The idea here is to avoid a memory access and conditional branch per
probe site. Instead, the probe is represented by an "unreachable"
unconditional function call. asm goto is used to store the address of
the probe site (represented by a no-op sled) and the address of the
function call into a tracepoint record. Each SDT probe carries a list
of tracepoints.
When the probe is enabled, the no-op sled corresponding to each
tracepoint is overwritten with a jmp to the corresponding label. The
implementation uses smp_rendezvous() to park all other CPUs while the
instruction is being overwritten, as this can't be done atomically in
general. The compiler moves argument marshalling code and the
sdt_probe() function call out-of-line, i.e., to the end of the function.
Per gallatin@ in D43504, this approach has less overhead when probes are
disabled. To make the implementation a bit simpler, I removed support
for probes with 7 arguments; nothing makes use of this except a
regression test case. It could be re-added later if need be.
The approach taken in this patch enables some more improvements:
1. We can now automatically fill out the "function" field of SDT probe
names. The SDT macros let the programmer specify the function and
module names, but this is really a bug and shouldn't have been
allowed. The intent was to be able to have the same probe in
multiple functions and to let the user restrict which probes actually
get enabled by specifying a function name or glob.
2. We can avoid branching on SDT_PROBES_ENABLED() by adding the ability
to include blocks of code in the out-of-line path. For example:
if (SDT_PROBES_ENABLED()) {
int reason = CLD_EXITED;
if (WCOREDUMP(signo))
reason = CLD_DUMPED;
else if (WIFSIGNALED(signo))
reason = CLD_KILLED;
SDT_PROBE1(proc, , , exit, reason);
}
could be written
SDT_PROBE1_EXT(proc, , , exit, reason,
int reason;
reason = CLD_EXITED;
if (WCOREDUMP(signo))
reason = CLD_DUMPED;
else if (WIFSIGNALED(signo))
reason = CLD_KILLED;
);
In the future I would like to use this mechanism more generally, e.g.,
to remove branches and marshalling code used by hwpmc, and generally to
make it easier to add new tracepoint consumers without having to add
more conditional branches to hot code paths.
Reviewed by: Domagoj Stolfa, avg
MFC after: 2 months
Differential Revision: https://reviews.freebsd.org/D44483
808f5ac fixed an edge case invloving mlock() and superpage creation
by creating and inserting a leaf pagetable page for mlock'd superpages.
However, the code does not properly release the reference to the
pagetable page in the error handling path.
This commit fixes the issue by adding calls to 'pmap_abort_ptp'
in the error handling path.
Reported by: alc
Approved by: markj (mentor)
Fixes: 808f5ac
Differential Revision: https://reviews.freebsd.org/D45578
The parse_bias method returns a signed int, with a value of -1 when
the device tree reports nothing of the bias configuration. Convert the
local 'bias' from unsigned to signed to properly check this condition.
PR: 229721
Reviewed by: mhorne
MFC after: 3 days
Most of vmm.h is machine-independent. Simplify merging amd64 and arm64
vmm code by removing this machine-dependent routine from arm64's vmm.h.
No functional change intended.
Reviewed by: andrew
Differential Revision: https://reviews.freebsd.org/D45557
FreeBSD's boot times have decreased to the point where vm_page array
initialization represents a significant fraction of the total boot time.
For example, when booting FreeBSD in Firecracker (a VMM designed to
support lightweight VMs) with 128MB and 1GB of RAM, vm_page
initialization consumes 9% (3ms) and 37% (21.5ms) of the kernel boot
time, respectively. This is generally relevant in cloud environments,
where one wants to be able to spin up VMs as quickly as possible.
This patch implements lazy initialization of (most) page structures,
following a suggestion from cperciva@. The idea is to introduce a new
free pool, VM_FREEPOOL_LAZYINIT, into which all vm_page structures are
initially placed. For this to work, we need only initialize the first
free page of each chunk placed into the buddy allocator. Then, early
page allocations draw from the lazy init pool and initialize vm_page
chunks (up to 16MB, 4096 pages) on demand. Once APs are started, an
idle-priority thread drains the lazy init pool in the background to
avoid introducing extra latency in the allocator. With this scheme,
almost all of the initialization work is moved out of the critical path.
A couple of vm_phys operations require the pool to be drained before
they can run: vm_phys_find_range() and vm_phys_unfree_page(). However,
these are rare operations. I believe that
vm_phys_find_freelist_contig() does not require any special treatment,
as it only ever accesses the first page in a power-of-2-sized free page
chunk, which is always initialized.
For now the new pool is only used on amd64 and arm64, since that's where
I can easily test and those platforms would get the most benefit.
Reviewed by: alc, kib
Differential Revision: https://reviews.freebsd.org/D40403
This will be used when we add SVE support to reduce the registers
needed to be saved on context switch.
Reviewed by: imp
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D43305
When returning from an exception to userspace clear the saved td_frame.
On the next exception this should point to the frame, however this is
not guaranteed.
