aboutsummaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
9 daysada: Adding support to defer the addition of extra formals (part 2)Javier Miranda7-30/+56
Adjust previous patch to improve the support for AI05-0151-1/08. gcc/ada/ChangeLog: * exp_attr.adb (Rewrite_Attribute_Proc_Call): Add new parameter to calls to Create_Extra_Formals. (Expand_N_Attribute_Reference): Ditto. * exp_ch3.adb (Expand_Freeze_Record_Type): Ditto. * exp_ch6.adb (Expand_Call_Helper): Ditto. * exp_disp.adb (Expand_Dispatching_Call): Ditto. * freeze.adb (Check_Itype): Ditto. (Freeze_Expression): Ditto. * sem_ch6.adb (Analyze_Subprogram_Body_Helper): Ditto. (Create_Extra_Formals): Add new formal, and use it to determine if the creation of the extra formals can be deferred. Add the new parameter to calls to Create_Extra_Formals. (Is_Unsupported_Extra_Actuals_Call): Adjust the code to improve its performance when the result is known. (Is_Unsupported_Extra_Formals_Entity): Ditto. Add new formal * sem_ch6.ads (Create_Extra_Formals): Add new formal. (Is_Unsupported_Extra_Formals_Entity): Ditto.
9 daysada: Improve error message in the case of missing indicesViljar Indus1-2/+22
When both a missing index and an out of bounds error are present then we should also mention the error on out of bounds index as it suggests removing the index that is out of bounds rather than suggesting adding indices that are also out of bounds. gcc/ada/ChangeLog: * sem_aggr.adb (Resolve_Array_Aggregate): Indicate an out of bounds index error also in the case of a missing index.
9 daysada: Create a pragma to emit a misplaced Storage_Size aspect errorViljar Indus1-2/+16
We do not emit an error on misplaced Strorage_Size error when the aspect is not applied to a task (or an access type). We instead create an attribute definition which is ignored most of the time. Create a temporary pragma for the aspect specification to emit the same misplaced aspect error as we do for pragmas. gcc/ada/ChangeLog: * sem_ch13.adb (Analyze_Aspect_Definitions): Create a temporary pragma for the non-task and access type cases.
9 daysada: Remove useless statementsRonan Desplanques1-3/+0
The initial value of fields of type Elist_Id is No_Elist, therefore the statements this patch removes were useless. gcc/ada/ChangeLog: * sem_ch7.adb (New_Private_Type): Remove useless statements.
9 daysada: Replace Preelaborable_Initialization pragmas with equivalent aspect specsDenis Mazzucato86-231/+231
All Preelaborable_Initialization pragmas in the language-defined units are replaced by the equivalent Preelaborable_Initialization aspect. gcc/ada/ChangeLog: * targparm.adb (Get_Target_Parameters): Address type declaration doesn't ends anymore with a semicolon. * libgnat/a-cdlili.ads: Replace Preelaborable_Initialization. * libgnat/a-cidlli.ads: Likewise. * libgnat/a-cihama.ads: Likewise. * libgnat/a-cihase.ads: Likewise. * libgnat/a-cimutr.ads: Likewise. * libgnat/a-ciorma.ads: Likewise. * libgnat/a-ciormu.ads: Likewise. * libgnat/a-ciorse.ads: Likewise. * libgnat/a-cohama.ads: Likewise. * libgnat/a-cohase.ads: Likewise. * libgnat/a-coinho.ads: Likewise. * libgnat/a-coinho__shared.ads: Likewise. * libgnat/a-coinve.ads: Likewise. * libgnat/a-comutr.ads: Likewise. * libgnat/a-convec.ads: Likewise. * libgnat/a-coorma.ads: Likewise. * libgnat/a-coormu.ads: Likewise. * libgnat/a-coorse.ads: Likewise. * libgnat/a-crdlli.ads: Likewise. * libgnat/a-except.ads: Likewise. * libgnat/a-finali.ads: Likewise. * libgnat/a-ngcoty.ads: Likewise. * libgnat/a-strbou.ads: Likewise. * libgnat/a-stream.ads: Likewise. * libgnat/a-strmap.ads: Likewise. * libgnat/a-strunb.ads: Likewise. * libgnat/a-strunb__shared.ads: Likewise. * libgnat/a-ststio.ads: Likewise. * libgnat/a-stwibo.ads: Likewise. * libgnat/a-stwima.ads: Likewise. * libgnat/a-stwiun.ads: Likewise. * libgnat/a-stwiun__shared.ads: Likewise. * libgnat/a-stzbou.ads: Likewise. * libgnat/a-stzmap.ads: Likewise. * libgnat/a-stzunb.ads: Likewise. * libgnat/a-stzunb__shared.ads: Likewise. * libgnat/a-tags.ads: Likewise. * libgnat/i-cstrin.ads: Likewise. * libgnat/s-stopoo.ads: Likewise. * libgnat/s-stposu.ads: Likewise. * libgnat/system-aix.ads: Likewise. * libgnat/system-darwin-arm.ads: Likewise. * libgnat/system-darwin-ppc.ads: Likewise. * libgnat/system-darwin-x86.ads: Likewise. * libgnat/system-djgpp.ads: Likewise. * libgnat/system-dragonfly-x86_64.ads: Likewise. * libgnat/system-freebsd.ads: Likewise. * libgnat/system-gnu.ads: Likewise. * libgnat/system-hpux-ia64.ads: Likewise. * libgnat/system-hpux.ads: Likewise. * libgnat/system-linux-alpha.ads: Likewise. * libgnat/system-linux-arm.ads: Likewise. * libgnat/system-linux-hppa.ads: Likewise. * libgnat/system-linux-ia64.ads: Likewise. * libgnat/system-linux-loongarch.ads: Likewise. * libgnat/system-linux-m68k.ads: Likewise. * libgnat/system-linux-mips.ads: Likewise. * libgnat/system-linux-ppc.ads: Likewise. * libgnat/system-linux-riscv.ads: Likewise. * libgnat/system-linux-s390.ads: Likewise. * libgnat/system-linux-sh4.ads: Likewise. * libgnat/system-linux-sparc.ads: Likewise. * libgnat/system-linux-x86.ads: Likewise. * libgnat/system-lynxos178-ppc.ads: Likewise. * libgnat/system-lynxos178-x86.ads: Likewise. * libgnat/system-mingw.ads: Likewise. * libgnat/system-qnx-arm.ads: Likewise. * libgnat/system-rtems.ads: Likewise. * libgnat/system-solaris-sparc.ads: Likewise. * libgnat/system-solaris-x86.ads: Likewise. * libgnat/system-vxworks-ppc-kernel.ads: Likewise. * libgnat/system-vxworks-ppc-rtp-smp.ads: Likewise. * libgnat/system-vxworks-ppc-rtp.ads: Likewise. * libgnat/system-vxworks7-aarch64-rtp-smp.ads: Likewise. * libgnat/system-vxworks7-aarch64.ads: Likewise. * libgnat/system-vxworks7-arm-rtp-smp.ads: Likewise. * libgnat/system-vxworks7-arm.ads: Likewise. * libgnat/system-vxworks7-ppc-kernel.ads: Likewise. * libgnat/system-vxworks7-ppc-rtp-smp.ads: Likewise. * libgnat/system-vxworks7-ppc64-kernel.ads: Likewise. * libgnat/system-vxworks7-ppc64-rtp-smp.ads: Likewise. * libgnat/system-vxworks7-x86-kernel.ads: Likewise. * libgnat/system-vxworks7-x86-rtp-smp.ads: Likewise. * libgnat/system-vxworks7-x86_64-kernel.ads: Likewise. * libgnat/system-vxworks7-x86_64-rtp-smp.ads: Likewise.
