| Message ID | 14bbd13c5c93211ed4e36188624f429087a6a273.1521016003.git.berto@igalia.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On 2018-03-14 09:29, Alberto Garcia wrote: > The L2 and refcount caches have default sizes that can be overridden > using the l2-cache-size and refcount-cache-size (an additional > parameter named cache-size sets the combined size of both caches). > > Unless forced by one of the aforementioned parameters, QEMU will set > the unspecified sizes so that the L2 cache is 4 times larger than the > refcount cache. > > This is based on the premise that the refcount metadata needs to be > only a fourth of the L2 metadata to cover the same amount of disk > space. This is incorrect for two reasons: > > a) The amount of disk covered by an L2 table depends solely on the > cluster size, but in the case of a refcount block it depends on > the cluster size *and* the width of each refcount entry. > The 4/1 ratio is only valid with 16-bit entries (the default). > > b) When we talk about disk space and L2 tables we are talking about > guest space (L2 tables map guest clusters to host clusters), > whereas refcount blocks are used for host clusters (including > L1/L2 tables and the refcount blocks themselves). On a fully > populated (and uncompressed) qcow2 file, image size > virtual size > so there are more refcount entries than L2 entries. > > Problem (a) could be fixed by adjusting the algorithm to take into > account the refcount entry width. Problem (b) could be fixed by > increasing a bit the refcount cache size to account for the clusters > used for qcow2 metadata. Yeah, well, but I wouldn't give too much thought to (b). In case of internal snapshots, you won't write to them, so those clusters don't matter. So, yes, you get a bit of overhead for the metadata, but come on, that's not really much. And (a) would be trivial, as you yourself say. > However this patch takes a completely different approach and instead > of keeping a ratio between both cache sizes it assigns as much as > possible to the L2 cache and the remainder to the refcount cache. Seems reasonable in a sense, but I'm wary still. I especially don't like that it just changes the default when nothing has been specified. I can see that we want to limit the refcount cache size, so if you want to give really much RAM to the L2 tables, then you don't want a 25 % overhead for the refcount cache, but do we really need to limit it to the minimum size? OTOH, four clusters doesn't seem extremely limited... Did you do your benchmark on an HDD? Max > The reason is that L2 tables are used for every single I/O request > from the guest and the effect of increasing the cache is significant > and clearly measurable. Refcount blocks are however only used for > cluster allocation and internal snapshots and in practice are accessed > sequentially in most cases, so the effect of increasing the cache is > negligible (even when doing random writes from the guest). > > So, make the refcount cache as small as possible unless the user > explicitly asks for a larger one. > > Signed-off-by: Alberto Garcia <berto@igalia.com> > Reviewed-by: Eric Blake <eblake@redhat.com> > --- > block/qcow2.c | 31 +++++++++++++++++++------------ > block/qcow2.h | 4 ---- > tests/qemu-iotests/137.out | 2 +- > 3 files changed, 20 insertions(+), 17 deletions(-) > > diff --git a/block/qcow2.c b/block/qcow2.c > index 7472af6931..8342b0186f 100644 > --- a/block/qcow2.c > +++ b/block/qcow2.c > @@ -802,23 +802,30 @@ static void read_cache_sizes(BlockDriverState *bs, QemuOpts *opts, > } else if (refcount_cache_size_set) { > *l2_cache_size = combined_cache_size - *refcount_cache_size; > } else { > - *refcount_cache_size = combined_cache_size > - / (DEFAULT_L2_REFCOUNT_SIZE_RATIO + 1); > - *l2_cache_size = combined_cache_size - *refcount_cache_size; > + uint64_t virtual_disk_size = bs->total_sectors * BDRV_SECTOR_SIZE; > + uint64_t max_l2_cache = virtual_disk_size / (s->cluster_size / 8); > + uint64_t min_refcount_cache = > + (uint64_t) MIN_REFCOUNT_CACHE_SIZE * s->cluster_size; > + > + /* Assign as much memory as possible to the L2 cache, and > + * use the remainder for the refcount cache */ > + if (combined_cache_size >= max_l2_cache + min_refcount_cache) { > + *l2_cache_size = max_l2_cache; > + *refcount_cache_size = combined_cache_size - *l2_cache_size; > + } else { > + *refcount_cache_size = > + MIN(combined_cache_size, min_refcount_cache); > + *l2_cache_size = combined_cache_size - *refcount_cache_size; > + } > } > } else { > - if (!l2_cache_size_set && !refcount_cache_size_set) { > + if (!l2_cache_size_set) { > *l2_cache_size = MAX(DEFAULT_L2_CACHE_BYTE_SIZE, > (uint64_t)DEFAULT_L2_CACHE_CLUSTERS > * s->cluster_size); > - *refcount_cache_size = *l2_cache_size > - / DEFAULT_L2_REFCOUNT_SIZE_RATIO; > - } else if (!l2_cache_size_set) { > - *l2_cache_size = *refcount_cache_size > - * DEFAULT_L2_REFCOUNT_SIZE_RATIO; > - } else if (!refcount_cache_size_set) { > - *refcount_cache_size = *l2_cache_size > - / DEFAULT_L2_REFCOUNT_SIZE_RATIO; > + } > + if (!refcount_cache_size_set) { > + *refcount_cache_size = MIN_REFCOUNT_CACHE_SIZE * s->cluster_size; > } > } > > diff --git a/block/qcow2.h b/block/qcow2.h > index ccb92a9696..cdf41055ae 100644 > --- a/block/qcow2.h > +++ b/block/qcow2.h > @@ -77,10 +77,6 @@ > #define DEFAULT_L2_CACHE_CLUSTERS 8 /* clusters */ > #define DEFAULT_L2_CACHE_BYTE_SIZE 1048576 /* bytes */ > > -/* The refblock cache needs only a fourth of the L2 cache size to cover as many > - * clusters */ > -#define DEFAULT_L2_REFCOUNT_SIZE_RATIO 4 > - > #define DEFAULT_CLUSTER_SIZE 65536 > > > diff --git a/tests/qemu-iotests/137.out b/tests/qemu-iotests/137.out > index e28e1eadba..96724a6c33 100644 > --- a/tests/qemu-iotests/137.out > +++ b/tests/qemu-iotests/137.out > @@ -22,7 +22,7 @@ refcount-cache-size may not exceed cache-size > L2 cache size too big > L2 cache entry size must be a power of two between 512 and the cluster size (65536) > L2 cache entry size must be a power of two between 512 and the cluster size (65536) > -L2 cache size too big > +Refcount cache size too big > Conflicting values for qcow2 options 'overlap-check' ('constant') and 'overlap-check.template' ('all') > Unsupported value 'blubb' for qcow2 option 'overlap-check'. Allowed are any of the following: none, constant, cached, all > Unsupported value 'blubb' for qcow2 option 'overlap-check'. Allowed are any of the following: none, constant, cached, all >
On Fri 13 Apr 2018 05:00:48 PM CEST, Max Reitz wrote: > On 2018-03-14 09:29, Alberto Garcia wrote: >> The L2 and refcount caches have default sizes that can be overridden >> using the l2-cache-size and refcount-cache-size (an additional >> parameter named cache-size sets the combined size of both caches). >> >> Unless forced by one of the aforementioned parameters, QEMU will set >> the unspecified sizes so that the L2 cache is 4 times larger than the >> refcount cache. >> >> This is based on the premise that the refcount metadata needs to be >> only a fourth of the L2 metadata to cover the same amount of disk >> space. This is incorrect for two reasons: >> >> a) The amount of disk covered by an L2 table depends solely on the >> cluster size, but in the case of a refcount block it depends on >> the cluster size *and* the width of each refcount entry. >> The 4/1 ratio is only valid with 16-bit entries (the default). >> >> b) When we talk about disk space and L2 tables we are talking about >> guest space (L2 tables map guest clusters to host clusters), >> whereas refcount blocks are used for host clusters (including >> L1/L2 tables and the refcount blocks themselves). On a fully >> populated (and uncompressed) qcow2 file, image size > virtual size >> so there are more refcount entries than L2 entries. >> >> Problem (a) could be fixed by adjusting the algorithm to take into >> account the refcount entry width. Problem (b) could be fixed by >> increasing a bit the refcount cache size to account for the clusters >> used for qcow2 metadata. > > Yeah, well, but I wouldn't give too much thought to (b). In case of > internal snapshots, you won't write to them, so those clusters don't > matter. So, yes, you get a bit of overhead for the metadata, but come > on, that's not really much. With that text I just wanted to point out that the calculation mixes host clusters and guest clusters. I agree that the ratio we're using is a good approximation (apart from the refcount width issue). >> However this patch takes a completely different approach and instead >> of keeping a ratio between both cache sizes it assigns as much as >> possible to the L2 cache and the remainder to the refcount cache. > > Seems reasonable in a sense, but I'm wary still. I especially don't > like that it just changes the default when nothing has been specified. > > I can see that we want to limit the refcount cache size, so if you > want to give really much RAM to the L2 tables, then you don't want a > 25 % overhead for the refcount cache, but do we really need to limit > it to the minimum size? > > OTOH, four clusters doesn't seem extremely limited... Did you do your > benchmark on an HDD? I tried, I see no difference, and I don't see why there would be any. Refcount entries are used to reference host clusters, and host clusters are always(*) allocated sequentially, so caching refcount blocks doesn't do much. You're always loading the same refcount block until it's full, then you move on to the next one. As I mentioned in a previous e-mail, having a very large refcount cache would even be detrimental because it would make cache hits slower (noticeable under tmpfs). (*) when a cluster gets deallocated (e.g. when the refcount table needs to grow) then there's a hole in the file and QEMU starts looking for free clusters again from the beginning of the file. In that case old refcount blocks are loaded again, but the cache wouldn't help here either (unless we would have all refcount blocks cached). Berto
