Upon re-consideration in commit 569d50f1c6 ("x86: further refine SSE
check (SSE4a, SHA, GFNI)") I went too far: Mixing of SSE and AVX insns
doesn't suffer as bad a penalty on AMD CPUs as on Intel ones. SSE4a
being an AMD-only extension, it shouldn't be part of the ISA extensions
set for which the diagnostic may get issued. Undo that part.
2020-06-15 Max Filippov <jcmvbkbc@gmail.com>
bfd/
* elf32-xtensa.c (XSHAL_ABI, XTHAL_ABI_UNDEFINED)
(XTHAL_ABI_WINDOWED, XTHAL_ABI_CALL0): New macros.
(elf32xtensa_abi): New global variable.
(xtensa_abi_choice): New function.
(elf_xtensa_create_plt_entry): Use xtensa_abi_choice instead of
XSHAL_ABI to select PLT code.
gas/
* config/tc-xtensa.c (XTHAL_ABI_WINDOWED, XTHAL_ABI_CALL0): New
macros.
(elf32xtensa_abi): New declaration.
(option_abi_windowed, option_abi_call0): New enum constants.
(md_longopts): Add entries for --abi-windowed and --abi-call0.
(md_parse_option): Add handlers for --abi-windowed and
--abi-call0.
(xtensa_add_config_info): Use xtensa_abi_choice instead of
XSHAL_ABI to format ABI tag.
* doc/as.texi (Target Xtensa options): Add --abi-windowed and
--abi-call0 to the list of options.
* doc/c-xtensa.texi: Add description for options --abi-windowed
and --abi-call0.
* testsuite/gas/xtensa/abi-call0.d: New test definition.
* testsuite/gas/xtensa/abi-windowed.d: New test definition.
* testsuite/gas/xtensa/abi.s: New test source.
include/
* elf/xtensa.h (xtensa_abi_choice): New declaration.
ld/
* emultempl/xtensaelf.em (XSHAL_ABI): Remove macro definition.
(XTHAL_ABI_UNDEFINED, XTHAL_ABI_WINDOWED, XTHAL_ABI_CALL0): New
macros.
(elf32xtensa_abi): New declaration.
(xt_config_info_unpack_and_check): Set elf32xtensa_abi if it is
undefined. Use xtensa_abi_choice instead of XSHAL_ABI to test
ABI tag consistency.
(xtensa_add_config_info): Use xtensa_abi_choice instead of
XSHAL_ABI to format ABI tag.
(PARSE_AND_LIST_PROLOGUE): Define OPTION_ABI_WINDOWED,
OPTION_ABI_CALL0 and declare elf32xtensa_abi.
(PARSE_AND_LIST_LONGOPTS): Add entries for --abi-windowed and
--abi-call0.
(PARSE_AND_LIST_OPTIONS): Add help text for --abi-windowed and
--abi-call0.
(PARSE_AND_LIST_ARGS_CASES): Add handlers for --abi-windowed and
--abi-call0.
* ld.texi: Add description for options --abi-windowed and
--abi-call0.
There is a conflict between v1.9 and v1.9.1 - CSR MISA address. MISA is
0xf10 in v1.9, but change to 0x301 in v1.9.1. The change made MISA writable,
but may also cause risk of compatibility. Binutils already support the
-mpriv-spec options and ELF priv attributes, which can used to choose what
privileged spec you want, and then give a correponding CSR name and address
to use. But Gdb and other tools don't have the simialr mechanism for now.
However, there are two things can be confirmed,
1. If we don't have a way to control the priv specs, then the changes, like
MISA, will cause risk and hard to maintain.
2. We get the guarantee that the CSR address won't be reused in the future
specs, even if it is dropped.
I'm not sure if Gdb needs to care about the priv spec versions, it is still
discussing. But drop the priv spec v1.9, and make sure that we won't reuse
the CSR address is a useful solution for now. Also, we might drop the v1.9.1
in a year or two. After that, specs above v1.10 should be compatible anyway.
gas/
* testsuite/gas/riscv/priv-reg-fail-version-1p9.d: Removed.
* testsuite/gas/riscv/priv-reg-fail-version-1p9.l: Likewise.
* testsuite/gas/riscv/priv-reg-version-1p9.d: Likewise.
include/
* opcode/riscv-opc.h: Update the defined versions of CSR from
PRIV_SPEC_CLASS_1P9 to PRIV_SPEC_CLASS_1P9P1. Also, drop the
MISA DECLARE_CSR_ALIAS since it's aborted version is v1.9.