9 daysada: Typo fix in commentMarc Poulhiès1-1/+1
gcc/ada/ChangeLog: * s-pack.adb.tmpl: Typo fix in comment.
9 daysada: Avoid marking a ghost pragma twiceViljar Indus1-6/+6
Previously all assertion level pragmas were marked based on the current scope. Additionally some pragmas were marked afterwards based on the relevant entity e.g. Pre. This would lead to cases where a pragma was marked as both ignored and checked ghost pragma. Each pragma should get marked only once based on their individual semantics. gcc/ada/ChangeLog: * sem_prag.adb (Analyze_Pragma): Set Mark_Ghost_Code individually based on the semantics of each pragma.
9 daysada: Avoid renamed declaration becoming both ignored and checkedViljar Indus1-4/+33
gcc/ada/ChangeLog: * ghost.adb (Mark_Ghost_Declaration_Or_Body): Mark ghost entity explicitly as ignored or checked.
9 daysada: The policy of a generic can be different within a instantiationViljar Indus1-2/+4
gcc/ada/ChangeLog: * ghost.adb (Check_Ghost_Policy): ignore ghost policy changes within instantiation statements.
9 daysada: Fix ghost policy change detection for procedure callsViljar Indus1-3/+10
gcc/ada/ChangeLog: * ghost.adb (Check_Ghost_Policy): Use the policy in affect for the identifier at the current moment instead of the region around it when checking a policy change for a procedure call.
9 daysada: Fix ghost policy detection for declarationsViljar Indus1-33/+20
Declarations should take the ghost policy from the region if they do not have an explicit ghost aspect/pragam themselves. gcc/ada/ChangeLog: * ghost.adb (Mark_And_Set_Ghost_Declaration): apply the ghost policy and level from the declaration only if the declaration has an explicit ghost aspect/pragma.
9 daysada: Fix inheritance of Is_Primitive and legality check for nonoverridable ↵Denis Mazzucato2-14/+230
aspects This patch fixes the identification of inherited subprograms as primitive operations via the Is_Primitive flag. This is essential in the context of the new legality check which makes sure that, if any subprogram denoted by a nonoverridable aspect of a type T with formal or return of either type T or access T, then all denoted subprograms should be primitive operations. Note that all valid interpretations of a subprogram are denoted by the aspect under evaluation, all of these needs to be primitive then. This is a respin of eng/toolchain/gnat!2039 gcc/ada/ChangeLog: * sem_ch13.adb (Check_Nonoverridable_Aspect_Subprograms): Add the new legality check in Check_Nonoverridable_Aspect_Subprograms for nonoverridable aspects to check whether the denoted subprograms satisfy MR 13.1.1(18.4/6), otherwise we emit an error. Fix spacing. * sem_ch6.adb (New_Overloaded_Entity): Set Is_Primitive flag for inherited primitives, and filter out homonym candidates without a function specification as parents.
9 daysada: Reject pragma Attach_Handler on procedures in protected bodiesPiotr Trojanek1-2/+4
Pragma Attach_Handler and Interrupt_Handler are only legal for procedures declared in protected definition. When given for procedures declared in protected body, they were wrongly accepted and triggered an odd error message from expansion. gcc/ada/ChangeLog: * sem_prag.adb (Check_Interrupt_Or_Attach_Handler): Refine test for protected procedures; fix typo in comment.
9 daysada: Avoid performing policy checks on ignored ghost pragmasViljar Indus1-0/+7
These pragamas are already disabled if the relative ghost entity is disabled. gcc/ada/ChangeLog: * ghost.adb (Is_OK_Pragma): mark the context of ignored ghost pragmas as OK.
9 daysada: Fix resolution of assertion levelsViljar Indus2-26/+9
Unsuccessful preanalyze_and_resolve still marks identifiers as Any_Id. Override that result if an identifier matched an assertion level. gcc/ada/ChangeLog: * ghost.adb (Assertion_Level_From_Arg): Ensure that assertion level is stored as the entity for its reference. (Enables_Ghostness): Derive the result from whether or not the an argument indicated an assertion level. * tbuild.adb (Make_Assertion_Level): ensure that assertion levels have a standard scope.
9 daysada: Check restriction No_Secondary_Stack for overflow eliminationPiotr Trojanek1-0/+2
When overflow checks are eliminated using System.Bignums package, we must check if secondary stack that is necessary for this package is available. gcc/ada/ChangeLog: * checks.adb (Make_Bignum_Block): Check restriction No_Secondary_Stack.
9 daysada: Add special handling for Runtime and Static in Policy_In_EffectViljar Indus1-4/+15
When one of those levels is present then we should not look for the policy in the policy stack but rather determine the policy immidiately like we do in Check_Applicable_Policy. gcc/ada/ChangeLog: * sem_util.adb (Policy_In_Effect): Add special handling for Runtime and Static values.