On 2018-04-16 15:56, Alberto Garcia wrote: > On Fri 13 Apr 2018 05:00:48 PM CEST, Max Reitz wrote: >> On 2018-03-14 09:29, Alberto Garcia wrote: >>> The L2 and refcount caches have default sizes that can be overridden >>> using the l2-cache-size and refcount-cache-size (an additional >>> parameter named cache-size sets the combined size of both caches). >>> >>> Unless forced by one of the aforementioned parameters, QEMU will set >>> the unspecified sizes so that the L2 cache is 4 times larger than the >>> refcount cache. >>> >>> This is based on the premise that the refcount metadata needs to be >>> only a fourth of the L2 metadata to cover the same amount of disk >>> space. This is incorrect for two reasons: >>> >>> a) The amount of disk covered by an L2 table depends solely on the >>> cluster size, but in the case of a refcount block it depends on >>> the cluster size *and* the width of each refcount entry. >>> The 4/1 ratio is only valid with 16-bit entries (the default). >>> >>> b) When we talk about disk space and L2 tables we are talking about >>> guest space (L2 tables map guest clusters to host clusters), >>> whereas refcount blocks are used for host clusters (including >>> L1/L2 tables and the refcount blocks themselves). On a fully >>> populated (and uncompressed) qcow2 file, image size > virtual size >>> so there are more refcount entries than L2 entries. >>> >>> Problem (a) could be fixed by adjusting the algorithm to take into >>> account the refcount entry width. Problem (b) could be fixed by >>> increasing a bit the refcount cache size to account for the clusters >>> used for qcow2 metadata. >> >> Yeah, well, but I wouldn't give too much thought to (b). In case of >> internal snapshots, you won't write to them, so those clusters don't >> matter. So, yes, you get a bit of overhead for the metadata, but come >> on, that's not really much. > > With that text I just wanted to point out that the calculation mixes > host clusters and guest clusters. I agree that the ratio we're using is > a good approximation (apart from the refcount width issue). > >>> However this patch takes a completely different approach and instead >>> of keeping a ratio between both cache sizes it assigns as much as >>> possible to the L2 cache and the remainder to the refcount cache. >> >> Seems reasonable in a sense, but I'm wary still. I especially don't >> like that it just changes the default when nothing has been specified. >> >> I can see that we want to limit the refcount cache size, so if you >> want to give really much RAM to the L2 tables, then you don't want a >> 25 % overhead for the refcount cache, but do we really need to limit >> it to the minimum size? >> >> OTOH, four clusters doesn't seem extremely limited... Did you do your >> benchmark on an HDD? > > I tried, I see no difference, and I don't see why there would be any. > > Refcount entries are used to reference host clusters, and host clusters > are always(*) allocated sequentially, so caching refcount blocks doesn't > do much. You're always loading the same refcount block until it's full, > then you move on to the next one. As I mentioned in a previous e-mail, > having a very large refcount cache would even be detrimental because it > would make cache hits slower (noticeable under tmpfs). Ah, right. I'm not sure whether I want to ask you whether you have tested internal snapshots. I suppose it can be detrimental for them because when taking or deleting a snapshot you need to update the refcounts of a whole bunch of clusters that might be spread randomly across the image. But I suspect that it would be rather time-consuming to produce an image with such a configuration; and I suppose if people want to make heavy use of internal snapshots they can adapt the cache size themselves? Max > (*) when a cluster gets deallocated (e.g. when the refcount table needs > to grow) then there's a hole in the file and QEMU starts looking for > free clusters again from the beginning of the file. In that case old > refcount blocks are loaded again, but the cache wouldn't help here > either (unless we would have all refcount blocks cached).
