Use same calculation as getrusage, not some ad-hoc arithmetic of
internal scheduler parameters that are periodically rewound.
PR kern/57512
XXX pullup-8
XXX pullup-9
XXX pullup-10
Pointer casts from an integer type to a floating-point type and vice
versa get a 'maybe troublesome' warning now. The previous assumption
that all types of the same bit-size are convertible may have been valid
from a technical point of view, but still such code should get more
attention.
The rules for struct and union types could be made more fine-grained
later, if the need arises. To suppress this warning, it's always
possible to cast to an intermediate 'void *'.
Might help us to see if we're off by just a little bit (maybe a tiny
jitter between the hardclock timer and the monotonic clock), or if
something is seriously amiss when the timerfd_block test fails
sporadically.
This reverts the change from tree.c 1.547 from 2023-07-03. Back then, I
didn't know that the actual value from a type's 'portable size in bits'
was not supposed to be used.
The existing warning was only issued for function declarations, not for
function definitions.
The interesting change in the tests is in msg_351.c. Many other tests
use non-static functions due to their syntactic brevity. In these
tests, the warning is disabled individually, to allow new functions to
be added without generating warning 351.
Previously, in -w mode, any warning suppressed further 'unused'
warnings, even though there was no need to do that. This can be seen in
the test gcc_attribute_var.c, where only the last unused variable from a
function was marked as unused, the others slipped through.
Fixed by counting the errors and the warnings separately and only
combining them if actually desired.
1. Consolidate logic into a single helper.mk to reduce duplication.
2. Set NO* variables, not MK* variables which are reserved for user.
3. Avoid eager X!= in favour of lazy ${X:sh}.
4. Mark _g.a set list entries obsolete. Never should've been built!
PR misc/57462
- Test cartesian product of a sampling of sizes and a sampling of
alignments.
- Verify all the edge cases I could find in posix_memalign and
aligned_alloc, including failure modes.
- Test an unreasonably large (but aligned) allocation size.
- Use ATF_CHECK_* instead of ATF_REQUIRE_* so all failures will be
reported, not just the first one.
- While here, build with -fno-builtin-aligned_alloc and with
-fno-builtin-posix_memalign to make sure the compiler doesn't try
any shenanigans.
XXX pullup-10
Since tree.c 1.546 from 2023-07-03, lint no longer warned about possible
loss of accuracy when converting from 'long' to 'int' on an ILP32
platform that uses 'unsigned long' for size_t, when run in portable mode
(-p), which is enabled by default in the NetBSD build.
The integer constraints avoid false-positive warnings by looking at the
actual values an expression can take. The function can_represent is
guarded by a condition that uses the portable_size_in_bits, but then
internally used the opposite size_in_bits, which led to inconsistent
results.
The warning looks confusing though, as on an ILP32 platform, 'int' and
'long' have the same size and representation, therefore there cannot be
an actual loss of accuracy. The warning may need to be reworded to
explicitly mention the portability mode, in which sizeof(int) is assumed
to be 3 instead of 4, to catch possible loss of accuracy on other
platforms.
The previous prefix 'DK_' (declaration level kind) had a conflict with
the 'DK_' (designator kind) in init.c, so change the prefix to 'DLK_'.
The new name for dinfo_t is decl_level, which is more expressive.
No functional change.
The support for unnamed struct/union members that was added in decl.c
1.60 from 2015-10-13 was simple but wrong. It didn't cover initializers
of these structures and computed wrong sizes for structures containing
anonymous unions. At that time, the handling of initializers was broken
as well, it was fixed 6 years later in init.c 1.229 from 2021-12-22.
Real-life examples for code that lint couldn't handle are:
* external/bsd/jemalloc/dist/src/jemalloc.c
* external/mit/xorg/lib/dri.old/Makefile
When bit-fields in packed structs were added on 2009-10-02, lint assumed
that they would only use 'signed int' or 'unsigned int' as storage unit,
even though C99 also allows _Bool.
The cleanup commit for decl.c 1.225 from 2021-08-28 accidentally changed
the rounding mode for bit-field storage units from round-up to
round-down.
When running the tests via ATF, the filename was an absolute filename,
while the preprocessing line in the test file uses a relative filename.
These two didn't match.
Previously, 'indent run-equals-prev-output' was allowed even when there
was no 'indent run' section above. This created an ambiguity, since
'previous output' could mean either the 'indent run' section or the
'indent run-equals-input' section.
The assumption that in an expression of the form '(a * anything)', the
'*' marks a pointer type was too simple-minded.