* opcode/riscv.h (enum riscv_priv_spec_class): Remove
PRIV_SPEC_CLASS_1P9.
opcodes/
* riscv-opc.c (priv_specs): Remove v1.9 and PRIV_SPEC_CLASS_1P9.
All of data size, rep, lock, and rex prefixes are invalid with VEX- and
alike encoded insns. Make sure they get printed explicitly in all cases,
to signal the anomaly. With this, do away with "rex_ignored" - if there
is a rex prefix, we want to print it anyway for VEX etc (and there's
nothing "ignored" about it in the first place - such an instruction will
raise #UD).
Various AVX insns utilizing the X macro fail to reject F3/F2 embedded
prefix encodings. As the PREFIX_OPCODE attribute wasn't used by any
non-legacy-encoded insns so far, re-use it to achieve the intended
effect.
An unwanted side effect of 5990e377e5 ("x86-64: Intel64 adjustments
for insns dealing with far pointers") was that with -mintel64 LCALL and
LJMP would now default to 64-bit operand size. Since 64-bit far branches
aren't portable, the default operand size should still be 32-bit.
However, since the 64-bit variant is permitted, an ambiguous operand
warning should be issued.
As to the actual code change, please note that the conditional
surrounding the switch() that gets adjusted covers several cases which
are of no interest to or benign in 64-bit mode, hence the new
conditional added can be quite a bit less involved.
{vex}, {vex3}, and {evex} are mandatory prefixes, and hence should not
be randomly ignored. Fix this for insns without operands as well as for
insns referencing the high 16 [XYZ]MM registers. To achieve the former,
re-purpose VEX_check_operands(), renaming it to VEX_check_encoding() and
moving its only operand check to check_VecOperands().
This involves fixing a testcase relying on {vex2} to get ignored.
With the introduction of what right now is the very first conditional
in the function's loop (commit dc821c5f9a ["x86: replace Reg8, Reg16,
Reg32, and Reg64"]), the last if() in the same loop has become
pointless - retain just its body.
First of all, these registers have never been available on any 64-bit
CPU, and hence should not be recognized in 64-bit mode. But even before
that they had already disappeared - also don't recognize them when 586
or 686 architectures were explicitly set.
While putting together the previous patch I noticed that this test,
forever since its introduction, was dead. Update it so it will pass,
and enable it.
In 0e0eea7820 ("x86: x87-related adjustments") I screwed up CFI
directives with FPU support disabled, by moving the conditional there
across a check of "allow_pseudo_reg". Add the missing check.
Register aliases (created e.g. via .set) check their target register at
the time of creation of the alias. While this makes sense, it's not
enough: The underlying register must also be "visible" at the time of
use. Wrong use of such aliases would lead to internal errors in e.g.
add_prefix() or build_modrm_byte().
Split the checking part of parse_real_register() into a new helper
function and use it also from the latter part of parse_register() (at
the same time replacing a minor open coded part of it).
Since parse_register() returning NULL already has a meaning, a fake new
"bad register" indicator gets added, which all callers need to check
for.
These relocations should have had REL in their names, to reflect the
fact that they are pc-relative. Fix that now by adding _PCREL.
I've added some back-compatibility code to support anyone using
.reloc with the old relocations.
include/
* elf/ppc64.h (elf_ppc64_reloc_type): Rename
R_PPC64_GOT_TLSGD34 to R_PPC64_GOT_TLSGD_PCREL34,
R_PPC64_GOT_TLSLD34 to R_PPC64_GOT_TLSLD_PCREL34,
R_PPC64_GOT_TPREL34 to R_PPC64_GOT_TPREL_PCREL34, and
R_PPC64_GOT_DTPREL34 to R_PPC64_GOT_DTPREL_PCREL34.
bfd/
* reloc.c: Rename
BFD_RELOC_PPC64_GOT_TLSGD34 to BFD_RELOC_PPC64_GOT_TLSGD_PCREL34,
BFD_RELOC_PPC64_GOT_TLSLD34 to BFD_RELOC_PPC64_GOT_TLSLD_PCREL34,
BFD_RELOC_PPC64_GOT_TPREL34 to BFD_RELOC_PPC64_GOT_TPREL_PCREL34,
BFD_RELOC_PPC64_GOT_DTPREL34 to BFD_RELOC_PPC64_GOT_DTPREL_PCREL34.
* elf64-ppc.c: Update throughout for reloc renaming.
(ppc64_elf_reloc_name_lookup): Handle old reloc names.
* libbfd.h: Regenerate.