On Mon 16 Apr 2018 04:05:21 PM CEST, Max Reitz wrote: >> Refcount entries are used to reference host clusters, and host >> clusters are always(*) allocated sequentially, so caching refcount >> blocks doesn't do much. You're always loading the same refcount block >> until it's full, then you move on to the next one. As I mentioned in >> a previous e-mail, having a very large refcount cache would even be >> detrimental because it would make cache hits slower (noticeable under >> tmpfs). > > Ah, right. I'm not sure whether I want to ask you whether you have > tested internal snapshots. I suppose it can be detrimental for them > because when taking or deleting a snapshot you need to update the > refcounts of a whole bunch of clusters that might be spread randomly > across the image. But I suspect that it would be rather > time-consuming to produce an image with such a configuration; and I > suppose if people want to make heavy use of internal snapshots they > can adapt the cache size themselves? Exactly. Berto
On 2018-03-14 09:29, Alberto Garcia wrote: > The L2 and refcount caches have default sizes that can be overridden > using the l2-cache-size and refcount-cache-size (an additional > parameter named cache-size sets the combined size of both caches). > > Unless forced by one of the aforementioned parameters, QEMU will set > the unspecified sizes so that the L2 cache is 4 times larger than the > refcount cache. > > This is based on the premise that the refcount metadata needs to be > only a fourth of the L2 metadata to cover the same amount of disk > space. This is incorrect for two reasons: > > a) The amount of disk covered by an L2 table depends solely on the > cluster size, but in the case of a refcount block it depends on > the cluster size *and* the width of each refcount entry. > The 4/1 ratio is only valid with 16-bit entries (the default). > > b) When we talk about disk space and L2 tables we are talking about > guest space (L2 tables map guest clusters to host clusters), > whereas refcount blocks are used for host clusters (including > L1/L2 tables and the refcount blocks themselves). On a fully > populated (and uncompressed) qcow2 file, image size > virtual size > so there are more refcount entries than L2 entries. > > Problem (a) could be fixed by adjusting the algorithm to take into > account the refcount entry width. Problem (b) could be fixed by > increasing a bit the refcount cache size to account for the clusters > used for qcow2 metadata. > > However this patch takes a completely different approach and instead > of keeping a ratio between both cache sizes it assigns as much as > possible to the L2 cache and the remainder to the refcount cache. > > The reason is that L2 tables are used for every single I/O request > from the guest and the effect of increasing the cache is significant > and clearly measurable. Refcount blocks are however only used for > cluster allocation and internal snapshots and in practice are accessed > sequentially in most cases, so the effect of increasing the cache is > negligible (even when doing random writes from the guest). > > So, make the refcount cache as small as possible unless the user > explicitly asks for a larger one. > > Signed-off-by: Alberto Garcia <berto@igalia.com> > Reviewed-by: Eric Blake <eblake@redhat.com> > --- > block/qcow2.c | 31 +++++++++++++++++++------------ > block/qcow2.h | 4 ---- > tests/qemu-iotests/137.out | 2 +- > 3 files changed, 20 insertions(+), 17 deletions(-) Reviewed-by: Max Reitz <mreitz@redhat.com>
diff --git a/block/qcow2.c b/block/qcow2.c index 7472af6931..8342b0186f 100644 --- a/block/qcow2.c +++ b/block/qcow2.c @@ -802,23 +802,30 @@ static void read_cache_sizes(BlockDriverState *bs, QemuOpts *opts, } else if (refcount_cache_size_set) { *l2_cache_size = combined_cache_size - *refcount_cache_size; } else { - *refcount_cache_size = combined_cache_size - / (DEFAULT_L2_REFCOUNT_SIZE_RATIO + 1); - *l2_cache_size = combined_cache_size - *refcount_cache_size; + uint64_t virtual_disk_size = bs->total_sectors * BDRV_SECTOR_SIZE; + uint64_t max_l2_cache = virtual_disk_size / (s->cluster_size / 8); + uint64_t min_refcount_cache = + (uint64_t) MIN_REFCOUNT_CACHE_SIZE * s->cluster_size; + + /* Assign as much memory as possible to the L2 cache, and + * use the remainder for the refcount cache */ + if (combined_cache_size >= max_l2_cache + min_refcount_cache) { + *l2_cache_size = max_l2_cache; + *refcount_cache_size = combined_cache_size - *l2_cache_size; + } else { + *refcount_cache_size = + MIN(combined_cache_size, min_refcount_cache); + *l2_cache_size = combined_cache_size - *refcount_cache_size; + } } } else { - if (!l2_cache_size_set && !refcount_cache_size_set) { + if (!l2_cache_size_set) { *l2_cache_size = MAX(DEFAULT_L2_CACHE_BYTE_SIZE, (uint64_t)DEFAULT_L2_CACHE_CLUSTERS * s->cluster_size); - *refcount_cache_size = *l2_cache_size - / DEFAULT_L2_REFCOUNT_SIZE_RATIO; - } else if (!l2_cache_size_set) { - *l2_cache_size = *refcount_cache_size - * DEFAULT_L2_REFCOUNT_SIZE_RATIO; - } else if (!refcount_cache_size_set) { - *refcount_cache_size = *l2_cache_size - / DEFAULT_L2_REFCOUNT_SIZE_RATIO; + } + if (!refcount_cache_size_set) { + *refcount_cache_size = MIN_REFCOUNT_CACHE_SIZE * s->cluster_size; } } diff --git a/block/qcow2.h b/block/qcow2.h index ccb92a9696..cdf41055ae 100644 --- a/block/qcow2.h +++ b/block/qcow2.h @@ -77,10 +77,6 @@ #define DEFAULT_L2_CACHE_CLUSTERS 8 /* clusters */ #define DEFAULT_L2_CACHE_BYTE_SIZE 1048576 /* bytes */ -/* The refblock cache needs only a fourth of the L2 cache size to cover as many - * clusters */ -#define DEFAULT_L2_REFCOUNT_SIZE_RATIO 4 - #define DEFAULT_CLUSTER_SIZE 65536 diff --git a/tests/qemu-iotests/137.out b/tests/qemu-iotests/137.out index e28e1eadba..96724a6c33 100644 --- a/tests/qemu-iotests/137.out +++ b/tests/qemu-iotests/137.out @@ -22,7 +22,7 @@ refcount-cache-size may not exceed cache-size L2 cache size too big L2 cache entry size must be a power of two between 512 and the cluster size (65536) L2 cache entry size must be a power of two between 512 and the cluster size (65536) -L2 cache size too big +Refcount cache size too big Conflicting values for qcow2 options 'overlap-check' ('constant') and 'overlap-check.template' ('all') Unsupported value 'blubb' for qcow2 option 'overlap-check'. Allowed are any of the following: none, constant, cached, all Unsupported value 'blubb' for qcow2 option 'overlap-check'. Allowed are any of the following: none, constant, cached, all
The L2 and refcount caches have default sizes that can be overridden using the l2-cache-size and refcount-cache-size (an additional parameter named cache-size sets the combined size of both caches). Unless forced by one of the aforementioned parameters, QEMU will set the unspecified sizes so that the L2 cache is 4 times larger than the refcount cache. This is based on the premise that the refcount metadata needs to be only a fourth of the L2 metadata to cover the same amount of disk space. This is incorrect for two reasons: a) The amount of disk covered by an L2 table depends solely on the cluster size, but in the case of a refcount block it depends on the cluster size *and* the width of each refcount entry. The 4/1 ratio is only valid with 16-bit entries (the default). b) When we talk about disk space and L2 tables we are talking about guest space (L2 tables map guest clusters to host clusters), whereas refcount blocks are used for host clusters (including L1/L2 tables and the refcount blocks themselves). On a fully populated (and uncompressed) qcow2 file, image size > virtual size so there are more refcount entries than L2 entries. Problem (a) could be fixed by adjusting the algorithm to take into account the refcount entry width. Problem (b) could be fixed by increasing a bit the refcount cache size to account for the clusters used for qcow2 metadata. However this patch takes a completely different approach and instead of keeping a ratio between both cache sizes it assigns as much as possible to the L2 cache and the remainder to the refcount cache. The reason is that L2 tables are used for every single I/O request from the guest and the effect of increasing the cache is significant and clearly measurable. Refcount blocks are however only used for cluster allocation and internal snapshots and in practice are accessed sequentially in most cases, so the effect of increasing the cache is negligible (even when doing random writes from the guest). So, make the refcount cache as small as possible unless the user explicitly asks for a larger one. Signed-off-by: Alberto Garcia <berto@igalia.com> Reviewed-by: Eric Blake <eblake@redhat.com> --- block/qcow2.c | 31 +++++++++++++++++++------------ block/qcow2.h | 4 ---- tests/qemu-iotests/137.out | 2 +- 3 files changed, 20 insertions(+), 17 deletions(-)