We don't expose it as an AEAD yet because the logic for that is complex
due to overlap issues. For #69981 we will make a cipher.AEAD wrapper
outside the FIPS module, but maybe a v2 interface will make it easier,
and then we'll be able to use this method more directly.
Updates #69981
For #69536
Change-Id: Id88191c01443b0dec89ff0d6c4a6289f519369d1
Reviewed-on: https://go-review.googlesource.com/c/go/+/624916
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
TestEmptyPlaintext and TestCryptBlocks were folded into cryptotest.
Change-Id: I6131ab8582eb0e6d3a1b24bab1147a145d9766ac
Reviewed-on: https://go-review.googlesource.com/c/go/+/624738
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
[ ]
[ It has been [ 0 ] days since Filippo broke a TestAllocations. ]
[ ]
Concentrate all the skips in one place, so we don't have to re-discover
always the same ones via trial and error.
This might over-skip fixable allocations, but all these targets are not
fast anyway, so they are not worth going back for.
Removed the sysrand TestAllocations because it causes an import loop
with cryptotest and it's covered by TestAllocations in crypto/rand.
Change-Id: Icd40e97f9128e037f567147f8c9dafa758a47fac
Reviewed-on: https://go-review.googlesource.com/c/go/+/626438
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
We're going to use that package as the passive entropy source for the
FIPS module, and we need to import it from a package that will be
imported by crypto/rand.
Since there is no overridable Reader now, introduced a mechanism to test
the otherwise impossible failure of the OS entropy source.
For #69536
Change-Id: I558687ed1ec896dba05b99b937970bb809de3fe7
Reviewed-on: https://go-review.googlesource.com/c/go/+/624976
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
Reviewed-by: Roland Shoemaker <roland@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
We will need to import this functionality from the FIPS module, and we
can't import AES from there. Plan 9 is not going to be FIPS validated
anyway, so we can use non-approved cryptography.
For #69536
Change-Id: I7921ec0829b576de2e80f3a7d0a9a776ff387684
Reviewed-on: https://go-review.googlesource.com/c/go/+/624975
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
The crypto/aes <-> crypto/cipher interfaces and the hardware support
upgrades were layered over the years, and had grown unwieldily.
Before: conditionally wrap the private crypto/aes type in private types
that implement an interface that's interface-upgraded by crypto/cipher
to replace the generic implementation in crypto/cipher.
crypto/aes depended on crypto/cipher, which is backwards.
After: provide concrete exported implementations of modes in
crypto/internal/fips/aes that crypto/cipher returns if the input Block
is the crypto/internal/fips/aes concrete implementation.
crypto/aes and crypto/cipher both depend on crypto/internal/fips/aes.
Also, made everything follow go.dev/wiki/TargetSpecific by only putting
the minimal code necessary and no exported functions in build-tagged
files.
The GCM integration still uses an interface upgrade, because the generic
implementation is complex enough that it was not trivial to duplicate.
This will be fixed in a future CL to make review easier.
For #69536
Change-Id: I21c2b93a498edb31c562b1aca824e21e8457fdff
Reviewed-on: https://go-review.googlesource.com/c/go/+/624395
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Roland Shoemaker <roland@golang.org>
Auto-Submit: Filippo Valsorda <filippo@golang.org>
The implementation runs up to 8 AES instructions in different registers
one after another in ASM code. Because CPU has instruction pipelining
and the instructions do not depend on each other, they can run in
parallel with this layout of code. This results in significant speedup
compared to the regular implementation in which blocks are processed in
the same registers so AES instructions do not run in parallel.
GCM mode already utilizes the approach.
The ASM implementation of ctrAble has most of its code in XORKeyStreamAt
method which has an additional argument, offset. It allows to use it
in a stateless way and to jump to any location in the stream. The method
does not exist in pure Go and boringcrypto implementations.
[ Mailed as CL 413594, then edited by filippo@ to manage the counter
with bits.Add64, remove bounds checks, make the assembly interface more
explicit, and to port the amd64 to Avo. Squeezed another -6.38% out. ]
goos: linux
goarch: amd64
pkg: crypto/cipher
cpu: AMD Ryzen 7 PRO 8700GE w/ Radeon 780M Graphics
│ 19df80d792 │ c8b0409d40 │
│ sec/op │ sec/op vs base │
AESCTR/50-8 64.68n ± 0% 26.89n ± 0% -58.42% (p=0.000 n=10)
AESCTR/1K-8 1145.0n ± 0% 135.8n ± 0% -88.14% (p=0.000 n=10)
AESCTR/8K-8 9145.0n ± 0% 917.5n ± 0% -89.97% (p=0.000 n=10)
geomean 878.2n 149.6n -82.96%
│ 19df80d792 │ c8b0409d40 │
│ B/s │ B/s vs base │
AESCTR/50-8 737.2Mi ± 0% 1773.3Mi ± 0% +140.54% (p=0.000 n=10)
AESCTR/1K-8 848.5Mi ± 0% 7156.6Mi ± 0% +743.40% (p=0.000 n=10)
AESCTR/8K-8 853.8Mi ± 0% 8509.9Mi ± 0% +896.70% (p=0.000 n=10)
geomean 811.4Mi 4.651Gi +486.94%
Fixes#20967
Updates #39365
Updates #26673
Co-authored-by: Filippo Valsorda <filippo@golang.org>
Change-Id: Iaeea29fb93a56456f2e54507bc25196edb31b84b
Reviewed-on: https://go-review.googlesource.com/c/go/+/621958
Auto-Submit: Filippo Valsorda <filippo@golang.org>
Reviewed-by: Roland Shoemaker <roland@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
This tweaks the inlining cost knob for closures
specifically, they receive a doubled budget. The
rationale for this is that closures have a lot of
"crud" in their IR that will disappear after inlining,
so the standard budget penalizes them unnecessarily.
This is also the cause of these bugs -- looking at the
code involved, these closures "should" be inlineable,
therefore tweak the parameters until behavior matches
expectations. It's not costly in binary size, because
the only-called-from-one-site case is common (especially
for rangefunc iterators).
I can imagine better fixes and I am going to try to
get that done, but this one is small and makes things
better.
Fixes#69411, #69539.
Change-Id: I8a892c40323173a723799e0ddad69dcc2724a8f9
Reviewed-on: https://go-review.googlesource.com/c/go/+/629195
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: Keith Randall <khr@golang.org>
Reviewed-by: Robert Griesemer <gri@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
CL 627478 inadvertently returns a non-nil error
containing a syscall.Errno(0).
Change-Id: I1d6a9d0575d3ed651ddc02f30505437d0d266bb3
Reviewed-on: https://go-review.googlesource.com/c/go/+/629515
Auto-Submit: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Commit-Queue: Ian Lance Taylor <iant@google.com>
The updates are:
- API documentation changes.
- Removal of the old package documentation discouraging linkname.
- Addition of new package documentation with some advice.
- Renaming of weak.Pointer.Strong -> weak.Pointer.Value.
Fixes#67552.
Change-Id: Ifad7e629b6d339dacaf2ca37b459d7f903e31bf8
Reviewed-on: https://go-review.googlesource.com/c/go/+/628455
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Carlos Amedee <carlos@golang.org>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Creating and deleting users is tricky to get right, and it's not
something we want to do on a developer machine. This change skips the
tests that create users when not running on a Go builder.
This will fix#70396, although I still don't understand why the test
user couldn't be recreated.
Fixes#70396
Change-Id: Ie7004dc209f94e72152c7d6bd8ec95cc12c79757
Reviewed-on: https://go-review.googlesource.com/c/go/+/627877
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Instead, have the runtime build the gc bitmaps on demand
at runtime.
Change-Id: If7a245bc62e4bce3ce80972410b0ed307d921abe
Reviewed-on: https://go-review.googlesource.com/c/go/+/616255
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
Reviewed-by: Keith Randall <khr@google.com>
The linker knows the types of the global variables. We can use those
types to build the GC programs that describe the data and bss pointer masks.
That way we don't use the GC programs of the constituent types.
This is part of an effort to remove GC programs from the runtime.
There's a major complication in that when we're linking against a
shared library (typically, libstd.so), the relocations we need to
break apart arrays and structs into constituent types are difficult to
find. Load that additional data when linking against shared libraries.
Change-Id: I8516b24a0604479895c7b8a8a358d3cd8d421530
Reviewed-on: https://go-review.googlesource.com/c/go/+/546216
Reviewed-by: Keith Randall <khr@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
This change adds a GOEXPERIMENT, synchashtriemap, which replaces the
internals of a sync.Map with internal/sync.HashTrieMap[any, any]. The
main purpose behind this change is improved performance. Across almost
every benchmark, HashTrieMap[any, any] performs better than Map.
Also, relax TestMapClearNoAllocations to allow for one allocation.
Currently, the HashTrieMap allocates a new empty root node and stores
it: that's the whole clear operation. At the cost of some complexity, we
could allow Clear to have zero allocations by clearing the root node.
The complexity comes down to allowing threads to race to install a new
root node *or* creating a top-level mutex for installing a root node.
But I'm not sure this is worth it. Whether Clear or some other operation
takes the hit for allocating a single node almost certainly doesn't
matter. And Clear is still much, much faster in the new implementation
than the old, so I don't consider this a regression.
Change-Id: I939aa70a0edf2e850cedbea239aaf29a11a77b79
Reviewed-on: https://go-review.googlesource.com/c/go/+/608335
Auto-Submit: Michael Knyszek <mknyszek@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
We observe the CompareAndSwap and Swap can both be substantially faster
if the value in each entry node is mutable. This change modifies the
map entry node to store the value indirectly, allowing us to perform
swaps for existing nodes and compare-and-swaps without taking the
parent node's lock.
Change-Id: I371343aa81a843d3a7e6bc5ac87b8a96c12ca3a8
Reviewed-on: https://go-review.googlesource.com/c/go/+/606462
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Also, rename Map benchmarks to make them easier to single out via
regexp.
Change-Id: I4dcb066745aba1c340f56050d08539ae2976274d
Reviewed-on: https://go-review.googlesource.com/c/go/+/606461
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This implementation is trivial given that All already exists. The
purpose of Range is to conform to sync.Map's API.
Change-Id: Icfe755b9986c46b88c7201644e562b1631a02b66
Reviewed-on: https://go-review.googlesource.com/c/go/+/606460
Reviewed-by: David Chase <drchase@google.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
This change adds the Clear operation to HashTrieMap to align it with
sync.Map.
Change-Id: I46069b018725d7fe3a07d667876235534bca86f1
Reviewed-on: https://go-review.googlesource.com/c/go/+/606459
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This change adds the Delete operation (with the same semantics as
sync.Map's Delete) to HashTrieMap.
Change-Id: If90376ff735256444538719d670ed07e9e42870b
Reviewed-on: https://go-review.googlesource.com/c/go/+/606458
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This change does a minor refactor of the HashTrieMap tests to be better
organized into suites for each operation. It might be worthwhile to
deduplicate some code in here, but it's also helpful to see exactly
what's happening when something goes wrong.
Change-Id: I138515ee6de5aec5f3d38afe8a3bf3972afb2800
Reviewed-on: https://go-review.googlesource.com/c/go/+/606457
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
This change adds the LoadAndDelete operation (with the same semantics as
sync.Map's LoadAndDelete) to HashTrieMap.
Change-Id: Id6777dffcd3ebc98490aa51f0e85e59a56f63074
Reviewed-on: https://go-review.googlesource.com/c/go/+/606456
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Reviewed-by: David Chase <drchase@google.com>
This change adds the Store operation (with the same semantics as
sync.Map's Store) to HashTrieMap.
Change-Id: Ida55421f28d8cc6527f983f7f1c2fd55bb280925
Reviewed-on: https://go-review.googlesource.com/c/go/+/606455
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This change adds the Swap operation (with the same semantics as
sync.Map's Swap) to HashTrieMap.
Change-Id: I8697a0c8c2eb761e2452a41b868b590ccbfa5c03
Reviewed-on: https://go-review.googlesource.com/c/go/+/594064
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This change adds the CompareAndSwap operation (with the same semantics
as sync.Map's CompareAndSwap) to HashTrieMap.
Change-Id: I86153799fc47304784333d17f0c6a7ad7682f04a
Reviewed-on: https://go-review.googlesource.com/c/go/+/594063
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This improves ergonomics a little and aligns the HashTrieMap with
sync.Map.
Change-Id: Idb2b981a4f59a35f8670c6b5038e2bd207484483
Reviewed-on: https://go-review.googlesource.com/c/go/+/594062
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Currently the HashTrieMap requires both keys and values to be
comparable, but it's actually OK if the value is not comparable. Some
operations may fail, but others will not, and we can check comparability
dynamically on map initialization. This makes the implementation
substantially more flexible.
Change-Id: Idc9c30dfa273d80ae4d46a9eefb5c155294408aa
Reviewed-on: https://go-review.googlesource.com/c/go/+/594061
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Reviewed-by: David Chase <drchase@google.com>
There's are unnecessary calls to the key's equal function -- we can just
leverage the language here. Leave the values alone for now, we want to
relax that constraint.
