At present sandbox sets non-blocking I/O as soon as any input is read
from the terminal. However it does not restore the previous state on
exit. Fix this and drop the old os_read_no_block() function.
This means that we always enable blocking I/O in sandbox (if input is a
terminal) whereas previously it would only happen on the first call to
tstc() or getc(). However, the difference is likely not important.
Signed-off-by: Simon Glass <sjg@chromium.org>
When debugging sandbox it is sometimes annoying that the memory file is
deleted early on. If sandbox later crashes or we quit (using the
debugger), it is not possible to run it again with the same state since
the memory file is gone.
Remove the old memory file when sandbox exits, instead. Also add debugging
showing the memory filename.
Signed-off-by: Simon Glass <sjg@chromium.org>
It is useful to be able to set the default log level from the command line
when running sandbox. Add a new -L command-line flag for this. The log
level is set using the enum log_level_t in log.h. At present a number must
be specified, e.g. -L7 for debug.
Signed-off-by: Simon Glass <sjg@chromium.org>
At present we support booting from SPL to U-Boot proper. Add support for
the previous stage too, so sandbox can be started with TPL.
Signed-off-by: Simon Glass <sjg@chromium.org>
For debugging it is sometimes useful to write out data for inspection
using an external tool. Add a function which can write this data to a
given file.
Signed-off-by: Simon Glass <sjg@chromium.org>
On my Ubuntu 18.04.1 machine two driver-model bus tests have started
failing recently. The problem appears to be that the DATA region of the
executable is protected. This does not seem correct, but perhaps there
is a reason.
To work around it, unprotect the regions in these tests before accessing
them.
Signed-off-by: Simon Glass <sjg@chromium.org>
With efi_loader, we may want to execute payloads from RAM. By default,
permissions on the RAM region don't allow us to execute from there though.
So let's change the default allocation scheme for RAM to also allow
execution from it. That way payloads that live in U-Boot RAM can be
directly executed.
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
In sandbox, longjmp returns to itself in an endless loop because
os_longjmp() calls into longjmp() which is provided by U-Boot which
again calls os_longjmp().
Setjmp on the other hand must not return because otherwise the
return freees up stack elements that we need during longjmp().
The only straight forward fix that doesn't involve nasty hacks I
could find is to directly link against the system setjmp/longjmp
implementations. That means we just provide the compiler with
hints that the symbol will be available and actually fill them
out with versions from libc.
This approach should be reasonably platform agnostic
Signed-off-by: Alexander Graf <agraf@suse.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
At present map_sysmem() maps an address into the sandbox RAM buffer,
return a pointer, while map_to_sysmem() goes the other way.
The mapping is currently just 1:1 since a case was not found where a more
flexible mapping was needed. PCI does have a separate and more complex
mapping, but uses its own mechanism.
However this arrange cannot handle one important case, which is where a
test declares a stack variable and passes a pointer to it into a U-Boot
function which uses map_to_sysmem() to turn it into a address. Since the
pointer is not inside emulated DRAM, this will fail.
Add a mapping feature which can handle any such pointer, mapping it to a
simple tag value which can be passed around in U-Boot as an address.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This function is useful to signal that the application needs to exit
immediate. It can be caught with a debugger (e.g. gdb). Add a stub for it
so that it can be called from within sandbox when an internal error
occurs.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
At present the sandbox RAM buffer is not aligned to any particular
address boundary. This makes the internal pointers somewhat random with
respect to the associated RAM buffer addresses.
Align the buffer to the page size of the machine to help with this. Note
that there is a header at the start of the allocated pointer. To avoid
returning a pointer which is not aligned to a page boundary, we waste
almost an entire page of memory for each allocation.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
This partially reverts commit 7e21fbca26.
That change broke sandbox EFI support for unknown reasons. It also changes
sandbox to use--gc-sections which we don't want.
For now I am just reverting the sandbox portion as presumably this change
is safe on other architectures.
Fixes: 7e21fbca26 (efi_loader: Rename sections to allow for implicit data)
Signed-off-by: Simon Glass <sjg@chromium.org>
Something went wrong when writing the sandbox linker scripts and so we
ended up with a .bss section marker right before the efi runtime sections.
That obviously is a terrible idea, as it may result in overwriting efi
runtime code and data. So let's move the .bss identifier behind the efi
sections.
Signed-off-by: Alexander Graf <agraf@suse.de>
Ask the OS for each of its interfaces and for each one, bind a U-Boot
device and then probe it. This will allocate the priv data structure
that is then populated.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
With systemd stable interface names, eth0 will almost never exist.