To ensure the trap frame pointer is either valid or NULL clear it
before returning to userspace in the EL0 synchronous exception handler.
Reviewed by: kib, markj
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D44807
Adjust the mair_el1 macro indentation to be consistent with the
surrounding macros.
Reviewed by: emaste
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45524
Defer the bti lookup until after page table page allocation is complete.
We sometimes release the pmap lock and sleep during page table page
allocation. Consequently, the result of a bti lookup from before
page table page allocation could be stale when we finally create the
mapping based on it.
Modify pmap_bti_same() to update the prototype PTE at the same time as
checking the address range. This eliminates the need for calling
pmap_pte_bti() in addition to pmap_bti_same(). pmap_bti_same() was
already doing most of the work of pmap_pte_bti().
Reviewed by: markj
Differential Revision: https://reviews.freebsd.org/D45502
Add a method rangeset_next to find the first range that starts at or
after a given value. Use it to rewrite pmap_pkru_same and
pmap_bti_same to avoid walking a page at a time over pages in no
range.
Reviewed by: andrew, kib
Differential Revision: https://reviews.freebsd.org/D45511
When we enable checking for BTI on arm64 we need to include an ELF
note in all object files linked into a module.
As using objcopy from a binary to an ELF object file doesn't add the
note switch to using .incbin from an assembly file. This allows us to
add the needed note without affecting the included object.
Reviewed by: imp, kib, emaste
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45468
LLD has the -zbti-report=error argument to check if the BTI note is
present when linking. To allow for this to be used when linking the
kernel and modules:
- Add the BTI note to the remaining assembly files
- Mark ptrauth.c as protected by BTI
- Disable -zbti-report for vmm hypervisor switching code as it's not
used there.
The linux64 module doesn't build with the flag as it includes vdso code
that doesn't include the note.
Reviewed by: imp, kib, emaste
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45466
More precisely, implement L3C (64KB/2MB, depending on base page size)
promotion in pmap_enter_quick()'s helper function,
pmap_enter_quick_locked(). At the same time, use the recently
introduced flag VM_PROT_NO_PROMOTE from pmap_enter_object() to
pmap_enter_quick_locked() to avoid L3C promotion attempts that will
fail.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D45445
The optional 'table' pointer is a legacy part of the interface, which
has been replaced by devmap_register_table()/devmap_add_entry(). The few
in-tree callers have already adapted to this, so it can be removed.
The 'l1pt' argument is already entirely unused within the function.
Reviewed by: andrew, markj
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D45319
As a convenience to callers, who might allocate the array on the stack.
An empty/zero-valued range indicates the end of the physmap entries.
Remove the now-redundant calls to bzero() at the call site.
Reviewed by: andrew
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D45318
This will be used to implement parts of bhyve's gdb stub.
Three VM capabilities are added, similar to amd64 without monitor mode.
Two cause breakpoint and single-step exceptions to be raised to EL2 and
then down to bhyve. One lets the gdb stub mask hardware interrupts
while single-stepping, since otherwise the guest will handle a timer
interrupt before executing the target instruction and thus fail
to make progress.
Reviewed by: bnovkov, andrew
Sponsored by: Innovate UK
Differential Revision: https://reviews.freebsd.org/D44739
Implement a simple heuristic to skip pointless promotion attempts by
pmap_enter_quick_locked() and moea64_enter(). Specifically, when
vm_fault() calls pmap_enter_quick() to map neighboring pages at the end
of a copy-on-write fault, there is no point in attempting promotion in
pmap_enter_quick_locked() and moea64_enter(). Promotion will fail
because the base pages have differing protection.
Reviewed by: kib
Differential Revision: https://reviews.freebsd.org/D45431
MFC after: 1 week
The kernel source contains several definitions of an ilog2 function;
some are slower than necessary, and one of them is incorrect.
Elimininate them all and define an ilog2 macro in libkern to replace
them, in a way that is fast, correct for all argument types, and, in a
GENERIC kernel, includes a check for an invalid zero parameter.
Folks at Microsoft have verified that having a correct ilog2
definition for their MANA driver doesn't break it.
Reviewed by: alc, markj, mhorne (older version), jhibbits (older version)
Differential Revision: https://reviews.freebsd.org/D45170
Differential Revision: https://reviews.freebsd.org/D45235
There is no reason for the ATTR_SW_NO_PROMOTE checks in
pmap_update_{entry,strided}() to be panic()s instead of KASSERT()s.
Requested by: markj
Reviewed by: markj
Differential Revision: https://reviews.freebsd.org/D45424
I cannot find a time where the function was not named this.
Reviewed by: kib, markj
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D45383
This commit introduces the MINIDUMP_STARTUP_PAGE_TRACKING symbol and
uses it to simplify several instances of a complex preprocessor conditional
for adding pages allocated when bootstraping the kernel to minidumps.
Reviewed by: markj, mhorne
Approved by: markj (mentor)
Differential Revision: https://reviews.freebsd.org/D45085
This commit refactors the UMA small alloc code and
removes most UMA machine-dependent code.
