Joshua Kinard [Wed, 8 Oct 2014 01:33:04 +0000 (21:33 -0400)]
MIPS: IP22/IP32: Add line to arch/mips/Makefile archhelp about vmlinux.32
Building a 64bit kernel for the SGI O2 (IP32) and the SGI Indy (IP22) uses
the 'vmlinux.32' target, which converts the output 64-bit 'vmlinux' image
into a 32-bit wrapped image. This is needed for certain revisions of the
IP22 and IP32 ARCS PROMs to boot correctly, but this target is missing
from the 'archhelp' info that is emitted by 'make help'.
Signed-off-by: Joshua Kinard <kumba@gentoo.org> Cc: Linux MIPS List <linux-mips@linux-mips.org>
Patchwork: https://patchwork.linux-mips.org/patch/7991/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Having #ifdefs just to guard comments is not really helpful
so drop them. Moreover, the code wasn't really reached anyway
since there is a #ifndef CONFIG_CPU_MIPSR2 on the top of the file.
We should not need to read fuses during normal operation, also the current
code has issues with that (not safe for concurrent access). Since there
are no in-kernel users for these, just delete them. Drivers should
not need such OCTEON_HAS_FEATURE mechanism in any case, instead the
information should be passed via device tree.
Signed-off-by: Aaro Koskinen <aaro.koskinen@nsn.com> Cc: David Daney <david.daney@cavium.com> Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/7665/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Aaro Koskinen [Fri, 17 Oct 2014 15:10:26 +0000 (18:10 +0300)]
MIPS: oprofile: Backtrace: don't fail on leaf functions
Continue the backtrace if we cannot find SP adjustment and RA save. In
that case, just assume the current RA. This allows us to get samples of
frequent callers of e.g. GLIBC memset().
Aaro Koskinen [Fri, 17 Oct 2014 15:10:25 +0000 (18:10 +0300)]
MIPS: oprofile: Enable backtrace on timer-based profiling
Allow unsupported CPU types to use backtrace with timer-based profiling.
Some CPUs (notably OCTEON) lack architecture-specific oprofile driver. In
such case oprofile can fallback to timer-based mode, and arch code can
still provide the backtrace functionality. So just set up the backtrace
hook always.
Add device-tree support for the MIPS GIC. Update the GIC irqdomain's
xlate() callback to handle the three-cell specifier described in the
MIPS GIC binding document.
Signed-off-by: Andrew Bresticker <abrestic@chromium.org> Acked-by: Arnd Bergmann <arnd@arndb.de> Acked-by: Jason Cooper <jason@lakedaemon.net> Cc: Rob Herring <robh+dt@kernel.org> Cc: Pawel Moll <pawel.moll@arm.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Ian Campbell <ijc+devicetree@hellion.org.uk> Cc: Kumar Gala <galak@codeaurora.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: John Crispin <blogic@openwrt.org> Cc: David Daney <ddaney.cavm@gmail.com> Cc: Qais Yousef <qais.yousef@imgtec.com> Cc: James Hogan <james.hogan@imgtec.com> Cc: linux-mips@linux-mips.org Cc: devicetree@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/8422/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
The Global Interrupt Controller (GIC) present on certain MIPS systems
can be used to route external interrupts to individual VPEs and CPU
interrupt vectors. It also supports a timer and software-generated
interrupts.
Signed-off-by: Andrew Bresticker <abrestic@chromium.org> Acked-by: Arnd Bergmann <arnd@arndb.de> Cc: Rob Herring <robh+dt@kernel.org> Cc: Pawel Moll <pawel.moll@arm.com> Cc: Mark Rutland <mark.rutland@arm.com> Cc: Ian Campbell <ijc+devicetree@hellion.org.uk> Cc: Kumar Gala <galak@codeaurora.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Jason Cooper <jason@lakedaemon.net> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: John Crispin <blogic@openwrt.org> Cc: David Daney <ddaney.cavm@gmail.com> Cc: Qais Yousef <qais.yousef@imgtec.com> Cc: James Hogan <james.hogan@imgtec.com> Cc: linux-mips@linux-mips.org Cc: devicetree@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/8420/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Alban Bedel [Fri, 7 Nov 2014 11:44:35 +0000 (12:44 +0100)]
MIPS: ath79: Use the firmware lib to parse the kernel command line
No need to duplicate code that is available in the firmware library.
It also give us access to the firmware environment which is needed
to read the initrd address and size.
Sergey Ryazanov [Tue, 28 Oct 2014 23:18:49 +0000 (03:18 +0400)]
ath5k: update dependencies
- Use config symbol defined in the driver instead of arch specific one for
conditional compilation.
- Rename the ATHEROS_AR231X config symbol to ATH25.
- Fix include (ar231x_platform.h -> ath25_platform.h).
- Some of AR231x SoCs (e.g. AR2315) have PCI bus support, so remove !PCI
dependency, which block AHB support build.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com> Acked-by: John W. Linville <linville@tuxdriver.com> Cc: Jiri Slaby <jirislaby@gmail.com> Cc: Nick Kossifidis <mickflemm@gmail.com> Cc: "Luis R. Rodriguez" <mcgrof@do-not-panic.com> Cc: linux-wireless@vger.kernel.org Cc: ath5k-devel@lists.ath5k.org Cc: Linux MIPS <linux-mips@linux-mips.org>
Patchwork: https://patchwork.linux-mips.org/patch/8248/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
AHB bus code has been removed, since we did not have support Atheros
AR231x SoC, required for building the AHB version of ath5k. Now that
support WiSoC chips added we can restore functionality back.
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com> Acked-by: John W. Linville <linville@tuxdriver.com> Cc: Jiri Slaby <jirislaby@gmail.com> Cc: Nick Kossifidis <mickflemm@gmail.com> Cc: "Luis R. Rodriguez" <mcgrof@do-not-panic.com> Cc: linux-wireless@vger.kernel.org Cc: ath5k-devel@lists.ath5k.org Cc: Linux MIPS <linux-mips@linux-mips.org>
Patchwork: https://patchwork.linux-mips.org/patch/8247/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Sergey Ryazanov [Tue, 28 Oct 2014 23:18:46 +0000 (03:18 +0400)]
MIPS: ath25: register AR5312 flash controller
AR5312 SoC flash controller maps the flash content to memory and
translates the memory access operations to the flash access operations.
Such controller is fully supported by the physmap-flash driver.
Sergey Ryazanov [Tue, 28 Oct 2014 23:18:44 +0000 (03:18 +0400)]
MIPS: ath25: add board configuration detection
All boards based on AR5312/AR2315 SoC have a special structure located
at the end of flash. This structure contains board-specific data such as
Ethernet and Wireless MAC addresses. The flash is mapped to the memmory
at predefined location.
Rewrite TURBOchannel error handling to use a common failure path, making
sure put_device is called for devices that failed initialization. While
at it update printk calls to use pr_err rather than KERN_ERR.
Signed-off-by: Maciej W. Rozycki <macro@linux-mips.org> Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/6701/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
John Crispin [Thu, 9 Oct 2014 02:02:53 +0000 (04:02 +0200)]
MIPS: ralink: cleanup the soc specific pinmux data
Before we had a pinctrl driver we used a custom OF api. This patch converts the
soc specific pinmux data to a new set of structs. We also add some new pinmux
setings.
Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/8009/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
John Crispin [Wed, 8 Oct 2014 21:28:51 +0000 (23:28 +0200)]
MIPS: ralink: copy the commandline from the devicetree
This is a regression caused by:
commit afb46f7996e91aeb36e07bc92cf96e8045bec00e
Author: Rob Herring <robh@kernel.org>
Date: Wed Apr 2 19:07:24 2014 -0500
mips: ralink: convert to use unflatten_and_copy_device_tree
Make the of init code reuse the cmdline defined inside the dts.
Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/8008/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
John Crispin [Wed, 8 Oct 2014 20:40:02 +0000 (22:40 +0200)]
MIPS: ralink: allow manual memory override
RT5350 relies on the bootloader setting up the memc correctly. On some boards
the setup is incorrect leading to 32 MB being available but only 16 MB being
recognized. Allow these boards to manually override the memory range.
Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/8004/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
John Crispin [Fri, 10 Oct 2014 21:14:29 +0000 (23:14 +0200)]
MIPS: lantiq: copy the commandline from the devicetree
This is a regression caused by:
commit afb46f7996e91aeb36e07bc92cf96e8045bec00e
Author: Rob Herring <robh@kernel.org>
Date: Wed Apr 2 19:07:24 2014 -0500
mips: ralink: convert to use unflatten_and_copy_device_tree
Make the of init code reuse the cmdline defined inside the dts.
Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/8048/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
John Crispin [Wed, 10 Sep 2014 20:39:19 +0000 (22:39 +0200)]
MIPS: lantiq: add support for xrx200 firmware depending on soc type
VR9 needs different firmware files for the various phy/soc revisions. Some
boards are ship with older and newer SoC revisions. To be able to boot a single
image on all versions we need to define both firmware files inside the
devicetree.
John Crispin [Wed, 13 Mar 2013 09:04:01 +0000 (10:04 +0100)]
MIPS: lantiq: handle vmmc memory reservation
The Lantiq SoCs have a 2nd mips core called "voice mips macro core (vmmc)"
which is used to run the voice firmware. This driver allows us to register
a chunk of memory that the voice driver can later use for the 2nd core.
Signed-off-by: John Crispin <blogic@openwrt.org>
Patchwork: http://patchwork.linux-mips.org/patch/8042/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
clocksource: mips-gic: Use CPU notifiers to setup the timer
Instead of requiring an explicit call to gic_clockevent_init in the SMP
startup path, use CPU notifiers to register and enable the GIC timer on
CPU startup.
Signed-off-by: Andrew Bresticker <abrestic@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Jason Cooper <jason@lakedaemon.net> Cc: Paul Burton <paul.burton@imgtec.com> Cc: Qais Yousef <qais.yousef@imgtec.com> Cc: John Crispin <blogic@openwrt.org> Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/8139/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
clocksource: mips-gic: Move gic_frequency to clocksource driver
There's no reason for gic_frequency to be global any more and it
certainly doesn't belong in the GIC irqchip driver, so move it to
the GIC clocksource driver.
Signed-off-by: Andrew Bresticker <abrestic@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Jason Cooper <jason@lakedaemon.net> Cc: Paul Burton <paul.burton@imgtec.com> Cc: Qais Yousef <qais.yousef@imgtec.com> Cc: John Crispin <blogic@openwrt.org> Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/8137/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
clocksource: mips-gic: Combine with GIC clockevent driver
Combine the GIC clocksource driver with the GIC clockevent driver from
arch/mips/kernel/cevt-gic.c and remove the clockevent driver's separate
Kconfig symbol.
Signed-off-by: Andrew Bresticker <abrestic@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Jason Cooper <jason@lakedaemon.net> Cc: Andrew Bresticker <abrestic@chromium.org> Cc: Paul Burton <paul.burton@imgtec.com> Cc: Qais Yousef <qais.yousef@imgtec.com> Cc: John Crispin <blogic@openwrt.org> Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/8132/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
There's no reason for the pending and masked interrupt bitmasks
to be global. Just declare them on the stack in gic_get_int()
since they only consume (256*2)/8 = 64 bytes.
Signed-off-by: Andrew Bresticker <abrestic@chromium.org> Cc: Daniel Lezcano <daniel.lezcano@linaro.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Jason Cooper <jason@lakedaemon.net> Cc: Paul Burton <paul.burton@imgtec.com> Cc: Qais Yousef <qais.yousef@imgtec.com> Cc: John Crispin <blogic@openwrt.org> Cc: linux-mips@linux-mips.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/8131/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Kevin Cernekee [Tue, 21 Oct 2014 04:28:05 +0000 (21:28 -0700)]
MIPS: bcm3384: Initial commit of bcm3384 platform support
This supports SMP Linux running on the BCM3384 Zephyr (BMIPS5000)
application processor, with fully functional UART and USB 1.1/2.0.
Device Tree is used to configure the following items:
- All peripherals
- Early console base address
- SMP or UP mode
- MIPS counter frequency
- Memory size / regions
- DMA offset
- Kernel command line
The DT-enabled bootloader and build instructions are posted at
https://github.com/Broadcom/aeolus
Kevin Cernekee [Tue, 21 Oct 2014 04:28:01 +0000 (21:28 -0700)]
MIPS: BMIPS: Add PRId for BMIPS5200 (Whirlwind)
This is a dual core (quad thread) BMIPS5000. It needs a little extra
code to boot the second core (CPU2/CPU3), but for now we can treat it the
same as a single core BMIPS5000.
Kevin Cernekee [Tue, 21 Oct 2014 04:28:00 +0000 (21:28 -0700)]
MIPS: BMIPS: Add special cache handling in c-r4k.c
BMIPS435x and BMIPS438x have a single shared L1 D$ and load/store unit,
so it isn't necessary to raise IPIs to keep both CPUs coherent.
BMIPS5000 has VIPT L1 caches that handle aliases in hardware, and its I$
fills from D$. But a special sequence with 2 SYNCs and 32 NOPs is needed
to ensure coherency.
Kevin Cernekee [Tue, 21 Oct 2014 04:27:59 +0000 (21:27 -0700)]
MIPS: BMIPS: Let each platform customize the CPU1 IRQ mask
On some chips like bcm3384, "other stuff" gets wired up to CPU1's IE_IRQ1
input, generating spurious IRQs. In this case we want the platform code
to be able to mask it off.
Kevin Cernekee [Tue, 21 Oct 2014 04:27:58 +0000 (21:27 -0700)]
MIPS: BMIPS: Select the appropriate L1_CACHE_SHIFT for 438x and 5000 CPUs
BMIPS438x has a 64-byte D$ line size and BMIPS5000 has a 128-byte L2
line size. If L1_CACHE_SHIFT is undersized, DMA buffers will not be
cacheline-aligned and terrible things will happen.
Kevin Cernekee [Tue, 21 Oct 2014 04:27:57 +0000 (21:27 -0700)]
MIPS: Allow MIPS_CPU_SCACHE to be used with different line sizes
CONFIG_MIPS_CPU_SCACHE determines whether to build sc-mips.c. However,
it is currently hardwired to use an L1_SHIFT of 6 (64 bytes). Move the
L1_SHIFT selection into the CPU or SoC section so that other SoCs can
select different values.
Kevin Cernekee [Tue, 21 Oct 2014 04:27:56 +0000 (21:27 -0700)]
MIPS: BMIPS: Explicitly configure reset vectors prior to secondary boot
The secondary CPU's reset vector needs to be set to KSEG1 for a cold
boot (release from reset), or KSEG0 for a warm restart. On a cold boot
KSEG0 may be unavailable (BMIPS4380), and on a warm restart KSEG1 may
be unavailable (XKS01 mode on 4380 or 5000).
Jon Fraser [Tue, 21 Oct 2014 04:27:55 +0000 (21:27 -0700)]
MIPS: BMIPS: Mask off timer IRQs when hot-unplugging a CPU
CPU interrupts need to be disabled on a cpu being taken down.
When a cpu is hot-plugged out of the system the following sequence occurs.
On the CPU where the hotplug sequence was initiated:
cpu_down
_cpu_down {
__cpu_notify(CPU_DOWN_PREPARE
__stop_machine(take_cpu_down
wait for cpu to run disable code.
__cpu_die
}
On the CPU being disabled:
take_cpu_down
__cpu_disable {
mp_ops->cpu_disable
bmips_cpu_disable
clear_c0_status(IE_IRQ5) (added)
cpu_notify(CPU_DYING...
}
Before the cpu_notifier is called with CPU_DYING, all interrupts on the
dying cpu must be disabled. This guarantees that before tick_notify is
called with the CPU_DYING event and sets the clock device pointer to
NULL, there can not be any more clock interrupts.
When this wasn't done, an unfortunately-timed timer interrupt sometimes
caused hangs immediately prior to system suspend:
Debug PM is not enabled. To enable partial suspend, rebuild kernel with CONFIG_PM_DEBUG
Pass 1 out of 1,PM: Syncing filesystems ... mode=none, tp1=done.
1, flags=5, cycle_tp=, sleep=
Freezing user space processes ... (elapsed 0.01 seconds) done.
Freezing remaining freezable tasks ... (elapsed 0.01 seconds) done.
PM: suspend of devices complete after 54.199 msecs
PM: late suspend of devices complete after 0.172 msecs
Disabling non-boot CPUs ...
SMP: CPU1 is offline
INFO: rcu_sched detected stalls on CPUs/tasks: { 3} (detected by 0, t=62537 jiffies)
Call Trace:
[<804baa78>] dump_stack+0x8/0x34
[<8008a2d8>] __rcu_pending+0x4b8/0x55c
[<8008adf4>] rcu_check_callbacks+0x78/0x180
[<80037830>] update_process_times+0x40/0x6c
[<80072fe4>] tick_sched_timer+0x74/0xe4
[<80050180>] __run_hrtimer.clone.30+0x64/0x140
[<80051150>] hrtimer_interrupt+0x19c/0x4bc
[<8000cdb8>] c0_compare_interrupt+0x50/0x88
[<80081b18>] handle_irq_event_percpu+0x5c/0x2f4
[<80086490>] handle_percpu_irq+0x8c/0xc0
[<800811b4>] generic_handle_irq+0x34/0x54
[<800067dc>] do_IRQ+0x18/0x2c
[<8000375c>] plat_irq_dispatch+0xd0/0x128
[<80004a04>] ret_from_irq+0x0/0x4
[<80004c40>] r4k_wait+0x20/0x40
[<80006b6c>] cpu_idle+0x98/0xf0
[<805d3988>] start_kernel+0x424/0x440
Jon Fraser [Tue, 21 Oct 2014 04:27:54 +0000 (21:27 -0700)]
MIPS: BMIPS: Allow BMIPS3300 to utilize SMP ebase relocation code
BMIPS3300 processors do not have the hardware to support SMP, but with a
small tweak, the SMP ebase relocation code allows BMIPS3300-based
platforms to reuse the S2/S3 power management code from BMIPS4380-based
chips. Normally this is as simple as adding one line to prom_init():
Kevin Cernekee [Tue, 21 Oct 2014 04:27:53 +0000 (21:27 -0700)]
MIPS: BMIPS: Introduce helper function to change the reset vector
This will need to be called from a few different places, and the logic
is starting to get a bit hairy (with the need for IPIs, CPU bug
workarounds, and hazards).
