linux.git/arch/arm64, branch v3.7-rc2 Linux kernel source tree arm64: fix alignment padding in assembly code 2012-10-20T10:12:01+00:00 Marc Zyngier Marc.Zyngier@arm.com 2012-10-19T16:33:27+00:00 aeed41a9371ee02257b608eb06a9058507a7d0f4 An interesting effect of using the generic version of linkage.h is that the padding is defined in terms of x86 NOPs, which can have even more interesting effects when the assembly code looks like this: ENTRY(func1) mov x0, xzr ENDPROC(func1) // fall through ENTRY(func2) mov x0, #1 ret ENDPROC(func2) Admittedly, the code is not very nice. But having code from another architecture doesn't look completely sane either. The fix is to add arm64's version of linkage.h, which causes the insertion of proper AArch64 NOPs. Signed-off-by: Marc Zyngier <marc.zyngier@arm.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> 
An interesting effect of using the generic version of linkage.h
is that the padding is defined in terms of x86 NOPs, which can have
even more interesting effects when the assembly code looks like this:

ENTRY(func1)
	mov	x0, xzr
ENDPROC(func1)
	// fall through
ENTRY(func2)
	mov	x0, #1
	ret
ENDPROC(func2)

Admittedly, the code is not very nice. But having code from another
architecture doesn't look completely sane either.

The fix is to add arm64's version of linkage.h, which causes the insertion
of proper AArch64 NOPs.

Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
arm64: ptrace: use HW_BREAKPOINT_EMPTY type for disabled breakpoints 2012-10-18T19:15:34+00:00 Will Deacon will.deacon@arm.com 2012-10-18T14:17:00+00:00 8f34a1da35aed7b438a2de8ac27723a5472e8399 If a debugger tries to zero a hardware debug control register, the kernel will try to infer both the type and length of the breakpoint in order to sanity-check against the requested regset type. This will fail because the encoding will appear as a zero-length breakpoint. This patch changes the control register setting so that disabled breakpoints are treated as HW_BREAKPOINT_EMPTY and no further sanity-checking is required. Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> 
If a debugger tries to zero a hardware debug control register, the
kernel will try to infer both the type and length of the breakpoint
in order to sanity-check against the requested regset type. This will
fail because the encoding will appear as a zero-length breakpoint.

This patch changes the control register setting so that disabled
breakpoints are treated as HW_BREAKPOINT_EMPTY and no further
sanity-checking is required.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
arm64: ptrace: make structure padding explicit for debug registers 2012-10-18T19:15:31+00:00 Will Deacon will.deacon@arm.com 2012-10-11T11:10:57+00:00 7797d17c591ae62c6f43c6de4fdb8beeb50eb692 The user_hwdebug_state structure contains implicit padding to conform to the alignment requirements of the AArch64 ABI (namely that aggregates must be aligned to their most aligned member). This patch fixes the ptrace functions operating on struct user_hwdebug_state so that the padding is handled correctly. Signed-off-by: Will Deacon <will.deacon@arm.com> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> 
The user_hwdebug_state structure contains implicit padding to conform to
the alignment requirements of the AArch64 ABI (namely that aggregates
must be aligned to their most aligned member).

This patch fixes the ptrace functions operating on struct
user_hwdebug_state so that the padding is handled correctly.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
arm64: No need to set the x0-x2 registers in start_thread() 2012-10-18T19:14:01+00:00 Catalin Marinas catalin.marinas@arm.com 2012-10-16T16:07:46+00:00 16dd46bb781a1d37eeb2377e8e48276e9d14d15d For historical reasons, ARM used to set r0-r2 in start_thread() to the first values on the user stack when starting a new user application. The same logic has been inherited in AArch64. The x0 register is overridden by the sys_execve() return value so it's always zero on success. The x1 and x2 registers are ignored by AArch64 and EABI AArch32 applications, so we can safely remove the register setting for both native and compat user space. This also fixes a potential fault with the kernel accessing user space stack directly. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Reported-by: Al Viro <viro@zeniv.linux.org.uk> 
For historical reasons, ARM used to set r0-r2 in start_thread() to the
first values on the user stack when starting a new user application. The
same logic has been inherited in AArch64. The x0 register is overridden
by the sys_execve() return value so it's always zero on success. The x1
and x2 registers are ignored by AArch64 and EABI AArch32 applications,
so we can safely remove the register setting for both native and compat
user space.