9 daysada: Fix crash with global No_Tasking and async delaysDenis Mazzucato2-6/+4
When optimizations are enabled, the runtime for delays is inlined and expanded. If No_Tasking is set globally, for instance via a configuration file, then the initialization of _Master, _Chain, and _Task_Name formals is skipped for task entities. Later during expansion, these identifiers are expected but won't be found, crashing the compiler. This patch fixes Init_Formals by removing the check on Global_No_Tasking. If a No_Tasking restriction applies, then a violation error will be raised when analyzing the task type. gcc/ada/ChangeLog: * exp_ch3.adb (Init_Formals): Remove the check on Global_No_Tasking. * sem.adb: Fix typo.
9 daysada: Compiler crash on container aggregate with constant element choiceGary Dismukes1-4/+2
The compiler fails when compiling a container aggregate with an element association with a key choice that denotes a constant object. The code for getting the value of the choice was only accounting for the possibility of integer and enumeration literals, and is corrected to handle static expressions generally. gcc/ada/ChangeLog: * exp_aggr.adb (Build_Container_Aggr_Code.To_Int): Replace existing conditional expression with call to Expr_Value.
9 daysada: Add default assertion level for assertionsViljar Indus2-0/+4
To standardize the comparisson between the levels of ghost entities and levels of assertion pragmasTo standardize the comparisson between the levels of ghost entities and levels of assertion pragmas. gcc/ada/ChangeLog: * sem_ch13.adb (Analyze_Aspect_Specifications): add default assertion level to assertion aspects. * sem_prag.adb (Analyze_Pragma): Likewise.
9 daysada: exp_ch6.adb: perform less checks in CodePeer_ModeGhjuvan Lacambre1-11/+18
As explained in the comments, the CodePeer_Mode AST isn't well-formed for a GCC back-end, but is fine for a CodePeer one. The checks ensuring that the AST is well-formed thus need to be disabled. gcc/ada/ChangeLog: * exp_ch6.adb (Check_BIP_Actuals, Process_Node): Disable checks.
9 daysada: Unsigned_Base_Range aspectJavier Miranda29-80/+830
This patch adds support for a new GNAT aspect/pragma for integer type definitions to explicitly enforce the use of an unsigned base type. gcc/ada/ChangeLog: * aspects.ads (Aspect_Unsigned_Base_Range): New aspect. * checks.adb (Determine_Range): Handle types with unsigned base range aspect. (Enable_Overflow_Check): ditto (Apply_Arithmetic_Overflow_Strict): ditto * debug.adb (d_o): Document new usage. * einfo.ads (Has_Unsigned_Base_Range_Aspect): New flag. * exp_attr.adb (Expand_N_Attribute_Reference): No action since it has been already handled at this stage. * exp_ch4.adb (Expand_N_Op_Add): Generate aritmetic overflow check on unsigned base range type operands. (Expand_N_Op_Subtract): ditto (Expand_N_Op_Multiply): ditto (Expand_N_Op_Minus): ditto * gen_il-fields.ads (Has_Unsigned_Base_Range_Aspect): New flag. * gen_il-gen-gen_entities.adb (Has_Unsigned_Base_Range_Aspect): New flag. * gen_il-internals.adb (Has_Unsigned_Base_Range_Aspect): New flag. * gnat1drv.adb (Adjust_Global_Switches): Handle -gnatd_o * par-prag.adb (Pragma_Unsigned_Base_Range): No action since it will be entirely handled by the semantic analyzer. * rtsfind.ads (RE_Id): Add RE_Uns_[Add|Subtract|Multiply]_With_ Ovflo_Check * sem_attr.ads (Attribute_Unsigned_Base_Range): Added to the set of implementation defined attributes. * sem_attr.adb (Analyze_Attribute): Analyze attribute Unsigned_Base_Range. (Eval_Attribute): Evaluate attribute Unsigned_Base_Range. * sem_ch13.adb (Analyze_One_Aspect): Defer checks for this aspect to the analysis of the corresponding pragma. * sem_ch3.ads (Unsigned_Base_Range_Type_Declaration): New subprogram. * sem_ch3.adb (Build_Derived_Numeric_Type): Inherit flag Has_Unsigned_Base_Range_Aspect. (Unsigned_Base_Range_Type_Declaration): New subprogram. (Has_Pragma_Unsigned_Base_Range): New subprogram. * sem_prag.adb (Analyze_Pragma): Handle Pragma_Unsigned_Base_Range. * snames.adb-tmpl (Get_Pragma_Id): Handle Name_Unsigned_Base_Range. (Is_Pragma_Name): ditto. * snames.ads-tmpl (Name_Unsigned_Base_Range): New name. (Attribute_Unsigned_Base_Range): New attribute. (Pragma_Unsigned_Base_Range): New pragma. * libgnat/s-aridou.ads (Add_With_Ovflo_Check): New routine for Double_Uns. (Subtract_With_Ovflo_Check): ditto. (Multiply_With_Ovflo_Check): ditto. * libgnat/s-aridou.adb (Add_With_Ovflo_Check): ditto. (Subtract_With_Ovflo_Check): ditto. (Multiply_With_Ovflo_Check): ditto. * libgnat/s-arit64.ads (Uns_Add_With_Ovflo_Check64): New subprogram. (Uns_Subtract_With_Ovflo_Check64): ditto. (Uns_Multiply_With_Ovflo_Check64): ditto. * libgnat/s-arit64.adb (Uns_Add_With_Ovflo_Check64): New subprogram. (Uns_Subtract_With_Ovflo_Check64): ditto. (Uns_Multiply_With_Ovflo_Check64): ditto. * libgnat/s-arit128.ads (Uns_Add_With_Ovflo_Check128): New subprogram. (Uns_Subtract_With_Ovflo_Check128): ditto. (Uns_Multiply_With_Ovflo_Check128): ditto. * libgnat/s-arit128.adb (Uns_Add_With_Ovflo_Check128): New subprogram. (Uns_Subtract_With_Ovflo_Check128): ditto. (Uns_Multiply_With_Ovflo_Check128): ditto. * doc/gnat_rm/gnat_language_extensions.rst: Document unsigned base range. * gnat_rm.texi: Regenerate. * gnat_ugn.texi: Regenerate.
