public inbox for libstdc++-cvs@sourceware.org
help / color / mirror / Atom feed
* [gcc r10-8780] libstdc++: handle small max_blocks_per_chunk in pool resources [PR 94160]
@ 2020-09-21 20:18 Jonathan Wakely
0 siblings, 0 replies; only message in thread
From: Jonathan Wakely @ 2020-09-21 20:18 UTC (permalink / raw)
To: gcc-cvs, libstdc++-cvs
https://gcc.gnu.org/g:9ae110d4f8edd61cf56ee5fea62cadc8e781e8dc
commit r10-8780-g9ae110d4f8edd61cf56ee5fea62cadc8e781e8dc
Author: Jonathan Wakely <jwakely@redhat.com>
Date: Thu Sep 10 15:39:15 2020 +0100
libstdc++: handle small max_blocks_per_chunk in pool resources [PR 94160]
When a pool resource is constructed with max_blocks_per_chunk=1 it ends
up creating a pool with blocks_per_chunk=0 which means it never
allocates anything. Instead it returns null pointers, which should be
impossible.
To avoid this problem, round the max_blocks_per_chunk value to a
multiple of four, so it's never smaller than four.
libstdc++-v3/ChangeLog:
PR libstdc++/94160
* src/c++17/memory_resource.cc (munge_options): Round
max_blocks_per_chunk to a multiple of four.
(__pool_resource::_M_alloc_pools()): Simplify slightly.
* testsuite/20_util/unsynchronized_pool_resource/allocate.cc:
Check that valid pointers are returned when small values are
used for max_blocks_per_chunk.
(cherry picked from commit 30b41cfbb2dade63e52465234a725d1d02fe70aa)
Diff:
---
libstdc++-v3/src/c++17/memory_resource.cc | 21 +++++++++++++++------
.../unsynchronized_pool_resource/allocate.cc | 20 ++++++++++++++++++++
2 files changed, 35 insertions(+), 6 deletions(-)
diff --git a/libstdc++-v3/src/c++17/memory_resource.cc b/libstdc++-v3/src/c++17/memory_resource.cc
index 95352b23537..75ae24e0a06 100644
--- a/libstdc++-v3/src/c++17/memory_resource.cc
+++ b/libstdc++-v3/src/c++17/memory_resource.cc
@@ -873,7 +873,18 @@ namespace pmr
}
else
{
- // TODO round to preferred granularity ?
+ // Round to preferred granularity.
+ if (opts.max_blocks_per_chunk < size_t(-4))
+ {
+ // round up
+ opts.max_blocks_per_chunk += 3;
+ opts.max_blocks_per_chunk &= ~size_t(3);
+ }
+ else
+ {
+ // round down
+ opts.max_blocks_per_chunk &= ~size_t(3);
+ }
}
if (opts.max_blocks_per_chunk > chunk::max_blocks_per_chunk())
@@ -1013,11 +1024,9 @@ namespace pmr
: pool_sizes[i];
// Decide on initial number of blocks per chunk.
- // Always have at least 16 blocks per chunk:
- const size_t min_blocks_per_chunk = 16;
- // But for smaller blocks, use a larger initial size:
- size_t blocks_per_chunk
- = std::max(1024 / block_size, min_blocks_per_chunk);
+ // At least 16 blocks per chunk seems reasonable,
+ // more for smaller blocks:
+ size_t blocks_per_chunk = std::max(size_t(16), 1024 / block_size);
// But don't exceed the requested max_blocks_per_chunk:
blocks_per_chunk
= std::min(blocks_per_chunk, _M_opts.max_blocks_per_chunk);
diff --git a/libstdc++-v3/testsuite/20_util/unsynchronized_pool_resource/allocate.cc b/libstdc++-v3/testsuite/20_util/unsynchronized_pool_resource/allocate.cc
index 5bf20cf262c..ef5f921211d 100644
--- a/libstdc++-v3/testsuite/20_util/unsynchronized_pool_resource/allocate.cc
+++ b/libstdc++-v3/testsuite/20_util/unsynchronized_pool_resource/allocate.cc
@@ -239,6 +239,25 @@ test06()
}
}
+void
+test08()
+{
+ std::pmr::pool_options opts;
+ opts.largest_required_pool_block = 64;
+
+ // PR libstdc++/94160
+ // max_blocks_per_chunk=1 causes pool resources to return null pointers
+ for (int i = 0; i < 8; ++i)
+ {
+ opts.max_blocks_per_chunk = i;
+ std::pmr::unsynchronized_pool_resource upr(opts);
+ auto* p = (int*)upr.allocate(4);
+ VERIFY( p != nullptr );
+ *p = i;
+ upr.deallocate(p, 4);
+ }
+}
+
int
main()
{
@@ -248,4 +267,5 @@ main()
test04();
test05();
test06();
+ test08();
}
^ permalink raw reply [flat|nested] only message in thread
only message in thread, other threads:[~2020-09-21 20:18 UTC | newest]
Thread overview: (only message) (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-09-21 20:18 [gcc r10-8780] libstdc++: handle small max_blocks_per_chunk in pool resources [PR 94160] Jonathan Wakely
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).