| ID |
Summary |
State |
Reported In Release |
Target Release |
Workaround |
Release Notes |
| CODEGEN-3650 |
Linker crashes when printing module list to map file |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
Add linker option --mapfile_contents=all,nomodules |
When creating a linker map file, and a summary of modules is written to the linker map file, and the entry point is undefined, the linker may crash. |
| CODEGEN-3617 |
gmtime incorrect when invoked in timezones east of GMT |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
Set the timezone to 0 (UTC) until the fix is in place |
|
| CODEGEN-3613 |
Some vector types computations are not carried out |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
It's possible that compiling with --opt_level=1 or greater will avoid the problem. It does in our test cases, but there are reports of cases for which that doesn't help.
The problem occurs with "*pv++ = ..." where pv is a pointer to a vector. Using "pv[i] = ..." or "*pv = ...; ++pv;" seem to effectively avoid it. |
|
| CODEGEN-3597 |
ELF header field e_phoff should be 0 when no program header is present |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
|
ELF header field e_phoff should be 0 when no program header is present, and ELF header field e_shoff should be 0 when no section headers are present. |
| CODEGEN-3571 |
Temp filename collisions on Windows in linear assembler |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
Disable parallel builds of linear assembly source files. |
When compiling multiple linear assembly jobs in parallel, it is possible to get the error message:
Unable to open temp macro library : %TEMP%\tempTI - Aborting
|
| CODEGEN-2401 |
Compiler crashes during software pipelining |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
Use option --disable_software_pipeline, or
do not use option --interrupt_threshold |
For a given loop for which we attempt software pipelining, if every attempt to pipeline the loop fails, there is a chance the compiler could perform and out-of-bounds array read, leading to unpredictable results, up to and including the compiler crashing with an internal error of some sort. The bug can only happen at --opt_level=2 or higher. This can only happen if every software pipelining attempt fails; this likelyhood tends to be aggravated by using the option --interrupt_threshold When encountering this bug, this program is more likely to crash on Windows than on Linux. |
| CODEGEN-2346 |
Compiler produces incorrect code at -o2 optimization |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
None known. |
Compiler may produce incorrect code at -o2 or greater optimization when given irregular loops. |
| CODEGEN-2286 |
palign(8) of .init_array messes up __TI_INITARRAY_Limit address |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
In the linker command file, replace
.init_array > FLASH, palign(8), fill = 0xffffffff
with the following GROUP statement:
GROUP
{
.init_array
} > FLASH, palign(8)
The palign(8) on GROUP will ensure that any required padding is added after .init_array. However, both the size of .init_array and the value of __TI_INITARRAY_Limit remain unchanged. |
Applying palign(8) to .init_array caused __TI_INIT_ARRAY_Limit to be set to the end of .init_array including the padding. This broke RTS startup code responsible for calling constructors because the table of constructors now includes invalid data. This bug has been fixed and __TI_INIT_ARRAY_Limit is no longer affected by padding. |
| CODEGEN-2238 |
Compiler incorrectly emits MISRA 10.1 for function argument |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
|
|
| CODEGEN-2214 |
Compiler optimizes away function which calls assert |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.4 |
Make sure there is some other visible effect in the function -- a call, a use of a volatile variable, something like that. |
A function whose only visible effect is to call assert() may be optimised away at --opt_level=0 or higher. |
| CODEGEN-2113 |
Hex utility mishandles space in directory name of output file |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
Use directory names without spaces for output files. |
The hex utility did not correctly handle spaces in output directory and file names. |
| CODEGEN-2098 |
Temp filename collisions on Windows with many parallel invocations |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.4 |
|
In some cases with a large number of compilations in parallel on Windows, the compiler could create temporary files with colliding names, resulting in strange compilation failures. |
| CODEGEN-2053 |
Compiler incorrectly reorders struct assign for small, volatile structs with bit-fields |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.4 |
Use optimization level 0 or off (option --opt_level) |
For a very specific optimization, the optimizer may drop a volatile qualifier from a struct assignment. In some cases, the compiler may later perform an incorrect optimization, most commonly incorrectly reordering volatile assignments.
This bug can only happen in a function compiled with optimization level 1 or higher which contains both of the following: 1) A struct assign where the destination is volatile, and the source is a known constant, and the struct contains a bit-field, and the struct is of size "int" or smaller. 2) Any bit-field access (read or write) other than as part of a struct assign or initialization expression. That is, the name of the bit-field is present in the access expression. Note that the optimizer can create this situation by inlining functions, so 1 and 2 might be in different functions in the source code.