Instead of using that name in the sandbox.dts, use an index.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Instead of doing a simple string compare against "lo", look for the flag
that indicates a localhost interface.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
If we let descriptors equal 0, we can end up closing STDIN. Make sure
they start out as -1.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
In open, the socket is correctly checked to be -1 in the error case.
In send and recv, we checked for 0, but that is a valid socket number.
Correct this by checking for -1 as a bad socket everywhere.
Signed-off-by: Joe Hershberger <joe.hershberger@ni.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
Some times gcc may generate data that is then used within code that may
be part of an efi runtime section. That data could be jump tables,
constants or strings.
In order to make sure we catch these, we need to ensure that gcc emits
them into a section that we can relocate together with all the other
efi runtime bits. This only works if the -ffunction-sections and
-fdata-sections flags are passed and the efi runtime functions are
in a section that starts with ".text".
Up to now we had all efi runtime bits in sections that did not
interfere with the normal section naming scheme, but this forces
us to do so. Hence we need to move the efi_loader text/data/rodata
sections before the global *(.text*) catch-all section.
With this patch in place, we should hopefully have an easier time
to extend the efi runtime functionality in the future.
Signed-off-by: Alexander Graf <agraf@suse.de>
[agraf: Fix x86_64 breakage]
The EFI loader code requires certain linker sections to exist. Add these
for sandbox so that the EFI loader code will link.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
Add an implementation of setjmp() and longjmp() which rely on the
underlying host C library. Since we cannot know how large the jump buffer
needs to be, pick something that should be suitable and check it at
runtime. At present we need access to the underlying struct as well.
Signed-off-by: Simon Glass <sjg@chromium.org>
Signed-off-by: Alexander Graf <agraf@suse.de>
When U-Boot started using SPDX tags we were among the early adopters and
there weren't a lot of other examples to borrow from. So we picked the
area of the file that usually had a full license text and replaced it
with an appropriate SPDX-License-Identifier: entry. Since then, the
Linux Kernel has adopted SPDX tags and they place it as the very first
line in a file (except where shebangs are used, then it's second line)
and with slightly different comment styles than us.
In part due to community overlap, in part due to better tag visibility
and in part for other minor reasons, switch over to that style.
This commit changes all instances where we have a single declared
license in the tag as both the before and after are identical in tag
contents. There's also a few places where I found we did not have a tag
and have introduced one.
Signed-off-by: Tom Rini <trini@konsulko.com>
Thomas reported U-Boot failed to build host tools if libfdt-devel
package is installed because tools include libfdt headers from
/usr/include/ instead of using internal ones.
This commit moves the header code:
include/libfdt.h -> include/linux/libfdt.h
include/libfdt_env.h -> include/linux/libfdt_env.h
and replaces include directives:
#include <libfdt.h> -> #include <linux/libfdt.h>
#include <libfdt_env.h> -> #include <linux/libfdt_env.h>
Reported-by: Thomas Petazzoni <thomas.petazzoni@bootlin.com>
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
To debug device tree issues involving 32- and 64-bit platforms, it is useful to
have a generic 64-bit platform available.
Add a version of the sandbox that uses 64-bit integers for its physical
addresses as well as a modified device tree.
Signed-off-by: Mario Six <mario.six@gdsys.cc>
Added CONFIG_SYS_TEXT_BASE to configs/sandbox64_defconfig
Signed-off-by: Simon Glass <sjg@chromium.org>
While sandbox works OK without the special-case code, it does result in
console output being stored in the pre-console buffer while sandbox starts
up. If there is a crash or a problem then there is no indication of what
is going on.
For ease of debugging it seems better to revert this change.
This reverts commit 47b98ad0f6.
Signed-off-by: Simon Glass <sjg@chromium.org>
Realloc does not free the old memory area if it fails.
Identified by cppcheck.
Signed-off-by: Heinrich Schuchardt <xypron.glpk@gmx.de>
Reviewed-by: Simon Glass <sjg@chromium.org>
Convert the sandbox architecture to make use of the new asm-generic/io.h
to provide address mapping functions. As sandbox actually performs
non-identity mapping between physical & virtual addresses we can't
simply make use of the generic mapping functions, but are able to
implement phys_to_virt() & make use of it from map_physmem().
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Simon Glass <sjg@chromium.org>
Acked-by: Simon Glass <sjg@chromium.org>
They are unused since commit d8c6fb8ced ("sandbox: Drop special
case console code for sandbox").
Signed-off-by: Masahiro Yamada <yamada.masahiro@socionext.com>
Acked-by: Simon Glass <sjg@chromium.org>
SPL and normal u-boot stage use different malloc pool size
configuration before relocation, so use CONFIG_VAL(SYS_MALLOC_F_LEN)
to fit different boot stage.
