svn commit: r352832 - head/sys/powerpc/booke
Author: jhibbits Date: Sat Sep 28 03:33:07 2019 New Revision: 352832 URL: https://svnweb.freebsd.org/changeset/base/352832 Log: powerpc/booke64: Align initial stack setting to match that of aim64's Clang9/LLD9 appears to get quite confused with the instruction stream used to obtain the tmpstack pointer, almost as though it thinks this is a C function, so tries to optimize it. Since the AIM64 method doesn't use the TOC to obtain the tmpstack, just follow that model, and lld won't get confused. Reported by: bdragon MFC after:2 weeks Modified: head/sys/powerpc/booke/locore.S Modified: head/sys/powerpc/booke/locore.S == --- head/sys/powerpc/booke/locore.S Sat Sep 28 02:49:46 2019 (r352831) +++ head/sys/powerpc/booke/locore.S Sat Sep 28 03:33:07 2019 (r352832) @@ -298,16 +298,18 @@ done_mapping: ld %r1,0(%r2) add %r2,%r1,%r2 mtspr SPR_SPRG8, %r2 + nop /* Get load offset */ ld %r31,-0x8000(%r2) /* First TOC entry is TOC base */ subf%r31,%r31,%r2 /* Subtract from real TOC base to get base */ /* Set up the stack pointer */ - addis %r1,%r2,TOC_REF(tmpstack)@ha - ld %r1,TOC_REF(tmpstack)@l(%r1) - addi%r1,%r1,TMPSTACKSZ-96 - add %r1,%r1,%r31 + bl 1f + .llong tmpstack + TMPSTACKSZ - 96 - . +1: mflr%r3 + ld %r1,0(%r3) + add %r1,%r1,%r3 bl 1f .llong _DYNAMIC-. 1: mflr%r3 ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352831 - head/sys/contrib/ncsw/Peripherals/BM
Author: jhibbits Date: Sat Sep 28 02:49:46 2019 New Revision: 352831 URL: https://svnweb.freebsd.org/changeset/base/352831 Log: dpaa(4): Fix memcpy size for threshold copy in NCSW contrib On 64-bit platforms uintptr_t makes the copy twice as large as it should be. This code isn't actually used in FreeBSD, since it's for guest mode only, not hypervisor mode, but fixing it for completeness sake. Reported by: bdragon (clang9 build) Modified: head/sys/contrib/ncsw/Peripherals/BM/bm.c Modified: head/sys/contrib/ncsw/Peripherals/BM/bm.c == --- head/sys/contrib/ncsw/Peripherals/BM/bm.c Sat Sep 28 02:15:23 2019 (r352830) +++ head/sys/contrib/ncsw/Peripherals/BM/bm.c Sat Sep 28 02:49:46 2019 (r352831) @@ -258,7 +258,7 @@ t_Error BmSetPoolThresholds(t_Handle h_Bm, uint8_t bpi memset(, 0, sizeof(t_BmIpcMsg)); ipcPoolThresh.bpid = bpid; -memcpy(ipcPoolThresh.thresholds, thresholds, sizeof(uintptr_t) * MAX_DEPLETION_THRESHOLDS); +memcpy(ipcPoolThresh.thresholds, thresholds, sizeof(uint32_t) * MAX_DEPLETION_THRESHOLDS); msg.msgId = BM_SET_POOL_THRESH; memcpy(msg.msgBody, , sizeof(t_BmIpcPoolThreshParams)); if ((errCode = XX_IpcSendMessage(p_Bm->h_Session, ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352829 - head/sys/kern
Author: markj
Date: Sat Sep 28 01:42:59 2019
New Revision: 352829
URL: https://svnweb.freebsd.org/changeset/base/352829
Log:
Fix some problems with the SPARSE_MAPPING option in the kernel linker.
- Ensure that the end of the mapping passed to vm_page_wire() is
page-aligned. vm_page_wire() expects this.
- Wire pages before reading data into them.
- Apply protections specified in the segment descriptor using
vm_map_protect() once relocation processing is done.
- On amd64, ensure that we load KLDs above KERNBASE, since they
are compiled with the "kernel" memory model by default.
Reviewed by: kib
MFC after:2 weeks
Sponsored by: Netflix
Differential Revision:https://reviews.freebsd.org/D21756
Modified:
head/sys/kern/link_elf.c
Modified: head/sys/kern/link_elf.c
==
--- head/sys/kern/link_elf.cSat Sep 28 01:31:33 2019(r352828)
+++ head/sys/kern/link_elf.cSat Sep 28 01:42:59 2019(r352829)
@@ -40,6 +40,9 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
+#ifdef SPARSE_MAPPING
+#include
+#endif
#include
#include
#include
@@ -419,7 +422,7 @@ link_elf_init(void* arg)
ef->address = 0;
#endif
#ifdef SPARSE_MAPPING
- ef->object = 0;
+ ef->object = NULL;
#endif
ef->dynamic = dp;
@@ -772,7 +775,7 @@ link_elf_link_preload(linker_class_t cls,
ef->modptr = modptr;
ef->address = *(caddr_t *)baseptr;
#ifdef SPARSE_MAPPING
- ef->object = 0;
+ ef->object = NULL;
#endif
dp = (vm_offset_t)ef->address + *(vm_offset_t *)dynptr;
ef->dynamic = (Elf_Dyn *)dp;
@@ -826,7 +829,7 @@ link_elf_load_file(linker_class_t cls, const char* fil
struct nameidata nd;
struct thread* td = curthread; /* XXX */
Elf_Ehdr *hdr;
- caddr_t firstpage;
+ caddr_t firstpage, segbase;
int nbytes, i;
Elf_Phdr *phdr;
Elf_Phdr *phlimit;
@@ -993,47 +996,61 @@ link_elf_load_file(linker_class_t cls, const char* fil
error = ENOMEM;
goto out;
}
- ef->address = (caddr_t) vm_map_min(kernel_map);
+#ifdef __amd64__
+ mapbase = (caddr_t)KERNBASE;
+#else
+ mapbase = (caddr_t)vm_map_min(kernel_map);
+#endif
+ /*
+* Mapping protections are downgraded after relocation processing.
+*/
error = vm_map_find(kernel_map, ef->object, 0,
- (vm_offset_t *) >address, mapsize, 0, VMFS_OPTIMAL_SPACE,
+ (vm_offset_t *), mapsize, 0, VMFS_OPTIMAL_SPACE,
VM_PROT_ALL, VM_PROT_ALL, 0);
if (error != 0) {
vm_object_deallocate(ef->object);
- ef->object = 0;
+ ef->object = NULL;
goto out;
}
#else
- ef->address = malloc(mapsize, M_LINKER, M_EXEC | M_WAITOK);
+ mapbase = malloc(mapsize, M_LINKER, M_EXEC | M_WAITOK);
#endif
- mapbase = ef->address;
+ ef->address = mapbase;
/*
* Read the text and data sections and zero the bss.
*/
for (i = 0; i < nsegs; i++) {
- caddr_t segbase = mapbase + segs[i]->p_vaddr - base_vaddr;
- error = vn_rdwr(UIO_READ, nd.ni_vp,
- segbase, segs[i]->p_filesz, segs[i]->p_offset,
- UIO_SYSSPACE, IO_NODELOCKED, td->td_ucred, NOCRED,
- , td);
- if (error != 0)
- goto out;
- bzero(segbase + segs[i]->p_filesz,
- segs[i]->p_memsz - segs[i]->p_filesz);
+ segbase = mapbase + segs[i]->p_vaddr - base_vaddr;
#ifdef SPARSE_MAPPING
/*
-* Wire down the pages
+* Consecutive segments may have different mapping permissions,
+* so be strict and verify that their mappings do not overlap.
*/
+ if (((vm_offset_t)segbase & PAGE_MASK) != 0) {
+ error = EINVAL;
+ goto out;
+ }
+
error = vm_map_wire(kernel_map,
- (vm_offset_t) segbase,
- (vm_offset_t) segbase + segs[i]->p_memsz,
- VM_MAP_WIRE_SYSTEM|VM_MAP_WIRE_NOHOLES);
+ (vm_offset_t)segbase,
+ (vm_offset_t)segbase + round_page(segs[i]->p_memsz),
+ VM_MAP_WIRE_SYSTEM | VM_MAP_WIRE_NOHOLES);
if (error != KERN_SUCCESS) {
error = ENOMEM;
goto out;
}
#endif
+
+ error = vn_rdwr(UIO_READ, nd.ni_vp,
+ segbase, segs[i]->p_filesz, segs[i]->p_offset,
+ UIO_SYSSPACE, IO_NODELOCKED, td->td_ucred, NOCRED,
+ , td);
+ if (error != 0)
+ goto out;
+ bzero(segbase +
Re: svn commit: r352795 - head/lib/libc/sys
On 9/27/19, Warner Losh wrote:
> On Fri, Sep 27, 2019 at 2:38 PM Mateusz Guzik wrote:
>
>> On 9/27/19, Konstantin Belousov wrote:
>> > On Fri, Sep 27, 2019 at 08:32:20PM +0200, Mateusz Guzik wrote:
>> >> On 9/27/19, Warner Losh wrote:
>> >> > Document varadic args as int, since you can't have short varadic
>> args
>> >> > (they are
>> >> > promoted to ints).
>> >> >
>> >> > - `mode_t` is `uint16_t` (`sys/sys/_types.h`)
>> >> > - `openat` takes variadic args
>> >> > - variadic args cannot be 16-bit, and indeed the code uses int
>> >> > - the manpage currently kinda implies the argument is 16-bit by
>> >> > saying
>> >> > `mode_t`
>> >> >
>> >> But opengroup says it is mode_t. Perhaps it is mode_t which needs
>> >> to be changed?
>> >
>> > Yes, users must pass mode_t, and the man page is written for users.
>> > Implementation needs to be aware of the implicit promotion and handle
>> > it accordingly.
>> >
>> > In theory, mode_t might be wider than int.
>> >
>>
>> So I think the change should be reverted. Whatever workaround is being
>> in place in rust should remain for the current codebase.
>>
>
> Rust needs to understand that it's not C. It's mistake was assuming it was
> just like C and this is a case where the languages differ because C is so
> quirky.
>
>
>> If anyone is to fixed the problem they should bump mode_t to uint32_t,
>> to match Linux. This is ABI breakage, I don't know how that's handled.
>>
>
> That's not going to happen. And there's no need. It would cause more
> heartache than it's worth.
>
>
In isolation, sure. Someone(tm) should do a type comprehensive type
check against Linux. There are probably many cases where something
has a different size, but software hardcodes what happens to work on
Linux (instead of using the type documented by opengroup or whatever
else is applicable).
>> I have no interest in handling any of this, but the change committed
>> is definitely wrong.
>>
>
> I tend to agree, but the manual was/is incomplete. The arg *IS* promoted to
> an int, per normal C rules, so that part is right and there's no
> type-checking against truncation or the wrong type being used as would be
> the case if it weren't varadic (so don't pass a long here).
>
But the fact there is any need for promotion in the first place is only
an implementation wart.
> However, type purity aside, that's not how things are implemented. Open is
> expecting an int (as is openat):
>
> int
> open(const char *path, int flags, ...)
> {
> va_list ap;
> int mode;
>
> if ((flags & O_CREAT) != 0) {
> va_start(ap, flags);
> mode = va_arg(ap, int);
> va_end(ap);
> } else {
> mode = 0;
> }
> return (((int (*)(int, const char *, int, ...))
> __libc_interposing[INTERPOS_openat])(fd, path, flags, mode));
> }
>
> so the change, from that perspective, actually documents the interface (so
> isn't definitely wrong, and my guarded 'tend to agree'). So if you did
> change the type of mode_t, the above code might be wrong afterwards (hence
> my can of worms comment). And then we're passing it again through a varadic
> function pointer...
>
> So while POSIX says one thing, we implement something else. Should we
> document POSIX or what we implement? Or do we fix our implementation to
> match the docs? For all programs that don't pass in a 'long' or a pointer,
> the difference is zero, however.
>
> To be honest, though, quibbling over how it should be implemented aside, I
> think we should actually do the following:
>
> diff --git a/lib/libc/sys/open.2 b/lib/libc/sys/open.2
> index a771461e2e49..aa912b797f74 100644
> --- a/lib/libc/sys/open.2
> +++ b/lib/libc/sys/open.2
> @@ -61,7 +61,7 @@ In this case
> and
> .Fn openat
> require an additional argument
> -.Fa "int mode" ,
> +.Fa "mode_t mode" ,
> and the file is created with mode
> .Fa mode
> as described in
> @@ -615,3 +615,8 @@ permits searches.
> The present implementation of the
> .Fa openat
> checks the current permissions of directory instead.
> +.Pp
> +The
> +.Fa mode
> +argument is varadic and may result in different calling conventions
> +than might otherwise be expected.
>
> Is what I was thinking of committing instead. It's in the BUGS section, and
> is useful to know if you are debugging code that has this in the call path
> (since values may be on the stack instead of in registers, depending on the
> calling convention for the underlying architecture).
>
I think this is fine. I mostly object to telling people to pass int instead of
mode_t.
--
Mateusz Guzik
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352826 - in head/sys: arm64/arm64 arm64/include riscv/include riscv/riscv
Author: markj
Date: Fri Sep 27 23:37:01 2019
New Revision: 352826
URL: https://svnweb.freebsd.org/changeset/base/352826
Log:
Implement pmap_page_is_mapped() correctly on arm64 and riscv.
We must also check for large mappings. pmap_page_is_mapped() is
mostly used in assertions, so the problem was not very noticeable.
Reviewed by: alc
MFC after:1 week
Differential Revision:https://reviews.freebsd.org/D21824
Modified:
head/sys/arm64/arm64/pmap.c
head/sys/arm64/include/pmap.h
head/sys/riscv/include/pmap.h
head/sys/riscv/riscv/pmap.c
Modified: head/sys/arm64/arm64/pmap.c
==
--- head/sys/arm64/arm64/pmap.c Fri Sep 27 22:18:50 2019(r352825)
+++ head/sys/arm64/arm64/pmap.c Fri Sep 27 23:37:01 2019(r352826)
@@ -4241,6 +4241,27 @@ restart:
}
/*
+ * Returns true if the given page is mapped individually or as part of
+ * a 2mpage. Otherwise, returns false.
+ */
+bool
+pmap_page_is_mapped(vm_page_t m)
+{
+ struct rwlock *lock;
+ bool rv;
+
+ if ((m->oflags & VPO_UNMANAGED) != 0)
+ return (false);
+ lock = VM_PAGE_TO_PV_LIST_LOCK(m);
+ rw_rlock(lock);
+ rv = !TAILQ_EMPTY(>md.pv_list) ||
+ ((m->flags & PG_FICTITIOUS) == 0 &&
+ !TAILQ_EMPTY(_to_pvh(VM_PAGE_TO_PHYS(m))->pv_list));
+ rw_runlock(lock);
+ return (rv);
+}
+
+/*
* Destroy all managed, non-wired mappings in the given user-space
* pmap. This pmap cannot be active on any processor besides the
* caller.