For now, fix the obvious cases and leave the others for later. If
needed, they can be worked around using the '-T' option.
slower platforms causes the net/ipsec tests to take as much as 30% of
the total time to run all of the ATF tests. Reduce the number of TCs
in various net/ipsec TPs by iterating over *_ALGORITHMS_MINIMUM rather
than *_ALGORITHMS. Various of the net/ipsec tests already use the
smaller lists, so change the rest of them to do so as well.
(1) If an initial-exec relocation was used for a non-local symbol
(i.e. the definition of the symbol is in a different DSO), the
computation of the static TLS offset used the wrong DSO.
This would effectively mean the wrong address was computed
(PR toolchain/50277, PR pkg/57445).
Fix this by forcing the computation of the correct DSO (the one defining
the symbol).
This code uses __UNCONST to avoid the vast interface changes for this
special case.
(2) If symbols from a DSO loaded via dlopen are used with both
global-dynamic/local-dynamic and initial-exec relocations AND
a initial-exec relocation was resolved first in a thread, a split brain
situation could exist where the dynamic relocations would use one memory
block (separate allocation) and the initial-exec relocations the static
per-thread TLS space.
(3) If the initial-exec relocation in (2) is seen after any thread has
already used a GD/LD allocation, bail out. Since IE relocations are used
only in the GOT, this will prevent the dlopen. This is a bit more
aggressive than necessary, but a full blown reference counting doesn't
seem to be justified.
Use ${CC_WNO_MAYBE_UNINITIALIZED} instead of
the older style more complex expressions.
Remove workarounds if they were for a specific
version of gcc < 10.
Rename compiler-warning-disable variables from
GCC_NO_warning
to
CC_WNO_warning
where warning is the full warning name as used by the compiler.
GCC_NO_IMPLICIT_FALLTHRU is CC_WNO_IMPLICIT_FALLTHROUGH
Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.
Provide a single variable
CC_WNO_ADDRESS_OF_PACKED_MEMBER
with options for both clang and gcc, to replace
CLANG_NO_ADDR_OF_PACKED_MEMBER
CC_NO_ADDR_OF_PACKED_MEMBER
GCC_NO_ADDR_OF_PACKED_MEMBER
Using the convention CC_compilerflag, where compilerflag
is based on the full compiler flag name.
One library defines a symbol and _doesn't_ use it, so it has no
indication of whether the symbol is for static TLS or dynamic TLS,
and then two other libraries use it in different ways.
XXX pullup-10
Only print the 'token' buffer in debug mode if it is interesting, group
the blocks in handling of '(' tokens by topic, remove obsolete comment
from test.
This reduces the amount of wizardry needed to interpret the files. For
a more gentle introduction of the file format, see ../lint2/msg_000.ln.
The whitespace that is removed is very similar to the one in the lint2
tests, the difference is that '%<space>' and '%#' need to be preserved
in abbreviated printf and scanf format strings.
Lua reports more details when os.execute fails, which is useful when
running old versions of indent for comparison. The new test driver also
supports multiple test files in the same run.
Validate that O_NONBLOCK and O_CLOEXEC are actually set by
posix_openpt(3), as until circa 9.99.101 they were not.
If/when other flags are added like close-on-fork, this test could be
adjusted. The current concern is with supporting the expectations of
components like vte3, used by various graphical terminal clients.
Previously, the 'INDENT OFF' comments were interpreted when the newline
token from the line above the comment was processed, which was earlier
than could be reasonably expected.
The 'INDENT ON' comments were interpreted equally early, which led to
the situation that the 'INDENT OFF' comments were preserved literally
but the 'INDENT ON' comments weren't.
The indenter did not handle these backslashes well, interpreting them as
unary operators, and they are an edge case anyway. Line continuations
in string literals and character constants are kept.
Remove the need to explicitly initialize the buffers. To avoid
subtracting null pointers or comparing them using '<', migrate the
buffers from the (start, end) form to the (start, len) form. This form
also avoids inconsistencies in whether 'buf.e == buf.s' or 'buf.s ==
buf.e' is used.
Make buffer.st const, to avoid accidental modification of the buffer's
content.
Replace '*buf.e++ = ch' with buf_add_char, to avoid having to keep track
how much unwritten space is left in the buffer. Remove all safety
margins, that is, no more unchecked access to buf.st[-1] or appending
using '*buf.e++'.
Fix line number counting in lex_word for words that contain line breaks.
No functional change.
A comment is not supposed to change the state of the 'blank line after
declaration', but it did. The initialization of saved_just_saw_decl was
wrong though since it tried to capture the previous value after it had
already been overwritten.
This fixes several bugs where blank lines were erroneously added or
removed, treating these old bugs for new bugs in different places.
These new bugs are expected to be easier to fix, as the old bugs will
not interfere anymore.
rillig