Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
On Wednesday 09 May 2012, Christoph Hellwig wrote: > On Wed, May 09, 2012 at 01:59:40PM +, Arnd Bergmann wrote: > > My feeling is that we should just treat every (REQ_SYNC | REQ_READ) > > request the same and let them interrupt long-running writes, > > independent of whether it's REQ_META or demand paging. > > It's funny that the CFQ scheduler used to boost metadata reads that > have REQ_META set - in fact it still does for those filesystems using > the now split out REQ_PRIO. That certainly sounds more sensible than the opposite. Of course, this is somewhat unrelated to the question of prioritizing reads over any writes that are already started. IMHO It would be pointless to only stop the write in order to do a REQ_PRIO read but not any other read. Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
On Mon, May 07, 2012 at 10:16:30PM +0530, S, Venkatraman wrote: [..] > This feature doesn't fiddle with the I/O scheduler's ability to balance > read vs write requests or handling requests from various process queues (CFQ). > Does this feature work with CFQ? As CFQ does not submit sync IO (for idling queues) while async IO is pending and vice a versa (cfq_may_dispatch()). Thanks Vivek -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
On Wed, May 09, 2012 at 01:59:40PM +, Arnd Bergmann wrote: > My feeling is that we should just treat every (REQ_SYNC | REQ_READ) > request the same and let them interrupt long-running writes, > independent of whether it's REQ_META or demand paging. It's funny that the CFQ scheduler used to boost metadata reads that have REQ_META set - in fact it still does for those filesystems using the now split out REQ_PRIO. -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
On Wednesday 09 May 2012, Dave Chinner wrote: > > In low end flash devices, some requests might take too long than normal > > due to background device maintenance (i.e flash erase / reclaim procedure) > > kicking in in the context of an ongoing write, stalling them by several > > orders of magnitude. > > And thereby stalling what might be writes critical to operation. > Indeed, how does this affect the system when it starts swapping > heavily? If you keep stalling writes, the system won't be able to > swap and free memory... The point here is that reads have a consistent latency, e.g. 500 microseconds for a small access, while writes have a latency that can easily become 1000x the read latency (e.g. 500 ms of blocking the device) depending on the state of the device. Most of the time, writes are fast as well, but sometimes (when garbage collection happens in the device), they are extremely slow and block everything else. This is the only time we ever want to interrupt a write: keeping the system running interactively while eventually getting to do the writeback. There is a small penalty for interrupting the garbage collection, but the device should be able to pick up its work at the point where we interrupt it, so we can still make forward progress. > > > This really seems like functionality that belongs in an IO > > > scheduler so that write starvation can be avoided, not in high-level > > > data read paths where we have no clue about anything else going on > > > in the IO subsystem > > > > Indeed, the feature is built mostly in the low level device driver and > > minor changes in the elevator. Changes above the block layer are only > > about setting > > attributes and transparent to their operation. > > The problem is that the attribute you are setting covers every > single data read that is done by all users. If that's what you want > to have happen, then why do you even need a new flag at this layer? > Just treat every non-REQ_META read request as a demand paged IO and > you've got exactly the same behaviour without needing to tag at the > higher layer My feeling is that we should just treat every (REQ_SYNC | REQ_READ) request the same and let them interrupt long-running writes, independent of whether it's REQ_META or demand paging. Arnd -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