Modified: head/sys/arm64/include/pmap.h
==
--- head/sys/arm64/include/pmap.h Fri Sep 27 22:18:50 2019
(r352825)
+++ head/sys/arm64/include/pmap.h Fri Sep 27 23:37:01 2019
(r352826)
@@ -150,6 +150,7 @@ vm_paddr_t pmap_kextract(vm_offset_t va);
void pmap_kremove(vm_offset_t);
void pmap_kremove_device(vm_offset_t, vm_size_t);
void *pmap_mapdev_attr(vm_offset_t pa, vm_size_t size, vm_memattr_t ma);
+bool pmap_page_is_mapped(vm_page_t m);
bool pmap_ps_enabled(pmap_t pmap);
void *pmap_mapdev(vm_offset_t, vm_size_t);
@@ -166,8 +167,6 @@ boolpmap_get_tables(pmap_t, vm_offset_t, pd_entry_t
*
intpmap_fault(pmap_t, uint64_t, uint64_t);
struct pcb *pmap_switch(struct thread *, struct thread *);
-
-#definepmap_page_is_mapped(m) (!TAILQ_EMPTY(&(m)->md.pv_list))
static inline int
pmap_vmspace_copy(pmap_t dst_pmap __unused, pmap_t src_pmap __unused)
Modified: head/sys/riscv/include/pmap.h
==
--- head/sys/riscv/include/pmap.h Fri Sep 27 22:18:50 2019
(r352825)
+++ head/sys/riscv/include/pmap.h Fri Sep 27 23:37:01 2019
(r352826)
@@ -146,6 +146,7 @@ voidpmap_kenter_device(vm_offset_t, vm_size_t,
vm_pad
vm_paddr_t pmap_kextract(vm_offset_t va);
void pmap_kremove(vm_offset_t);
void pmap_kremove_device(vm_offset_t, vm_size_t);
+bool pmap_page_is_mapped(vm_page_t m);
bool pmap_ps_enabled(pmap_t);
void *pmap_mapdev(vm_offset_t, vm_size_t);
@@ -158,8 +159,6 @@ voidpmap_unmap_io_transient(vm_page_t *,
vm_offset_t
bool pmap_get_tables(pmap_t, vm_offset_t, pd_entry_t **, pd_entry_t **,
pt_entry_t **);
-
-#definepmap_page_is_mapped(m) (!TAILQ_EMPTY(&(m)->md.pv_list))
int pmap_fault_fixup(pmap_t, vm_offset_t, vm_prot_t);
Modified: head/sys/riscv/riscv/pmap.c
==
--- head/sys/riscv/riscv/pmap.c Fri Sep 27 22:18:50 2019(r352825)
+++ head/sys/riscv/riscv/pmap.c Fri Sep 27 23:37:01 2019(r352826)
@@ -3541,6 +3541,27 @@ restart:
return (count);
}
+/*
+ * Returns true if the given page is mapped individually or as part of
+ * a 2mpage. Otherwise, returns false.
+ */
+bool
+pmap_page_is_mapped(vm_page_t m)
+{
+ struct rwlock *lock;
+ bool rv;
+
+ if ((m->oflags & VPO_UNMANAGED) != 0)
+ return (false);
+ lock = VM_PAGE_TO_PV_LIST_LOCK(m);
+ rw_rlock(lock);
+ rv = !TAILQ_EMPTY(>md.pv_list) ||
+ ((m->flags & PG_FICTITIOUS) == 0 &&
+ !TAILQ_EMPTY(_to_pvh(VM_PAGE_TO_PHYS(m))->pv_list));
+ rw_runlock(lock);
+ return (rv);
+}
+
static void
pmap_remove_pages_pv(pmap_t pmap, vm_page_t m, pv_entry_t pv,
struct spglist *free, bool superpage)
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352795 - head/lib/libc/sys
On Fri, Sep 27, 2019 at 2:38 PM Mateusz Guzik wrote:
> On 9/27/19, Konstantin Belousov wrote:
> > On Fri, Sep 27, 2019 at 08:32:20PM +0200, Mateusz Guzik wrote:
> >> On 9/27/19, Warner Losh wrote:
> >> > Document varadic args as int, since you can't have short varadic
> args
> >> > (they are
> >> > promoted to ints).
> >> >
> >> > - `mode_t` is `uint16_t` (`sys/sys/_types.h`)
> >> > - `openat` takes variadic args
> >> > - variadic args cannot be 16-bit, and indeed the code uses int
> >> > - the manpage currently kinda implies the argument is 16-bit by
> >> > saying
> >> > `mode_t`
> >> >
> >> But opengroup says it is mode_t. Perhaps it is mode_t which needs
> >> to be changed?
> >
> > Yes, users must pass mode_t, and the man page is written for users.
> > Implementation needs to be aware of the implicit promotion and handle
> > it accordingly.
> >
> > In theory, mode_t might be wider than int.
> >
>
> So I think the change should be reverted. Whatever workaround is being
> in place in rust should remain for the current codebase.
>
Rust needs to understand that it's not C. It's mistake was assuming it was
just like C and this is a case where the languages differ because C is so
quirky.
> If anyone is to fixed the problem they should bump mode_t to uint32_t,
> to match Linux. This is ABI breakage, I don't know how that's handled.
>
That's not going to happen. And there's no need. It would cause more
heartache than it's worth.
> I have no interest in handling any of this, but the change committed
> is definitely wrong.
>
I tend to agree, but the manual was/is incomplete. The arg *IS* promoted to
an int, per normal C rules, so that part is right and there's no
type-checking against truncation or the wrong type being used as would be
the case if it weren't varadic (so don't pass a long here).
However, type purity aside, that's not how things are implemented. Open is
expecting an int (as is openat):
int
open(const char *path, int flags, ...)
{
va_list ap;
int mode;
if ((flags & O_CREAT) != 0) {
va_start(ap, flags);
mode = va_arg(ap, int);
va_end(ap);
} else {
mode = 0;
}
return (((int (*)(int, const char *, int, ...))
__libc_interposing[INTERPOS_openat])(fd, path, flags, mode));
}
so the change, from that perspective, actually documents the interface (so
isn't definitely wrong, and my guarded 'tend to agree'). So if you did
change the type of mode_t, the above code might be wrong afterwards (hence
my can of worms comment). And then we're passing it again through a varadic
function pointer...
So while POSIX says one thing, we implement something else. Should we
document POSIX or what we implement? Or do we fix our implementation to
match the docs? For all programs that don't pass in a 'long' or a pointer,
the difference is zero, however.
To be honest, though, quibbling over how it should be implemented aside, I
think we should actually do the following:
diff --git a/lib/libc/sys/open.2 b/lib/libc/sys/open.2
index a771461e2e49..aa912b797f74 100644
--- a/lib/libc/sys/open.2
+++ b/lib/libc/sys/open.2
@@ -61,7 +61,7 @@ In this case
and
.Fn openat
require an additional argument
-.Fa "int mode" ,
+.Fa "mode_t mode" ,
and the file is created with mode
.Fa mode
as described in
@@ -615,3 +615,8 @@ permits searches.
The present implementation of the
.Fa openat
checks the current permissions of directory instead.
+.Pp
+The
+.Fa mode
+argument is varadic and may result in different calling conventions
+than might otherwise be expected.
Is what I was thinking of committing instead. It's in the BUGS section, and
is useful to know if you are debugging code that has this in the call path
(since values may be on the stack instead of in registers, depending on the
calling convention for the underlying architecture).
Warner
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352823 - in head/sys: amd64/vmm/amd amd64/vmm/intel dev/qlnx/qlnxe
Author: markj
Date: Fri Sep 27 21:04:33 2019
New Revision: 352823
URL: https://svnweb.freebsd.org/changeset/base/352823
Log:
Correct the scope of several global variables.
They are accessed from multiple compilation units. No functional change
intended.
MFC after:1 week
Sponsored by: Netflix
Modified:
head/sys/amd64/vmm/amd/amdvi_hw.c
head/sys/amd64/vmm/intel/vmx.c
head/sys/dev/qlnx/qlnxe/qlnx_os.c
Modified: head/sys/amd64/vmm/amd/amdvi_hw.c
==
--- head/sys/amd64/vmm/amd/amdvi_hw.c Fri Sep 27 20:56:49 2019
(r352822)
+++ head/sys/amd64/vmm/amd/amdvi_hw.c Fri Sep 27 21:04:33 2019
(r352823)
@@ -96,7 +96,7 @@ SYSCTL_INT(_hw_vmm_amdvi, OID_AUTO, host_ptp, CTLFLAG_
TUNABLE_INT("hw.vmm.amdvi.host_ptp", _host_ptp);
/* Page table level used <= supported by h/w[v1=7]. */
-static int amdvi_ptp_level = 4;
+int amdvi_ptp_level = 4;
SYSCTL_INT(_hw_vmm_amdvi, OID_AUTO, ptp_level, CTLFLAG_RDTUN,
_ptp_level, 0, NULL);
TUNABLE_INT("hw.vmm.amdvi.ptp_level", _ptp_level);
Modified: head/sys/amd64/vmm/intel/vmx.c
==
--- head/sys/amd64/vmm/intel/vmx.c Fri Sep 27 20:56:49 2019
(r352822)
+++ head/sys/amd64/vmm/intel/vmx.c Fri Sep 27 21:04:33 2019
(r352823)
@@ -189,10 +189,10 @@ static u_int vpid_alloc_failed;
SYSCTL_UINT(_hw_vmm_vmx, OID_AUTO, vpid_alloc_failed, CTLFLAG_RD,
_alloc_failed, 0, NULL);
-static int guest_l1d_flush;
+int guest_l1d_flush;
SYSCTL_INT(_hw_vmm_vmx, OID_AUTO, l1d_flush, CTLFLAG_RD,
_l1d_flush, 0, NULL);
-static int guest_l1d_flush_sw;
+int guest_l1d_flush_sw;
SYSCTL_INT(_hw_vmm_vmx, OID_AUTO, l1d_flush_sw, CTLFLAG_RD,
_l1d_flush_sw, 0, NULL);
Modified: head/sys/dev/qlnx/qlnxe/qlnx_os.c
==
--- head/sys/dev/qlnx/qlnxe/qlnx_os.c Fri Sep 27 20:56:49 2019
(r352822)
+++ head/sys/dev/qlnx/qlnxe/qlnx_os.c Fri Sep 27 21:04:33 2019
(r352823)
@@ -232,10 +232,9 @@ MODULE_DEPEND(if_qlnxev, ether, 1, 1, 1);
MALLOC_DEFINE(M_QLNXBUF, "qlnxbuf", "Buffers for qlnx driver");
-
-static char qlnx_dev_str[128];
-static char qlnx_ver_str[VER_SIZE];
-static char qlnx_name_str[NAME_SIZE];
+char qlnx_dev_str[128];
+char qlnx_ver_str[VER_SIZE];
+char qlnx_name_str[NAME_SIZE];
/*
* Some PCI Configuration Space Related Defines
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352821 - head/tools/tools/nanobsd
Author: imp
Date: Fri Sep 27 20:56:44 2019
New Revision: 352821
URL: https://svnweb.freebsd.org/changeset/base/352821
Log:
Push and pop xtrace correctly for run_early_customize
run_early_customize is run as a shell list, not as a subshell, so that the
side
effects of setting variables can affect later stages of the build (for better
or
worse, it's been like this since it was introduced). It therefore has the side
effect of turning off xtrace always, which limits the usefulness of sh -x
nanobsd.sh. Remember the old setting and only turn off tracing after the
command
if tracing was off before. All the other places where we do similar things we
use
a subshell, so we don't need to do this.
Modified:
head/tools/tools/nanobsd/defaults.sh
Modified: head/tools/tools/nanobsd/defaults.sh
==
--- head/tools/tools/nanobsd/defaults.shFri Sep 27 20:56:31 2019
(r352820)
+++ head/tools/tools/nanobsd/defaults.shFri Sep 27 20:56:44 2019
(r352821)
@@ -445,7 +445,11 @@ run_early_customize ( ) {
pprint 2 "early customize \"$c\""
pprint 3 "log: ${NANO_LOG}/_.early_cust.$c"
pprint 4 "`type $c`"
- { set -x ; $c ; set +x ; } >${NANO_LOG}/_.early_cust.$c 2>&1
+ { t=$(set -o | awk '$1 == "xtrace" && $2 == "off" { print "set
+x"}');
+ set -x ;
+ $c ;
+ eval $t
+ } >${NANO_LOG}/_.early_cust.$c 2>&1
done
}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352822 - head/tools/tools/nanobsd
Author: imp
Date: Fri Sep 27 20:56:49 2019
New Revision: 352822
URL: https://svnweb.freebsd.org/changeset/base/352822
Log:
Use set -o xtrace in preference to set -x for consistency with
the rest of nanobsd.sh.
Modified:
head/tools/tools/nanobsd/defaults.sh
Modified: head/tools/tools/nanobsd/defaults.sh
==
--- head/tools/tools/nanobsd/defaults.shFri Sep 27 20:56:44 2019
(r352821)
+++ head/tools/tools/nanobsd/defaults.shFri Sep 27 20:56:49 2019
(r352822)
@@ -445,8 +445,8 @@ run_early_customize ( ) {
pprint 2 "early customize \"$c\""
pprint 3 "log: ${NANO_LOG}/_.early_cust.$c"
pprint 4 "`type $c`"
- { t=$(set -o | awk '$1 == "xtrace" && $2 == "off" { print "set
+x"}');
- set -x ;
+ { t=$(set -o | awk '$1 == "xtrace" && $2 == "off" { print "set
+o xtrace"}');
+ set -o xtrace ;
$c ;
eval $t
} >${NANO_LOG}/_.early_cust.$c 2>&1
@@ -466,7 +466,7 @@ run_customize ( ) (
pprint 2 "customize \"$c\""
pprint 3 "log: ${NANO_LOG}/_.cust.$c"
pprint 4 "`type $c`"
- ( set -x ; $c ) > ${NANO_LOG}/_.cust.$c 2>&1
+ ( set -o xtrace ; $c ) > ${NANO_LOG}/_.cust.$c 2>&1
done
)
@@ -481,7 +481,7 @@ run_late_customize ( ) (
pprint 2 "late customize \"$c\""
pprint 3 "log: ${NANO_LOG}/_.late_cust.$c"
pprint 4 "`type $c`"
- ( set -x ; $c ) > ${NANO_LOG}/_.late_cust.$c 2>&1
+ ( set -o xtrace ; $c ) > ${NANO_LOG}/_.late_cust.$c 2>&1
done
)
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352820 - head/tools/tools/nanobsd
Author: imp
Date: Fri Sep 27 20:56:31 2019
New Revision: 352820
URL: https://svnweb.freebsd.org/changeset/base/352820
Log:
Remove workaround for building on FreeBSD hosts prior to FreeBSD 10.
rm -x was introduced in the FreeBSD 10 time frame. 4 years ago I added a
function to cope with building nanobsd images on hosts as old FreeBSD 7 that
lacked rm -x. The workaround is no longer needed as FreeBSD 9 hasn't been
supported for almost 3 years. Eliminate the wrapper and use rm -x directly
again.
