This test failed if ipfw was loaded (as well as pf). pf used the same
tag as dummynet to indicate a packet had already gone through dummynet.
However, ipfw removes this tag, so pf didn't realise the packet had
already gone through dummynet.
Introduce a separate flag, in the existing pf mtag rather than re-using
the ipfw tag. There were no free flag bits, but PF_TAG_FRAGCACHE is no
longer used so its bit can be re-purposed.
MFC after: 2 weeks
Sponsored by: Rubicon Communications, LLC ("Netgate")
Differential Revision: https://reviews.freebsd.org/D33087
In e5c4987e3f we fixed issues with nat and dummynet, but only changed
the IPv4 code. Make the same change for IPv6 as well.
Reviewed by: glebius
MFC after: 3 weeks
Sponsored by: Rubicon Communications, LLC ("Netgate")
Differential Revision: https://reviews.freebsd.org/D33086
It is used by late ifunc resolvers so needs to be at an earlier stage
of the boot. Previously it was at the same stage so may not have run
before the ifunc resolvers.
Sponsored by: The FreeBSD Foundation
The last consumer of if_com_alloc() is firewire. It never fails
to allocate. Most likely the if_com_alloc() KPI will go away
together with if_fwip(), less likely new consumers of if_com_alloc()
will be added, but they would need to follow the no fail KPI.
With this change if_index can become static. There is nothing
that if_debug.c would want to isolate from if.c. Potentially
if.c wants to share everything with if_debug.c.
Move Bjoern's copyright to if.c.
Reviewed by: bz
There there are two changes here. First, ofreebsd32_sigreturn
is declared to take a struct osigcontext rather than a struct
ia32_sigcontext3. This type is incorrect, but harmlessly so.
Second, the name of the unimplemented ogetkerninfo changes in
freebsd32_syscallnames.
This avoids the need to keep a freebsd32-specific syscalls.master
in sync with the default ABI. As evidenced by the number of commits
required to sync the two, it is extremely easy for them to get out
of sync due to misunderstandings and user errors.
Reviewed by: kevans, kib
Add _Contains_ annotations indicating that the data pointed to by a
pointer argument contains types that vary between FreeBSD ABIs. The
supported set is long (including size_t), pointer (including
intptr_t), and time_t. The first two vary between 32- and 64-bit
ABIs. The laste betwen i386 and everything else.
These will be used to detect which syscalls require handling on
particular ABIs.
Reviewed by: kevans, kib
While we can detect most ABI changes through analysis of
syscalls.master with suitable annotations, to cases are handled
in the core implementation and others have changes that can not be
infered. Add two new config variables syscall_abi_change and
syscall_no_abi_change which override the detected value. Both are
space-seperated lists of syscall names.
Reviewed by: kevans
Use pattern matching including matches of _Contains_*_ argument
annotations to (mostly) determine which system calls require
ABI-specific handling. Automatically treat syscalls as NOPROTO
if no ABI changes are present.
Reviewed by: kevans
The obsol and unimpl config variables are space-seperated lists of
syscalls that should treated as being declared OBSOL and UNIMPL.
The allows an ABI to exclude select system calls listed in
syscalls.master.
Reviewed by: kevans
This eliminates the need for ifdefs in syscalls.master and contains the
largest set of diff to generated files on the way to switching to using
the default ABI's syscalls.master.
Reviewed by: kevans
On 32-bit architectures, 64-bit arguments are passed in pairs of
registers. On non-x86 architectures these arguments must be in evenly
aligned registers which necessiciates inserting a pad register into the
argument list. This has historically been supported by adding ifdefs
around padded and unpadded syscall defintions in syscalls.master.
In order to enable generation of 32-bit support files from the base
syscalls.master, pull this support in to makesyscalls.lua enabled by
adding pair_64bit to abi_flags.
The changes to sys_proto.h simply add #ifdef PAD64_REQUIRED
around pad arguments in struct <syscall>_args. In systrace_args(),
replace static syscall index values with post-incremented indexs
allowing a simple ifdef around the argument. Under -O1 or higher
code generation is identical. systrace_entry_setargdesc() is a bit
more complicated as we switch on argument indices. Solve this
with some use of define/undef pairs to compute the correct indices.
Reviewed by: kevans
Replace long-derived types with their abi equivalent where
required by the target ABI. There are two cases:
- All pointers to types that go from 64-bit to 32-bit between the
default ABI and the target ABI.
- Signed arguments that go from 64-bit to 32-bit (these require
sign-extension before passing to general kernel ABIs).
This adds four new config variables: abi_long, semid_t, abi_size_t,
and abi_u_long which default to long, size_t, and u_long respectively.
Reviewed by: kevans
Translate instances of intptr_t to the config value abi_intptr_t
(defaults to "intptr_t"). Used in CheriABI to translate intptr_t
to intcap_t for hybrid kernels.
Reviewed by: kevans
When the string %%ABI_HEADERS%% is found in syscalls.master, replace
it with the contents of the abi_headers config variable. This allows
an ABI-specific syscalls.conf to add lines like:
#include <compat/freebsd32/freebsd32.h>
when working from a shared syscalls.master.
Reviewed by: kevans
Optionally return errors when truncating dev_t, ino_t, and nlink_t.
In the interest of code reuse, use freebsd11_cvtstat() to perform the
truncation and error handling and then convert the resulting struct
freebsd11_stat to struct nstat.
Add missing freebsd32 compat syscalls. These syscalls require
translation because struct nstat contains four instances of struct
timespec which in turn contains a time_t and a long.
Reviewed by: kib
Strictly speaking, it takes a virtual address and doesn't touch the
object directly, but this is consistant with other aio_*() syscalls.
Reviewed by: kib