9 daysada: Multiple levels of ghost codeViljar Indus28-838/+2654
Adds support for the new language feature that allows ghost enties and assertion pragmas and aspects to be associated with a new entity called assertion level. Added support for a new pragma Assertion_Level that declares new assertion levels. This pragma consists of the level name and assertion levels it depends on. pragma Assertion_Level (L1); pragma Assertion_Level (L2); pragma Assertion_Level (L3, Depends => [L1, L2]); There are two special assertion levels that are considered to be declared in the Standard package that have unique properties. Assertion level Runtime is always considered to be Checked. Its assertion policy cannot be changed and it is considered that all other assertion levels depend on this level. Assertion level Static is always considered to be Ignored. Its assertion policy cannot be changed. All assertion levels that depend on this level can also never be activated. Aspect Ghost now supports the assertion level as a possible argument: ... with Ghost => Level; All pragmas and aspects which were considered to be valid assertion kinds for pragma Assertion_Policy now support assertion level associations. The association consists of an assertion level and a set of existing arguments. Note that you can have multiple assertion level associations in a given pragma or aspect. e.g. pragma Assert (Level1 => (Check => Expr1, Message => "Msg1", Level2 => (Check => Expr2, Message => "Msg2")); It is possible to set an explicit Assertion_Policy for those levels which can control the policies for all entities associated with those levels. Setting the policy Check for a given level means also that policy Check is set for all of the policies that it depends on. From the previous example. pragma Assertion_Policy (L3 => Check); is equivalent to: pragma Assertion_Policy (L1 => Check); pragma Assertion_Policy (L2 => Check); pragma Assertion_Policy (L3 => Check); Setting the policy Ignore for a given level means that the policy Ignore is also applied to all the levels that depend on it. e.g pragma Assertion_Policy (L2 => Ignore); is equivalent to: pragma Assertion_Policy (L2 => Ignore); pragma Assertion_Policy (L3 => Ignore); Since now ghost regions can contain other ghost regions with a different assertion policy then new rules needed to be added for those situations to ensure valid compilation. Additionally all rules where we checked for compatible assertion policies have an additional check for a compatible assertion level dependencies. Ghost entities A and B are considered assertion level dependent when * A or B does not have an associated assertion level. * Both A and B have an assertion level and either * the level of A is or depends on the level of B. * the level of B cannot be enabled (is or depends on Static) gcc/ada/ChangeLog: * atree.adb (Mark_New_Ghost_Node): Store the assertion level on the entity. * contracts.adb (Analyze_Package_Contract): Add support for multiple pragma Initial_Condition orginating from multiple assertion levels. * cstand.adb (Make_Assertion_Level_Definition): New function that creates a new Assertion_Level and adds it to the Assertion_Levels table. (Create_Standard): Add definitions for assertion levels defined in Standard. (Print_Standard): Add assertion level pragmas do the output. * exp_ch6.adb (Check_Subprogram_Variant): Add support for multiple Subprogram_Variant pragmas created by assertion levels. * einfo.ads: add info for the new nodes and attributes. * exp_prag.adb (Consequence_Error): Fix error message string corruption caused by another call to the internal strings during the call to Make_Procedure_Call_Statement. (Expand_Pragma_Initial_Condition): Ensure all ghost related attributes are copied to the new pragma. (Expand_Pragma_Loop_Variant): Likewise. (Expand_Pragma_Subprogram_Variant). Likewise. Additionally create a new Subprogram_Variant function for each pragma associated with an assertion level. * exp_util.adb (Add_DIC_Check): Ensure all ghost related attributes are copied to the new pragma. (Build_DIC_Procedure_Body): Add support for mutliple DIC pragmas created from assertion levels. * gen_il-fields.ads: (Aspect_Ghost_Assertion_Level): New field. (Original_Aspect): New field. (Original_Pragma): New field. (Pragma_Ghost_Assertion_Level): New field. (Child_Levels): New field. (Ghost_Assertion_Level): New field. (Parent_Levels): New field. * gen_il-gen-gen_entities.adb: Add Ghost_Assertion_Level field for all entities Add new E_Assertion_Level entity for storing assertion levels. * gen_il-gen-gen_nodes.adb: Add Aspect_Ghost_Assertion_Level for N_Aspect to store the assertion level associated with the aspect. Add Original_Aspect to store the original aspect where the aspect that was transformed from an aspect with an assertion level origninated from. Add Pragma_Ghost_Assertion_Level and Original_Pragma to store the same information for N_Prama nodes. * gen_il-types.ads: Add new entity kind E_Assertion_Level * ghost.adb (Assertion_Level_Error_Msg): Create constant for error messages using the same main error message. (Ghost_Policy_Error_Msg): Likewise. (Assertion_Level_To_Name): New subprogram. (Check_Valid_Ghost_Declaration): New subprogram. (Get_Ghost_Aspect): New subprogram. (Get_Ghost_Assertion_Level): New subprogram. (Ghost_Policy_In_Effect): New subprogram. (Install_Ghost_Region): New subprogram. (Mark_And_Set_Ghost_Region): New subprogram. (Mark_Ghost_Declaration_Or_Body): Add new argument for assertion levels. (Check_Ghost_Completion): Update ghost policy calculation with assertion levels. Refactor error message. (Is_OK_Statement): Add new checks for valid assertion policies and assertion levels. (Is_OK_Pragma): Refactor the calculation of valid ghost pragmas. (Check_Ghost_Policy): Make the checks ghost region based. (Check_Ghost_Context): Refactor the order of checks. (Check_Ghost_Formal_Procedure_Or_Package): Relax the checks for overriding procedures. Now only ignored subprograms cannot be overridden by checked or non-ghost subprograms. (Check_Ghost_Primitive): Relax conditions for primitve operations. Now only checked primitive subprograms are considered invalid for ignored tagged types. Add assertion