Modified:
head/tools/tools/nanobsd/defaults.sh
Modified: head/tools/tools/nanobsd/defaults.sh
==
--- head/tools/tools/nanobsd/defaults.shFri Sep 27 20:53:31 2019
(r352819)
+++ head/tools/tools/nanobsd/defaults.shFri Sep 27 20:56:31 2019
(r352820)
@@ -241,16 +241,6 @@ nano_global_make_env ( ) (
[ -z "${NANO_CPUTYPE}" ] || echo TARGET_CPUTYPE="${NANO_CPUTYPE}"
)
-# rm doesn't know -x prior to FreeBSD 10, so cope with a variety of build
-# hosts for now. This will go away when support in the base goes away.
-rm ( ) {
-echo "NANO RM $*"
- case $(uname -r) in
- 7*|8*|9*) command rm $* ;;
- *) command rm -x $* ;;
- esac
-}
-
#
# Create empty files in the target tree, and record the fact. All paths
# are relative to NANO_WORLDDIR.
@@ -274,7 +264,7 @@ tgt_dir2symlink ( ) (
symlink=$2
cd "${NANO_WORLDDIR}"
- rm -rf "$dir"
+ rm -xrf "$dir"
ln -s "$symlink" "$dir"
if [ -n "$NANO_METALOG" ]; then
echo "./${dir} type=link mode=0777 link=${symlink}" >>
${NANO_METALOG}
@@ -294,9 +284,9 @@ CR0 ( ) {
clean_build ( ) (
pprint 2 "Clean and create object directory (${MAKEOBJDIRPREFIX})"
- if ! rm -rf ${MAKEOBJDIRPREFIX}/ > /dev/null 2>&1 ; then
+ if ! rm -xrf ${MAKEOBJDIRPREFIX}/ > /dev/null 2>&1 ; then
chflags -R noschg ${MAKEOBJDIRPREFIX}/
- rm -r ${MAKEOBJDIRPREFIX}/
+ rm -xr ${MAKEOBJDIRPREFIX}/
fi
)
@@ -348,17 +338,17 @@ build_kernel ( ) (
clean_world ( ) (
if [ "${NANO_OBJ}" != "${MAKEOBJDIRPREFIX}" ]; then
pprint 2 "Clean and create object directory (${NANO_OBJ})"
- if ! rm -rf ${NANO_OBJ}/ > /dev/null 2>&1 ; then
+ if ! rm -xrf ${NANO_OBJ}/ > /dev/null 2>&1 ; then
chflags -R noschg ${NANO_OBJ}
- rm -r ${NANO_OBJ}/
+ rm -xr ${NANO_OBJ}/
fi
mkdir -p "${NANO_OBJ}" "${NANO_WORLDDIR}"
printenv > ${NANO_LOG}/_.env
else
pprint 2 "Clean and create world directory (${NANO_WORLDDIR})"
- if ! rm -rf "${NANO_WORLDDIR}/" > /dev/null 2>&1 ; then
+ if ! rm -xrf "${NANO_WORLDDIR}/" > /dev/null 2>&1 ; then
chflags -R noschg "${NANO_WORLDDIR}"
- rm -rf "${NANO_WORLDDIR}/"
+ rm -xrf "${NANO_WORLDDIR}/"
fi
mkdir -p "${NANO_WORLDDIR}"
fi
@@ -531,7 +521,7 @@ setup_nanobsd ( ) (
cd usr/local/etc
find . -print | cpio -dumpl ../../../etc/local
cd ..
- rm -rf etc
+ rm -xrf etc
)
fi
@@ -780,7 +770,7 @@ cust_pkgng ( ) (
mount -t nullfs -o noatime -o ro ${NANO_PACKAGE_DIR}
${NANO_WORLDDIR}/_.p
mount -t devfs devfs ${NANO_WORLDDIR}/dev
- trap "umount ${NANO_WORLDDIR}/dev; umount ${NANO_WORLDDIR}/_.p ; rm -rf
${NANO_WORLDDIR}/_.p" 1 2 15 EXIT
+ trap "umount ${NANO_WORLDDIR}/dev; umount ${NANO_WORLDDIR}/_.p ; rm
-xrf ${NANO_WORLDDIR}/_.p" 1 2 15 EXIT
# Install pkg-* package
CR "${PKGCMD} add /_.p/${_NANO_PKG_PACKAGE}"
@@ -807,7 +797,7 @@ cust_pkgng ( ) (
trap - 1 2 15 EXIT
umount ${NANO_WORLDDIR}/dev
umount ${NANO_WORLDDIR}/_.p
- rm -rf ${NANO_WORLDDIR}/_.p
+ rm -xrf ${NANO_WORLDDIR}/_.p
)
###
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352819 - head/usr.bin/top
Author: dim
Date: Fri Sep 27 20:53:31 2019
New Revision: 352819
URL: https://svnweb.freebsd.org/changeset/base/352819
Log:
Allow entering fractional delays in top(1) interactive mode.
This uses the same logic as with the -s option, first validating the
entered value, then storing the result in a struct timeval.
MFC after:3 days
X-MFC-With: r352818
Modified:
head/usr.bin/top/top.c
Modified: head/usr.bin/top/top.c
==
--- head/usr.bin/top/top.c Fri Sep 27 20:20:21 2019(r352818)
+++ head/usr.bin/top/top.c Fri Sep 27 20:53:31 2019(r352819)
@@ -886,12 +886,22 @@ restart:
case CMD_delay: /* new seconds delay */
new_message(MT_standout, "Seconds to delay: ");
- if ((i = readline(tempbuf1, 8, true)) > 0)
+ if ((i = readline(tempbuf1, 8, false)) > 0)
{
- delay.tv_sec = i;
- delay.tv_usec = 0;
+ double delay_d = strtod(tempbuf1, );
+ if (nptr == tempbuf1 || delay_d <= 0)
+ {
+ new_message(MT_standout, " Invalid
delay");
+ putchar('\r');
+ no_command = true;
+ }
+ else
+ {
+ delay.tv_sec = delay_d;
+ delay.tv_usec = (delay_d -
delay.tv_sec) * 1e6;
+ clear_message();
+ }
}
- clear_message();
break;
case CMD_displays: /* change display count */
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352795 - head/lib/libc/sys
On 9/27/19, Konstantin Belousov wrote: > On Fri, Sep 27, 2019 at 08:32:20PM +0200, Mateusz Guzik wrote: >> On 9/27/19, Warner Losh wrote: >> > Document varadic args as int, since you can't have short varadic args >> > (they are >> > promoted to ints). >> > >> > - `mode_t` is `uint16_t` (`sys/sys/_types.h`) >> > - `openat` takes variadic args >> > - variadic args cannot be 16-bit, and indeed the code uses int >> > - the manpage currently kinda implies the argument is 16-bit by >> > saying >> > `mode_t` >> > >> But opengroup says it is mode_t. Perhaps it is mode_t which needs >> to be changed? > > Yes, users must pass mode_t, and the man page is written for users. > Implementation needs to be aware of the implicit promotion and handle > it accordingly. > > In theory, mode_t might be wider than int. > So I think the change should be reverted. Whatever workaround is being in place in rust should remain for the current codebase. If anyone is to fixed the problem they should bump mode_t to uint32_t, to match Linux. This is ABI breakage, I don't know how that's handled. I have no interest in handling any of this, but the change committed is definitely wrong. -- Mateusz Guzik ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352795 - head/lib/libc/sys
On Fri, 27 Sep 2019, Konstantin Belousov wrote: On Fri, Sep 27, 2019 at 08:32:20PM +0200, Mateusz Guzik wrote: On 9/27/19, Warner Losh wrote: Author: imp Date: Fri Sep 27 16:11:47 2019 New Revision: 352795 URL: https://svnweb.freebsd.org/changeset/base/352795 Log: Document varadic args as int, since you can't have short varadic args (they are promoted to ints). - `mode_t` is `uint16_t` (`sys/sys/_types.h`) - `openat` takes variadic args - variadic args cannot be 16-bit, and indeed the code uses int - the manpage currently kinda implies the argument is 16-bit by saying `mode_t` Prompted by Rust things: https://github.com/tailhook/openat/issues/21 Submitted by: Greg V at unrelenting Differential Revision: https://reviews.freebsd.org/D21816 Modified: head/lib/libc/sys/mq_open.2 head/lib/libc/sys/open.2 Modified: head/lib/libc/sys/mq_open.2 == --- head/lib/libc/sys/mq_open.2 Fri Sep 27 15:28:30 2019(r352794) +++ head/lib/libc/sys/mq_open.2 Fri Sep 27 16:11:47 2019(r352795) @@ -133,7 +133,7 @@ Create a message queue. It requires two additional arguments: .Fa mode , which is of type -.Vt mode_t , +.Vt int , and .Fa attr , which is a pointer to an Modified: head/lib/libc/sys/open.2 == --- head/lib/libc/sys/open.2Fri Sep 27 15:28:30 2019(r352794) +++ head/lib/libc/sys/open.2Fri Sep 27 16:11:47 2019(r352795) @@ -61,7 +61,7 @@ In this case and .Fn openat require an additional argument -.Fa "mode_t mode" , +.Fa "int mode" , and the file is created with mode .Fa mode as described in But opengroup says it is mode_t. Perhaps it is mode_t which needs to be changed? POSIX needed to be changed before it became standard in 1988, but it is too late now. Types shorter than int shouldn't be used in APIs since they cause ABI and API problems. Especially in 1988 when non-prototyped functions were common. Non-prototyped functions use the default promotions much like variadic functions. open() is variadic, so its mode_t arg is always promoted, but 'int chmod(const char *path, mode_t mode)' is just wrong since in 1988 prototypes were not required and the prototype matching the K API was 'int chmod(const char *path, __default_promotion_of(mode_t) mode)'. Yes, users must pass mode_t, and the man page is written for users. Implementation needs to be aware of the implicit promotion and handle it accordingly. In theory, mode_t might be wider than int. Indeed. mode_t can also be the same size as int, but unsigned. This happens naturally if int is 16 bits, which POSIX allowed before 2001. In V7, mode_t was really a 16-bit u_int, but it was type-punned to int. When mode_t is u_int or wider than int, then the promotion is null, and variadic and K do it consistently (but the arg type before the call must be precisely mode_t, not int). The POSIX list recently discussed a related problem with variadic pid_t args for fcntl(). For F_SETOWN, the original arg type should be pid_t as for mode_t for open(). Howver, F_GETOWN only returns int, so pids larger than INT_MAX cannot work and making pid_t different from int is just foot shooting. Bruce ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352818 - head/usr.bin/top
Author: dim
Date: Fri Sep 27 20:20:21 2019
New Revision: 352818
URL: https://svnweb.freebsd.org/changeset/base/352818
Log:
Make fractional delays for top(1) work for interactive mode.
In r334906, the -s option was changed to allow fractional times, but
this only functioned correctly for batch mode. In interactive mode, any
delay below 1.0 would get floored to zero. This would put top(1) into a
tight loop, which could be difficult to interrupt.
Fix this by storing the -s option value (after validation) into a struct
timeval, and using that struct consistently for delaying with select(2).
Next up is to allow interactive entry of a fractional delay value.
MFC after:3 days
Modified:
head/usr.bin/top/top.1
head/usr.bin/top/top.c
Modified: head/usr.bin/top/top.1
==
--- head/usr.bin/top/top.1 Fri Sep 27 20:09:43 2019(r352817)
+++ head/usr.bin/top/top.1 Fri Sep 27 20:20:21 2019(r352818)
@@ -147,7 +147,7 @@ no information is available about the percentage of ti
.It Fl s Ar time
Set the delay between screen updates to
.Ar time
-seconds.
+seconds, which may be fractional.
The default delay between updates is 1 second.
.It Fl o Ar field
Sort the process display area on the specified field.
Modified: head/usr.bin/top/top.c
==
--- head/usr.bin/top/top.c Fri Sep 27 20:09:43 2019(r352817)
+++ head/usr.bin/top/top.c Fri Sep 27 20:20:21 2019(r352818)
@@ -233,7 +233,7 @@ main(int argc, const char *argv[])
static char tempbuf2[50];
sigset_t old_sigmask, new_sigmask;
int topn = Infinity;
-double delay = 2;
+struct timeval delay = { 2, 0 };
int displays = 0; /* indicates unspecified */
int sel_ret = 0;
time_t curr_time;
@@ -377,21 +377,27 @@ main(int argc, const char *argv[])
break;
}
- case 's':
- delay = strtod(optarg, );
- if (nptr == optarg) {
- warnx("warning: invalid delay");
- delay = 2;
- warnings++;
- }
- if (delay < 0) {
- warnx("warning: seconds delay should be
positive -- using default");
- delay = 2;
- warnings++;
- }
+ case 's':
+ {
+ double delay_d = strtod(optarg, );
+ if (nptr == optarg)
+ {
+ warnx("warning: invalid delay");
+ warnings++;
+ }
+ else if (delay_d <= 0)
+ {
+ warnx("warning: seconds delay should be positive -- using
default");
+ warnings++;
+ }
+ else
+ {
+ delay.tv_sec = delay_d;
+ delay.tv_usec = (delay_d - delay.tv_sec) * 1e6;
+ }
+ break;
+ }
- break;
-
case 'q': /* be quick about it */
errno = 0;
i = setpriority(PRIO_PROCESS, 0, PRIO_MIN);
@@ -704,7 +710,8 @@ restart:
no_command = true;
if (!interactive)
{
- usleep(delay * 1e6);
+ timeout = delay;
+ select(0, NULL, NULL, NULL, );
if (leaveflag) {
end_screen();
exit(0);
@@ -718,8 +725,7 @@ restart:
/* set up arguments for select with timeout */
FD_ZERO();
FD_SET(0, );/* for standard input */
- timeout.tv_sec = delay;
- timeout.tv_usec = 0;
+ timeout = delay;
if (leaveflag) {
end_screen();
@@ -880,12 +886,10 @@ restart:
case CMD_delay: /* new seconds delay */
new_message(MT_standout, "Seconds to delay: ");
- if ((i = readline(tempbuf1, 8, true)) > -1)
+ if ((i = readline(tempbuf1, 8, true)) > 0)
{
- if ((delay = i) == 0)
- {
- delay = 1;
- }
+ delay.tv_sec = i;
+ delay.tv_usec = 0;
}
clear_message();
break;
___
svn-src-head@freebsd.org
svn commit: r352816 - in head/sys: kern sys
Author: gallatin
Date: Fri Sep 27 20:08:19 2019
New Revision: 352816
URL: https://svnweb.freebsd.org/changeset/base/352816
Log:
kTLS: Fix a bug where we would not encrypt anon data inplace.
Software Kernel TLS needs to allocate a new destination crypto
buffer when encrypting data from the page cache, so as to avoid
overwriting shared clear-text file data with encrypted data
specific to a single socket. When the data is anonymous, eg, not
tied to a file, then we can encrypt in place and avoid allocating
a new page. This fixes a bug where the existing code always
assumes the data is private, and never encrypts in place. This
results in unneeded page allocations and potentially more memory
bandwidth consumption when doing socket writes.
When the code was written at Netflix, ktls_encrypt() looked at
private sendfile flags to determine if the pages being encrypted
where part of the page cache (coming from sendfile) or
anonymous (coming from sosend). This was broken internally at
Netflix when the sendfile flags were made private, and the
M_WRITABLE() check was added. Unfortunately, M_WRITABLE() will
always be false for M_NOMAP mbufs, since one cannot just mtod()
them.
This change introduces a new flags field to the mbuf_ext_pgs
struct by stealing a byte from the tls hdr. Note that the current
header is still 2 bytes larger than the largest header we
support: AES-CBC with explicit IV. We set MBUF_PEXT_FLAG_ANON
when creating an unmapped mbuf in m_uiotombuf_nomap() (which is
the path that socket writes take), and we check for that flag in
ktls_encrypt() when looking for anon pages.
Reviewed by: jhb
Sponsored by: Netflix
Differential Revision:https://reviews.freebsd.org/D21796
Modified:
head/sys/kern/kern_mbuf.c
head/sys/kern/uipc_ktls.c
head/sys/kern/uipc_mbuf.c
head/sys/sys/mbuf.h
Modified: head/sys/kern/kern_mbuf.c
==
--- head/sys/kern/kern_mbuf.c Fri Sep 27 19:26:52 2019(r352815)
+++ head/sys/kern/kern_mbuf.c Fri Sep 27 20:08:19 2019(r352816)
@@ -1171,6 +1171,7 @@ mb_alloc_ext_pgs(int how, bool pkthdr, m_ext_free_t ex
ext_pgs->nrdy = 0;
ext_pgs->first_pg_off = 0;
ext_pgs->last_pg_len = 0;
+ ext_pgs->flags = 0;
ext_pgs->hdr_len = 0;
ext_pgs->trail_len = 0;
ext_pgs->tls = NULL;
Modified: head/sys/kern/uipc_ktls.c
==
--- head/sys/kern/uipc_ktls.c Fri Sep 27 19:26:52 2019(r352815)
+++ head/sys/kern/uipc_ktls.c Fri Sep 27 20:08:19 2019(r352816)
@@ -1363,7 +1363,7 @@ ktls_encrypt(struct mbuf_ext_pgs *pgs)
* (from sendfile), anonymous wired pages are
* allocated and assigned to the destination iovec.
*/
- is_anon = M_WRITABLE(m);
+ is_anon = (pgs->flags & MBUF_PEXT_FLAG_ANON) != 0;
off = pgs->first_pg_off;
for (i = 0; i < pgs->npgs; i++, off = 0) {
@@ -1416,6 +1416,9 @@ retry_page:
/* Use the basic free routine. */
m->m_ext.ext_free = mb_free_mext_pgs;
+
+ /* Pages are now writable. */
+ pgs->flags |= MBUF_PEXT_FLAG_ANON;
}
/*
Modified: head/sys/kern/uipc_mbuf.c
==
--- head/sys/kern/uipc_mbuf.c Fri Sep 27 19:26:52 2019(r352815)
+++ head/sys/kern/uipc_mbuf.c Fri Sep 27 20:08:19 2019(r352816)
@@ -1664,6 +1664,7 @@ m_uiotombuf_nomap(struct uio *uio, int how, int len, i
prev->m_next = mb;
prev = mb;
pgs = mb->m_ext.ext_pgs;
+ pgs->flags = MBUF_PEXT_FLAG_ANON;
needed = length = MIN(maxseg, total);
for (i = 0; needed > 0; i++, needed -= PAGE_SIZE) {
retry_page:
Modified: head/sys/sys/mbuf.h
==
--- head/sys/sys/mbuf.h Fri Sep 27 19:26:52 2019(r352815)
+++ head/sys/sys/mbuf.h Fri Sep 27 20:08:19 2019(r352816)
@@ -312,7 +312,7 @@ struct socket;
* - 21 (AES-CBC with explicit IV)
* - 13 (AES-GCM with 8 byte explicit IV)
*/
-#defineMBUF_PEXT_HDR_LEN 24
+#defineMBUF_PEXT_HDR_LEN 23
/*
* TLS records for TLS 1.0-1.2 can have the following maximum trailer
@@ -333,6 +333,8 @@ struct socket;
#defineMBUF_PEXT_MAX_BYTES
\
(MBUF_PEXT_MAX_PGS * PAGE_SIZE + MBUF_PEXT_HDR_LEN + MBUF_PEXT_TRAIL_LEN)
+#define MBUF_PEXT_FLAG_ANON1 /* Data can be encrypted in place. */
+
/*
* This struct is 256 bytes in size and is arranged so that the most
* common case (accessing the first 4
svn commit: r352815 - head/tools/tools/controlelf
Author: emaste Date: Fri Sep 27 19:26:52 2019 New Revision: 352815 URL: https://svnweb.freebsd.org/changeset/base/352815 Log: controlelf: update man page Some minor corrections, clarifications or rewording. Modified: head/tools/tools/controlelf/controlelf.1 Modified: head/tools/tools/controlelf/controlelf.1 == --- head/tools/tools/controlelf/controlelf.1Fri Sep 27 19:17:40 2019 (r352814) +++ head/tools/tools/controlelf/controlelf.1Fri Sep 27 19:26:52 2019 (r352815) @@ -26,12 +26,12 @@ .\" .\" $FreeBSD$ .\" -.Dd February 21, 2019 +.Dd September 27, 2019 .Dt CONTROLELF 1 .Os .Sh NAME .Nm controlelf -.Nd change an ELF binary's control features +.Nd change an ELF binary's feature control note .Sh SYNOPSIS .Nm .Op Fl h | Fl -help @@ -41,33 +41,33 @@ .Sh DESCRIPTION The .Nm -utility modifies an ELF binary to change it's control features. +utility modifies feature flags in the feature control note in an ELF binary. .Pp The options are as follows: .Bl -tag -width indent .It Fl h | Fl -help Print a usage message and exit. .It Fl l -Writes the list of all known ELF types to standard output. +List known ELF feature flags. .It Fl e Ar featurelist Edit features from the given comma separated list .Ar featurelist . .Ar featurelist -starts with one of the three modes: +starts with one of the three operations: .Dq Li + to turn on the features, .Dq Li - to turn off the features, .Dq Li = to only turn on the given features. -Followed by a comma separated features. +A comma separated list of feature names follows the operation. .El .Pp If -.Fl e , -option wasn't specified, +.Fl e +is not specified .Nm -will display the current control features for the files named by the arguments +displays the status of each feature in the ELF note in each .Ar . .Sh EXIT STATUS Exit status is 0 on success, and 1 if the command @@ -83,10 +83,12 @@ controlelf file controlelf -e +aslr file .Ed .Sh HISTORY -The .Nm -manual page first appeared in +first appeared in .Fx 13.0 . .Sh AUTHORS -This manual page was written by -.An Bora Ozarslan Mt borako.ozars...@gmail.com . +.Nm +was written by +.An Bora Ozarslan Mt borako.ozars...@gmail.com +under sponsorship from the +.Fx Foundation. ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352814 - in head/sys: kern net opencrypto sys
Author: gallatin
Date: Fri Sep 27 19:17:40 2019
New Revision: 352814
URL: https://svnweb.freebsd.org/changeset/base/352814
Log:
kTLS support for TLS 1.3
TLS 1.3 requires a few changes because 1.3 pretends to be 1.2
with a record type of application data. The "real" record type is
then included at the end of the user-supplied plaintext
data. This required adding a field to the mbuf_ext_pgs struct to
save the record type, and passing the real record type to the
sw_encrypt() ktls backend functions.
Reviewed by: jhb, hselasky
Sponsored by: Netflix
Differential Revision:D21801
Modified:
head/sys/kern/uipc_ktls.c
head/sys/net/iflib.c
head/sys/opencrypto/ktls_ocf.c
head/sys/sys/ktls.h
head/sys/sys/mbuf.h
Modified: head/sys/kern/uipc_ktls.c
==
--- head/sys/kern/uipc_ktls.c Fri Sep 27 19:14:03 2019(r352813)
+++ head/sys/kern/uipc_ktls.c Fri Sep 27 19:17:40 2019(r352814)
@@ -389,14 +389,14 @@ ktls_create_session(struct socket *so, struct tls_enab
if (en->tls_vmajor != TLS_MAJOR_VER_ONE)
return (EINVAL);
if (en->tls_vminor < TLS_MINOR_VER_ZERO ||
- en->tls_vminor > TLS_MINOR_VER_TWO)
+ en->tls_vminor > TLS_MINOR_VER_THREE)
return (EINVAL);
if (en->auth_key_len < 0 || en->auth_key_len > TLS_MAX_PARAM_SIZE)
return (EINVAL);
if (en->cipher_key_len < 0 || en->cipher_key_len > TLS_MAX_PARAM_SIZE)
return (EINVAL);
- if (en->iv_len < 0 || en->iv_len > TLS_MAX_PARAM_SIZE)
+ if (en->iv_len < 0 || en->iv_len > sizeof(tls->params.iv))
return (EINVAL);
/* All supported algorithms require a cipher key. */
@@ -425,7 +425,10 @@ ktls_create_session(struct socket *so, struct tls_enab
}
if (en->auth_key_len != 0)
return (EINVAL);
- if (en->iv_len != TLS_AEAD_GCM_LEN)
+ if ((en->tls_vminor == TLS_MINOR_VER_TWO &&
+ en->iv_len != TLS_AEAD_GCM_LEN) ||
+ (en->tls_vminor == TLS_MINOR_VER_THREE &&
+ en->iv_len != TLS_1_3_GCM_IV_LEN))
return (EINVAL);
break;
case CRYPTO_AES_CBC:
@@ -477,8 +480,22 @@ ktls_create_session(struct socket *so, struct tls_enab
tls->params.tls_hlen = sizeof(struct tls_record_layer);
switch (en->cipher_algorithm) {
case CRYPTO_AES_NIST_GCM_16:
- tls->params.tls_hlen += 8;
+ /*
+* TLS 1.2 uses a 4 byte implicit IV with an explicit 8 byte
+* nonce. TLS 1.3 uses a 12 byte implicit IV.
+*/
+ if (en->tls_vminor < TLS_MINOR_VER_THREE)
+ tls->params.tls_hlen += sizeof(uint64_t);
tls->params.tls_tlen = AES_GMAC_HASH_LEN;
+
+ /*
+* TLS 1.3 includes optional padding which we
+* do not support, and also puts the "real" record
+* type at the end of the encrypted data.
+*/
+ if (en->tls_vminor == TLS_MINOR_VER_THREE)
+ tls->params.tls_tlen += sizeof(uint8_t);
+
tls->params.tls_bs = 1;
break;
case CRYPTO_AES_CBC:
@@ -539,7 +556,6 @@ ktls_create_session(struct socket *so, struct tls_enab
* of the IV are generated in ktls_frame() and ktls_seq().
*/
if (en->iv_len != 0) {
- MPASS(en->iv_len <= sizeof(tls->params.iv));
tls->params.iv_len = en->iv_len;
error = copyin(en->iv, tls->params.iv, en->iv_len);
if (error)
@@ -1188,8 +1204,21 @@ ktls_frame(struct mbuf *top, struct ktls_session *tls,
/* Populate the TLS header. */
tlshdr = (void *)pgs->hdr;
tlshdr->tls_vmajor = tls->params.tls_vmajor;
- tlshdr->tls_vminor = tls->params.tls_vminor;
- tlshdr->tls_type = record_type;
+
+ /*
+* TLS 1.3 masquarades as TLS 1.2 with a record type
+* of TLS_RLTYPE_APP.
+*/
+ if (tls->params.tls_vminor == TLS_MINOR_VER_THREE &&
+ tls->params.tls_vmajor == TLS_MAJOR_VER_ONE) {
+ tlshdr->tls_vminor = TLS_MINOR_VER_TWO;
+ tlshdr->tls_type = TLS_RLTYPE_APP;
+ /* save the real record type for later */
+ pgs->record_type = record_type;
+ } else {
+ tlshdr->tls_vminor = tls->params.tls_vminor;
+ tlshdr->tls_type = record_type;
+ }
tlshdr->tls_length = htons(m->m_len - sizeof(*tlshdr));
/*
@@ -1365,7 +1394,8 @@ retry_page:
svn commit: r352813 - head/sys/kern
Author: mjg Date: Fri Sep 27 19:14:03 2019 New Revision: 352813 URL: https://svnweb.freebsd.org/changeset/base/352813 Log: cache: decrease ncnegfactor to 5 The current mechanism is bogus in several ways: - the limit is a percentage of total entries added, which means negative entries get evicted all the time even if there are plenty of resources - evicting code is almost not concurrent, which makes it unable to remove entries fast enough when doing something as simple as -j 104 buildworld - there is no support for performing mass removal if necessary Vast majority of negative entries never get any hits. Only evicting them when the filesystem demands it results in a significant growth of the namecache with almost no improvement in the hit ratio. Sample result about afer 90 minutes of poudriere -j 104: currentno evict % of the original numneg 219737 2013157916 numneghits 266711906 263544562 98 [1] [1] this may look funny but there is a certain dose of variation to the build The number was chosen as something which mostly eliminates spurious evictions during lighter workloads but still keeps the total at bay. Sponsored by: The FreeBSD Foundation Modified: head/sys/kern/vfs_cache.c Modified: head/sys/kern/vfs_cache.c == --- head/sys/kern/vfs_cache.c Fri Sep 27 19:13:22 2019(r352812) +++ head/sys/kern/vfs_cache.c Fri Sep 27 19:14:03 2019(r352813) @@ -200,7 +200,7 @@ static __read_mostly LIST_HEAD(nchashhead, namecache) static u_long __read_mostlynchash; /* size of hash table */ SYSCTL_ULONG(_debug, OID_AUTO, nchash, CTLFLAG_RD, , 0, "Size of namecache hash table"); -static u_long __read_mostlyncnegfactor = 12; /* ratio of negative entries */ +static u_long __read_mostlyncnegfactor = 5; /* ratio of negative entries */ SYSCTL_ULONG(_vfs, OID_AUTO, ncnegfactor, CTLFLAG_RW, , 0, "Ratio of negative namecache entries"); static u_long __exclusive_cache_line numneg; /* number of negative entries allocated */ ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352812 - head/sys/kern
Author: mjg
Date: Fri Sep 27 19:13:22 2019
New Revision: 352812
URL: https://svnweb.freebsd.org/changeset/base/352812
Log:
cache: stop requeuing negative entries on the hot list
Turns out it does not improve hit ratio, but it does come with a cost
induces stemming from dirtying hit entries.