Consider a tree x = y where x and y are of type struct S. If the value of y is known at compile time (e.g. a const value), the optimizer will try to turn the struct constant into an integer constant (possibly combining bit-fields) and rewrite the tree to look like this: "(unsigned *)x = 32;" However, if y is volatile, that should be "*(volatile unsigned *)x = 32;" Because the access is not volatile, instruction scheduling could cause this instruction to drift past nearby volatile accesses. |
| CODEGEN-2025 |
Linker crash message incorrectly suggests submitting a preprocessed file |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.4 |
|
|
| CODEGEN-1979 |
Statements before declarations with no white space (aggravated by macros) may cause incorrect parser error |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
Do not start any statement in the left-most column
Rearrange your code so that there are no statements before declarations. |
C99 and C++ allow statements before declarations in functions. This is not allowed by the C89 language, but as an extension, the TI compiler allows such statements in relaxed mode. However, in certain circumstances, the compiler may emit the 'error: expected "}"' for otherwise legal code which has statements before declarations.
This problem can only occur in relaxed C89 mode (which is the default mode), and 1) you have a function with statements that start in the left-most column, or 2) you use macros where the macro body contains C code with statements before declarations. |
| CODEGEN-1976 |
Value of __cplusplus is wrong |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.3 |
If possible, use the -ps or --strict_ansi options. This mode will use the strict definition of __cplusplus, which is 199711L. |
Our parser mimicked G++ behavior for the value of this macro in relaxed ANSI mode. This reproduced a bug in G++ versions v.4.7 and v.4.3 that has since been fixed. |
| CODEGEN-1703 |
Designated initializer plus struct hack hangs compiler |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.3 |
Avoiding using string constants to initialize objects with flexible array members. Instead, use a brace-initialized array. For example:
struct { int a; char b[]; } mystruct = {0, {'h', 'e', 'l', 'l', 'o'} }; |
Fixed a compiler hang caused by initializing flexible array members with string constants.
struct {int a; char b[]; } mystruct = {0, "hello"} /* Would cause the compiler to hang and/or crash */ |
| CODEGEN-1658 |
SPLOOP Missed Stall Exception with Debugger |
Fixed |
C6000_8.1.1 |
C6000_8.1.3 |
None. |
A hardware bug on C6000 prevents the debugger from working correctly within certain instruction sequences in an SPLOOP. |
| CODEGEN-1649 |
Warning "INLINE recursion limit exceeded" during RTS build |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.2 |
None, for the automatic library builds. Elsewhere, one can use the hidden --inline_recursion_limit option to increase the limit above the default 10. |
The compiler may emit the warning "INLINE recursion limit exceeded," but there is no way to suppress the warning, and it really isn't useful for the typical compiler user. These changes remove the warning. |
| CODEGEN-1640 |
MISRA 19.1 misreported: #include statements should only be preceded by other preprocessor directives |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.3 |
No practical workaround |
|
| CODEGEN-1639 |
MISRA-C:2004 17.6/R false positive when assigning local struct member to static struct member |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.5 |
|
|
| CODEGEN-1634 |
MISRA 7.1 misreported: octal tokens in token paste |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.3 |
No practical workaround |
|
| CODEGEN-1632 |
MISRA 15.2 misreported: switch clause unconditional break |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.3 |
Put an explicit break at the end of the compound statement |
|
| CODEGEN-1441 |
Compiler generates software pipelined loop that has multiple assignment code, and is not interrupt safe |
Fixed |
C6000_8.1.1 |
C6000_8.1.2 |
If the specific code sequence causing the issue can be located in the assembly source, then a user may insert a DINT instruction in front of the multi-cycle instruction and a RINT instruction after the last delay slot of the multi-cycle instruction. |
The compiler will mistakenly define and use a register that is defined
by a preceeding multi-cycle instruction while the definition of the value
created by the multi-cycle instruction is still in the process of being
written to that register. If an interrupt occurs between the delay slot
definition of that register and the delay slot use of that register, then
an incorrect value will arrive in the register for the delay slot use of
the register.
For example,
MPYLI .M1X A4,B6,A5:A4
MVK .S1 2380,A3
|| ADD .L1X A3,B8,A5
ADD .L1 A3,A5,A9
LDHU .D1T1 *A9,A5
NOP 1
...