Signed-off-by: Andy Yan <andyshrk@gmail.com>
Changes in v3:
- use CONFIG_VAL(), which suggested by Simon
Changes in v2: None
arch/sandbox/cpu/start.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
Acked-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Reviewed-by: Philipp Tomsich <philipp.tomsich@theobroma-systems.com>
Currently we have code which prints out platform data at the start of SPL.
Now that we have tests for dtoc this is probably not necessary. Drop it.
Update test_ofplatdata to check for empty output since it is useful to
check that sandbox_spl works as expected.
Signed-off-by: Simon Glass <sjg@chromium.org>
In os_dirent_get_typename() we are checking that type falls within the
known values of the enum os_dirent_t. With clang-3.8 testing this value
as being >= 0 results in a warning as it will always be true. This
assumes of course that we are only given valid data. Given that we want
to sanity check the input, we change this to check that it falls within
the range of the first to the last entry in the given enum.
Cc: Simon Glass <sjg@chromium.org>
Signed-off-by: Tom Rini <trini@konsulko.com>
Reviewed-by: Simon Glass <sjg@chromium.org>
At present we provide a default version of this function for use by
bootstage. However it uses the system timer and therefore likely requires
driver model. This makes it impossible to time driver-model init.
Drop the function and require boards to provide their own. Add a sandbox
version also. There is a default implememtation in lib/time.c for boards
which use CONFIG_SYS_TIMER_COUNTER.
Signed-off-by: Simon Glass <sjg@chromium.org>
Running a new test should reset the sandbox state to avoid tests
interferring with each other. Move the existing state-reset code into a
function so it can be used from tests.
Also update the code to reset the SPI devices and adjust the test code to
call it.
Signed-off-by: Simon Glass <sjg@chromium.org>
This header includes things that are needed to make driver build. Adjust
existing users to include that always, even if other dm/ includes are
present
Signed-off-by: Simon Glass <sjg@chromium.org>
We don't need a special hook for sandbox as one of the later ones will do
just as well. We can print error messages about bad options after we
print the banner. In fact, it seems better.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Stefan Roese <sr@denx.de>
There is no need to have this call in the generic init sequence and no
other architecture has needed it in the time it has been there. Move it
into sandbox's private code.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Stefan Roese <sr@denx.de>
Reviewed-by: Stefan Roese <sr@denx.de>
It is useful to name each method so that we can print out this name when
using the method. Currently this happens using a separate function. In
preparation for unifying this, add a name to each method.
The name is only available if we have libcommon support (i.e can use
printf()).
Signed-off-by: Simon Glass <sjg@chromium.org>
The "hostfs ls" command prefixes each directory entry with either DIR,
LNK or " " if it is a directory, symlink resp. regular file, or
"???" for any other or unknown type.
The latter only works if the type is set correctly, as the entry defaults
to OS_FILET_REG and e.g. socket files show up as regular files.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Acked-by: Simon Glass <sjg@chromium.org>
Using readdir_r limits the maximum file name length and may even be
unsafe, and is thus deprecated in since glibc 2.24.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Acked-by: Simon Glass <sjg@chromium.org>
The readdir linux manpage explicitly states (quoting POSIX.1) that
sizeof(d_name) is not correct for determining the required size, but to
always use strlen. Grow the buffer if needed.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Acked-by: Simon Glass <sjg@chromium.org>
Previously, after reading/creating the second dirent, the second entry
would be chained to the first entry and the first entry would be linked
to head. Instead, immediately link the first entry to head.
Signed-off-by: Stefan Brüns <stefan.bruens@rwth-aachen.de>
Acked-by: Simon Glass <sjg@chromium.org>
Rather than having a global variable, pass the spl_image as a parameter.
This avoids BSS use, and makes it clearer what the function is actually
doing.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Add a linker list declaration for this method and remove the explicit
switch() code. Update existing users.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
At present some spl_xxx_load_image() functions take a parameter and some
don't. Of those that do, most take an integer but one takes a string.
Convert this parameter into a struct so that we can pass all functions the
same thing. This will allow us to use a common function signature.
Signed-off-by: Simon Glass <sjg@chromium.org>
Reviewed-by: Tom Rini <trini@konsulko.com>
Start up the test devices. These print out of-platdata contents, providing a
check that the of-platdata feature is working correctly.
The device-tree changes are made to sandbox.dts rather than test.dts. since
the former controls the of-platdata generation.
Signed-off-by: Simon Glass <sjg@chromium.org>
Simon Glass
Alexander Graf
Joe Hershberger
Tom Rini
Masahiro Yamada
Mario Six
Simon Goldschmidt
Heinrich Schuchardt
Paul Burton
Bin Meng
Andy Yan
Stefan Brüns