Sample result: hit counts of evicted entries after 2 buildworlds
before:
value - Distribution - count
-1 | 0
0 |@180865
1 |@@@ 49150
2 |@@@ 19067
4 |@9825
8 |@7340
16 |@5952
32 |@5243
64 |@4446
128 | 3556
256 | 3035
512 | 1705
1024 | 1078
2048 | 365
4096 | 95
8192 | 34
16384 | 26
32768 | 23
65536 | 8
131072 | 6
262144 | 0
after:
value - Distribution - count
-1 | 0
0 |@184004
1 |@@ 47577
2 |@@@ 19446
4 |@10093
8 |@7470
16 |@5544
32 |@5475
64 |@5011
128 | 3451
256 | 3002
512 | 1729
1024 | 1086
2048 | 363
4096 | 86
8192 | 26
16384 | 25
32768 | 24
65536 | 7
131072 | 5
262144 | 0
Sponsored by: The FreeBSD Foundation
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Fri Sep 27 19:12:43 2019(r352811)
+++ head/sys/kern/vfs_cache.c Fri Sep 27 19:13:22 2019(r352812)
@@ -92,10 +92,10 @@ SDT_PROBE_DEFINE1(vfs, namecache, purge_negative, done
SDT_PROBE_DEFINE1(vfs, namecache, purgevfs, done, "struct mount *");
SDT_PROBE_DEFINE3(vfs, namecache, zap, done, "struct vnode *", "char *",
"struct vnode *");
-SDT_PROBE_DEFINE3(vfs, namecache, zap_negative, done, "struct vnode *",
-"char *", "int");
-SDT_PROBE_DEFINE3(vfs, namecache, shrink_negative, done, "struct vnode *",
-"char *", "int");
+SDT_PROBE_DEFINE2(vfs, namecache, zap_negative, done, "struct vnode *",
+"char *");
+SDT_PROBE_DEFINE2(vfs, namecache, shrink_negative, done, "struct vnode *",
+"char *");
/*
* This structure describes the elements in the cache of recent
@@ -109,7 +109,6 @@ struct namecache {
struct vnode *nc_dvp; /* vnode of parent of name */
union {
struct vnode *nu_vp; /* vnode the name refers to */
- u_int nu_neghits; /* negative entry hits */
} n_un;
u_char nc_flag;/* flag bits */
u_char nc_nlen;/* length of name */
@@ -132,7 +131,6 @@ struct namecache_ts {
};
#definenc_vp n_un.nu_vp
-#definenc_neghits n_un.nu_neghits
/*
* Flags in namecache.nc_flag
@@ -220,9 +218,6 @@ SYSCTL_UINT(_vfs, OID_AUTO, ncsizefactor, CTLFLAG_RW,
static u_int __read_mostly
svn commit: r352811 - head/sys/kern
Author: mjg
Date: Fri Sep 27 19:12:43 2019
New Revision: 352811
URL: https://svnweb.freebsd.org/changeset/base/352811
Log:
cache: make negative list shrinking a little bit concurrent
Continue protecting demotion from the hotlist and selection of the
target list with the ncneg_shrink_lock lock, but drop it before
relocking to zap the node.
While here count how many times we skipped shrinking due to the lock
being already taken.
Sponsored by: The FreeBSD Foundation
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Fri Sep 27 19:12:20 2019(r352810)
+++ head/sys/kern/vfs_cache.c Fri Sep 27 19:12:43 2019(r352811)
@@ -380,8 +380,6 @@ STATNODE_COUNTER(numfullpathfail2,
"Number of fullpath search errors (VOP_VPTOCNP failures)");
STATNODE_COUNTER(numfullpathfail4, "Number of fullpath search errors
(ENOMEM)");
STATNODE_COUNTER(numfullpathfound, "Number of successful fullpath calls");
-static long numneg_evicted; STATNODE_ULONG(numneg_evicted,
-"Number of negative entries evicted when adding a new entry");
STATNODE_COUNTER(zap_and_exit_bucket_relock_success,
"Number of successful removals after relocking");
static long zap_and_exit_bucket_fail; STATNODE_ULONG(zap_and_exit_bucket_fail,
@@ -392,6 +390,10 @@ static long cache_lock_vnodes_cel_3_failures;
STATNODE_ULONG(cache_lock_vnodes_cel_3_failures,
"Number of times 3-way vnode locking failed");
STATNODE_ULONG(numhotneg, "Number of hot negative entries");
+STATNODE_COUNTER(numneg_evicted,
+"Number of negative entries evicted when adding a new entry");
+STATNODE_COUNTER(shrinking_skipped,
+"Number of times shrinking was already in progress");
static void cache_zap_locked(struct namecache *ncp, bool neg_locked);
static int vn_fullpath1(struct thread *td, struct vnode *vp, struct vnode
*rdir,
@@ -809,8 +811,10 @@ cache_negative_zap_one(void)
struct rwlock *blp;
if (mtx_owner(_shrink_lock) != NULL ||
- !mtx_trylock(_shrink_lock))
+ !mtx_trylock(_shrink_lock)) {
+ counter_u64_add(shrinking_skipped, 1);
return;
+ }
mtx_lock(_hot.nl_lock);
ncp = TAILQ_FIRST(_hot.nl_list);
@@ -831,8 +835,9 @@ cache_negative_zap_one(void)
shrink_list_turn = 0;
if (ncp == NULL && shrink_list_turn == 0)
cache_negative_shrink_select(shrink_list_turn, , );
+ mtx_unlock(_shrink_lock);
if (ncp == NULL)
- goto out;
+ return;
MPASS(ncp->nc_flag & NCF_NEGATIVE);
dvlp = VP2VNODELOCK(ncp->nc_dvp);
@@ -845,19 +850,15 @@ cache_negative_zap_one(void)
if (ncp != ncp2 || dvlp != VP2VNODELOCK(ncp2->nc_dvp) ||
blp != NCP2BUCKETLOCK(ncp2) || !(ncp2->nc_flag & NCF_NEGATIVE)) {
ncp = NULL;
- goto out_unlock_all;
+ } else {
+ SDT_PROBE3(vfs, namecache, shrink_negative, done, ncp->nc_dvp,
+ ncp->nc_name, ncp->nc_neghits);
+ cache_zap_locked(ncp, true);
+ counter_u64_add(numneg_evicted, 1);
}
- SDT_PROBE3(vfs, namecache, shrink_negative, done, ncp->nc_dvp,
- ncp->nc_name, ncp->nc_neghits);
-
- cache_zap_locked(ncp, true);
- numneg_evicted++;
-out_unlock_all:
mtx_unlock(>nl_lock);
rw_wunlock(blp);
mtx_unlock(dvlp);
-out:
- mtx_unlock(_shrink_lock);
cache_free(ncp);
}
@@ -2016,6 +2017,8 @@ nchinit(void *dummy __unused)
numfullpathfail4 = counter_u64_alloc(M_WAITOK);
numfullpathfound = counter_u64_alloc(M_WAITOK);
zap_and_exit_bucket_relock_success = counter_u64_alloc(M_WAITOK);
+ numneg_evicted = counter_u64_alloc(M_WAITOK);
+ shrinking_skipped = counter_u64_alloc(M_WAITOK);
}
SYSINIT(vfs, SI_SUB_VFS, SI_ORDER_SECOND, nchinit, NULL);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352810 - head/sys/kern
Author: mjg
Date: Fri Sep 27 19:12:20 2019
New Revision: 352810
URL: https://svnweb.freebsd.org/changeset/base/352810
Log:
cache: stop recalculating upper limit each time a new entry is added
Sponsored by: The FreeBSD Foundation
Modified:
head/sys/kern/vfs_cache.c
Modified: head/sys/kern/vfs_cache.c
==
--- head/sys/kern/vfs_cache.c Fri Sep 27 19:07:11 2019(r352809)
+++ head/sys/kern/vfs_cache.c Fri Sep 27 19:12:20 2019(r352810)
@@ -214,7 +214,7 @@ SYSCTL_ULONG(_debug, OID_AUTO, numcache, CTLFLAG_RD, &
static u_long __exclusive_cache_line numcachehv;/* number of cache entries
with vnodes held */
SYSCTL_ULONG(_debug, OID_AUTO, numcachehv, CTLFLAG_RD, , 0,
"Number of namecache entries with vnodes held");
-u_int __read_mostlyncsizefactor = 2;
+u_int ncsizefactor = 2;
SYSCTL_UINT(_vfs, OID_AUTO, ncsizefactor, CTLFLAG_RW, , 0,
"Size factor for namecache");
static u_int __read_mostly ncpurgeminvnodes;
@@ -223,6 +223,7 @@ SYSCTL_UINT(_vfs, OID_AUTO, ncpurgeminvnodes, CTLFLAG_
static u_int __read_mostly ncneghitsrequeue = 8;
SYSCTL_UINT(_vfs, OID_AUTO, ncneghitsrequeue, CTLFLAG_RW, , 0,
"Number of hits to requeue a negative entry in the LRU list");
+static u_int __read_mostly ncsize; /* the size as computed on creation or
resizing */
struct nchstatsnchstats; /* cache effectiveness
statistics */
@@ -1713,7 +1714,7 @@ cache_enter_time(struct vnode *dvp, struct vnode *vp,
* Avoid blowout in namecache entries.
*/
lnumcache = atomic_fetchadd_long(, 1) + 1;
- if (__predict_false(lnumcache >= desiredvnodes * ncsizefactor)) {
+ if (__predict_false(lnumcache >= ncsize)) {
atomic_add_long(, -1);
return;
}
@@ -1969,6 +1970,7 @@ nchinit(void *dummy __unused)
NULL, NULL, NULL, NULL, UMA_ALIGNOF(struct namecache_ts),
UMA_ZONE_ZINIT);
+ ncsize = desiredvnodes * ncsizefactor;
nchashtbl = hashinit(desiredvnodes * 2, M_VFSCACHE, );
ncbuckethash = cache_roundup_2(mp_ncpus * mp_ncpus) - 1;
if (ncbuckethash < 7) /* arbitrarily chosen to avoid having one lock */
@@ -2024,8 +2026,10 @@ cache_changesize(int newmaxvnodes)
u_long new_nchash, old_nchash;
struct namecache *ncp;
uint32_t hash;
+ int newncsize;
int i;
+ newncsize = newmaxvnodes * ncsizefactor;
newmaxvnodes = cache_roundup_2(newmaxvnodes * 2);
if (newmaxvnodes < numbucketlocks)
newmaxvnodes = numbucketlocks;
@@ -2055,6 +2059,7 @@ cache_changesize(int newmaxvnodes)
LIST_INSERT_HEAD(NCHHASH(hash), ncp, nc_hash);
}
}
+ ncsize = newncsize;
cache_unlock_all_buckets();
cache_unlock_all_vnodes();
free(old_nchashtbl, M_VFSCACHE);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352809 - head/tools/tools/controlelf
Author: emaste
Date: Fri Sep 27 19:07:11 2019
New Revision: 352809
URL: https://svnweb.freebsd.org/changeset/base/352809
Log:
controlelf: exit with error if file endianness does not match host
We need to add support for cross-endian operation, but until that's done
just exit with an error rather than misbehaving.
Modified:
head/tools/tools/controlelf/controlelf.c
Modified: head/tools/tools/controlelf/controlelf.c
==
--- head/tools/tools/controlelf/controlelf.cFri Sep 27 18:49:13 2019
(r352808)
+++ head/tools/tools/controlelf/controlelf.cFri Sep 27 19:07:11 2019
(r352809)
@@ -30,6 +30,7 @@
#include
#include
+#include
#include
#include
@@ -70,6 +71,12 @@ static struct option long_opts[] = {
{ NULL, 0, NULL, 0 }
};
+#if BYTE_ORDER == LITTLE_ENDIAN
+#define SUPPORTED_ENDIAN ELFDATA2LSB
+#else
+#define SUPPORTED_ENDIAN ELFDATA2MSB
+#endif
+
int
main(int argc, char **argv)
{
@@ -140,6 +147,15 @@ main(int argc, char **argv)
if (gelf_getehdr(elf, ) == NULL) {
warnx("gelf_getehdr: %s", elf_errmsg(-1));
+ retval = 1;
+ goto fail;
+ }
+ /*
+* XXX need to support cross-endian operation, but for now
+* exit on error rather than misbehaving.
+*/
+ if (ehdr.e_ident[EI_DATA] != SUPPORTED_ENDIAN) {
+ warnx("file endianness must match host");
retval = 1;
goto fail;
}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352808 - head/tools/tools/controlelf
Author: emaste
Date: Fri Sep 27 18:49:13 2019
New Revision: 352808
URL: https://svnweb.freebsd.org/changeset/base/352808
Log:
controlelf: simplify feature string parsing
Also add error handling on failure to seek/write updated value.
Modified:
head/tools/tools/controlelf/controlelf.c
Modified: head/tools/tools/controlelf/controlelf.c
==
--- head/tools/tools/controlelf/controlelf.cFri Sep 27 18:43:36 2019
(r352807)
+++ head/tools/tools/controlelf/controlelf.cFri Sep 27 18:49:13 2019
(r352808)
@@ -56,7 +56,7 @@ static void usage(void);
struct ControlFeatures {
const char *alias;
- unsigned long featureVal;
+ unsigned long value;
const char *desc;
};
@@ -189,30 +189,19 @@ usage(void)
static bool
convert_to_feature_val(char *feature_str, u_int32_t *feature_val)
{
- char *feature_input, *feature;
+ char *feature;
int i, len;
u_int32_t input;
- bool add, set;
+ char operation;
- add = set = false;
input = 0;
-
- if (feature_str[0] == '+')
- add = true;
- else if (feature_str[0] == '=')
- set = true;
- else if (feature_str[0] != '-') {
- warnx("'%c' not an operator - use '+', '-', '='",
- feature_str[0]);
- return (false);
- }
-
- feature_input = feature_str + 1;
+ operation = *feature_str;
+ feature_str++;
len = nitems(featurelist);
- while ((feature = strsep(_input, ",")) != NULL) {
+ while ((feature = strsep(_str, ",")) != NULL) {
for (i = 0; i < len; ++i) {
if (strcmp(featurelist[i].alias, feature) == 0) {
- input |= featurelist[i].featureVal;
+ input |= featurelist[i].value;
break;
}
}
@@ -222,12 +211,16 @@ convert_to_feature_val(char *feature_str, u_int32_t *f
}
}
- if (add) {
+ if (operation == '+') {
*feature_val |= input;
- } else if (set) {
+ } else if (operation == '=') {
*feature_val = input;
+ } else if (operation == '-') {
+ *feature_val &= ~input;
} else {
- *feature_val -= (*feature_val) & input;
+ warnx("'%c' not an operator - use '+', '-', '='",
+ feature_str[0]);
+ return (false);
}
return (true);
}
@@ -246,8 +239,12 @@ edit_file_features(Elf *elf, int phcount, int fd, char
if (!convert_to_feature_val(val, ))
return (false);
- lseek(fd, off, SEEK_SET);
- write(fd, , sizeof(u_int32_t));
+ if (lseek(fd, off, SEEK_SET) == -1 ||
+ write(fd, , sizeof(features)) <
+ (ssize_t)sizeof(features)) {
+ warnx("error writing feature value");
+ return (false);
+ }
return (true);
}
@@ -277,7 +274,7 @@ print_file_features(Elf *elf, int phcount, int fd, cha
printf("%s\t\t'%s' is ", featurelist[i].alias,
featurelist[i].desc);
- if ((featurelist[i].featureVal & features) == 0)
+ if ((featurelist[i].value & features) == 0)
printf("un");
printf("set.\n");
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352795 - head/lib/libc/sys
On Fri, Sep 27, 2019 at 08:32:20PM +0200, Mateusz Guzik wrote: > On 9/27/19, Warner Losh wrote: > > Author: imp > > Date: Fri Sep 27 16:11:47 2019 > > New Revision: 352795 > > URL: https://svnweb.freebsd.org/changeset/base/352795 > > > > Log: > > Document varadic args as int, since you can't have short varadic args > > (they are > > promoted to ints). > > > > - `mode_t` is `uint16_t` (`sys/sys/_types.h`) > > - `openat` takes variadic args > > - variadic args cannot be 16-bit, and indeed the code uses int > > - the manpage currently kinda implies the argument is 16-bit by saying > > `mode_t` > > > > Prompted by Rust things: https://github.com/tailhook/openat/issues/21 > > Submitted by: Greg V at unrelenting > > Differential Revision: https://reviews.freebsd.org/D21816 > > > > Modified: > > head/lib/libc/sys/mq_open.2 > > head/lib/libc/sys/open.2 > > > > Modified: head/lib/libc/sys/mq_open.2 > > == > > --- head/lib/libc/sys/mq_open.2 Fri Sep 27 15:28:30 2019 > > (r352794) > > +++ head/lib/libc/sys/mq_open.2 Fri Sep 27 16:11:47 2019 > > (r352795) > > @@ -133,7 +133,7 @@ Create a message queue. > > It requires two additional arguments: > > .Fa mode , > > which is of type > > -.Vt mode_t , > > +.Vt int , > > and > > .Fa attr , > > which is a pointer to an > > > > Modified: head/lib/libc/sys/open.2 > > == > > --- head/lib/libc/sys/open.2Fri Sep 27 15:28:30 2019 > > (r352794) > > +++ head/lib/libc/sys/open.2Fri Sep 27 16:11:47 2019 > > (r352795) > > @@ -61,7 +61,7 @@ In this case > > and > > .Fn openat > > require an additional argument > > -.Fa "mode_t mode" , > > +.Fa "int mode" , > > and the file is created with mode > > .Fa mode > > as described in > > > > But opengroup says it is mode_t. Perhaps it is mode_t which needs > to be changed? Yes, users must pass mode_t, and the man page is written for users. Implementation needs to be aware of the implicit promotion and handle it accordingly. In theory, mode_t might be wider than int. ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352807 - in head/sys: amd64/amd64 amd64/vmm arm/arm arm64/arm64 i386/i386 kern mips/mips powerpc/powerpc riscv/riscv sparc64/sparc64 sys vm
Author: kib
Date: Fri Sep 27 18:43:36 2019
New Revision: 352807
URL: https://svnweb.freebsd.org/changeset/base/352807
Log:
Improve MD page fault handlers.