On Mon, May 07, 2012 at 10:16:30PM +0530, S, Venkatraman wrote: > Mon, May 7, 2012 at 5:01 AM, Dave Chinner wrote: > > On Thu, May 03, 2012 at 07:53:00PM +0530, Venkatraman S wrote: > >> From: Ilan Smith > >> > >> Add attribute to identify demand paging requests. > >> Mark readpages with demand paging attribute. > >> > >> Signed-off-by: Ilan Smith > >> Signed-off-by: Alex Lemberg > >> Signed-off-by: Venkatraman S > >> --- > >> fs/mpage.c | 2 ++ > >> include/linux/bio.h | 7 +++ > >> include/linux/blk_types.h | 2 ++ > >> 3 files changed, 11 insertions(+) > >> > >> diff --git a/fs/mpage.c b/fs/mpage.c > >> index 0face1c..8b144f5 100644 > >> --- a/fs/mpage.c > >> +++ b/fs/mpage.c > >> @@ -386,6 +386,8 @@ mpage_readpages(struct address_space *mapping, struct > >> list_head *pages, > >> &last_block_in_bio, &map_bh, > >> &first_logical_block, > >> get_block); > >> + if (bio) > >> + bio->bi_rw |= REQ_RW_DMPG; > > > > Have you thought about the potential for DOSing a machine > > with this? That is, user data reads can now preempt writes of any > > kind, effectively stalling writeback and memory reclaim which will > > lead to OOM situations. Or, alternatively, journal flushing will get > > stalled and no new modifications can take place until the read > > stream stops. > > This feature doesn't fiddle with the I/O scheduler's ability to balance > read vs write requests or handling requests from various process queues (CFQ). And for schedulers like no-op that don't do any read/write balancing? Also, I thought the code was queuing such demand paged requests at the front of the queues, too, so bypassing most of the read/write balancing logic of the elevators... > Also, for block devices which don't implement the ability to preempt (and even > for older versions of MMC devices which don't implement this feature), > the behaviour > falls back to waiting for write requests to complete before issuing the read. Sure, but my point is that you are adding a flag that will be set for all user data read IO, and then making it priviledged in the lower layers. > In low end flash devices, some requests might take too long than normal > due to background device maintenance (i.e flash erase / reclaim procedure) > kicking in in the context of an ongoing write, stalling them by several > orders of magnitude. And thereby stalling what might be writes critical to operation. Indeed, how does this affect the system when it starts swapping heavily? If you keep stalling writes, the system won't be able to swap and free memory... > This implementation (See 14/16) does have several checks and > timers to see that it's not triggered very often. In my tests, > where I usually have a generous preemption time window, the abort > happens < 0.1% of the time. Yes, but seeing as the user has direct control of the pre-emption vector, it's not hard to imagine someone using it for a timing attack... > > This really seems like functionality that belongs in an IO > > scheduler so that write starvation can be avoided, not in high-level > > data read paths where we have no clue about anything else going on > > in the IO subsystem > > Indeed, the feature is built mostly in the low level device driver and > minor changes in the elevator. Changes above the block layer are only > about setting > attributes and transparent to their operation. The problem is that the attribute you are setting covers every single data read that is done by all users. If that's what you want to have happen, then why do you even need a new flag at this layer? Just treat every non-REQ_META read request as a demand paged IO and you've got exactly the same behaviour without needing to tag at the higher layer Cheers, Dave. -- Dave Chinner da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
On Tue, May 8, 2012 at 11:58 AM, mani wrote: > How about adding the AS_DMPG flag in the file -> address_space when getting > a filemap_fault() > so that we can treat the page fault pages as the high priority pages over > normal read requests. > How about changing below lines for the support of the pages those are > requested for the page fault ? > > > --- a/fs/mpage.c 2012-05-04 12:59:12.0 +0530 > +++ b/fs/mpage.c 2012-05-07 13:13:49.0 +0530 > @@ -408,6 +408,8 @@ mpage_readpages(struct address_space *ma > &last_block_in_bio, &map_bh, > &first_logical_block, > get_block); > + if(test_bit(AS_DMPG, &mapping->flags) && bio) > > + bio->bi_rw |= REQ_RW_DMPG > } > page_cache_release(page); > } > --- a/include/linux/pagemap.h 2012-05-04 12:57:35.0 +0530 > +++ b/include/linux/pagemap.h 2012-05-07 