The issues that broke the usage of VIRTIO_PCI on MIPS when we are
running uncached (i.e. with CONFIG_MMU=n) seem to have been fixed.
As such, allow this option to be set for those configurations.
Signed-off-by: Denis Orlov
---
drivers/virtio/Kconfig | 1 -
1 file changed, 1 deletion(-)
diff
The issues with VIRTIO_PCI on MIPS seems to have been fixes, so enable
this option. And, while at it, also enable a few more for which there is
no reason not to have them enabled.
Denis Orlov (2):
virtio: remove VIRTIO_PCI dependance on MMU for MIPS
MIPS: enable more options in malta
On Tue, 6 Jun 2023 at 11:03, Ahmad Fatoum wrote:
>
> On 05.06.23 22:10, Denis Orlov wrote:
> > KSEG macro is not available when compiling with CONFIG_64BIT enabled, so
> > use CKSEG instead.
>
> If they're interchangeable, why do we need KSEG?
Hmm, I think maybe we don
Hello Ahmad,
On Tue, 6 Jun 2023 at 11:02, Ahmad Fatoum wrote:
>
> Hello Denis,
>
> On 05.06.23 22:10, Denis Orlov wrote:
> > The existing 64BIT support for MIPS was somewhat incomplete with no
> > board having MIPS64 CPUs specified to be available as targets. Define
>
On Tue, 6 Jun 2023 at 12:04, Ahmad Fatoum wrote:
>
> On 05.06.23 22:10, Denis Orlov wrote:
> > The code that uses TEXT_BASE will fail on reading truncated 32-bit
> > address if running with 64BIT enabled. As we only support running from
> > compatibility segments (i.e.
On Tue, 6 Jun 2023 at 11:13, Ahmad Fatoum wrote:
>
> On 05.06.23 22:10, Denis Orlov wrote:
> > This allows to write generic assembly code that will compile under both
> > o32 and n64 ABIs, as otherwise the register definitions would conflict.
> >
> > Taken from Linu
After relocating the code we jump straight to it, with the function
having an 'unreachable()' call at the end. So add an appropriate
attribute.
Signed-off-by: Denis Orlov
---
arch/mips/lib/reloc.c | 4 ++--
1 file changed, 2 insertions(+), 2 deletions(-)
diff --git a/arch/mips/lib/reloc.c b
Deduplicate some of the definitions for 32/64-bit targets and make KSEG*
32-bit defines available for 64-bit builds, as those might still be
useful in certain cases.
Signed-off-by: Denis Orlov
---
arch/mips/include/asm/addrspace.h | 45 ++-
1 file changed, 20
Denis Orlov (21):
MIPS: addrspace: rectify ksseg segment naming
MIPS: addrspace: simplify the header
MIPS: main_entry-pbl: remove unused variable
MIPS: barebox.lds: remove extra whitespace
MIPS: reloc: mark relocate_code() as noreturn
MIPS: cpuinfo: use appropriate format specifiers
This register contains the failing address on Address Error and some
TLB-related exceptions, and is useful for debugging purposes.
Signed-off-by: Denis Orlov
---
arch/mips/include/asm/stackframe.h | 4
arch/mips/lib/traps.c | 3 ++-
2 files changed, 6 insertions(+), 1 deletion
Use 'mips_barebox_10h' macro instead. Remove the comment about the same
address being reserved for a Board ID as we don't really do anything
about it anyway.
Signed-off-by: Denis Orlov
---
arch/mips/boards/qemu-malta/lowlevel.S | 11 +--
1 file changed, 1 insertion(+), 10 deletions
The various definitions related to GT64120 system controller were split
between two headers: 'asm/gt64120.h' and 'mach/mach-gt64120.h'. It seems
that they were taken from Linux in this form. No reason to keep it like
that though, so unify them into 'mach/gt64120.h'.
Signed-off-by: Denis Orlov
of those actions and ensure that it's used in every board's entry
code.
Signed-off-by: Denis Orlov
---
arch/mips/boards/8devices-lima/lowlevel.S | 2 ++
arch/mips/boards/dlink-dir-320/lowlevel.S | 2 +-
arch/mips/boards/img-ci20/lowlevel.S | 2 +-
arch/mips/boards/loongson-ls1b
Signed-off-by: Denis Orlov
---
arch/mips/lib/cpuinfo.c | 6 +++---
1 file changed, 3 insertions(+), 3 deletions(-)
diff --git a/arch/mips/lib/cpuinfo.c b/arch/mips/lib/cpuinfo.c
index fd27920f9b..41ec7b8d53 100644
--- a/arch/mips/lib/cpuinfo.c
+++ b/arch/mips/lib/cpuinfo.c
@@ -25,12 +25,12
GAS allows to use '.asciz'/'.asciiz' instead of manually writing
'.byte 0' to null-terminate the string specified in '.ascii' directive.