* bfd-in2.h: Regenerate.
gas/
* config/tc-ppc.c: Update throughout for reloc renaming.
elfcpp/
* powerpc.h: Rename
R_PPC64_GOT_TLSGD34 to R_PPC64_GOT_TLSGD_PCREL34,
R_PPC64_GOT_TLSLD34 to R_PPC64_GOT_TLSLD_PCREL34,
R_PPC64_GOT_TPREL34 to R_PPC64_GOT_TPREL_PCREL34, and
R_PPC64_GOT_DTPREL34 to R_PPC64_GOT_DTPREL_PCREL34.
gold/
* powerpc.cc: Update throughout for reloc renaming.
The GAS struct frag ends with a field `fr_literal' whose purpose is to
mark the begining of the frag's data:
struct frag {
...
/* Data begins here. */
char fr_literal[1];
};
The code in gas/config/tc-bpf.c recently committed:
where = fixP->fx_frag->fr_literal + fixP->fx_where;
where[1] = target_big_endian ? 0x01 : 0x10;
triggers the stringop-overflow warning in GCC 10+, since the compiler
assumes the size of the modified buffer is 1 byte. This patch
slightly modifies the code to make tc-bpf.c buildable with GCC 10+.
2020-06-05 Jose E. Marchesi <jose.marchesi@oracle.com>
* config/tc-bpf.c (md_apply_fix): Avoid GCC 10 warning
stringop-overflow.
gas/
* config/tc-riscv.c (explicit_csr): New static boolean.
Used to indicate CSR are explictly used.
(riscv_ip): Set explicit_csr to TRUE if any CSR is used.
(riscv_write_out_attrs): If we already have set elf priv
attributes, then generate them. Otherwise, don't generate
them when no CSR are used.
* testsuite/gas/riscv/attribute-01.d: Remove the priv attributes.
* testsuite/gas/riscv/attribute-02.d: Likewise.
* testsuite/gas/riscv/attribute-03.d: Likewise.
* testsuite/gas/riscv/attribute-04.d: Likewise.
* testsuite/gas/riscv/attribute-05.d: Likewise.
* testsuite/gas/riscv/attribute-06.d: Likewise.
* testsuite/gas/riscv/attribute-07.d: Likewise.
* testsuite/gas/riscv/attribute-08.d: Likewise.
* testsuite/gas/riscv/attribute-09.d: Likewise.
* testsuite/gas/riscv/attribute-10.d: Likewise.
* testsuite/gas/riscv/attribute-unknown.d: Likewise.
* testsuite/gas/riscv/attribute-11.s: New testcase.
* testsuite/gas/riscv/attribute-11.d: New testcase. The CSR is
used, so we should output the ELF priv attributes.
* testsuite/gas/riscv/attribute-12.d: New testcase. The CSR is
used, so output the priv attributes according to the -mpriv-spec.
* testsuite/gas/riscv/attribute-13.d: New testcase. The CSR isn't
used, so ignore the -mpriv-spec setting.
ld/
* testsuite/ld-riscv-elf/attr-merge-arch-01.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-arch-02.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-arch-03.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-stack-align.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-strict-align-01.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-strict-align-02.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-strict-align-03.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-strict-align-04.d: Likewise.
* testsuite/ld-riscv-elf/attr-merge-strict-align-05.d: Likewise.
* testsuite/ld-riscv-elf/call-relax.d: Add -mno-arch-attr.
Fix ip2k-elf and xstormy16-elf build due to
commit e9bffec9af
Author: Jose E. Marchesi <jose.marchesi@oracle.com>
Date: Thu Jun 4 16:15:53 2020 +0200
opcodes: discriminate endianness and insn-endianness in CGEN ports
* config/tc-ip2k. (ip2k_apply_fix): Pass endianness to
cgen_get_insn_value.
* config/tc-xstormy16.c (xstormy16_md_apply_fix): Pass
endianness to cgen_get_insn_value and cgen_put_insn_value.
cpu/ChangeLog:
2020-06-02 Jose E. Marchesi <jose.marchesi@oracle.com>
* bpf.cpu (define-bpf-isa): Set base-insn-bitsize to 64.
* bpf.opc (bpf_print_insn): Do not set endian_code here.
gas/ChangeLog:
2020-06-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* config/tc-bpf.c (md_begin): Pass CGEN_CPU_OPEN_INSN_ENDIAN to
bpf_cgen_cpu_open.
(md_assemble): Remove no longer needed hack.
opcodes/ChangeLog:
2020-06-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* disassemble.c (disassemble_init_for_target): Set endian_code for
bpf targets.
* bpf-desc.c: Regenerate.
* bpf-opc.c: Likewise.
* bpf-dis.c: Likewise.
