
> >> offset = sector & (SECTORS_PER_PAGE - 1) << SECTOR_SHIFT;
> >>
> >> --
> >> 1.9.1
> >>
> > --
> > To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> > the body of a message to majordomo@...r.kernel.org
> > More majordomo info at http://vger.kernel.org/majordomo-info.html
> > Please read the FAQ at http://www.tux.org/lkml/
>
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists