flame
diff --git a/‎CREDITS
Lines changed: 1 addition & 0 deletions b/‎CREDITS
Lines changed: 1 addition & 0 deletions
diff --git a/‎Makefile
Lines changed: 1 addition & 0 deletions b/‎Makefile
Lines changed: 1 addition & 0 deletions
diff --git a/‎README.md
Lines changed: 25 additions & 6 deletions b/‎README.md
Lines changed: 25 additions & 6 deletions
diff --git a/‎build/config.mk.in
Lines changed: 3 additions & 0 deletions b/‎build/config.mk.in
Lines changed: 3 additions & 0 deletions
diff --git a/‎common.mk
Lines changed: 27 additions & 7 deletions b/‎common.mk
Lines changed: 27 additions & 7 deletions
diff --git a/‎configure
Lines changed: 66 additions & 16 deletions b/‎configure
Lines changed: 66 additions & 16 deletions
@@ -37,6 +37,7 @@ but many others have contributed code and feedback, including
   Roman Gareev             @gareevroman
   Richard Goldschmidt      @SuperFluffy
   Chris Goodyer
+  Alexander Grund          @Flamefire
   John Gunnels             @jagunnels          (IBM, T.J. Watson Research Center)
   Ali Emre Gülcü           @Lephar
   Jeff Hammond             @jeffhammond        (Intel)
 
@@ -1161,6 +1161,7 @@ showconfig: check-env
 	@echo "install includedir:         $(INSTALL_INCDIR)"
 	@echo "install sharedir:           $(INSTALL_SHAREDIR)"
 	@echo "debugging status:           $(DEBUG_TYPE)"
+	@echo "enable AddressSanitizer?    $(MK_ENABLE_ASAN)"
 	@echo "enabled threading model(s): $(THREADING_MODEL)"
 	@echo "enable BLAS API?            $(MK_ENABLE_BLAS)"
 	@echo "enable CBLAS API?           $(MK_ENABLE_CBLAS)"
 
@@ -3,6 +3,8 @@
 [![Build Status](https://api.travis-ci.com/flame/blis.svg?branch=master)](https://app.travis-ci.com/github/flame/blis)
 [![Build Status](https://ci.appveyor.com/api/projects/status/github/flame/blis?branch=master&svg=true)](https://ci.appveyor.com/project/shpc/blis/branch/master)
 
+[<img alt="Discord logo" title="Join us on Discord!" height="32px" src="docs/images/discord.svg" />](docs/Discord.md)
+
 Contents
 --------
 
@@ -97,6 +99,17 @@ all of which are available for free via the [edX platform](http://www.edx.org/).
 What's New
 ----------
 
+ * **Join us on Discord!** In 2021, we soft-launched our [Discord](https://discord.com/)
+server by privately inviting current and former collaborators, attendees of our BLIS
+Retreat, as well as other participants within the BLIS ecosystem. We've been thrilled by
+the results thus far, and are happy to announce that our new community is now open to
+the broader public! If you'd like to hang out with other BLIS users and developers,
+ask a question, discuss future features, or just say hello, please feel free to join us!
+We've put together a [step-by-step guide](docs/Discord.md) for creating an account and
+joining our cozy enclave. We even have a monthly "BLIS happy hour" event where people
+can casually come together for a video chat, Q&A, brainstorm session, or whatever it
+happens to unfold into!
+
  * **Addons feature now available!** Have you ever wanted to quickly extend BLIS's
 operation support or define new custom BLIS APIs for your application, but were
 unsure of how to add your source code to BLIS? Do you want to isolate your custom
@@ -417,6 +430,9 @@ If/when you have time, we *strongly* encourage you to read the detailed
 walkthrough of the build system found in our [Build System](docs/BuildSystem.md)
 guide.
 
+If you are still having trouble, you are welcome to [join us on Discord](docs/Discord.md)
+for further information and/or assistance.
+
 Example Code
 ------------
 
@@ -500,6 +516,10 @@ empirically measured performance of `gemm` on select hardware architectures
 within BLIS and other BLAS libraries when performing matrix problems where one
 or two dimensions is exceedingly small.
 
+ * **[Discord](docs/Discord.md).** This document describes how to: create an
+account on Discord (if you don't already have one); obtain a private invite
+link; and use that invite link to join our BLIS server on Discord.
+
  * **[Release Notes](docs/ReleaseNotes.md).** This document tracks a summary of
 changes included with each new version of BLIS, along with contributor credits
 for key features.
@@ -610,16 +630,15 @@ has Linux, OSX and Windows binary packages for x86_64.
 Discussion
 ----------
 
-You can keep in touch with developers and other users of the project by joining
-one of the following mailing lists:
+Most of the active discussions are now happening on our [Discord](https://discord.com/)
+server. Users and developers alike are welcome! Please see the
+[BLIS Discord guide](docs/Discord.md) for a walkthrough of how to join us.
+
+You can also still stay in touch by using either of the following mailing lists:
 
  * [blis-devel](https://groups.google.com/group/blis-devel): Please join and
 post to this mailing list if you are a BLIS developer, or if you are trying
 to use BLIS beyond simply linking to it as a BLAS library.
-**Note:** Most of the interesting discussions happen here; don't be afraid to
-join! If you would like to submit a bug report, or discuss a possible bug,
-please consider opening a [new issue](https://github.com/flame/blis/issues) on
-github.
 
  * [blis-discuss](https://groups.google.com/group/blis-discuss): Please join and
 post to this mailing list if you have general questions or feedback regarding
 
@@ -124,6 +124,9 @@ LDFLAGS_PRESET    := @ldflags_preset@
 # The level of debugging info to generate.
 DEBUG_TYPE        := @debug_type@
 
+# Whether to compile and link the AddressSanitizer library.
+MK_ENABLE_ASAN    := @enable_asan@
+
 # Whether operating system support was requested via --enable-system.
 ENABLE_SYSTEM     := @enable_system@
 
 
@@ -118,6 +118,7 @@ get-noopt-cxxflags-for   = $(strip $(CFLAGS_PRESET) \
 get-refinit-cflags-for   = $(strip $(call load-var-for,COPTFLAGS,$(1)) \
                                    $(call get-noopt-cflags-for,$(1)) \
                                    -DBLIS_CNAME=$(1) \
+                                   $(BUILD_ASANFLAGS) \
                                    $(BUILD_CPPFLAGS) \
                                    $(BUILD_SYMFLAGS) \
                                    -DBLIS_IN_REF_KERNEL=1 \
@@ -129,6 +130,7 @@ get-refkern-cflags-for   = $(strip $(call load-var-for,CROPTFLAGS,$(1)) \
                                    $(call get-noopt-cflags-for,$(1)) \
                                    $(COMPSIMDFLAGS) \
                                    -DBLIS_CNAME=$(1) \
+                                   $(BUILD_ASANFLAGS) \
                                    $(BUILD_CPPFLAGS) \
                                    $(BUILD_SYMFLAGS) \
                                    -DBLIS_IN_REF_KERNEL=1 \
@@ -137,12 +139,14 @@ get-refkern-cflags-for   = $(strip $(call load-var-for,CROPTFLAGS,$(1)) \
 
 get-config-cflags-for    = $(strip $(call load-var-for,COPTFLAGS,$(1)) \
                                    $(call get-noopt-cflags-for,$(1)) \
+                                   $(BUILD_ASANFLAGS) \
                                    $(BUILD_CPPFLAGS) \
                                    $(BUILD_SYMFLAGS) \
                             )
 
 get-frame-cflags-for     = $(strip $(call load-var-for,COPTFLAGS,$(1)) \
                                    $(call get-noopt-cflags-for,$(1)) \
+                                   $(BUILD_ASANFLAGS) \
                                    $(BUILD_CPPFLAGS) \
                                    $(BUILD_SYMFLAGS) \
                             )
@@ -201,11 +205,14 @@ get-sandbox-cxxflags-for = $(strip $(call load-var-for,COPTFLAGS,$(1)) \
 # Define a separate function that will return appropriate flags for use by
 # applications that want to use the same basic flags as those used when BLIS
 # was compiled. (NOTE: This is the same as the $(get-frame-cflags-for ...)
-# function, except that it omits two variables that contain flags exclusively
-# for use when BLIS is being compiled/built: BUILD_CPPFLAGS, which contains a
-# cpp macro that confirms that BLIS is being built; and BUILD_SYMFLAGS, which
-# contains symbol export flags that are only needed when a shared library is
-# being compiled/linked.)
+# function, except that it omits a few variables that contain flags exclusively
+# for use when BLIS is being compiled/built:
+# - BUILD_CPPFLAGS, which contains a cpp macro that confirms that BLIS
+#   is being built;
+# - BUILD_SYMFLAGS, which contains symbol export flags that are only
+#   needed when a shared library is being compiled/linked; and
+# - BUILD_ASANFLAGS, which contains a flag that causes the compiler to
+#   insert instrumentation for memory error detection.
 get-user-cflags-for      = $(strip $(call load-var-for,COPTFLAGS,$(1)) \
                                    $(call get-noopt-cflags-for,$(1)) \
                             )
@@ -563,6 +570,11 @@ ifeq ($(DEBUG_TYPE),sde)
 LDFLAGS    := $(filter-out $(LIBMEMKIND),$(LDFLAGS))
 endif
 
+# If AddressSanitizer is enabled, add the compiler flag to LDFLAGS.
+ifeq ($(MK_ENABLE_ASAN),yes)
+LDFLAGS    += -fsanitize=address
+endif
+
 # Specify the shared library's 'soname' field.
 # NOTE: The flag for creating shared objects is different for Linux and OS X.
 ifeq ($(OS_NAME),Darwin)
@@ -808,11 +820,19 @@ $(foreach c, $(CONFIG_LIST_FAM), $(eval $(call append-var-for,CXXLANGFLAGS,$(c))
 CPPROCFLAGS := -D_POSIX_C_SOURCE=200112L
 $(foreach c, $(CONFIG_LIST_FAM), $(eval $(call append-var-for,CPPROCFLAGS,$(c))))
 
+# --- AddressSanitizer flags ---
+
+ifeq ($(MK_ENABLE_ASAN),yes)
+BUILD_ASANFLAGS := -fsanitize=address
+else
+BUILD_ASANFLAGS :=
+endif
+
 # --- Threading flags ---
 
 # NOTE: We don't have to explicitly omit -pthread when --disable-system is given
-# since that option forces --enable-threading=none, and thus -pthread never gets
-# added to begin with.
+# since that option forces --enable-threading=single, and thus -pthread never
+# gets added to begin with.
 
 CTHREADFLAGS :=
 
 
@@ -224,12 +224,22 @@ print_usage()
 	echo " "
 	echo "   --enable-mem-tracing, --disable-mem-tracing"
 	echo " "
-	echo "                 Enable (disable by default) output to stdout that traces"
+	echo "                 Enable (disabled by default) output to stdout that traces"
 	echo "                 the allocation and freeing of memory, including the names"
 	echo "                 of the functions that triggered the allocation/freeing."
 	echo "                 Enabling this option WILL NEGATIVELY IMPACT PERFORMANCE."
 	echo "                 Please use only for informational/debugging purposes."
 	echo " "
+	echo "   --enable-asan, --disable-asan"
+	echo " "
+	echo "                 Enable (disabled by default) compiling and linking BLIS"
+	echo "                 framework code with the AddressSanitizer (ASan) library."
+	echo "                 Optimized kernels are NOT compiled with ASan support due"
+	echo "                 to limitations of register assignment in inline assembly."
+	echo "                 WARNING: ENABLING THIS OPTION WILL NEGATIVELY IMPACT"
+	echo "                 PERFORMANCE. Please use only for informational/debugging"
+	echo "                 purposes."
+	echo " "
 	echo "   -i SIZE, --int-size=SIZE"
 	echo " "
 	echo "                 Set the size (in bits) of internal BLIS integers and"
@@ -1325,6 +1335,17 @@ blacklistbu_add()
 	fi
 }
 
+blacklistos_add()
+{
+	# Check whether we've already blacklisted the given sub-config so
+	# we don't output redundant messages.
+	if [ $(is_in_list "$1" "${config_blist}") == "false" ]; then
+
+		echowarn "The operating system does not support building '$1'; adding to blacklist."
+		config_blist="${config_blist} $1"
+	fi
+}
+
 blacklist_init()
 {
 	config_blist=""
@@ -1979,6 +2000,13 @@ check_assembler()
 	fi
 }
 
+check_os()
+{
+	if [[ "$(uname -s)" == "Darwin" && "$(uname -m)" == "arm64" ]]; then
+		blacklistos_add "armsve"
+	fi
+}
+
 try_assemble()
 {
 	local cc cflags asm_src asm_base asm_bin rval
@@ -2451,6 +2479,9 @@ main()
 	debug_type=''
 	debug_flag=''
 
+	# A flag indicating whether AddressSanitizer should be used.
+	enable_asan='no'
+
 	# The system flag.
 	enable_system='yes'
 
@@ -2576,6 +2607,12 @@ main()
 						disable-debug)
 							debug_flag=0
 							;;
+						enable-asan)
+							enable_asan='yes'
+							;;
+						disable-asan)
+							enable_asan='no'
+							;;
 						enable-verbose-make)
 							enable_verbose='yes'
 							;;
@@ -2867,6 +2904,9 @@ main()
 	get_binutils_version
 	check_assembler
 
+    # Check if there is any incompatibility due to the operating system.
+	check_os
+
 	# Remove duplicates and whitespace from the blacklist.
 	blacklist_cleanup
 
@@ -3357,6 +3397,20 @@ main()
 		echo "${script_name}: no preset LDFLAGS detected."
 	fi
 
+	# Check if the verbose make flag was specified.
+	if [ "x${enable_verbose}" = "xyes" ]; then
+		echo "${script_name}: enabling verbose make output. (disable with 'make V=0'.)"
+	else
+		echo "${script_name}: disabling verbose make output. (enable with 'make V=1'.)"
+	fi
+
+	# Check if the ARG_MAX hack was requested.
+	if [ "x${enable_arg_max_hack}" = "xyes" ]; then
+		echo "${script_name}: enabling ARG_MAX hack."
+	else
+		echo "${script_name}: disabling ARG_MAX hack."
+	fi
+
 	# Check if the debug flag was specified.
 	if [ -n "${debug_flag}" ]; then
 		if [ "x${debug_type}" = "xopt" ]; then
@@ -3373,29 +3427,24 @@ main()
 		echo "${script_name}: debug symbols disabled."
 	fi
 
-	# Check if the verbose make flag was specified.
-	if [ "x${enable_verbose}" = "xyes" ]; then
-		echo "${script_name}: enabling verbose make output. (disable with 'make V=0'.)"
-	else
-		echo "${script_name}: disabling verbose make output. (enable with 'make V=1'.)"
-	fi
-
-	# Check if the ARG_MAX hack was requested.
-	if [ "x${enable_arg_max_hack}" = "xyes" ]; then
-		echo "${script_name}: enabling ARG_MAX hack."
+	# Check if the AddressSanitizer flag was specified.
+	if [ "x${enable_asan}" = "xyes" ]; then
+		echo "${script_name}: enabling AddressSanitizer support (except for optimized kernels)."
 	else
-		echo "${script_name}: disabling ARG_MAX hack."
+        enable_asan='no'
+		echo "${script_name}: AddressSanitizer support disabled."
 	fi
 
-	enable_shared_01=1
 	# Check if the static lib flag was specified.
 	if   [ "x${enable_static}" = "xyes" -a "x${enable_shared}" = "xyes" ]; then
 		echo "${script_name}: building BLIS as both static and shared libraries."
+		enable_shared_01=1
+	elif [ "x${enable_static}" = "xno"  -a "x${enable_shared}" = "xyes" ]; then
+		echo "${script_name}: building BLIS as a shared library (static library disabled)."
+		enable_shared_01=1
 	elif [ "x${enable_static}" = "xyes" -a "x${enable_shared}" = "xno"  ]; then
 		echo "${script_name}: building BLIS as a static library (shared library disabled)."
 		enable_shared_01=0
-	elif [ "x${enable_static}" = "xno"  -a "x${enable_shared}" = "xyes" ]; then
-		echo "${script_name}: building BLIS as a shared library (static library disabled)."
 	else
 		echo "${script_name}: Both static and shared libraries were disabled."
 		echo "${script_name}: *** Please enable one (or both) to continue."
@@ -3917,7 +3966,7 @@ main()
 	# Create a #define for the configuration family (config_name).
 	uconf=$(echo ${config_name} | tr '[:lower:]' '[:upper:]')
 	config_name_define="#define BLIS_FAMILY_${uconf}\n"
-	
+
 	# Create a list of #defines, one for each configuration in config_list.
 	config_list_defines=""
 	for conf in ${config_list}; do
@@ -4012,6 +4061,7 @@ main()
 		| sed -e "s/@libpthread@/${libpthread_esc}/g" \
 		| sed -e "s/@cflags_preset@/${cflags_preset_esc}/g" \
 		| sed -e "s/@ldflags_preset@/${ldflags_preset_esc}/g" \
+		| sed -e "s/@enable_asan@/${enable_asan}/g" \
 		| sed -e "s/@debug_type@/${debug_type}/g" \
 		| sed -e "s/@enable_system@/${enable_system}/g" \
 		| sed -e "s/@threading_model@/${threading_model}/g" \