level compatibility checks. (Check_Ghost_Refinement): Relax conditions for ghost refinements. Add assertion level compatibility checks for refinements. (Install_Ghost_Region): Store the current region and the assertion for that region in the ghost config. (Enables_Ghostness): Refactor implementation to support assertion levels. (Is_Subject_To_Ghost): Simplify implementation. (Mark_And_Set_Ghost_Assignment): Refactor implementation. (Mark_And_Set_Ghost_Body): Add support for assertion levels. (Mark_And_Set_Ghost_Completion): Likewise. (Mark_And_Set_Ghost_Declaration): Likwise. (Mark_And_Set_Ghost_Instantiation): Likwise. (Mark_And_Set_Ghost_Procedure_Call): Refactor implementation. (Mark_Ghost_Declaration_Or_Body): Add support for assertion levels. (Set_Ghost_Mode): Likwise. * ghost.ads (Assertion_Level_From_Arg): New subprogram. (Install_Ghost_Region): Add argument Level for assertrion levels. (Is_Assertion_Level_Dependent): New subprogram. * lib-xref.ads: Add new mapping for E_Assertion_Level entities. * opt.ads (Ghost_Config_Type): Add new members Ghost_Assertion_Mode and Current_Region to the structure. * par-prag.adb (Prag): Add new pragma name Assertion_Level. * rtsfind.adb (Load_RTU): Update the arguments for the call to Install_Ghost_Region. * sem.adb (Do_Analyze): Likewise. * sem_ch13.adb (Convert_Aspect_With_Assertion_Levels): New subprogram. (Make_Aitem_Pragma): Copy ghost mode attributes from the aspect to the pragma. (Analyze_Aspect_Specifications): Convert aspects that have an assertion level association in the aspects without the association and the original supported syntax and with the assertion level stored on the aspect node. Updated duplicate detection to avoid duplicates being called on aspects with assertion levels that orginated from the same aspect. * sem_prag.adb (Apply_Check_Policy): New subprogram. (Get_Applicable_Policy): New subprogram. (Mark_Is_Checked): New subprogram. (Mark_Is_Disabled): New subprogram. (Mark_Is_Ignored): New subprgram. (Check_Arg_Is_One_Of): Remove versions that had a specific number of arguments and replace them with a list one. (Create_Pragma_Without_Assertion_Level): New subprogram. (Assertion_Level_Pragma_Comes_From_Source): New subprogram. (Analyze_Pragma): Replace aspects that have an assertion level with aspects without them where the level is stored on the pragma node. (Abstract_State): Add support for assertion levels in ghost Abstract_State pragmas. (Assert): Update argument handling for Assert like pragmas. (Assertion_Level): Add a new section to support the analysis of pragma Assertion_Level. (Assertion_Policy): Add support for setting the policy for assertion levels. (Check): Update argument handling. Update the assertion policy application process. (Check_Policy): Add support for assertion levels. Add check_policy pragmas for assertion_level dependencis also to the stack of known Check_Policy pragmas. (Default_Initial_Condition): Reject the use of DIC with assertion levels. Update duplication checks. (Ghost): Add support for assertion levels. Fix issue where assertion levels with Ghost => False were treated as ghost. (Predicate): Update the policy handling of Ghost_Predicate. (Analyze_Refined_State_In_Decl_Part): Create a new ghost region for analyzing Refined_State. (Check_Applicable_Policy): Refactor the implementation. Break it down to Get_Applicable_Policy and Apply_Check_Policy. (Check_Kind): Removed. Replaced by Get_Applicable_Policy and Apply_Check_Policy. (Initialize): Initialize the table storing all know assertion levels. * sem_prag.ads (Find_Assertion_Level): New subprogram. (Insert_Assertion_Level): New subprogram. (Check_Applicable_Policy): Add new argument Level. (Check_Kind): Removed. Merged with Get_Applicable_Policy. (Get_Assertion_Level): New subprogram. (Is_Valid_Assertion_Level): New subprogram. * sem_util.adb (Copy_Assertion_Policy_Attributes): New function for copying the ghost related attributes from one pragma to another. (Copy_Subprogram_Spec): Additionally copy the level from the spec. (Depends_On_Level): New function for checking if one level depends on another level. (From_Same_Aspect): New function for checking whether the aspects orignate from the same original aspect. (From_Same_Pragma): New function for checking whether the pragmas originate from the same original aspect or pragma. (Get_Subprogram_Entity): Avoid crash when being called when the entity has not been set for the subprogram. (Has_Assertion_Level_Argument): New function for checking whether the aspect or a pragma has an argument that is using an assertion level association. (Policy_In_Effect): add an additional argument for the level that should be checked along with the assertion name. * sem_util.ads (Copy_Assertion_Policy_Attributes): New function. (Depends_On_Level): Likewise. (From_Same_Aspect): Likewise. (From_Same_Pragma): Likewise. (Has_Assertion_Level_Argument): Likewise. (Is_Same_Or_Depends_On_Level): Likewise. (Policy_In_Effect): Add new argument Level. * sinfo.ads: Add documentation for all the new attributes that were added to the nodes and entities. * snames.ads-tmpl: Add new entries for Name_Assertion_Level, Name_uDefault_Assertion_Level and Pragma_Assertion_Level. * stand.ads: Add new entities for the predefined assertion levels. (Standard_Level_Static): Definition for the predefined Static level that is always ignored. (Standard_Level_Runtime): Defintion for the predefined Runtime level that is always checked. (Standard_Level_Default): Definition for the implicit Default level that is given for ghost entities that were not associated with an assertion level (e.g. Ghost => True). * tbuild.adb (Make_Assertion_Level): New function for constructin an assertion level. * tbuild.ads (Make_Assertion_Level): Likewise.
9 dayslibstdc++: Replace "Tunables" link with one that does not have an anchorArsen Arsenović2-2/+2
This works around a dblatex bug, where dblatex emits invalid TeX, by placing a raw "#" in a \href. Fixes: f6ff531d907d ("libstdc++: Update link to "Tunables" in Glibc manual") libstdc++-v3/ChangeLog: * doc/xml/manual/using_exceptions.xml: Replace "Tunables" link with one that does not have an anchor. * doc/html/manual/using_exceptions.html: Regenerate.
9 daysc++: Fix mangling of _Float16 template args [PR121801]Matthias Kretz2-1/+27
Signed-off-by: Matthias Kretz <m.kretz@gsi.de> gcc/testsuite/ChangeLog: PR c++/121801 * g++.dg/abi/pr121801.C: New test. gcc/cp/ChangeLog: PR c++/121801 * mangle.cc (write_real_cst): Handle 16-bit real and assert that reals have 16 bits or a multiple of 32 bits.
10 daysx86: Enable SSE4.1 ceil/floor/trunc for -OsH.J. Lu3-5/+54
Enable SSE4.1 ceil/floor/trunc for -Os to replace a function call with roundss or roundsd by dropping the !flag_trapping_math check. gcc/ PR target/121861 * config/i386/i386.cc (ix86_optab_supported_p): Drop !flag_trapping_math check for floor_optab, ceil_optab and btrunc_optab. gcc/testsuite/ PR target/121861 * gcc.target/i386/pr121861-1a.c: New file. * gcc.target/i386/pr121861-1b.c: Likewise. Signed-off-by: H.J. Lu <hjl.tools@gmail.com>
10 daysUse vpermil{ps,pd} instead of vperm{d,q} when permutation is in-lane.liuhongt7-14/+78
gcc/ChangeLog: * config/i386/i386-expand.cc (expand_vec_perm_vpermil): Extend to handle V8SImode. * config/i386/i386.cc (avx_vpermilp_parallel): Extend to handle vector integer modes with same vector size and same component size. * config/i386/sse.md (<sse2_avx_avx512f>_vpermilp<mode><mask_name>): Ditto. (V48_AVX): New mode iterator. (ssefltmodesuffix): Extend for V16SI/V8DI/V16SF/V8DF. gcc/testsuite/ChangeLog: * gcc.target/i386/avx256_avoid_vec_perm-3.c: New test. * gcc.target/i386/avx256_avoid_vec_perm-4.c: New test. * gcc.target/i386/avx512bw-vpalignr-4.c: Adjust testcase. * gcc.target/i386/avx512vl-vpalignr-4.c: Ditto.
10 daysExclude fake cross-lane permutation from avx256_avoid_vec_perm.liuhongt4-3/+103
SLP may take a broadcast as kind of vec_perm, the patch checks the permutation index to exclude those false positive. gcc/ChangeLog: * config/i386/i386.cc (ix86_vector_costs::add_stmt_cost): Check permutation index for vec_perm, don't count it if we know it's not a cross-lane permutation. gcc/testsuite/ChangeLog: * gcc.target/i386/avx256_avoid_vec_perm.c: Adjust testcase. * gcc.target/i386/avx256_avoid_vec_perm-2.c: New test. * gcc.target/i386/avx256_avoid_vec_perm-5.c: New test.
10 daysDaily bump.GCC Administrator8-1/+366
10 dayslibstdc++: Rename template parameter of std::constant_wrapperJonathan Wakely1-3/+3
This fixes: FAIL: 17_intro/badnames.cc -std=gnu++26 (test for excess errors) libstdc++-v3/ChangeLog: * include/std/type_traits (constant_wrapper): Rename template parameter to avoid BADNAME.
10 daysc: Update TLS model after processing a TLS variableH.J. Lu1-2/+14
Set a tentative TLS model in grokvardecl and update TLS mode with the default TLS access model after a TLS variable has been fully processed if the default TLS access model is stronger. gcc/c/ PR c/107419 * c-decl.cc (c_decl_attributes): Update TLS model with the default TLS access model if the default TLS access model is stronger. (grokdeclarator): Set a tentative TLS model which will be updated by c_decl_attributes later. Signed-off-by: H.J. Lu <hjl.tools@gmail.com>
10 dayslibstdc++: Fix docs for --enable-vtable-verify [PR120698]Jonathan Wakely2-2/+2
libstdc++-v3/ChangeLog: PR libstdc++/120698 * doc/xml/manual/configure.xml: Do not claim that vtv is enabled by default. * doc/html/manual/configure.html: Regenerate.
10 dayslibstdc++: fix makefile.in for syncbuf.loNathan Myers1-1/+1
This patch fixes Makefile.in to include mention of new syncbuf.lo. libstdc++-v3/Changelog: * src/c++20/Makefile.in: Mention syncbuf.lo.
10 daysAda: Make -fdump-ada-spec deal with pointers to anonymous structureEric Botcazou1-3/+23
This is about -fdump-ada-spec not generating the definition of the structure for pointers to anonymous structure as structure elements. gcc/c-family: PR ada/121544 * c-ada-spec.cc (dump_ada_node) <POINTER_TYPE>: Dump the name of anonymous tagged pointed-to types specially. (dump_nested_type) <POINTER_TYPE>: Recurse on anonymous pointed-to types declared in the same file. Set TREE_VISITED on the underlying DECL of the field type, if any.
10 daysTestsuite: Fix spurious of ACATS-4 test cxai033Eric Botcazou1-2/+2
This tentatively applies the same tweak as in other similar cases. gcc/testsuite/ PR ada/121532 * ada/acats-4/tests/cxa/cxai033.a: Use Long_Switch_To_New_Task constant instead of Switch_To_New_Task in delay statements.
10 dayslibstdc++: Make syncbuf _S_get_mutex definition extern.Nathan Myers5-18/+56
This patch creates a global function __syncbuf_get_mutex, gated by _GLIBCXX_HAS_GTHREADS, replacing a static instantiated member _S_get_mutex used in syncbuf<> construction, and makes the global symbol visible. A static local table of 16 mutexes is shared among all specializations of syncbuf<>, chosen on construction by a hash of the wrapped streambuf's address. It detaches the implementation of _S_get_mutex from the C++20 ABI. libstdc++-v3/ChangeLog: * include/std/syncstream: (syncbuf<>::__mutex) Remove _S_get_mutex, use extern function instead. * src/c++20/syncbuf.cc: Define global __syncbuf_get_mutex. * src/c++20/Makefile.am: Mention syncbuf.cc. * src/c++20/Makefile.in: Regenerate. * config/abi/pre/gnu.ver: Mention mangled __syncbuf_get_mutex.