In the above code, if an interrupt occurs in the delay slots of the
MPYLI instruction, between the definition of A5 and the subsequent use
of A5, then the value in A5 used by the second ADD instruction will be
the value that was written to A5 by the MPYLI instruction instead of the
first ADD instruction. |
| CODEGEN-1429 |
Software pipelined loop generates different results than loop not pipelined |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.2 |
Avoid unsigned expressions in subscripts or in computing subscripts, or compile with -o1 or -o0. |
|
| CODEGEN-1333 |
Structure assignment causes compiler to fail with INTERNAL ERROR: Decomposition error |
Fixed |
C6000_8.1.0 |
C6000_8.1.2 |
Replace struct assignments involving packed structures with a memcpy() call to copy the contents of the RHS of the struct assign to the LHS. |
|
| SDSCM00052734 |
Internal error in trampoline generation when user defines data symbol named "signal" |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.2 |
It is a user error to have two different objects in the system with the same name. Rename one of the objects. |
|
| CODEGEN-1272 |
Compiling with optimization level o2 and higher triggers an exception |
Fixed |
C6000_8.1.0 |
C6000_8.1.3 |
Compile affected C/C++ source file with "--disable:pc" compiler option |
In some cases where the compiler generates a non-SPLOOP software pipelined loop which has a large iteration interval, the compiler may determine that it is OK to speculatively execute a load instruction when it shouldn't. This can lead to the compiler generating code which, when executed, will try to access an illegal data address with a load instruction whose address operand is pointing to an illegal address. The result at run time is a hardware exception. |
| SDSCM00052864 |
struct field in register is not sign-extended to 32 bits before 32-bit compare |
Fixed |
C6000_8.1.0 |
C6000_8.1.2 |
Declare the local struct as volatile |
The compiler can place small structs in registers for fast access. In this case, the compiler fails to sign-extend a 16-bit struct member to 32-bits before a 32-bit comparison. This can happen for structs 32 bits or less, with any integer field 16 bits or less. |
| SDSCM00052889 |
uint16_t value not truncated as it ought to be in if test using -O2 |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.1 |
|
|
| SDSCM00052887 |
Compiler produces incorrect code for C674x at -O2 |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.1 |
Place #pragma UNROLL(1) immediately before the loop |
|
| SDSCM00052833 |
Linker INTERNAL ERROR with object files with DWARF information compiled by IAR compiler |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.1 |
Add option --compress_dwarf=off |
|
| SDSCM00052805 |
Decomposition error on while(*ptr++) where ptr points to volatile |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.1 |
No workaround. |
|
| SDSCM00052721 |
lnk6x automatic RTS selection does not work when thread-safe version of RTS is needed |
Fixed |
C6000_8.1.0 |
C6000_8.1.1 |
Explicitly specify the thread-safe version of the RTS library that is to be included in the link instead of using libc.a on the link command line or in the linker command file. |
|
| SDSCM00052699 |
Compiler discards write to a volatile local struct member |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.1 |
No workaround. |
|
| SDSCM00052529 |
Optimizer assumes enum variable can only have value of enumeration constants |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.1 |
Don't make enumerated types with constants only for 0 and 1 if other values may be used. Either add a third value, or use values different than 0 and 1, or avoid using values that aren't the enumeration constants. |
|
| SDSCM00052339 |
demangler --output option does not work at all |
Fixed |
C6000_8.1.0 |
C6000_8.1.0 |
armdem file.asm > file.dis |
|
| SDSCM00052301 |
Linker crashes when --ecc:ecc_error is used |
Fixed |
C6000_8.1.0 |
C6000_8.1.0 |
None. |
|
| SDSCM00052292 |
Certain cases of array-typed struct member followed by non-array member cause incorrect code |
Fixed |
C6000_8.1.0 |
C6000_8.1.0 |
Put all array-typed members of the struct at the end. The problem is with array members that appear immediately before non-array members. |
|
| SDSCM00052265 |
Flexible array with initializer not handle correctly |
Fixed |
C6000_8.1.0 |
C6000_8.1.0 |
None. |
|
| SDSCM00051485 |
Incorrect reordering of nested op= with ++ |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.2 |
Don't embed "X op= Y++" under another assignment operator. |
An expression like "a += b += c++" will produce the wrong answer. The problem is specific to sub-expressions of the form "X op= Y++" that occur under another assignment operator. |
| SDSCM00050005 |
Compiler mistakenly issues MISRA diagnostic 12.9 for a float point type |
Fixed |
C6000_8.1.0B1 |
C6000_8.1.2 |
|
|
| SDSCM00008465 |
INTERNAL ERROR for user code with void pointer dereference |
Fixed |
|
C6000_8.1.5 |
Don't dereference pointers to type void |
In relaxed mode, it is possible to write an expression that dereferences a void *. This doesn't mean anything, but if the user's code dereferences a volatile void *, the compiler could crash with an internal error. |