Change-Id: Iccfaef030a2a29b6a24a7da41e5e816b70091c7c
Reviewed-on: https://go-review.googlesource.com/c/go/+/594060
Auto-Submit: Michael Knyszek <mknyszek@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
This lookup will be reused for other operations, like CompareAndSwap.
Change-Id: I6698e3c99f7ef6d2b82b9ef489ba8a0be2a71d61
Reviewed-on: https://go-review.googlesource.com/c/go/+/594059
Auto-Submit: Michael Knyszek <mknyszek@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
This change moves internal/concurrent.HashTrieMap from
internal/concurrent into internal/sync just to clean up the packages a
bit. This is all in anticipation of using HashTrieMap from the sync
package.
Change-Id: I18c007a301f83979d72f5d6bea600c42eaf2421e
Reviewed-on: https://go-review.googlesource.com/c/go/+/594058
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
This change uses linkname for the one random function
internal/concurrent needs to avoid taking a dependency on math/rand/v2.
This lowers the bar to using this package.
Change-Id: I9dba1121b66ba35f56521643937f220936ea5321
Reviewed-on: https://go-review.googlesource.com/c/go/+/594057
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
This CL refactors sync.Mutex such that its implementation lives in the
new internal/sync package. The purpose of this change is to eventually
reverse the dependency edge between internal/concurrent and sync, such
that sync can depend on internal/concurrent (or really, its contents,
which will likely end up in internal/sync).
The only change made to the sync.Mutex code is the frame skip count for
mutex profiling, so that the internal/sync frames are omitted in the
profile.
Change-Id: Ib3603d30e8e71508c4ea883a584ae2e51ce40c3f
Reviewed-on: https://go-review.googlesource.com/c/go/+/594056
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Change-Id: I9aa34951f2005c204aafd9da74e78c033c9b64ec
Reviewed-on: https://go-review.googlesource.com/c/go/+/629315
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Auto-Submit: Ian Lance Taylor <iant@google.com>
Commit-Queue: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Quim Muntal <quimmuntal@gmail.com>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Doing this because the slices functions are slightly faster and
slightly easier to use. It also removes one dependency layer.
We did this outside of bootstrap tools in CL 587655.
Now that the bootstrap compiler is 1.22, we can do this in more code.
Change-Id: I9ed2dd473758cacd14f76a0639368523ccdff72f
Reviewed-on: https://go-review.googlesource.com/c/go/+/626038
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Damien Neil <dneil@google.com>
FreeBSD and Dragonfly have used the sysctl method for years, while
NetBSD has read the name of the executable from /proc. Unfortunately,
some folks are hitting errors when building Go software in a sandbox
that lacks a mounted /proc filesystem.
Switch NetBSD to use the same implementation as FreeBSD and Dragonfly.
Unfortunately, the order of the arguments in the MIB is also
OS-dependent.
Change-Id: I6fd774904af417ccd127e3779af45a20dc8696ca
Reviewed-on: https://go-review.googlesource.com/c/go/+/629035
Reviewed-by: Tobias Klauser <tobias.klauser@gmail.com>
Run-TryBot: Benny Siegert <bsiegert@gmail.com>
TryBot-Bypass: Ian Lance Taylor <iant@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
TryBot-Result: Gopher Robot <gobot@golang.org>
Reviewed-by: Ian Lance Taylor <iant@google.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Auto-Submit: Ian Lance Taylor <iant@golang.org>
Following CLs will refactor Mutex and change the internals of Map. This
ends up breaking tests in x/tools for the copylock vet check, because
the error message changes. Let's insulate ourselves from such things
permanently by adding an explicit noCopy field. We'll update the vet
check to accept that as the problem, rather than depend on less explicit
internals.
We capture Once here too to clean up the error message as well.
Change-Id: Iead985fc8ec9ef3ea5ff615f26dde17bb03aeadb
Reviewed-on: https://go-review.googlesource.com/c/go/+/627777
Reviewed-by: Cherry Mui <cherryyz@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Michael Knyszek <mknyszek@google.com>
Reviewed-by: Tim King <taking@google.com>
This is a two-pronged approach. First, try to keep large objects
off the stack frame. Second, if they do manage to appear anyway,
use straight bitmasks instead of gc programs.
Generally probably a good idea to keep large objects out of stack frames.
