diff options
Diffstat (limited to 'TODO')
| -rw-r--r-- | TODO | 56 |
1 files changed, 48 insertions, 8 deletions
@@ -3,13 +3,11 @@ ** Automatic prototype discovery: *** Use debuginfo if available Alternatively, use debuginfo to generate configure file. -*** Demangled identifiers contain partial prototypes themselves +*** Mangled identifiers contain partial prototypes themselves + They don't contain return type info, which can change the + parameter passing convention. We could use it and hope for the + best. ** Automatically update list of syscalls? -** Update /etc/ltrace.conf - In particular, we could use a config directory, where packages - would install their ltrace config scripts. The config file could - be named after SONAME, and automatically read when corresponding - library is mapped. ** More operating systems (solaris?) ** Get rid of EVENT_ARCH_SYSCALL and EVENT_ARCH_SYSRET ** Implement displaced tracing @@ -21,6 +19,10 @@ reenablement. ** Create different ltrace processes to trace different children ** Config file syntax +*** mark some symbols as exported + For PLT hits, only exported prototypes would be considered. For + symtab entry point hits, all would be. + *** named arguments This would be useful for replacing the arg1, emt2 etc. @@ -57,9 +59,29 @@ Perhaps we should hook to something after all. +*** system call error returns + + This is closely related to above. Take the following syscall + prototype: + + | long read(int,+string0,ulong); + + string0 means the same as string(array(char, zero(retval))*). But + if read returns a negative value, that signifies errno. But zero + takes this at face value and is suspicious: + + | read@SYS(3 <no return ...> + | error: maximum array length seems negative + | , "\n\003\224\003\n", 4096) = -11 + + Ideally we would do what strace does, e.g.: + + | read@SYS(3, 0x12345678, 4096) = -EAGAIN + *** errno tracking Some calls result in setting errno. Somehow mark those, and on - failure, show errno. + failure, show errno. System calls return errno as a negative + value (see the previous point). *** second conversions? This definitely calls for some general scripting. The goal is to @@ -129,10 +151,28 @@ according to architecture rules. Maybe this could be achieved by a per-arch config file with typedefs such as: - | typedef ulong = uint8_t | + | typedef ulong = uint8_t; | + +** Support for ARM/AARCH64 types + - ARM and AARCH64 both support half-precision floating point + - there are two different half-precision formats, IEEE 754-2008 + and "alternative". Both have 10 bits of mantissa and 5 bits of + exponent, and differ only in how exponent==0x1F is handled. In + IEEE format, we get NaN's and infinities; in alternative + format, this encodes normalized value -1S × 2¹⁶ × (1.mant) + - The Floating-Point Control Register, FPCR, controls: — The + half-precision format where applicable, FPCR.AHP bit. + - AARCH64 supports fixed-point interpretation of {,double}words + - e.g. fixed(int, X) (int interpreted as a decimal number with X + binary digits of fraction). + - AARCH64 supports 128-bit quad words in SIMD ** Some more functions in vect might be made to take const* Or even marked __attribute__((pure)). +** pretty printer support + GDB supports python pretty printers. We migh want to hook this in + and use it to format certain types. + * BUGS ** After a clone(), syscalls may be seen as sysrets in s390 (see trace.c:syscall_p()) |