10 daystestsuite: Another fixup for fixed-point/bitint-1.c testXi Ruoyao1-1/+1
Besides r16-3595, there's another bug in this test: with -std=c23 the token _Sat isn't recognized as a keyword at all, thus an error massage different from the expected will be outputted. Fix it by using -std=gnu23 instead. gcc/testsuite: * gcc.dg/fixed-point/bitint-1.c (dg-options): Use -std=gnu23 instead of -std=c23.
10 daystree-optimization/121844 - IVOPTs and asm goto in latchRichard Biener2-6/+23
When there's an asm goto in the latch of a loop we may not use IP_END IVs since instantiating those would (need to) split the latch edge which in turn invalidates IP_NORMAL position handling. This is a revision of the PR107997 fix. PR tree-optimization/107997 PR tree-optimization/121844 * tree-ssa-loop-ivopts.cc (allow_ip_end_pos_p): Do not allow IP_END for latches ending with a control stmt. (create_new_iv): Do not split the latch edge, instead assert that's not necessary. * gcc.dg/torture/pr121844.c: New testcase.
10 daysRISC-V: Add pattern for vector-scalar widening floating-point addPaul-Antoine Arras12-3/+74
This pattern enables the combine pass (or late-combine, depending on the case) to merge a float_extend'ed vec_duplicate into a plus RTL instruction. Before this patch, we have four instructions, e.g.: fcvt.d.s fa0,fa0 vsetvli a5,zero,e64,m1,ta,ma vfmv.v.f v3,fa0 vfwadd.wv v1,v3,v2 After, we get only one: vfwadd.vf v1,v2,fa0 gcc/ChangeLog: * config/riscv/autovec-opt.md (*vfwadd_vf_<mode>): New pattern to combine float_extend + vec_duplicate + vfwadd.vv into vfwadd.vf. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f16.c: Add vfwadd. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf_binop.h (DEF_VF_BINOP_WIDEN_CASE_0): Fix OP. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfwadd-run-1-f16.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfwadd-run-1-f32.c: New test.
10 daysRISC-V: Adjust tt-ascalon-d8 branch costAnton Blanchard1-1/+1
If-conversion isn't being applied to this nbench code: #include <stdint.h> #define INTERNAL_FPF_PRECISION 4 typedef uint16_t u16; void ShiftMantLeft1(u16 *carry, u16 *mantissa) { int i; int new_carry; u16 accum; for(i=INTERNAL_FPF_PRECISION-1;i>=0;i--) { accum=mantissa[i]; new_carry=accum & 0x8000; accum=accum<<1; if(*carry) accum|=1; *carry=new_carry; mantissa[i]=accum; } return; } Bumping branch_cost from 3 to 4 triggers if-conversion, improving the nbench FP EMULATION result on Ascalon significantly. There's a risk that more aggressive use of conditional zero instructions will negatively impact workloads that predict well, but we haven't seen anything obvious. gcc/ChangeLog: * config/riscv/riscv.cc (tt_ascalon_d8_tune_info): Increase branch_cost from 3 to 4.
10 daystree-optimization/121830 - SLP cycle detection confused by nested cycleRichard Biener3-3/+16
The SLP reduc-index computation is confused by having an outer reduction inner loop nested cycle fed by another non-reduction nested cycle. Instead of undoing the unfortunate mixing of outer reduction inner cycles with general nested cycles the following instead distinguishes them by not setting STMT_VINFO_REDUC_DEF on the non-reduction nested cycles. PR tree-optimization/121830 * tree-vect-loop.cc (vect_analyze_scalar_cycles_1): Only set STMT_VINFO_REDUC_DEF on reductions. * tree-vect-slp.cc (vect_build_slp_tree_2): Identify reduction PHIs by a set STMT_VINFO_REDUC_DEF instead of their def type. * gcc.dg/vect/pr121830.c: New testcase.
10 daystree-optimization/121829 - bogus CFG with asm gotoRichard Biener2-15/+39
When the vectorizer removes a forwarder created earlier by split_edge it uses redirect_edge_pred for convenience and efficiency. That breaks down when the edge split is originating from an asm goto as that is a jump that needs adjustments from redirect_edge_and_branch. The following factores a simple vect_remove_forwarder handling this situation appropriately. PR tree-optimization/121829 * tree-vect-loop-manip.cc (vect_remove_forwarder): New function. (slpeel_tree_duplicate_loop_to_edge_cfg): Use it. * gcc.dg/torture/pr121829.c: New testcase.
10 daysdoc: Document the -folding option for -fdump-tree-* [PR114892]Alex Coplan1-0/+3
I noticed that the -fdump-tree-*-folding flag isn't documented in the Developer options section of invoke.texi; this patch fixes that. gcc/ChangeLog: PR tree-optimization/114892 * doc/invoke.texi (Developer Options): Document -folding option for -fdump-tree-*.
10 days[AutoFDO] Check count initialization to fix ICE with AutoFDOKugan Vivekanandarajah1-1/+2
Fix ICE with AutoFDO by adding initialization check before accessing IPA counts to avoid issues with uninitialized profile counts in self-recursive clone processing. gcc/ChangeLog: 2025-09-08 Kugan Vivekanandarajah <kvivekananda@nvidia.com> * ipa-cp.cc (gather_count_of_non_rec_edges): Check count initialization before adding to total. Signed-off-by: Kugan Vivekanandarajah <kvivekananda@nvidia.com>
10 daysRISC-V: Add pattern for vector-scalar single-width floating-point reverse subPaul-Antoine Arras17-0/+234
This pattern enables the combine pass (or late-combine, depending on the case) to merge a vec_duplicate into a minus RTL instruction. The vec_duplicate is the minuend operand. Before this patch, we have two instructions, e.g.: vfmv.v.f v2,fa0 vfsub.vv v1,v2,v1 After, we get only one: vfrsub.vf v1,v1,fa0 gcc/ChangeLog: * config/riscv/autovec-opt.md (*vfrsub_vf_<mode>): New pattern to combine vec_duplicate + vfsub.vv into vfrsub.vf. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f16.c: Add vfrsub. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf_binop_data.h: Add data for vfrsub. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfrsub-run-1-f16.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfrsub-run-1-f32.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfrsub-run-1-f64.c: New test.