But particularly keeping gc programs off the stack simplifies
runtime code a bit.
This CL sets the limit of most stack objects to 131072 bytes (on 64-bit archs).
There can still be large objects if allocated by a late pass, like order, or
they are required to be on the stack, like function arguments.
But the size for the bitmasks for these objects isn't a huge deal,
as we have already have (probably several) bitmasks for the frame
liveness map itself.
Change-Id: I6d2bed0e9aa9ac7499955562c6154f9264061359
Reviewed-on: https://go-review.googlesource.com/c/go/+/542815
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
We can reuse the same indirect storage when growing, so we don't
need an additional allocation.
Change-Id: I57adb406becfbec648188ec66f4bb2e94d4b9cab
Reviewed-on: https://go-review.googlesource.com/c/go/+/625902
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Michael Pratt <mpratt@google.com>
Reviewed-by: Keith Randall <khr@google.com>
It wasn't actually testing what it says it was testing.
A random permutation isn't cyclic. It only probably hits a few
elements before entering a cycle.
Use an algorithm that generates a random cyclic permutation instead.
Fixing the test makes the previous CL look less good. But it still helps.
(Theory: Fixing the test makes it less cache friendly, so there are
more misses all around. That makes the benchmark slower, suppressing
the differences seen. Also fixing the benchmark makes the loop
iteration count less predictable, which hurts the raw loop
implementation somewhat.)
(baseline = tip, experiment = tip+previous CL, noswiss = GOEXPERIMENT=noswissmap)
goos: darwin
goarch: arm64
pkg: runtime
cpu: Apple M2 Ultra
│ baseline │ experiment │
│ sec/op │ sec/op vs base │
MapCycle-24 20.59n ± 4% 18.99n ± 3% -7.77% (p=0.000 n=10)
khr@Mac-Studio src % benchstat noswiss experiment
goos: darwin
goarch: arm64
pkg: runtime
cpu: Apple M2 Ultra
│ noswiss │ experiment │
│ sec/op │ sec/op vs base │
MapCycle-24 16.12n ± 1% 18.99n ± 3% +17.83% (p=0.000 n=10)
Change-Id: I3a4edb814ba97fec020a6698c535ce3a87a9fc67
Reviewed-on: https://go-review.googlesource.com/c/go/+/625900
Reviewed-by: Michael Pratt <mpratt@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Keith Randall <khr@google.com>
XORBytes doesn't say anything about how it deals with destination and
source overlaps. Current implementations as written do work if the
destination overlaps perfectly with a source, but will unavoidably
return nonsensical results if the destination is ahead of the source.
Lock in the current behavior with tests, docs, and panics.
Note that this introduces a new panic, but if any applications run into
it we are potentially catching a security issue.
Also, expand the tests and move them outside the FIPS module per #69536
convention. (We want to minimize changes within the module boundary.)
Updates #53021
Change-Id: Ibb0875fd38da3818079e31b83b1a227b53755930
Reviewed-on: https://go-review.googlesource.com/c/go/+/622276
Reviewed-by: Roland Shoemaker <roland@golang.org>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Strings have relocations, and data relocations are bad.
Other literals are fine.
Fixes build failure in pending CL 628776.
Change-Id: I7a38bbff9776a365c5823d54c4a00e068dda5d9a
Reviewed-on: https://go-review.googlesource.com/c/go/+/628915
Reviewed-by: Cuong Manh Le <cuong.manhle.vn@gmail.com>
Reviewed-by: Robert Griesemer <gri@google.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Auto-Submit: Russ Cox <rsc@golang.org>
Excluding external test packages allows them to use
//go:embed, which requires data relocations in data.
(Obviously the external test code is testing the FIPS module,
not part of it, so this is reasonable.)
Change-Id: I4bae71320ccb5faf718c045540a9ba6dd93e378f
Reviewed-on: https://go-review.googlesource.com/c/go/+/628735
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Cherry Mui <cherryyz@google.com>
CFB and OFB are mostly unused, and not a performance target.
Updates #39365
Updates #69445
Change-Id: Ice6441e4fee2112a9e72607c63e49dbc50441ba6
Reviewed-on: https://go-review.googlesource.com/c/go/+/621957
Reviewed-by: Roland Shoemaker <roland@golang.org>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: Dmitri Shuralyov <dmitshur@google.com>
Reviewed-by: Daniel McCarney <daniel@binaryparadox.net>