13:15:24.0 +0530 > @@ -27,6 +27,7 @@ enum mapping_flags { > #if defined (CONFIG_BD_CACHE_ENABLED) > AS_DIRECT = __GFP_BITS_SHIFT + 4, /* DIRECT_IO specified on file op > */ > #endif > + AS_DMPG = __GFP_BITS_SHIFT + 5, /* DEMAND PAGE specified on file op > */ > }; > > static inline void mapping_set_error(struct address_space *mapping, int > error) > > --- a/mm/filemap.c 2012-05-04 12:58:49.0 +0530 > +++ b/mm/filemap.c 2012-05-07 13:15:03.0 +0530 > @@ -1646,6 +1646,7 @@ int filemap_fault(struct vm_area_struct > if (offset >= size) > return VM_FAULT_SIGBUS; > > + set_bit(AS_DMPG, &file->f_mapping->flags); > /* > * Do we have something in the page cache already? > */ > > Will these changes have any adverse effect ? > Thanks for the example but I can't judge which of the two is the most elegant or acceptable to maintainers. I can test with your change and inform if it works. > Thanks & Regards > Manish > > On Mon, May 7, 2012 at 5:01 AM, Dave Chinner wrote: >> >> On Thu, May 03, 2012 at 07:53:00PM +0530, Venkatraman S wrote: >> > From: Ilan Smith >> > >> > Add attribute to identify demand paging requests. >> > Mark readpages with demand paging attribute. >> > >> > Signed-off-by: Ilan Smith >> > Signed-off-by: Alex Lemberg >> > Signed-off-by: Venkatraman S >> > --- >> > fs/mpage.c | 2 ++ >> > include/linux/bio.h | 7 +++ >> > include/linux/blk_types.h | 2 ++ >> > 3 files changed, 11 insertions(+) >> > >> > diff --git a/fs/mpage.c b/fs/mpage.c >> > index 0face1c..8b144f5 100644 >> > --- a/fs/mpage.c >> > +++ b/fs/mpage.c >> > @@ -386,6 +386,8 @@ mpage_readpages(struct address_space *mapping, >> > struct list_head *pages, >> > &last_block_in_bio, &map_bh, >> > &first_logical_block, >> > get_block); >> > + if (bio) >> > + bio->bi_rw |= REQ_RW_DMPG; >> >> Have you thought about the potential for DOSing a machine >> with this? That is, user data reads can now preempt writes of any >> kind, effectively stalling writeback and memory reclaim which will >> lead to OOM situations. Or, alternatively, journal flushing will get >> stalled and no new modifications can take place until the read >> stream stops. >> >> This really seems like functionality that belongs in an IO >> scheduler so that write starvation can be avoided, not in high-level >> data read paths where we have no clue about anything else going on >> in the IO subsystem >> >> Cheers, >> >> Dave. >> -- >> Dave Chinner >> da...@fromorbit.com >> -- >> To unsubscribe from this list: send the line "unsubscribe linux-mmc" in >> the body of a message to majord...@vger.kernel.org >> More majordomo info at http://vger.kernel.org/majordomo-info.html > > -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
Mon, May 7, 2012 at 5:01 AM, Dave Chinner wrote: > On Thu, May 03, 2012 at 07:53:00PM +0530, Venkatraman S wrote: >> From: Ilan Smith >> >> Add attribute to identify demand paging requests. >> Mark readpages with demand paging attribute. >> >> Signed-off-by: Ilan Smith >> Signed-off-by: Alex Lemberg >> Signed-off-by: Venkatraman S >> --- >> fs/mpage.c | 2 ++ >> include/linux/bio.h | 7 +++ >> include/linux/blk_types.h | 2 ++ >> 3 files changed, 11 insertions(+) >> >> diff --git a/fs/mpage.c b/fs/mpage.c >> index 0face1c..8b144f5 100644 >> --- a/fs/mpage.c >> +++ b/fs/mpage.c >> @@ -386,6 +386,8 @@ mpage_readpages(struct address_space *mapping, struct >> list_head *pages, >> &last_block_in_bio, &map_bh, >> &first_logical_block, >> get_block); >> + if (bio) >> + bio->bi_rw |= REQ_RW_DMPG; > > Have you thought about the potential for DOSing a machine > with this? That is, user data reads can now preempt writes of any > kind, effectively stalling writeback and memory reclaim which will > lead to OOM situations. Or, alternatively, journal flushing will get > stalled and no new modifications can take place until the read > stream stops. This feature doesn't fiddle with the I/O scheduler's ability to balance read vs write requests or handling requests from various process queues (CFQ). Also, for block devices which don't implement the ability to preempt (and even for older versions of MMC devices which don't implement this feature), the behaviour falls back to waiting for write requests to complete before issuing the read. In low end flash devices, some requests might take too long than normal due to background device maintenance (i.e flash erase / reclaim procedure) kicking in in the context of an ongoing write, stalling them by several orders of magnitude. This implementation (See 14/16) does have several checks and timers to see that it's not triggered very often. In my tests, where I usually have a generous preemption time window, the abort happens < 0.1% of the time. > > This really seems like functionality that belongs in an IO > scheduler so that write starvation can be avoided, not in high-level > data read paths where we have no clue about anything else going on > in the IO subsystem Indeed, the feature is built mostly in the low level device driver and minor changes in the elevator. Changes above the block layer are only about setting attributes and transparent to their operation. > > Cheers, > > Dave. > -- > Dave Chinner > da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
Re: [PATCH v2 01/16] FS: Added demand paging markers to filesystem
On Thu, May 03, 2012 at 07:53:00PM +0530, Venkatraman S wrote: > From: Ilan Smith > > Add attribute to identify demand paging requests. > Mark readpages with demand paging attribute. > > Signed-off-by: Ilan Smith > Signed-off-by: Alex Lemberg > Signed-off-by: Venkatraman S > --- > fs/mpage.c|2 ++ > include/linux/bio.h |7 +++ > include/linux/blk_types.h |2 ++ > 3 files changed, 11 insertions(+) > > diff --git a/fs/mpage.c b/fs/mpage.c > index 0face1c..8b144f5 100644 > --- a/fs/mpage.c > +++ b/fs/mpage.c > @@ -386,6 +386,8 @@ mpage_readpages(struct address_space *mapping, struct > list_head *pages, > &last_block_in_bio, &map_bh, > &first_logical_block, > get_block); > + if (bio) > + bio->bi_rw |= REQ_RW_DMPG; Have you thought about the potential for DOSing a machine with this? That is, user data reads can now preempt writes of any kind, effectively stalling writeback and memory reclaim which will lead to OOM situations. Or, alternatively, journal flushing will get stalled and no new modifications can take place until the read stream stops. This really seems like functionality that belongs in an IO scheduler so that write starvation can be avoided, not in high-level data read paths where we have no clue about anything else going on in the IO subsystem Cheers, Dave. -- Dave Chinner da...@fromorbit.com -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html
[PATCH v2 01/16] FS: Added demand paging markers to filesystem
From: Ilan Smith Add attribute to identify demand paging requests. Mark readpages with demand paging attribute. Signed-off-by: Ilan Smith Signed-off-by: Alex Lemberg Signed-off-by: Venkatraman S --- fs/mpage.c|2 ++ include/linux/bio.h |7 +++ include/linux/blk_types.h |2 ++ 3 files changed, 11 insertions(+) diff --git a/fs/mpage.c b/fs/mpage.c index 0face1c..8b144f5 100644 --- a/fs/mpage.c +++ b/fs/mpage.c @@ -386,6 +386,8 @@ mpage_readpages(struct address_space *mapping, struct list_head *pages, &last_block_in_bio, &map_bh, &first_logical_block, get_block); + if (bio) + bio->bi_rw |= REQ_RW_DMPG; } page_cache_release(page); } diff --git a/include/linux/bio.h b/include/linux/bio.h index 4d94eb8..264e0ef 100644 --- a/include/linux/bio.h +++ b/include/linux/bio.h @@ -57,6 +57,13 @@ (bio)->bi_rw |= ((unsigned long) (prio) << BIO_PRIO_SHIFT); \ } while (0) +static inline bool bio_rw_flagged(struct bio *bio, unsigned long flag) +{ + return ((bio->bi_rw & flag) != 0); +} + +#define bio_dmpg(bio) bio_rw_flagged(bio, REQ_RW_DMPG) + /* * various member access, note that bio_data should of course not be used * on highmem page vectors diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h index 4053cbd..87feb80 100644 --- a/include/linux/blk_types.h +++ b/include/linux/blk_types.h @@ -150,6 +150,7 @@ enum rq_flag_bits { __REQ_FLUSH_SEQ,/* request for flush sequence */ __REQ_IO_STAT, /* account I/O stat */ __REQ_MIXED_MERGE, /* merge of different types, fail separately */ + __REQ_RW_DMPG, __REQ_NR_BITS, /* stops here */ }; @@ -191,5 +192,6 @@ enum rq_flag_bits { #define REQ_IO_STAT(1 << __REQ_IO_STAT) #define REQ_MIXED_MERGE(1 << __REQ_MIXED_MERGE) #define REQ_SECURE (1 << __REQ_SECURE) +#define REQ_RW_DMPG(1 << __REQ_RW_DMPG) #endif /* __LINUX_BLK_TYPES_H */ -- 1.7.10.rc2 -- To unsubscribe from this list: send the line "unsubscribe linux-omap" in the body of a message to majord...@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html