Do just that in 'mips_barebox_10h'. The malta-specific entry code
currently does the same.
Signed-off-by: Denis Orlov
---
arch/mips/include/asm/pbl_macros.h
Remove duplicated macro calls by moving it into a common macro, as is
done in ARM code. Some boards were missing this one for no good reason,
it seems. Well, now they don't.
Signed-off-by: Denis Orlov
---
arch/mips/boards/8devices-lima/lowlevel.S | 2 --
arch/mips/boards/dlink-dir-320
compatible exception handling scheme
(and we even hardcode that), so what are we trying to achieve there
anyway?
Signed-off-by: Denis Orlov
---
arch/mips/boot/main_entry.c | 4 +---
1 file changed, 1 insertion(+), 3 deletions(-)
diff --git a/arch/mips/boot/main_entry.c b/arch/mips/boot
It seems more reasonable to do that in PBL code that initializes all the
other appropriate CP0 register bits. This also makes a corresponding
call in barebox proper entry code redundant, paving the way to its
removal.
Signed-off-by: Denis Orlov
---
arch/mips/include/asm/pbl_macros.h | 1 +
1
This gets us rid of this logic repeated in 'flush_cache_all()' for each
cache and allows to use blast_*cache funcs elsewhere if such a need
arises.
Signed-off-by: Denis Orlov
---
arch/mips/lib/c-r4k.c | 41 -
1 file changed, 24 insertions(+), 17 deletions
-by: Denis Orlov
---
arch/mips/include/asm/addrspace.h | 10 +-
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/arch/mips/include/asm/addrspace.h
b/arch/mips/include/asm/addrspace.h
index 11f10e5011..7a2006a8d1 100644
--- a/arch/mips/include/asm/addrspace.h
+++ b/arch/mips
For some reason we do set 'ttb' global variable, however we never end up
using it.
Signed-off-by: Denis Orlov
---
arch/mips/boot/main_entry-pbl.c | 4
1 file changed, 4 deletions(-)
diff --git a/arch/mips/boot/main_entry-pbl.c b/arch/mips/boot/main_entry-pbl.c
index f75ec03851..78982fd995
()'.
Signed-off-by: Denis Orlov
---
arch/mips/boot/start.S | 46 --
arch/mips/boot/start.c | 17
2 files changed, 17 insertions(+), 46 deletions(-)
delete mode 100644 arch/mips/boot/start.S
create mode 100644 arch/mips/boot/start.c
diff
to copy now consists of just two instructions.
However it feels more safe to calculate that in code instead, so that we
don't have to update this value if some code is added there in the
future.
Signed-off-by: Denis Orlov
---
arch/mips/boot/main_entry.c | 39
Improves the code readability a tiny little bit.
Signed-off-by: Denis Orlov
---
arch/mips/lib/reloc.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/mips/lib/reloc.c b/arch/mips/lib/reloc.c
index 83b4040d95..a9078aa813 100644
--- a/arch/mips/lib/reloc.c
+++ b/arch/mips
'r4k_cache_init()' is already declared in the appropriate 'asm/cache.h'
header.
Signed-off-by: Denis Orlov
---
arch/mips/lib/c-r4k.c | 2 --
1 file changed, 2 deletions(-)
diff --git a/arch/mips/lib/c-r4k.c b/arch/mips/lib/c-r4k.c
index 2fb4e90138..44cf57d99b 100644
--- a/arch/mips/lib/c-r4k.c
Even though the original code was taken from Linux, it is quite old and
crufty. Replace the 'while (1)' loop with the 'for' one, use const for
constant variables and use ALIGN_DOWN macro instead of repeating its
code.
Signed-off-by: Denis Orlov
---
arch/mips/lib/c-r4k.c | 13 +
1
and calls to them. This will make it easier
to add support for boards that do happen to have secondary caches.