10 daysRISC-V: Add pattern for vector-scalar single-width floating-point subPaul-Antoine Arras19-7/+239
This pattern enables the combine pass (or late-combine, depending on the case) to merge a vec_duplicate into a minus RTL instruction. The vec_duplicate is the subtrahend operand. Before this patch, we have two instructions, e.g.: vfmv.v.f v2,fa0 vfsub.vv v1,v1,v2 After, we get only one: vfsub.vf v1,v1,fa0 gcc/ChangeLog: * config/riscv/autovec-opt.md (*vfsub_vf_<mode>): New pattern to combine vec_duplicate + vfsub.vv into vfsub.vf. * config/riscv/vector.md (@pred_<optab><mode>_scalar): Allow VLS modes. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/autovec/vls/floating-point-sub-2.c: Adjust scan dumps. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f16.c: Add vfsub. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf_binop_data.h: Add data for vfsub. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfsub-run-1-f16.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfsub-run-1-f32.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfsub-run-1-f64.c: New test.
10 daysRISC-V: Add pattern for vector-scalar single-width floating-point addPaul-Antoine Arras20-3/+236
This pattern enables the combine pass (or late-combine, depending on the case) to merge a vec_duplicate into a plus RTL instruction. Before this patch, we have two instructions, e.g.: vfmv.v.f v2,fa0 vfadd.vv v1,v1,v2 After, we get only one: vfadd.vf v1,v1,fa0 gcc/ChangeLog: * config/riscv/autovec-opt.md (*vfadd_vf_<mode>): New pattern to combine vec_duplicate + vfadd.vv into vfadd.vf. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/autovec/vls/floating-point-add-2.c: Adjust scan dump. * gcc.target/riscv/rvv/autovec/vls/floating-point-add-3.c: Likewise. * gcc.target/riscv/rvv/autovec/vls/floating-point-sub-3.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f16.c: Add vfadd. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf_binop_data.h: Add data for vfadd. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfadd-run-1-f16.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfadd-run-1-f32.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfadd-run-1-f64.c: New test.
10 daysRISC-V: Add pattern for vector-scalar widening floating-point multiplyPaul-Antoine Arras14-6/+139
This pattern enables the combine pass (or late-combine, depending on the case) to merge a float_extend'ed vec_duplicate into a mult RTL instruction. Before this patch, we have six instructions, e.g.: fcvt.d.s fa0,fa0 vsetvli a5,zero,e64,m1,ta,ma vfmv.v.f v3,fa0 vfwcvt.f.f.v v1,v2 vsetvli zero,zero,e64,m1,ta,ma vfmul.vv v1,v3,v1 After, we get only one: vfwmul.vf v1,v2,fa0 gcc/ChangeLog: * config/riscv/autovec-opt.md (*vfwmul_vf_<mode>): New pattern to combine float_extend + vec_duplicate + vfmul.vv into vfmul.vf. * config/riscv/vector.md (*@pred_dual_widen_<optab><mode>_scalar): Swap operands to match the RTL emitted by expand, i.e. first float_extend then vec_duplicate. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f16.c: Add vfwmul. * gcc.target/riscv/rvv/autovec/vx_vf/vf-1-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-2-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-3-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-4-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf_binop.h: Add support for widening variants. * gcc.target/riscv/rvv/autovec/vx_vf/vf_binop_widen_run.h: New test helper. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfwmul-run-1-f16.c: New test. * gcc.target/riscv/rvv/autovec/vx_vf/vf_vfwmul-run-1-f32.c: New test.
10 daysRISC-V: Add patterns for vector-scalar IEEE floating-point maxPaul-Antoine Arras13-10/+32
These patterns enable the combine pass (or late-combine, depending on the case) to merge a vec_duplicate into an unspec_vfmax RTL instruction. Before this patch, we have two instructions, e.g.: vfmv.v.f v2,fa0 vfmax.vv v1,v2,v1 After, we get only one: vfmax.vf v1,v1,fa0 In some cases, it also shaves off one vsetvli. gcc/ChangeLog: * config/riscv/autovec-opt.md (*vfmin_vf_ieee_<mode>): Rename into... (*v<ieee_fmaxmin_op>_vf_<mode>): New pattern to combine vec_duplicate + vf{max,min}.vv (unspec) into vf{max,min}.vf. gcc/testsuite/ChangeLog: * gcc.target/riscv/rvv/autovec/vx_vf/vf-5-f16.c: Add vfmax. * gcc.target/riscv/rvv/autovec/vx_vf/vf-5-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-5-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-6-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-6-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-6-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-7-f16.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-7-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-7-f64.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-8-f16.c: Add vfmax. Also add missing -fno-fast-math. * gcc.target/riscv/rvv/autovec/vx_vf/vf-8-f32.c: Likewise. * gcc.target/riscv/rvv/autovec/vx_vf/vf-8-f64.c: Likewise.
10 dayslibstdc++: Fix up <ext/pointer.h> [PR121827]Jakub Jelinek2-1/+1
During the tests mentioned in https://gcc.gnu.org/pipermail/gcc-patches/2025-August/692482.html (but dunno why I haven't noticed it back in August but only when testing https://gcc.gnu.org/pipermail/gcc-patches/2025-September/694527.html ) I've noticed two ext header problems. One is that #include <ext/pointer.h> got broken with the r13-3037-g18f176d0b25591e28 change and since then is no longer self-contained, as it includes iosfwd only if _GLIBCXX_HOSTED is defined but doesn't actually include bits/c++config.h to make sure it is defined, then includes a bunch of headers which do include bits/c++config.h and finally uses in #if _GLIBCXX_HOSTED guarded code what is declared in iosfwd. The other problem is that ext/cast.h is also not a self-contained header, but that one has /** @file ext/cast.h * This is an internal header file, included by other library headers. * Do not attempt to use it directly. @headername{ext/pointer.h} */ comment, so I think we just shouldn't include it in extc++.h and let ext/pointer.h include it. 2025-09-08 Jakub Jelinek <jakub@redhat.com> PR libstdc++/121827 * include/precompiled/extc++.h: Don't include ext/cast.h which is an internal header. * include/ext/pointer.h: Include bits/c++config.h before #if _GLIBCXX_HOSTED.
generated by cgit v1.1 at 2025-09-18 17:00:16 +0000