Centralize calculation of signal and ucode delivered on unhandled page
fault in new function vm_fault_trap(). MD trap_pfault() now almost
always uses the signal numbers and error codes calculated in
consistent MI way.
This introduces the protection fault compatibility sysctls to all
non-x86 architectures which did not have that bug, but apparently they
were already much more wrong in selecting delivered signals on
protection violations.
Change the delivered signal for accesses to mapped area after the
backing object was truncated. According to POSIX description for
mmap(2):
The system shall always zero-fill any partial page at the end of an
object. Further, the system shall never write out any modified
portions of the last page of an object which are beyond its
end. References within the address range starting at pa and
continuing for len bytes to whole pages following the end of an
object shall result in delivery of a SIGBUS signal.
An implementation may generate SIGBUS signals when a reference
would cause an error in the mapped object, such as out-of-space
condition.
Adjust according to the description, keeping the existing
compatibility code for SIGSEGV/SIGBUS on protection failures.
For situations where kernel cannot handle page fault due to resource
limit enforcement, SIGBUS with a new error code BUS_OBJERR is
delivered. Also, provide a new error code SEGV_PKUERR for SIGSEGV on
amd64 due to protection key access violation.
vm_fault_hold() is renamed to vm_fault(). Fixed some nits in
trap_pfault()s like mis-interpreting Mach errors as errnos. Removed
unneeded truncations of the fault addresses reported by hardware.
PR: 211924
Reviewed by: alc
Discussed with: jilles, markj
Sponsored by: The FreeBSD Foundation
MFC after:1 week
Differential revision:https://reviews.freebsd.org/D21566
Modified:
head/sys/amd64/amd64/trap.c
head/sys/amd64/vmm/vmm.c
head/sys/arm/arm/trap-v4.c
head/sys/arm/arm/trap-v6.c
head/sys/arm64/arm64/trap.c
head/sys/i386/i386/trap.c
head/sys/kern/sys_process.c
head/sys/mips/mips/trap.c
head/sys/powerpc/powerpc/trap.c
head/sys/riscv/riscv/trap.c
head/sys/sparc64/sparc64/trap.c
head/sys/sys/signal.h
head/sys/vm/vm_extern.h
head/sys/vm/vm_fault.c
head/sys/vm/vm_map.c
head/sys/vm/vm_param.h
Modified: head/sys/amd64/amd64/trap.c
==
--- head/sys/amd64/amd64/trap.c Fri Sep 27 18:39:05 2019(r352806)
+++ head/sys/amd64/amd64/trap.c Fri Sep 27 18:43:36 2019(r352807)
@@ -111,7 +111,7 @@ void __noinline trap(struct trapframe *frame);
void trap_check(struct trapframe *frame);
void dblfault_handler(struct trapframe *frame);
-static int trap_pfault(struct trapframe *, int);
+static int trap_pfault(struct trapframe *, bool, int *, int *);
static void trap_fatal(struct trapframe *, vm_offset_t);
#ifdef KDTRACE_HOOKS
static bool trap_user_dtrace(struct trapframe *,
@@ -155,10 +155,6 @@ static const char *const trap_msg[] = {
[T_DTRACE_RET] ="DTrace pid return trap",
};
-static int prot_fault_translation;
-SYSCTL_INT(_machdep, OID_AUTO, prot_fault_translation, CTLFLAG_RWTUN,
-_fault_translation, 0,
-"Select signal to deliver on protection fault");
static int uprintf_signal;
SYSCTL_INT(_machdep, OID_AUTO, uprintf_signal, CTLFLAG_RWTUN,
_signal, 0,
@@ -192,14 +188,11 @@ trap(struct trapframe *frame)
struct thread *td;
struct proc *p;
register_t addr, dr6;
- int signo, ucode;
+ int pf, signo, ucode;
u_int type;
td = curthread;
p = td->td_proc;
- signo = 0;
- ucode = 0;
- addr = 0;
dr6 = 0;
VM_CNT_INC(v_trap);
@@ -345,47 +338,18 @@ trap(struct trapframe *frame)
case T_PAGEFLT: /* page fault */
/*
-* Emulator can take care about this trap?
+* Can emulator handle this trap?
*/
if (*p->p_sysent->sv_trap != NULL &&
(*p->p_sysent->sv_trap)(td) == 0)
return;
- addr = frame->tf_addr;
- signo = trap_pfault(frame, TRUE);
- if (signo == -1)
+ pf = trap_pfault(frame, true, , );
+ if (pf == -1)
return;
- if (signo == 0)
+ if (pf == 0)
goto userret;
- if (signo == SIGSEGV) {
- ucode =
svn commit: r352806 - head/tools/tools/controlelf
Author: emaste
Date: Fri Sep 27 18:39:05 2019
New Revision: 352806
URL: https://svnweb.freebsd.org/changeset/base/352806
Log:
controlelf: tidy up option parsing
Sponsored by: The FreeBSD Foundation
Modified:
head/tools/tools/controlelf/controlelf.c
Modified: head/tools/tools/controlelf/controlelf.c
==
--- head/tools/tools/controlelf/controlelf.cFri Sep 27 17:28:25 2019
(r352805)
+++ head/tools/tools/controlelf/controlelf.cFri Sep 27 18:39:05 2019
(r352806)
@@ -65,7 +65,7 @@ static struct ControlFeatures featurelist[] = {
{ "protmax",NT_FREEBSD_FCTL_PROTMAX_DISABLE,"Disable implicit
PROT_MAX" },
};
-static struct option controlelf_longopts[] = {
+static struct option long_opts[] = {
{ "help", no_argument,NULL, 'h' },
{ NULL, 0, NULL, 0 }
};
@@ -76,38 +76,36 @@ main(int argc, char **argv)
GElf_Ehdr ehdr;
Elf *elf;
Elf_Kind kind;
- int ch, fd, listed, editfeatures, retval;
+ int ch, fd, editfeatures, retval;
char *features;
+ bool lflag;
- listed = 0;
+ lflag = 0;
editfeatures = 0;
retval = 0;
if (elf_version(EV_CURRENT) == EV_NONE)
errx(EXIT_FAILURE, "elf_version error");
- while ((ch = getopt_long(argc, argv, "hle:", controlelf_longopts,
- NULL)) != -1) {
+ while ((ch = getopt_long(argc, argv, "hle:", long_opts, NULL)) != -1) {
switch (ch) {
case 'l':
print_features();
- listed = 1;
+ lflag = true;
break;
case 'e':
features = optarg;
editfeatures = 1;
break;
case 'h':
- usage();
- break;
default:
usage();
}
}
argc -= optind;
argv += optind;
- if (!argc) {
- if (listed)
+ if (argc == 0) {
+ if (lflag)
exit(0);
else {
warnx("no file(s) specified");
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352795 - head/lib/libc/sys
On 9/27/19, Warner Losh wrote: > Author: imp > Date: Fri Sep 27 16:11:47 2019 > New Revision: 352795 > URL: https://svnweb.freebsd.org/changeset/base/352795 > > Log: > Document varadic args as int, since you can't have short varadic args > (they are > promoted to ints). > > - `mode_t` is `uint16_t` (`sys/sys/_types.h`) > - `openat` takes variadic args > - variadic args cannot be 16-bit, and indeed the code uses int > - the manpage currently kinda implies the argument is 16-bit by saying > `mode_t` > > Prompted by Rust things: https://github.com/tailhook/openat/issues/21 > Submitted by: Greg V at unrelenting > Differential Revision: https://reviews.freebsd.org/D21816 > > Modified: > head/lib/libc/sys/mq_open.2 > head/lib/libc/sys/open.2 > > Modified: head/lib/libc/sys/mq_open.2 > == > --- head/lib/libc/sys/mq_open.2 Fri Sep 27 15:28:30 2019 > (r352794) > +++ head/lib/libc/sys/mq_open.2 Fri Sep 27 16:11:47 2019 > (r352795) > @@ -133,7 +133,7 @@ Create a message queue. > It requires two additional arguments: > .Fa mode , > which is of type > -.Vt mode_t , > +.Vt int , > and > .Fa attr , > which is a pointer to an > > Modified: head/lib/libc/sys/open.2 > == > --- head/lib/libc/sys/open.2 Fri Sep 27 15:28:30 2019(r352794) > +++ head/lib/libc/sys/open.2 Fri Sep 27 16:11:47 2019(r352795) > @@ -61,7 +61,7 @@ In this case > and > .Fn openat > require an additional argument > -.Fa "mode_t mode" , > +.Fa "int mode" , > and the file is created with mode > .Fa mode > as described in > But opengroup says it is mode_t. Perhaps it is mode_t which needs to be changed? -- Mateusz Guzik ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352805 - head/tools/tools/controlelf
Author: emaste
Date: Fri Sep 27 17:28:25 2019
New Revision: 352805
URL: https://svnweb.freebsd.org/changeset/base/352805
Log:
controlelf: add protmax control
Sponsored by: The FreeBSD Foundation
Modified:
head/tools/tools/controlelf/controlelf.c
Modified: head/tools/tools/controlelf/controlelf.c
==
--- head/tools/tools/controlelf/controlelf.cFri Sep 27 17:11:21 2019
(r352804)
+++ head/tools/tools/controlelf/controlelf.cFri Sep 27 17:28:25 2019
(r352805)
@@ -62,6 +62,7 @@ struct ControlFeatures {
static struct ControlFeatures featurelist[] = {
{ "aslr", NT_FREEBSD_FCTL_ASLR_DISABLE, "Disable ASLR" },
+ { "protmax",NT_FREEBSD_FCTL_PROTMAX_DISABLE,"Disable implicit
PROT_MAX" },
};
static struct option controlelf_longopts[] = {
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352804 - head/usr.bin/top
Author: dim Date: Fri Sep 27 17:11:21 2019 New Revision: 352804 URL: https://svnweb.freebsd.org/changeset/base/352804 Log: Correct the final argument name in the top(1) manpage. The description talks about 'number', while the final argument was 'count'. Since 'count' is already used for the count of displays, change the final argument name to 'number'. MFC after:3 days Modified: head/usr.bin/top/top.1 Modified: head/usr.bin/top/top.1 == --- head/usr.bin/top/top.1 Fri Sep 27 16:57:32 2019(r352803) +++ head/usr.bin/top/top.1 Fri Sep 27 17:11:21 2019(r352804) @@ -15,7 +15,7 @@ .Op Fl s Ar time .Op Fl o Ar field .Op Fl p Ar pid -.Op Ar count +.Op Ar number .Sh DESCRIPTION .Nm displays the top ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352803 - head/tools/tools/controlelf
Author: emaste
Date: Fri Sep 27 16:57:32 2019
New Revision: 352803
URL: https://svnweb.freebsd.org/changeset/base/352803
Log:
controlelf: some style(9) cleanup
Submitted by: clang-format
Modified:
head/tools/tools/controlelf/controlelf.c
Modified: head/tools/tools/controlelf/controlelf.c
==
--- head/tools/tools/controlelf/controlelf.cFri Sep 27 16:46:08 2019
(r352802)
+++ head/tools/tools/controlelf/controlelf.cFri Sep 27 16:57:32 2019
(r352803)
@@ -14,7 +14,7 @@
* 2. Redistributions in binary form must reproduce the above copyright
*notice, this list of conditions and the following disclaimer in the
*documentation and/or other materials provided with the distribution.