The existing machine-dependent uma_small_alloc code is almost identical
across all architectures, except for powerpc where using the direct
map addresses involved extra steps in some cases.
The MI/MD split was replaced by a default uma_small_alloc
implementation that can be overridden by architecture-specific code by
defining the UMA_MD_SMALL_ALLOC symbol. Furthermore, UMA_USE_DMAP was
introduced to replace most UMA_MD_SMALL_ALLOC uses.
Reviewed by: markj, kib
Approved by: markj (mentor)
Differential Revision: https://reviews.freebsd.org/D45084
It is inherited from arm, where the global exists and is used. No
functional change.
Reviewed by: markj
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D45323
No functional change.
Reviewed by: markj
MFC after: 3 days
Sponsored by: The FreeBSD Foundation
Differential Revision: https://reviews.freebsd.org/D45322
On systems configured with 16KB pages, this change creates 1GB page
mappings in the direct map where possible. Previously, the largest page
size that was used to implement the direct map was 32MB. Similarly, on
systems configured with 4KB pages, this change creates 32MB page
mappings, instead of 2MB, in the direct map where 1GB is too large.
Implement demotion on L2C (32MB/1GB) page mappings within the DMAP.
Update sysctl vm.pmap.kernel_maps to report on L2C page mappings.
Reviewed by: markj
Tested by: gallatin, Eliot Solomon <ehs3@rice.edu>
Differential Revision: https://reviews.freebsd.org/D45224
These are missing changes after 1228b93b41 when ref_count was
removed from bus_dma_tag_common and 1e3f42b6ba, when the address
arguments were switched to pointers.
Reviewed by: jhb
MFC after: 3 days
Differential Revision: https://reviews.freebsd.org/D45289
We were missing brackets in GPIO_FLAGS_PINCTRL. Without them GCC
complains a use is ambiguous. Fix by adding the needed brackets.
Reviewed by: manu, brooks, imp, jhb, emaste
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45264
When building with gcc it complains the pointer authentication
registers aren't valid with the architecture level we are targeting.
Fix this by using the alternative spelling of these registers accesses
through MRS_REG_ALT_NAME.
Reviewed by: jhb
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45263
Add the pointer authentication registers to armreg.h. These will be
used to support pointer authentication in a kernel built with GCC.
Reviewed by: jhb
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45262
While clang can handle numbers with a UL suffix in assembly files
gcc/gas is unable to. Switch to use the UL macro for TCR_EL1 defines as
some are used in locore.S
Reviewed by: brooks, jhb
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45261
Moreover, if we attempt an L2 promotion on the kernel pmap from
pmap_enter_quick_locked(), this change eliminates the recomputation of
the L2 entry's address.
MFC after: 1 week
When booting the kernel with VHE it will be running at EL2. The current
config register values only enable the reaces at EL1 when tracing the
kernel.
Set the HMC flag to also trap from EL2.
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45121
CONTEXTIDR_EL1 is used in debug and trace features to identify the
current process or context.
Reviewed by: andrew
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45173
Bits [5:0] of PMBSR_MSS encodes either Buffer Status Code (BSC) or Fault
Status Code (FSC) depending on PMBSR_EC value.
Add PMBSR_MSS_{BSC,FSC} to cover this field.
Reviewed by: andrew
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45172
When the register is not defined in Armv8.0 i.e. added in a later
extension, like SPE added in v8.2, the alternative name format of:
S<op0>_<op1>_C<crn>_C<crm>_<op2>
should be used; otherwise, calls to {READ,WRITE}_SPECIALREG() will
fail.
Use the MRS_REG_ALT_NAME() macro for SPE changing hex to decimal as
required by the macro.
Reviewed by: andrew
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45171
To build with old toolchains use the *_REG macros to access the ID
registers. These become a name in the form S?_?_C?_C?_? where the '?'
values encode the op and CR values needed to access the register.
For consistency use these macros for all ID registers, even if most
toolchains understand them.
Reviewed by: Zachary Leaf <zachary.leaf@arm.com>
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45177
These can be used even when the compiler is too old for the register
to be included.
Reviewed by: Zachary Leaf <zachary.leaf@arm.com>
Sponsored by: Arm Ltd
Differential Revision: https://reviews.freebsd.org/D45176
The pmap_bti_same test in pmap_enter_l3c only happens in the
!ADDR_IS_KERNEL case; in the other case, a KASSERT fails. So move the
test into that case to save a bit of time when ADDR_IS_KERNEL.
Reviewed by: andrew
Differential Revision: https://reviews.freebsd.org/D45160
When the page size is 16K, use ATTR_CONTIGUOUS to map the kernel code
and data sections using 2M pages. Previously, they were mapped using
16K pages.
Reviewed by: markj
Tested by: markj
Differential Revision: https://reviews.freebsd.org/D45162
Define macros to perform pte to vm_page and vm_page to pte conversions
without composing two macros, and use the convenience macros wherever
possible.
Reviewed by: alc
Differential Revision: https://reviews.freebsd.org/D44699