The CGEN support code in opcodes accesses instruction contents using a
couple of functions defined in cgen-opc.c: cgen_get_insn_value and
cgen_put_insn_value. These functions use the "instruction endianness"
in the CPU description to order the read/written bytes.
The process of writing an instruction to the object file is:
a) cgen_put_insn_value ;; Writes out the opcodes.
b) ARCH_cgen_insert_operand
insert_normal
insert_1
cgen_put_insn_value ;; Writes out the bytes of the
;; operand.
Likewise, the process of reading an instruction from the object file
is:
a) cgen_get_insn_value ;; Reads the opcodes.
b) ARCH_cgen_extract_operand
extract_normal
extract_1
cgen_get_insn_value ;; Reads in the bytes of the
;; operand.
As can be seen above, cgen_{get,put}_insn_value are used to both
process the instruction opcodes (the constant fields conforming the
base instruction) and also the values of the instruction operands,
such as immediates.
This is problematic for architectures in which the endianness of
instructions is different to the endianness of data. An example is
BPF, where instructions are always encoded big-endian but the data may
be either big or little.
This patch changes the cgen_{get,put}_insn_value functions in order to
get an extra argument with the endianness to use, and adapts the
existin callers to these functions in order to provide cd->endian or
cd->insn_endian, whatever appropriate. Callers like extract_1 and
insert_1 pass cd->endian (since they are reading/writing operand
values) while callers reading/writing the base instruction pass
cd->insn_endian instead.
A few little adjustments have been needed in some existing CGEN based
ports:
* The BPF assembler uses cgen_put_insn_value. It has been adapted to
pass the new endian argument.
* The mep port has code in mep.opc that uses cgen_{get,put}_insn_value.
It has been adapted to pass the new endianargument. Ditto for a
call in the assembler.
Tested with --enable-targets=all.
Regested in all supported targets.
No regressions.
include/ChangeLog:
2020-06-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* opcode/cgen.h: Get an `endian' argument in both
cgen_get_insn_value and cgen_put_insn_value.
opcodes/ChangeLog:
2020-06-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* cgen-opc.c (cgen_get_insn_value): Get an `endian' argument.
(cgen_put_insn_value): Likewise.
(cgen_lookup_insn): Pass endianness to cgen_{get,put}_insn_value.
* cgen-dis.in (print_insn): Likewise.
* cgen-ibld.in (insert_1): Likewise.
(insert_1): Likewise.
(insert_insn_normal): Likewise.
(extract_1): Likewise.
* bpf-dis.c: Regenerate.
* bpf-ibld.c: Likewise.
* bpf-ibld.c: Likewise.
* cgen-dis.in: Likewise.
* cgen-ibld.in: Likewise.
* cgen-opc.c: Likewise.
* epiphany-dis.c: Likewise.
* epiphany-ibld.c: Likewise.
* fr30-dis.c: Likewise.
* fr30-ibld.c: Likewise.
* frv-dis.c: Likewise.
* frv-ibld.c: Likewise.
* ip2k-dis.c: Likewise.
* ip2k-ibld.c: Likewise.
* iq2000-dis.c: Likewise.
* iq2000-ibld.c: Likewise.
* lm32-dis.c: Likewise.
* lm32-ibld.c: Likewise.
* m32c-dis.c: Likewise.
* m32c-ibld.c: Likewise.
* m32r-dis.c: Likewise.
* m32r-ibld.c: Likewise.
* mep-dis.c: Likewise.
* mep-ibld.c: Likewise.
* mt-dis.c: Likewise.
* mt-ibld.c: Likewise.
* or1k-dis.c: Likewise.
* or1k-ibld.c: Likewise.
* xc16x-dis.c: Likewise.
* xc16x-ibld.c: Likewise.
* xstormy16-dis.c: Likewise.
* xstormy16-ibld.c: Likewise.
gas/ChangeLog:
2020-06-04 Jose E. Marchesi <jose.marchesi@oracle.com>
* cgen.c (gas_cgen_finish_insn): Pass the endianness to
cgen_put_insn_value.
(gas_cgen_md_apply_fix): Likewise.
(gas_cgen_md_apply_fix): Likewise.
* config/tc-bpf.c (md_apply_fix): Pass data endianness to
cgen_put_insn_value.
* config/tc-mep.c (mep_check_ivc2_scheduling): Pass endianness to
cgen_put_insn_value.
cpu/ChangeLog:
2020-06-02 Jose E. Marchesi <jose.marchesi@oracle.com>
* mep.opc (print_slot_insn): Pass the insn endianness to
cgen_get_insn_value.