- *
+ *
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
@@ -47,11 +47,11 @@
__FBSDID("$FreeBSD$");
-static bool convert_to_feature_val(char*, u_int32_t*);
-static bool edit_file_features(Elf *, int, int, char*);
+static bool convert_to_feature_val(char *, u_int32_t *);
+static bool edit_file_features(Elf *, int, int, char *);
static bool get_file_features(Elf *, int, int, u_int32_t *, u_int64_t *);
static void print_features(void);
-static bool print_file_features(Elf *, int, int, char*);
+static bool print_file_features(Elf *, int, int, char *);
static void usage(void);
struct ControlFeatures {
@@ -61,12 +61,12 @@ struct ControlFeatures {
};
static struct ControlFeatures featurelist[] = {
- { "aslr", NT_FREEBSD_FCTL_ASLR_DISABLE,
"Disable ASLR" }
+ { "aslr", NT_FREEBSD_FCTL_ASLR_DISABLE, "Disable ASLR" },
};
static struct option controlelf_longopts[] = {
- { "help", no_argument,NULL, 'h' },
- { NULL, 0, NULL, 0 }
+ { "help", no_argument,NULL, 'h' },
+ { NULL, 0, NULL, 0 }
};
int
@@ -117,8 +117,8 @@ main(int argc, char **argv)
while (argc) {
elf = NULL;
- if ((fd = open(argv[0], editfeatures ? O_RDWR :
- O_RDONLY, 0)) < 0) {
+ if ((fd = open(argv[0],
+ editfeatures ? O_RDWR : O_RDONLY, 0)) < 0) {
warn("error opening file %s", argv[0]);
retval = 1;
goto fail;
@@ -134,8 +134,7 @@ main(int argc, char **argv)
if (kind == ELF_K_AR)
warnx("file '%s' is an archive", argv[0]);
else
- warnx("file '%s' is not an ELF file",
- argv[0]);
+ warnx("file '%s' is not an ELF file", argv[0]);
retval = 1;
goto fail;
}
@@ -147,11 +146,13 @@ main(int argc, char **argv)
}
if (!editfeatures) {
- if (!print_file_features(elf, ehdr.e_phnum, fd,
argv[0])) {
+ if (!print_file_features(elf, ehdr.e_phnum, fd,
+ argv[0])) {
retval = 1;
goto fail;
}
- } else if (!edit_file_features(elf, ehdr.e_phnum, fd,
features)) {
+ } else if (!edit_file_features(elf, ehdr.e_phnum, fd,
+ features)) {
retval = 1;
goto fail;
}
@@ -169,7 +170,8 @@ fail:
return (retval);
}
-#defineUSAGE_MESSAGE "\
+#define USAGE_MESSAGE \
+ "\
Usage: %s [options] file...\n\
Set or display the control features for an ELF object.\n\n\
Supported options are:\n\
@@ -186,7 +188,7 @@ usage(void)
}
static bool
-convert_to_feature_val(char* feature_str, u_int32_t* feature_val)
+convert_to_feature_val(char *feature_str, u_int32_t *feature_val)
{
char *feature_input, *feature;
int i, len;
@@ -262,7 +264,8 @@ print_features(void)
}
static bool
-print_file_features(Elf *elf, int phcount, int fd, char *filename) {
+print_file_features(Elf *elf, int phcount, int fd, char *filename)
+{
u_int32_t features;
unsigned long i;
@@ -284,7 +287,8 @@ print_file_features(Elf *elf, int phcount, int fd, cha
}
static bool
-get_file_features(Elf *elf, int phcount, int fd, u_int32_t *features,
u_int64_t *off)
+get_file_features(Elf *elf, int phcount, int fd, u_int32_t *features,
+u_int64_t *off)
{
GElf_Phdr phdr;
Elf_Note note;
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe,
svn commit: r352802 - head/sys/vm
Author: markj
Date: Fri Sep 27 16:46:08 2019
New Revision: 352802
URL: https://svnweb.freebsd.org/changeset/base/352802
Log:
Fix a race in vm_page_swapqueue().
vm_page_swapqueue() atomically transitions a page between queues. To do
so, it must hold the page queue lock for the old queue. However, once
the queue index has been updated, the queue lock no longer protects the
page's queue state. Thus, we must speculatively remove the page from
the old queue before committing the queue state update, and roll back if
the update fails.
Reported and tested by: pho
Reviewed by: kib
Sponsored by: Intel, Netflix
Differential Revision:https://reviews.freebsd.org/D21791
Modified:
head/sys/vm/vm_page.c
Modified: head/sys/vm/vm_page.c
==
--- head/sys/vm/vm_page.c Fri Sep 27 16:44:29 2019(r352801)
+++ head/sys/vm/vm_page.c Fri Sep 27 16:46:08 2019(r352802)
@@ -3462,27 +3462,58 @@ void
vm_page_swapqueue(vm_page_t m, uint8_t oldq, uint8_t newq)
{
struct vm_pagequeue *pq;
+ vm_page_t next;
+ bool queued;
KASSERT(oldq < PQ_COUNT && newq < PQ_COUNT && oldq != newq,
("vm_page_swapqueue: invalid queues (%d, %d)", oldq, newq));
- KASSERT((m->oflags & VPO_UNMANAGED) == 0,
- ("vm_page_swapqueue: page %p is unmanaged", m));
vm_page_assert_locked(m);
- /*
-* Atomically update the queue field and set PGA_REQUEUE while
-* ensuring that PGA_DEQUEUE has not been set.
-*/
pq = _pagequeue_domain(m)->vmd_pagequeues[oldq];
vm_pagequeue_lock(pq);
- if (!vm_page_pqstate_cmpset(m, oldq, newq, PGA_DEQUEUE, PGA_REQUEUE)) {
+
+ /*
+* The physical queue state might change at any point before the page
+* queue lock is acquired, so we must verify that we hold the correct
+* lock before proceeding.
+*/
+ if (__predict_false(m->queue != oldq)) {
vm_pagequeue_unlock(pq);
return;
}
- if ((m->aflags & PGA_ENQUEUED) != 0) {
- vm_pagequeue_remove(pq, m);
+
+ /*
+* Once the queue index of the page changes, there is nothing
+* synchronizing with further updates to the physical queue state.
+* Therefore we must remove the page from the queue now in anticipation
+* of a successful commit, and be prepared to roll back.
+*/
+ if (__predict_true((m->aflags & PGA_ENQUEUED) != 0)) {
+ next = TAILQ_NEXT(m, plinks.q);
+ TAILQ_REMOVE(>pq_pl, m, plinks.q);
vm_page_aflag_clear(m, PGA_ENQUEUED);
+ queued = true;
+ } else {
+ queued = false;
}
+
+ /*
+* Atomically update the queue field and set PGA_REQUEUE while
+* ensuring that PGA_DEQUEUE has not been set.
+*/
+ if (__predict_false(!vm_page_pqstate_cmpset(m, oldq, newq, PGA_DEQUEUE,
+ PGA_REQUEUE))) {
+ if (queued) {
+ vm_page_aflag_set(m, PGA_ENQUEUED);
+ if (next != NULL)
+ TAILQ_INSERT_BEFORE(next, m, plinks.q);
+ else
+ TAILQ_INSERT_TAIL(>pq_pl, m, plinks.q);
+ }
+ vm_pagequeue_unlock(pq);
+ return;
+ }
+ vm_pagequeue_cnt_dec(pq);
vm_pagequeue_unlock(pq);
vm_page_pqbatch_submit(m, newq);
}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352801 - head/tools/tools/controlelf
Author: emaste Date: Fri Sep 27 16:44:29 2019 New Revision: 352801 URL: https://svnweb.freebsd.org/changeset/base/352801 Log: controlelf: install standard BSD 2 clause license Reported by: kaktus Sponsored by: The FreeBSD Foundation Modified: head/tools/tools/controlelf/controlelf.c Modified: head/tools/tools/controlelf/controlelf.c == --- head/tools/tools/controlelf/controlelf.cFri Sep 27 16:41:34 2019 (r352800) +++ head/tools/tools/controlelf/controlelf.cFri Sep 27 16:44:29 2019 (r352801) @@ -1,5 +1,5 @@ /*- - * SPDX-License-Identifier: BSD-3-Clause + * SPDX-License-Identifier: BSD-2-Clause-FreeBSD * * Copyright (c) 2019 The FreeBSD Foundation. * @@ -10,22 +10,22 @@ * modification, are permitted provided that the following conditions * are met: * 1. Redistributions of source code must retain the above copyright - *notice, this list of conditions and the following disclaimer - *in this position and unchanged. + *notice, this list of conditions and the following disclaimer. * 2. Redistributions in binary form must reproduce the above copyright *notice, this list of conditions and the following disclaimer in the *documentation and/or other materials provided with the distribution. - * - * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR - * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES - * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. - * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT, - * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT - * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, - * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY - * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT - * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF - * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND + * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE + * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE + * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE + * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL + * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS + * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) + * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT + * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY + * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF + * SUCH DAMAGE. */ #include ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352800 - head/sys/vm
Author: markj
Date: Fri Sep 27 16:41:34 2019
New Revision: 352800
URL: https://svnweb.freebsd.org/changeset/base/352800
Log:
Fix object locking in vm_object_unwire() after r352174.
Now, vm_page_busy_sleep() expects the page's object to be locked.
vm_object_unwire() does some unusual lazy locking of the object chain
and keeps objects locked until a busy page is encountered or the loop
terminates. When a busy page is encountered, rather than unlocking all
but the "bottom-level" object, we must instead skip the object to which
"tm" belongs.
Reported and tested by: pho
Reviewed by: kib
Discussed with: jeff
Sponsored by: Intel, Netflix
Differential Revision:https://reviews.freebsd.org/D21790
Modified:
head/sys/vm/vm_object.c
Modified: head/sys/vm/vm_object.c
==
--- head/sys/vm/vm_object.c Fri Sep 27 16:35:08 2019(r352799)
+++ head/sys/vm/vm_object.c Fri Sep 27 16:41:34 2019(r352800)
@@ -2169,10 +2169,11 @@ again:
m = TAILQ_NEXT(m, listq);
}
if (vm_page_xbusied(tm)) {
- for (tobject = object; locked_depth > 1;
+ for (tobject = object; locked_depth >= 1;
locked_depth--) {
t1object = tobject->backing_object;
- VM_OBJECT_RUNLOCK(tobject);
+ if (tm->object != tobject)
+ VM_OBJECT_RUNLOCK(tobject);
tobject = t1object;
}
vm_page_busy_sleep(tm, "unwbo", true);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352757 - in head: lib/libc/sys lib/libregex/tests stand/lua tests/sys/kern
> Author: kevans > Date: Thu Sep 26 16:19:22 2019 > New Revision: 352757 > URL: https://svnweb.freebsd.org/changeset/base/352757 > > Log: > Further normalize copyright notices > > - s/C/c/ where I've been inconsistent about it > - +SPDX tags > - Remove "All rights reserved" where possible > > Requested by: rgrimes (all rights reserved) Thank you Kyle! > Modified: > head/lib/libc/sys/shm_open.c > head/lib/libregex/tests/libregex_test.sh > head/stand/lua/config.lua > head/stand/lua/menu.lua > head/stand/lua/password.lua Regards, -- Rod Grimes rgri...@freebsd.org ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352799 - head/tools/tools/controlelf
Author: emaste
Date: Fri Sep 27 16:35:08 2019
New Revision: 352799
URL: https://svnweb.freebsd.org/changeset/base/352799
Log:
controlelf: clean up warnings
- use explicit ELF note name when not found
- no trailing . on warnings
- no \n
Sponsored by: The FreeBSD Foundation
Modified:
head/tools/tools/controlelf/controlelf.c
Modified: head/tools/tools/controlelf/controlelf.c
==
--- head/tools/tools/controlelf/controlelf.cFri Sep 27 16:32:44 2019
(r352798)
+++ head/tools/tools/controlelf/controlelf.cFri Sep 27 16:35:08 2019
(r352799)
@@ -132,9 +132,9 @@ main(int argc, char **argv)
if ((kind = elf_kind(elf)) != ELF_K_ELF) {
if (kind == ELF_K_AR)
- warnx("file '%s' is an archive.", argv[0]);
+ warnx("file '%s' is an archive", argv[0]);
else
- warnx("file '%s' is not an ELF file.",
+ warnx("file '%s' is not an ELF file",
argv[0]);
retval = 1;
goto fail;
@@ -201,7 +201,7 @@ convert_to_feature_val(char* feature_str, u_int32_t* f
else if (feature_str[0] == '=')
set = true;
else if (feature_str[0] != '-') {
- warnx("'%c' is not an operator. Use instead '+', '-', '='.",
+ warnx("'%c' not an operator - use '+', '-', '='",
feature_str[0]);
return (false);
}
@@ -216,7 +216,7 @@ convert_to_feature_val(char* feature_str, u_int32_t* f
}
}
if (i == len) {
- warnx("%s is not a valid feature.", feature);
+ warnx("%s is not a valid feature", feature);
return (false);
}
}
@@ -238,7 +238,7 @@ edit_file_features(Elf *elf, int phcount, int fd, char
u_int64_t off;
if (!get_file_features(elf, phcount, fd, , )) {
- warnx("No control features note on the file.\n");
+ warnx("NT_FREEBSD_FEATURE_CTL note not found");
return (false);
}
@@ -327,7 +327,7 @@ get_file_features(Elf *elf, int phcount, int fd, u_int
namesz = roundup2(note.n_namesz, 4);
name = malloc(namesz);
if (name == NULL) {
- warn("malloc() failed.\n");
+ warn("malloc() failed.");
return (false);
}
descsz = roundup2(note.n_descsz, 4);
@@ -339,7 +339,7 @@ get_file_features(Elf *elf, int phcount, int fd, u_int
note.n_type != NT_FREEBSD_FEATURE_CTL) {
/* Not the right note. Skip the description */
if (lseek(fd, descsz, SEEK_CUR) < 0) {
- warn("lseek() failed.\n");
+ warn("lseek() failed.");
free(name);
return (false);
}
@@ -360,7 +360,7 @@ get_file_features(Elf *elf, int phcount, int fd, u_int
* descriptor. This should respect descsz.
*/
if (note.n_descsz > sizeof(u_int32_t))
- warnx("Feature note is bigger than expected.");
+ warnx("Feature note is bigger than expected");
read(fd, features, sizeof(u_int32_t));
if (off != NULL)
*off = phdr.p_offset + read_total;
@@ -369,6 +369,6 @@ get_file_features(Elf *elf, int phcount, int fd, u_int
}
}
- warnx("Couldn't find a note header with control feature note.");
+ warnx("NT_FREEBSD_FEATURE_CTL note not found");
return (false);
}
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352798 - in head/sys: dev/nvdimm modules/nvdimm
Author: cem
Date: Fri Sep 27 16:32:44 2019
New Revision: 352798
URL: https://svnweb.freebsd.org/changeset/base/352798
Log:
nvdimm(4): Extract ACPI root bus driver
No functional change intended.
The intent is to add a "legacy" e820 pmem newbus bus for nvdimm device in a
subsequent revision, and it's a little more clear if the parent buses get
independent source files.
Quite a lot of ACPI-specific logic is left in nvdimm.c; disentangling that
is a much larger change (and probably not especially useful).