This also fixes a potential fault with the kernel accessing user space
stack directly.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Reported-by: Al Viro <viro@zeniv.linux.org.uk>
arm64: Ignore memory blocks below PHYS_OFFSET 2012-10-18T19:14:00+00:00 Catalin Marinas catalin.marinas@arm.com 2012-10-16T11:00:29+00:00 f71a1a42667f576ec736bb1200eba2118fee3a22 According to Documentation/arm64/booting.txt, the kernel image must be loaded at a pre-defined offset from the start of RAM so that the kernel can calculate PHYS_OFFSET based on this address. If the DT contains memory blocks below this PHYS_OFFSET, report them and ignore the corresponding memory range. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> 
According to Documentation/arm64/booting.txt, the kernel image must be
loaded at a pre-defined offset from the start of RAM so that the kernel
can calculate PHYS_OFFSET based on this address. If the DT contains
memory blocks below this PHYS_OFFSET, report them and ignore the
corresponding memory range.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
arm64: Fix the update_vsyscall() prototype 2012-10-18T19:14:00+00:00 Catalin Marinas catalin.marinas@arm.com 2012-10-16T10:44:53+00:00 c60b0c2817bd6a990b08a7651e9cf630414665f5 With commit 576094b7 (time: Introduce new GENERIC_TIME_VSYSCALL) the old update_vsyscall() prototype is no longer available. This patch updates the arm64 port. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Acked-by: Will Deacon <will.deacon@arm.com> Cc: John Stultz <john.stultz@linaro.org> 
With commit 576094b7 (time: Introduce new GENERIC_TIME_VSYSCALL) the old
update_vsyscall() prototype is no longer available. This patch updates
the arm64 port.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Will Deacon <will.deacon@arm.com>
Cc: John Stultz <john.stultz@linaro.org>
arm64: Select MODULES_USE_ELF_RELA 2012-10-18T19:11:54+00:00 Catalin Marinas catalin.marinas@arm.com 2012-10-16T10:26:57+00:00 fea2acaa5ca01eb2996226b2ae177aa66b222511 With commit 786d35d4 (make most arch asm/module.h files use asm-generic/module.h) arm64 needs to enable MODULES_USE_ELF_RELA for loadable modules. Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> Acked-by: Will Deacon <Will.Deacon@arm.com> 
With commit 786d35d4 (make most arch asm/module.h files use
asm-generic/module.h) arm64 needs to enable MODULES_USE_ELF_RELA for
loadable modules.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Acked-by: Will Deacon <Will.Deacon@arm.com>
arm64: Remove duplicate inclusion of mmu_context.h in smp.c 2012-10-16T16:17:45+00:00 Sachin Kamat sachin.kamat@linaro.org 2012-10-13T08:25:58+00:00 086e47b6c959ee557e9adefe72b8800c62d0ac34 asm/mmu_context.h was included twice. Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org> Signed-off-by: Catalin Marinas <catalin.marinas@arm.com> 
asm/mmu_context.h was included twice.

Signed-off-by: Sachin Kamat <sachin.kamat@linaro.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Merge tag 'arm64-uapi' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-aarch64 2012-10-13T02:20:04+00:00 Linus Torvalds torvalds@linux-foundation.org 2012-10-13T02:20:04+00:00 b4fe19f78e84813e720d0f0482da22194a7728e3 Pull arm64 uapi disintegration from Catalin Marinas: "UAPI headers for arm64 together with some clean-up to make it possible: - Do not export the COMPAT_* definitions to user - Simplify the compat unistd32.h definitions and remove the __SYSCALL_COMPAT guard - Disintegrate the arch/arm64/include/asm/* headers" * tag 'arm64-uapi' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-aarch64: UAPI: (Scripted) Disintegrate arch/arm64/include/asm arm64: Do not export the compat-specific definitions to the user arm64: Do not include asm/unistd32.h in asm/unistd.h arm64: Remove unused definitions from asm/unistd32.h 
Pull arm64 uapi disintegration from Catalin Marinas:
 "UAPI headers for arm64 together with some clean-up to make it
  possible:
   - Do not export the COMPAT_* definitions to user
   - Simplify the compat unistd32.h definitions and remove the
     __SYSCALL_COMPAT guard
   - Disintegrate the arch/arm64/include/asm/* headers"

* tag 'arm64-uapi' of git://git.kernel.org/pub/scm/linux/kernel/git/cmarinas/linux-aarch64:
  UAPI: (Scripted) Disintegrate arch/arm64/include/asm
  arm64: Do not export the compat-specific definitions to the user
  arm64: Do not include asm/unistd32.h in asm/unistd.h
  arm64: Remove unused definitions from asm/unistd32.h
vfs: define struct filename and have getname() return it 2012-10-13T00:14:55+00:00 Jeff Layton jlayton@redhat.com 2012-10-10T19:25:28+00:00 91a27b2a756784714e924e5e854b919273082d26 getname() is intended to copy pathname strings from userspace into a kernel buffer. The result is just a string in kernel space. It would however be quite helpful to be able to attach some ancillary info to the string. For instance, we could attach some audit-related info to reduce the amount of audit-related processing needed. When auditing is enabled, we could also call getname() on the string more than once and not need to recopy it from userspace. This patchset converts the getname()/putname() interfaces to return a struct instead of a string. For now, the struct just tracks the string in kernel space and the original userland pointer for it. Later, we'll add other information to the struct as it becomes convenient. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Al Viro <viro@zeniv.linux.org.uk> 
getname() is intended to copy pathname strings from userspace into a
kernel buffer. The result is just a string in kernel space. It would
however be quite helpful to be able to attach some ancillary info to
the string.

For instance, we could attach some audit-related info to reduce the
amount of audit-related processing needed. When auditing is enabled,
we could also call getname() on the string more than once and not
need to recopy it from userspace.

This patchset converts the getname()/putname() interfaces to return
a struct instead of a string. For now, the struct just tracks the
string in kernel space and the original userland pointer for it.

Later, we'll add other information to the struct as it becomes
convenient.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>