Right now several architectures allow their set_memory_*() family of
functions to fail, but callers may not be checking the return values.
If set_memory_*() returns with an error, call-site assumptions may be
infact wrong to assume that it would either succeed or not succeed at
all. Ideally, the failure of set_memory_*() should be passed up the call
stack, and callers should examine the failure and deal with it.
Need to fix the callers and add the __must_check attribute. They also
may not provide any level of atomicity, in the sense that the memory
protections may be left incomplete on failure. This issue likely has a
few steps on effects architectures:
1) Have all callers of set_memory_*() helpers check the return value.
2) Add __must_check to all set_memory_*() helpers so that new uses do
not ignore the return value.
3) Add atomicity to the calls so that the memory protections aren't
left in a partial state.
This series is part of step 1. Make sram driver check the return value
of set_memory_*().
Signed-off-by: Tianlin Li <tli@digitalocean.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Link: https://lore.kernel.org/r/20191217194528.16461-1-tli@digitalocean.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Currently the sram-exec functionality, which allows allocation of
executable memory and provides an API to move code to it, is only
selected in configs for the ARM architecture. Based on commit
5756e9dd0d ("ARM: 6640/1: Thumb-2: Symbol manipulation macros for
function body copying") simply copying a C function pointer address
using memcpy without consideration of alignment and Thumb is unsafe on
ARM platforms.
The aforementioned patch introduces the fncpy macro which is a safe way
to copy executable code on ARM platforms, so let's make use of that here
rather than the unsafe plain memcpy that was previously used by
sram_exec_copy. Now sram_exec_copy will move the code to "dst" and
return an address that is guaranteed to be safely callable.
In the future, architectures hoping to make use of the sram-exec
functionality must define an fncpy macro just as ARM has done to
guarantee or check for safe copying to executable memory before allowing
the arch to select CONFIG_SRAM_EXEC.
Acked-by: Tony Lindgren <tony@atomide.com>
Acked-by: Russell King <rmk+kernel@armlinux.org.uk>
Reviewed-by: Alexandre Belloni <alexandre.belloni@free-electrons.com>
Signed-off-by: Dave Gerlach <d-gerlach@ti.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
set_memory_* functions have moved to set_memory.h. Switch to this
explicitly.
Link: http://lkml.kernel.org/r/1488920133-27229-15-git-send-email-labbott@redhat.com
Signed-off-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Some platforms, like many ARM SoCs, require the ability to run code from
on-chip memory like SRAM for tasks like reconfiguring the SDRAM
controller or entering low-power sleep modes. In order to do this we
must be able to allocate memory that the code can be copied to but then
change the mapping to be read-only and executable so that no memory is
both writable and executable at the same time to avoid opening any
unneccesary security holes.
By using the existing "pool" partition type that the SRAM driver allows
we can create a memory space that will already be exposed by the
genalloc framework to allow for allocating memory but we must extend
this to meet the executable requirements. By making use of various
set_memory_* APIs we can change the attributes of pages to make them
writable for code upload but then read-only and executable when we want
to actually run code. Because SRAM is a shared resource we need a
centralized manager of these set memory calls. Because the SRAM driver
itself is responsible for allocating the memory we can introduce a
sram_copy_exec API for the driver that works like memcpy but also
manages the page attributes and locking to allow multiple users of the
same SRAM space to all copy their code over independent of other each
before starting execution.
It is maintained in a separate file from the core SRAM driver to allow
it to be selectively built depending on whether or not a platform has
the appropriate set_memory_* APIs. A future patch will integrate it with
the core SRAM driver.
Signed-off-by: Dave Gerlach <d-gerlach@ti.com>
Acked-by: Tony Lindgren <tony@atomide.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Tianlin Li