instead of the whole ds structure.
besides triggering a recently added assert in netbsd32, this stops
exposing kernel addresses.
copy the mode clamping to 0777 from sem to shm and msg.
while here, make sure that the compat callers to sysv_ipc clear
the contents of the compat structure before setting the result
members to ensure padding bytes are cleared.
don't set/copy _sem_base, _msg_first, _msg_last or _shm_internal.
even if used, which seems very dodgy, they leak KVAs as well.
possibly this may affect linux binaries, in particular, the
comments around _shm_internal ("XXX Oh well.") may mean apps
rely upon these but hopefully not -- the comments date back to
rev 1.1 in 1995.
the _key, _seq and _msg_cbytes members are exported as before as
i found multiple consumers of these (no less than ipcs(1), and
they appear to be useful for debugging and more.
XXX: the naming of compat functions have too many styles. there
are at least 3 different ones changed here.
this is not true for alpha, ia64 and arm32 ports, and the first two
were not building because of it, and the latter would be missing
the oabi support (likely not a big deal, but still wrong.)
add a makefile fragment that tells you if it is supported and include
it where needed to define COMPAT_NETBSD32 when building the normal
kernel (ie, modules & rump.)
fixes alpha build, probably fixes ia64 build.
XXX: still leaves some netbsd32 code in rf_netbsdkintf.c, that should
be moved into some hooks, but first the configuration setup
needs to be moved into a common function the netbsd32 code can
call into, vs living in the switch case itself.
The file will only be selected if "options COMPAT_80" is defined in
the config file. The COMPAT_80 macro is defined only if the option
is explicitly defined, and not if it is implicitly defined due to
inclusion of lesser-version COMPATs.
<subsystem>_<function>_<version>_hook
NFCI
XXX Note that although this introduces a change in the kernel-to-
XXX module interface, we are NOT bumping the kernel version number.
XXX We will bump the version number once the interface stabilizes.
+ Possible info leak: [len=80, leaked=10]
| #0 0xffffffff80bad7a7 in kleak_copyout
| #1 0xffffffff8048e71b in netbsd32___msgctl50
| #2 0xffffffff8022fb5b in netbsd32_syscall
| #3 0xffffffff802096dd in handle_syscall
(netbsd32_netbsd.c should have only simple shims. anything more
than copying arguments from one args struct to the other should
not be placed in this file. still a couple more to move out.)
These structs don't have padding but safer to keep the code
structured the same way between the various ntp_gettimes in case
anyone makes more copypasta of it for future updates.
- This conversion significantly simplifies the code and moves NPF to
a binary serialisation format (replacing the XML-like format).
- Fix some memory/reference leaks and possibly use-after-free bugs.
- Bump NPF_VERSION as this change makes libnpf incompatible with the
previous versions. Also, different serialisation format means NPF
connection/config saving and loading is not compatible with the
previous versions either.
Thanks to christos@ for extra testing.
These functions are defined on unsigned int. The generic name
min/max should not silently truncate to 32 bits on 64-bit systems.
This is purely a name change -- no functional change intended.
HOWEVER! Some subsystems have
#define min(a, b) ((a) < (b) ? (a) : (b))
#define max(a, b) ((a) > (b) ? (a) : (b))
even though our standard name for that is MIN/MAX. Although these
may invite multiple evaluation bugs, these do _not_ cause integer
truncation.
To avoid `fixing' these cases, I first changed the name in libkern,
and then compile-tested every file where min/max occurred in order to
confirm that it failed -- and thus confirm that nothing shadowed
min/max -- before changing it.
I have left a handful of bootloaders that are too annoying to
compile-test, and some dead code:
cobalt ews4800mips hp300 hppa ia64 luna68k vax
acorn32/if_ie.c (not included in any kernels)
macppc/if_gm.c (superseded by gem(4))
It should be easy to fix the fallout once identified -- this way of
doing things fails safe, and the goal here, after all, is to _avoid_
silent integer truncations, not introduce them.
Maybe one day we can reintroduce min/max as type-generic things that
never silently truncate. But we should avoid doing that for a while,
so that existing code has a chance to be detected by the compiler for
conversion to uimin/uimax without changing the semantics until we can
properly audit it all. (Who knows, maybe in some cases integer
truncation is actually intended!)
entry-point is either sys_nomodule or sys_nosys. Update the
makesyscalls.sh script to create a const array of bits to allow
syscall_disestablish() to properly restore the original entry-point.
Update all the initializers of struct emul to initialize the pointer
to the bit array struct emul.
XXX Regen of all files created by makesyscalls.sh will come soon,
XXX followed by a kernel version bump (since struct emul is being
XXX modified).
This commit should address PR kern/45781 and also removes the need
for the work-around for that PR in file
sys/arch/usermode/modules/syscallemu/syscallemu.c
This change:
* Removes "options PERFCTRS", the associated includes, and the associated
ifdefs. In doing so, it removes several XXXSMPs in the MI code, which is
good.
* Removes the PMC code of ARM XSCALE.
* Removes all the pmc.h files. They were all empty, except for ARM XSCALE.
* Reorders the x86 PMC code not to rely on the legacy pmc.h file. The
definitions are put in sysarch.h.
* Removes the kern/sys_pmc.c file, and along with it, the sys_pmc_control
and sys_pmc_get_info syscalls. They are marked as OBSOL in kern,
netbsd32 and rump.
* Removes the pmc_evid_t and pmc_ctr_t types.
* Removes all the associated man pages. The sets are marked as obsolete.
Cast minaddress from LP64-specific long (64-bit) to 32-bit long (int).
This value is usually 0 or 0+PAGE_SIZE. There is need to change type in
order to keep original code functional, without prompting the value with
'long long' or a similar approach.
There are more CTLTYPE_LONG with a need for fixes in the compat code:
uipc_usrreq.c
1975 CTLTYPE_LONG, "sendspace",
1981 CTLTYPE_LONG, "recvspace",
1987 CTLTYPE_LONG, "sendspace",
1993 CTLTYPE_LONG, "recvspace",
vfs_bio.c
1921 CTLTYPE_LONG, "bufmem",
1928 CTLTYPE_LONG, "bufmem_lowater",
1935 CTLTYPE_LONG, "bufmem_hiwater",
Sponsored by <The NetBSD Foundation>
Return VM_MAXUSER_ADDRESS32 instead of the 64-bit specific address.
Use CTLTYPE_INT instead of CTLTYPE_LONG in order to retain the same integer
type between 64-bit native kernel and 32-bit emulated program.
There probably should be CTLTYPE_LONG32 available for this purpose.
On NetBSD/i386:
vm.maxaddress=bfeff000
On NetBSD/amd64:
vm.maxaddress=7fbfdfeff000
On NetBSD/amd64 running i386 program:
vm.maxaddress=fffff000
A 32-bit program on the 64-bit kernel can use larger user space, this
difference is on purpose and expected.
Sponsored by <The NetBSD Foundation>
e_tracesig used to be implemented for Darwin compat. Nowadays the Darwin
compatiblity layer is gone and there are no other users.
This functionality isn't used where it shall be used in the existing
codebase.
If we want to emulate debugging interfaces in compat layers we would need
to implement that from scratch anyway. We would need to be bug compatible
with other OSes too.
Proposed on tech-kern@.
Welcome to NetBSD 8.99.16!
Sponsored by <The NetBSD Foundation>