Signed-off-by: Denis Orlov
---
arch/mips/include/asm/cacheops.h | 5 +
arch/mips/lib/c-r4k.c| 12 ++--
2 files changed, 11 insertions(+), 6 deletions(-)
diff --git
Signed-off-by: Denis Orlov
---
arch/mips/lib/barebox.lds.S | 2 --
1 file changed, 2 deletions(-)
diff --git a/arch/mips/lib/barebox.lds.S b/arch/mips/lib/barebox.lds.S
index 1259a4e94f..0720f9295d 100644
--- a/arch/mips/lib/barebox.lds.S
+++ b/arch/mips/lib/barebox.lds.S
@@ -3,7 +3,6
Make sure dma buffers are both properly allocated and freed (they were
not in a few places).
Denis Orlov (2):
usb: storage: fix missing calls to free()
usb: make sure dma buffers are properly allocated
drivers/usb/core/usb.c | 16 ---
drivers/usb/storage/transport.c | 74
. This is exactly what dma_alloc() is for, so use that for
buffers which we are going to map for dma.
Signed-off-by: Denis Orlov
---
drivers/usb/core/usb.c | 16 ---
drivers/usb/storage/transport.c | 74 +++--
drivers/usb/storage/usb.c | 17
3
Memory allocated with xzalloc() was not actually being freed in a few
functions, resulting in memory leaks.
Signed-off-by: Denis Orlov
---
drivers/usb/storage/usb.c | 17 -
1 file changed, 12 insertions(+), 5 deletions(-)
diff --git a/drivers/usb/storage/usb.c b/drivers/usb
The architecture was removed back in the commit efccc13513, however some
drivers that were available exclusively for NIOS2 were left over. As it
has been impossible to compile those since then, it seems reasonable to
just get rid of them.
Signed-off-by: Denis Orlov
---
drivers/serial/Kconfig
On 6/30/23 11:57, Ian Abbott wrote:
> On 29/06/2023 21:13, Denis Orlov wrote:
>> The architecture was removed back in the commit efccc13513, however some
>> drivers that were available exclusively for NIOS2 were left over. As it
>> has been impossible to compile those
of the latter function having __iomem attributes, we
should not be remapping them. So, remove those IOMEM calls and instead
do explicit remapping in bochs_isa_detect(), which was missing it.
Also fix bochs_hw_probe() declaration missing one __iomem.
Signed-off-by: Denis Orlov
---
drivers/video/bochs
This will be used by the DDR3 SPD data printing routine and is needed for
completeness, to cover all of the DDR3-related cases from the original
decode-dimms tool.
Signed-off-by: Denis Orlov
---
include/ddr_spd.h | 34 ++
1 file changed, 34 insertions(+)
diff
This extends the spd_decode command to correctly print out the contents
of the DDR3 SPD EEPROMs in addition to the DDR2 ones.
Denis Orlov (2):
ddr_spd: add DDR3 Load Reduced module specific data
ddr_spd: add routine for printing DDR3 SPD contents
common/ddr_spd.c | 556
as a hexadecimal value instead of a decoded string. The
logic is mostly the same too, but some changes in how calculations are
made were required so as not to use floating point arithmetic.
Signed-off-by: Denis Orlov
---
common/ddr_spd.c | 556 ++-
1
It was at odds with the list of options, with the former mentioning
RATE and the latter mentioning HZ as a parameter. It seems that every
other clk-related command uses HZ, so use it here too.
Signed-off-by: Denis Orlov
---
commands/clk.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion
Also reorder some of them to be more uniform.
Signed-off-by: Denis Orlov
---
arch/mips/include/asm/mipsregs.h | 12
1 file changed, 4 insertions(+), 8 deletions(-)
diff --git a/arch/mips/include/asm/mipsregs.h b/arch/mips/include/asm/mipsregs.h
index 9e7d656542..0b618320ff 100644
This adds support for identifying more CPUs emulatable by QEMU, so that
their names are shown when using 'cpuinfo' command.
Signed-off-by: Denis Orlov
---
arch/mips/include/asm/cpu.h | 30 +++
arch/mips/lib/cpu-probe.c | 59 -
2 files
Otherwise it would output as __cpu_name global variable ends up
being a null pointer.
Signed-off-by: Denis Orlov
---
arch/mips/lib/cpu-probe.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/arch/mips/lib/cpu-probe.c b/arch/mips/lib/cpu-probe.c
index cbde43a595..47c9c671eb
Denis Orlov (3):
mips: cpuinfo: fix cpu name output for unknown processors
mips: cpu-probe: identify more MIPS CPUs
mips: mipsregs: remove duplicate definitions for Config 1 bits
arch/mips/include/asm/cpu.h | 30 +---
arch/mips/include/asm/mipsregs.h | 12 +++
arch
101 - 142 of 142 matches
Mail list logo