This fixes some minor bugs in the docs for the .insn directive pointed out
by Frédéric Pétrot, and I added a few more cleanups since I was changing
the docs.
gas/
PR 26051
* doc/c-riscv.texi (RISC-V-Formats): Add missing I format using
simm12(rs1). Correct S format to use simm12(rs1). Drop SB and B
formats using simm12(rs1). Correct SB and B to use rs1 and rs2.
Move B before SB. Move J before UJ.
This patch fixes internal errors in (at least) arm and aarch64 GAS
when assembling code that attempts a negative .org. The bug appears
to be a regression introduced in binutils-2.29 by commit 9875b36538.
* write.c (relax_segment): Fix handling of negative offset when
relaxing an rs_org frag.
* testsuite/gas/aarch64/org-neg.d: New test.
* testsuite/gas/aarch64/org-neg.l: Error output for test.
* testsuite/gas/aarch64/org-neg.s: Input for test.
* testsuite/gas/arm/org-neg.d: New test.
* testsuite/gas/arm/org-neg.l: Error output for test.
* testsuite/gas/arm/org-neg.s: Input for test.
These failures were caused by the PDP11's mix of little-endian octets
in shorts but shorts in big endian order for long or quad so regexps
did not match. Also tests used addresses as values in .long which
required BRD_RELOC_32 that was not implemented.
* gas/config/tc-pdp11.c (md_number_to_chars): Implement .quad
* gas/testsuite/gas/all/gas.exp: Select alternate test scripts for
pdp11, skip octa test completely.
* gas/testsuite/gas/all/eqv-dot-pdp11.s: Identical to eqv-dot.s
* gas/testsuite/gas/all/eqv-dot-pdp11.d: Match different octet order.
* gas/testsuite/gas/all/cond-pdp11.l: Match different octet order.
* bfd/pdp11.c: Implement BRD_RELOC_32 to relocate the low 16 bits of
addreses in .long (used in testsuites) and .stab values.
PR gas/26001
* gas/config/tc-pdp11.c (parse_reg): Distinguish register names from
symbols that begin with a register name.
* gas/testsuite/gas/pdp11/pdp11.exp: Add test of such symbols.
* gas/testsuite/gas/pdp11/pr26001.s: Likewise.
* gas/testsuite/gas/pdp11/pr26001.d: Likewise.
The next pointer of struct riscv_csr_extra was not always initilized
to NULL or a valid pointer, causing the assembler to attempt to read
through an uninitialized pointer on startup.
gas/ChangeLog:
* config/tc-riscv.c (riscv_init_csr_hash): NULL initilize next
pointer when creating struct riscv_csr_extra.
Accept vector alignment hints on z13 although they are ignored there.
The advantage is that any binary compiled for architecture level z13 may
run on z14 or later and benefit from vector alignment hints.
gas/ChangeLog:
2020-05-18 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
* testsuite/gas/s390/zarch-z13.d: Add regexp checks for vector
load/store instruction variants with alignment hints.
* testsuite/gas/s390/zarch-z13.s: Emit new vector load/store
instruction variants with alignment hints.
opcodes/ChangeLog:
2020-05-18 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
* s390-opc.txt: Relocate vector load/store instructions with
additional alignment parameter and change architecture level
constraint from z14 to z13.
opcode/crx.h has
typedef enum
{
...
MAX_REG
}
reg;
typedef enum
{
c0 = MAX_REG,
}
copreg;
tc-crx.c has
static int
getreg_image (reg r)
{
...
/* Check whether the register is in registers table. */
if (r < MAX_REG)
rreg = &crx_regtab[r];
/* Check whether the register is in coprocessor registers table. */
else if (r < (int) MAX_COPREG)
rreg = &crx_copregtab[r-MAX_REG];
}
Change getreg_image's argument type to int and replace fragP->fr_literal
with &fragP->fr_literal[0] to silence GCC 10 warning.
PR gas/26044
* config/tc-crx.c (getreg_image): Change argument type to int.
(md_convert_frag): Replace fragP->fr_literal with
&fragP->fr_literal[0].
The frag code makes a distinction between inserting frags before the frag
chains are chained together and afterward. After chaining, we need to set
now_seg before creating the frag. tc-xtensa.c has a function called
fix_new_exp_in_seg that handles this right, but switches segments twice each
time it is called. In this case, we can inline it and pull the save and
restore out of the loop to get better code.
gas/
PR 26025
* config/tc-riscv.c (riscv_pre_output_hook): Change s type from const
asection to segT. New locals seg and subseg. Call subseg_set before
fix_new_exp. Call subseg_set after loop to restore original values.