Reviewed by: kib
Sponsored by: Dell EMC Isilon
Differential Revision:https://reviews.freebsd.org/D21813
Added:
head/sys/dev/nvdimm/nvdimm_acpi.c
- copied, changed from r352792, head/sys/dev/nvdimm/nvdimm.c
Modified:
head/sys/dev/nvdimm/nvdimm.c
head/sys/dev/nvdimm/nvdimm_var.h
head/sys/modules/nvdimm/Makefile
Modified: head/sys/dev/nvdimm/nvdimm.c
==
--- head/sys/dev/nvdimm/nvdimm.cFri Sep 27 16:27:52 2019
(r352797)
+++ head/sys/dev/nvdimm/nvdimm.cFri Sep 27 16:32:44 2019
(r352798)
@@ -43,10 +43,12 @@ __FBSDID("$FreeBSD$");
#include
#include
#include
+
#include
#include
#include
#include
+
#include
#define _COMPONENT ACPI_OEM
@@ -59,7 +61,6 @@ static struct uuid intel_nvdimm_dsm_uuid =
#define INTEL_NVDIMM_DSM_GET_LABEL_DATA 5
static devclass_t nvdimm_devclass;
-static devclass_t nvdimm_root_devclass;
MALLOC_DEFINE(M_NVDIMM, "nvdimm", "NVDIMM driver memory");
static int
@@ -325,8 +326,7 @@ nvdimm_attach(device_t dev)
nv = device_get_softc(dev);
handle = nvdimm_root_get_acpi_handle(dev);
- if (handle == NULL)
- return (EINVAL);
+ MPASS(handle != NULL);
nv->nv_dev = dev;
nv->nv_handle = nvdimm_root_get_device_handle(dev);
@@ -380,204 +380,6 @@ nvdimm_resume(device_t dev)
return (0);
}
-static ACPI_STATUS
-find_dimm(ACPI_HANDLE handle, UINT32 nesting_level, void *context,
-void **return_value)
-{
- ACPI_DEVICE_INFO *device_info;
- ACPI_STATUS status;
-
- status = AcpiGetObjectInfo(handle, _info);
- if (ACPI_FAILURE(status))
- return_ACPI_STATUS(AE_ERROR);
- if (device_info->Address == (uintptr_t)context) {
- *(ACPI_HANDLE *)return_value = handle;
- return_ACPI_STATUS(AE_CTRL_TERMINATE);
- }
- return_ACPI_STATUS(AE_OK);
-}
-
-static ACPI_HANDLE
-get_dimm_acpi_handle(ACPI_HANDLE root_handle, nfit_handle_t adr)
-{
- ACPI_HANDLE res;
- ACPI_STATUS status;
-
- res = NULL;
- status = AcpiWalkNamespace(ACPI_TYPE_DEVICE, root_handle, 1, find_dimm,
- NULL, (void *)(uintptr_t)adr, );
- if (ACPI_FAILURE(status))
- res = NULL;
- return (res);
-}
-
-static int
-nvdimm_root_create_devs(device_t dev, ACPI_TABLE_NFIT *nfitbl)
-{
- ACPI_HANDLE root_handle, dimm_handle;
- device_t child;
- nfit_handle_t *dimm_ids, *dimm;
- uintptr_t *ivars;
- int num_dimm_ids;
-
- root_handle = acpi_get_handle(dev);
- acpi_nfit_get_dimm_ids(nfitbl, _ids, _dimm_ids);
- for (dimm = dimm_ids; dimm < dimm_ids + num_dimm_ids; dimm++) {
- dimm_handle = get_dimm_acpi_handle(root_handle, *dimm);
- child = BUS_ADD_CHILD(dev, 100, "nvdimm", -1);
- if (child == NULL) {
- device_printf(dev, "failed to create nvdimm\n");
- return (ENXIO);
- }
- ivars = mallocarray(NVDIMM_ROOT_IVAR_MAX, sizeof(uintptr_t),
- M_NVDIMM, M_ZERO | M_WAITOK);
- device_set_ivars(child, ivars);
- nvdimm_root_set_acpi_handle(child, dimm_handle);
- nvdimm_root_set_device_handle(child, *dimm);
- }
- free(dimm_ids, M_NVDIMM);
- return (0);
-}
-
-static int
-nvdimm_root_create_spas(struct nvdimm_root_dev *dev, ACPI_TABLE_NFIT *nfitbl)
-{
- ACPI_NFIT_SYSTEM_ADDRESS **spas, **spa;
- struct SPA_mapping *spa_mapping;
- enum SPA_mapping_type spa_type;
- int error, num_spas;
-
- error = 0;
- acpi_nfit_get_spa_ranges(nfitbl, , _spas);
- for (spa = spas; spa < spas + num_spas; spa++) {
- spa_type = nvdimm_spa_type_from_uuid(
- (struct uuid *)(*spa)->RangeGuid);
- if (spa_type == SPA_TYPE_UNKNOWN)
- continue;
- spa_mapping = malloc(sizeof(struct SPA_mapping), M_NVDIMM,
- M_WAITOK | M_ZERO);
- error = nvdimm_spa_init(spa_mapping, *spa, spa_type);
- if (error != 0) {
- nvdimm_spa_fini(spa_mapping);
- free(spa, M_NVDIMM);
- break;
- }
- nvdimm_create_namespaces(spa_mapping, nfitbl);
-
svn commit: r352797 - head/tools/tools/controlelf
Author: emaste
Date: Fri Sep 27 16:27:52 2019
New Revision: 352797
URL: https://svnweb.freebsd.org/changeset/base/352797
Log:
Add tool to modify ELF binary feature control bits
This will allow feature control bits (e.g. for ASLR, PROT_MAX) to be
inspected or modified.
Some clean-up and additional work is likely still required, but we can
iterate on this in the tree.
Submitted by: Bora Özarslan
Sponsored by: The FreeBSD Foundation
Differential Revision:https://reviews.freebsd.org/D19290
Added:
head/tools/tools/controlelf/
head/tools/tools/controlelf/Makefile (contents, props changed)
head/tools/tools/controlelf/controlelf.1 (contents, props changed)
head/tools/tools/controlelf/controlelf.c (contents, props changed)
Added: head/tools/tools/controlelf/Makefile
==
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ head/tools/tools/controlelf/MakefileFri Sep 27 16:27:52 2019
(r352797)
@@ -0,0 +1,14 @@
+# $FreeBSD$
+
+ELFTCDIR= ${SRCTOP}/contrib/elftoolchain
+
+PROG= controlelf
+SRCS= controlelf.c
+
+LIBADD=elftc elf
+
+CFLAGS+=-I${ELFTCDIR}/libelftc -I${ELFTCDIR}/common
+
+WARNS?=6
+
+.include
Added: head/tools/tools/controlelf/controlelf.1
==
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ head/tools/tools/controlelf/controlelf.1Fri Sep 27 16:27:52 2019
(r352797)
@@ -0,0 +1,92 @@
+.\" Copyright 2019 The FreeBSD Foundation.
+.\"
+.\" This software was developed by Bora Ozarslan under sponsorship from
+.\" the FreeBSD Foundation.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\"notice, this list of conditions and the following disclaimer in the
+.\"documentation and/or other materials provided with the distribution.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS''
+.\" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\" $FreeBSD$
+.\"
+.Dd February 21, 2019
+.Dt CONTROLELF 1
+.Os
+.Sh NAME
+.Nm controlelf
+.Nd change an ELF binary's control features
+.Sh SYNOPSIS
+.Nm
+.Op Fl h | Fl -help
+.Op Fl l
+.Op Fl e Ar featurelist
+.Ar
+.Sh DESCRIPTION
+The
+.Nm
+utility modifies an ELF binary to change it's control features.
+.Pp
+The options are as follows:
+.Bl -tag -width indent
+.It Fl h | Fl -help
+Print a usage message and exit.
+.It Fl l
+Writes the list of all known ELF types to standard output.
+.It Fl e Ar featurelist
+Edit features from the given comma separated list
+.Ar featurelist .
+.Ar featurelist
+starts with one of the three modes:
+.Dq Li +
+to turn on the features,
+.Dq Li -
+to turn off the features,
+.Dq Li =
+to only turn on the given features.
+Followed by a comma separated features.
+.El
+.Pp
+If
+.Fl e ,
+option wasn't specified,
+.Nm
+will display the current control features for the files named by the arguments
+.Ar .
+.Sh EXIT STATUS
+Exit status is 0 on success, and 1 if the command
+fails if a file does not exist, is too short,
+or fails to find or edit features note.
+.Sh EXAMPLES
+The following is an example of a typical usage
+of the
+.Nm
+command:
+.Bd -literal -offset indent
+controlelf file
+controlelf -e +aslr file
+.Ed
+.Sh HISTORY
+The
+.Nm
+manual page first appeared in
+.Fx 13.0 .
+.Sh AUTHORS
+This manual page was written by
+.An Bora Ozarslan Mt borako.ozars...@gmail.com .
Added: head/tools/tools/controlelf/controlelf.c
==
--- /dev/null 00:00:00 1970 (empty, because file is newly added)
+++ head/tools/tools/controlelf/controlelf.cFri Sep 27 16:27:52 2019
(r352797)
@@ -0,0 +1,374 @@
+/*-
+ * SPDX-License-Identifier: BSD-3-Clause
+ *
+ * Copyright (c) 2019 The FreeBSD Foundation.
+ *
+ * This software was developed by Bora Ozarslan under sponsorship from
+ * the
svn commit: r352796 - head/sys/kern
Author: andrew
Date: Fri Sep 27 16:22:28 2019
New Revision: 352796
URL: https://svnweb.freebsd.org/changeset/base/352796
Log:
Check the vfs option length is valid before accessing through
When a VFS option passed to nmount is present but NULL the kernel will
place an empty option in its internal list. This will have a NULL
pointer and a length of 0. When we come to read one of these the kernel
will try to load from the last address of virtual memory. This is
normally invalid so will fault resulting in a kernel panic.
Fix this by checking if the length is valid before dereferencing.
MFC after:3 days
Sponsored by: DARPA, AFRL
Modified:
head/sys/kern/vfs_mount.c
Modified: head/sys/kern/vfs_mount.c
==
--- head/sys/kern/vfs_mount.c Fri Sep 27 16:11:47 2019(r352795)
+++ head/sys/kern/vfs_mount.c Fri Sep 27 16:22:28 2019(r352796)
@@ -646,7 +646,7 @@ vfs_donmount(struct thread *td, uint64_t fsflags, stru
*/
fstypelen = 0;
error = vfs_getopt(optlist, "fstype", (void **), );
- if (error || fstype[fstypelen - 1] != '\0') {
+ if (error || fstypelen <= 0 || fstype[fstypelen - 1] != '\0') {
error = EINVAL;
if (errmsg != NULL)
strncpy(errmsg, "Invalid fstype", errmsg_len);
@@ -654,7 +654,7 @@ vfs_donmount(struct thread *td, uint64_t fsflags, stru
}
fspathlen = 0;
error = vfs_getopt(optlist, "fspath", (void **), );
- if (error || fspath[fspathlen - 1] != '\0') {
+ if (error || fspathlen <= 0 || fspath[fspathlen - 1] != '\0') {
error = EINVAL;
if (errmsg != NULL)
strncpy(errmsg, "Invalid fspath", errmsg_len);
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352795 - head/lib/libc/sys
Author: imp Date: Fri Sep 27 16:11:47 2019 New Revision: 352795 URL: https://svnweb.freebsd.org/changeset/base/352795 Log: Document varadic args as int, since you can't have short varadic args (they are promoted to ints). - `mode_t` is `uint16_t` (`sys/sys/_types.h`) - `openat` takes variadic args - variadic args cannot be 16-bit, and indeed the code uses int - the manpage currently kinda implies the argument is 16-bit by saying `mode_t` Prompted by Rust things: https://github.com/tailhook/openat/issues/21 Submitted by: Greg V at unrelenting Differential Revision: https://reviews.freebsd.org/D21816 Modified: head/lib/libc/sys/mq_open.2 head/lib/libc/sys/open.2 Modified: head/lib/libc/sys/mq_open.2 == --- head/lib/libc/sys/mq_open.2 Fri Sep 27 15:28:30 2019(r352794) +++ head/lib/libc/sys/mq_open.2 Fri Sep 27 16:11:47 2019(r352795) @@ -133,7 +133,7 @@ Create a message queue. It requires two additional arguments: .Fa mode , which is of type -.Vt mode_t , +.Vt int , and .Fa attr , which is a pointer to an Modified: head/lib/libc/sys/open.2 == --- head/lib/libc/sys/open.2Fri Sep 27 15:28:30 2019(r352794) +++ head/lib/libc/sys/open.2Fri Sep 27 16:11:47 2019(r352795) @@ -61,7 +61,7 @@ In this case and .Fn openat require an additional argument -.Fa "mode_t mode" , +.Fa "int mode" , and the file is created with mode .Fa mode as described in ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352792 - head/contrib/compiler-rt/lib/sanitizer_common
Author: emaste
Date: Fri Sep 27 13:14:36 2019
New Revision: 352792
URL: https://svnweb.freebsd.org/changeset/base/352792
Log:
compiler-rt: correct RISC-V struct_kernel_stat64_sz
The value of struct_kernel_stat64_sz introduced by review D5021 for
RISC-V was incorrect.
Also add a __riscv_xlen == 64 conditional as the 32-bit ABI is not yet
finalized.
Submitted by: Luís Marques
Differential Revision:https://reviews.freebsd.org/D21684
Modified:
head/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
Modified:
head/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
==
---
head/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
Fri Sep 27 10:19:28 2019(r352791)
+++
head/contrib/compiler-rt/lib/sanitizer_common/sanitizer_platform_limits_posix.h
Fri Sep 27 13:14:36 2019(r352792)
@@ -79,10 +79,9 @@ namespace __sanitizer {
#elif defined(__powerpc64__)
const unsigned struct_kernel_stat_sz = 144;
const unsigned struct_kernel_stat64_sz = 104;
-#elif defined(__riscv)
- /* RISCVTODO: check that these values are correct */
+#elif defined(__riscv) && __riscv_xlen == 64
const unsigned struct_kernel_stat_sz = 128;
- const unsigned struct_kernel_stat64_sz = 128;
+ const unsigned struct_kernel_stat64_sz = 104;
#elif defined(__mips__)
const unsigned struct_kernel_stat_sz =
SANITIZER_ANDROID ? FIRST_32_SECOND_64(104, 128) :
___
svn-src-head@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/svn-src-head
To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
svn commit: r352791 - head/share/misc
Author: kaktus Date: Fri Sep 27 10:19:28 2019 New Revision: 352791 URL: https://svnweb.freebsd.org/changeset/base/352791 Log: Add myself (kaktus) as a src commiter. Reviewed by: kib (mentor) Approved by: kib (mentor), mjg (mentor) Differential Revision:https://reviews.freebsd.org/D21811 Modified: head/share/misc/committers-src.dot Modified: head/share/misc/committers-src.dot == --- head/share/misc/committers-src.dot Fri Sep 27 10:04:58 2019 (r352790) +++ head/share/misc/committers-src.dot Fri Sep 27 10:19:28 2019 (r352791) @@ -227,6 +227,7 @@ jtl [label="Jonathan T. Looney\n...@freebsd.org\n2015/ julian [label="Julian Elischer\njul...@freebsd.org\n1993/04/19"] jwd [label="John De Boskey\n...@freebsd.org\n2000/05/19"] kaiw [label="Kai Wang\nk...@freebsd.org\n2007/09/26"] +kaktus [label="Pawel Biernacki\nkak...@freebsd.org\n2019/09/26"] kan [label="Alexander Kabaev\n...@freebsd.org\n2002/07/21"] karels [label="Mike Karels\nkar...@freebsd.org\n2016/06/09"] ken [label="Ken Merry\n...@freebsd.org\n1998/09/08"] @@ -675,6 +676,7 @@ kib -> gjb kib -> jah kib -> jlh kib -> jpaetzel +kib -> kaktus kib -> lulf kib -> melifaro kib -> mmel @@ -719,6 +721,8 @@ mdf -> gleb mdodd -> jake mike -> das + +mjg -> kaktus mlaier -> benjsc mlaier -> dhartmei ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352707 - in head/sys: conf kern net sys
On Fri, Sep 27, 2019, 2:00 AM Niclas Zeising wrote: > On 2019-09-27 01:27, Warner Losh wrote: > > I talked to Gleb and this will move to opt_global.h. this will fix all > > these problems. > > > > Warner > > > > [ trimming CC list ] > So, in conclusion, the changes made to any out of tree kmods to handle > this should be reverted? > Yes. It was a momentary blip. Warner Regards > -- > Niclas Zeising > ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
Re: svn commit: r352707 - in head/sys: conf kern net sys
On 2019-09-27 01:27, Warner Losh wrote: I talked to Gleb and this will move to opt_global.h. this will fix all these problems. Warner [ trimming CC list ] So, in conclusion, the changes made to any out of tree kmods to handle this should be reverted? Regards -- Niclas Zeising ___ svn-src-head@freebsd.org mailing list https://lists.freebsd.org/mailman/listinfo/svn-src-head To unsubscribe, send any mail to "svn-src-head-unsubscr...@freebsd.org"