Kevin Cernekee [Tue, 21 Oct 2014 04:27:52 +0000 (21:27 -0700)]
MIPS: BMIPS: Align secondary boot sequence with latest firmware releases
On some older BMIPS5200 (dual core / quad thread) platforms, the
PROM code set up CPU2/CPU3 so they would be started through an NMI
instead of through the ACTION register. But this was incompatible with
some power management features that were later added, so the scheme was
changed so that Linux is fully responsible for booting CPU2/CPU3.
Kelvin Cheung [Fri, 10 Oct 2014 03:40:01 +0000 (11:40 +0800)]
MIPS: Loongson1B: Some fixes/updates for LS1B
- Fix hanging ethernet issue of LS1B v2.0 by adding pbl field in plat data.
(It seems that the MAC controller of LS1B v2.0 can only accept pbl=1)
- Add GMAC1 support and setup MUX in terms of PHY mode.
- Add CPUFreq support.
- Add MUX Register Definitions.
- Add PWM Register Definitions.
- Update clock register bitfields according to the latest spec.
- Update clock related stuff.
Rafał Miłecki [Thu, 30 Oct 2014 11:50:03 +0000 (12:50 +0100)]
MIPS: BCM47XX: Clean up nvram header
1) Move private defines to the .c file
2) Move SPROM helper to the sprom.c
3) Drop unused code
4) Rename magic to the NVRAM_MAGIC
5) Add const to the char pointer we never modify
Rafał Miłecki [Wed, 29 Oct 2014 09:05:06 +0000 (10:05 +0100)]
MIPS: BCM47XX: Use mtd as an alternative way/API to get NVRAM content
NVRAM can be read using magic memory offset, but after all it's just a
flash partition. On platforms where NVRAM isn't needed early we can get
it using mtd subsystem.
Paul Burton [Thu, 11 Sep 2014 07:30:23 +0000 (08:30 +0100)]
MIPS: Kconfig option to better exercise/debug hybrid FPRs
The hybrid FPR scheme exists to allow for compatibility between existing
FP32 code and newly compiled FP64A code. Such code should hopefully be
rare in the real world, and for the moment is difficult to come across.
All code except that built for the FP64 ABI can correctly execute using
the hybrid FPR scheme, so debugging the hybrid FPR implementation can
be eased by forcing all such code to use it. This is undesirable in
general due to the trap & emulate overhead of the hybrid FPR
implementation, but is a very useful option to have for debugging.
Signed-off-by: Paul Burton <paul.burton@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: linux-fsdevel@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/7680/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Paul Burton [Thu, 11 Sep 2014 07:30:22 +0000 (08:30 +0100)]
MIPS: ELF: Set FP mode according to .MIPS.abiflags
This patch reads the .MIPS.abiflags section when it is present, and sets
the FP mode of the task accordingly. Any loaded ELF files which do not
contain a .MIPS.abiflags section will continue to observe the previous
behaviour, that is FR=1 if EF_MIPS_FP64 is set else FR=0.
Signed-off-by: Paul Burton <paul.burton@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: linux-fsdevel@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/7681/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Paul Burton [Thu, 11 Sep 2014 07:30:21 +0000 (08:30 +0100)]
MIPS: ELF: Add definition for the .MIPS.abiflags section
New toolchains will generate a .MIPS.abiflags section, referenced by a
new PT_MIPS_ABIFLAGS program header. This section will provide
information about the requirements of the ELF, including the ISA level
the code is built for, the ASEs it requires, the size of various
registers and its expectations of the floating point mode. This patch
introduces a definition of the structure of this section and the program
header, for use in a subsequent patch.
Signed-off-by: Paul Burton <paul.burton@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: linux-fsdevel@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/7682/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Paul Burton [Thu, 11 Sep 2014 07:30:20 +0000 (08:30 +0100)]
MIPS: Support for hybrid FPRs
Hybrid FPRs is a scheme where scalar FP registers are 64b wide, but
accesses to odd indexed single registers use bits 63:32 of the
preceeding even indexed 64b register. In this mode all FP code
except that built for the plain FP64 ABI can execute correctly. Most
notably a combination of FP64A & FP32 code can execute correctly,
allowing for existing FP32 binaries to be linked with new FP64A binaries
that can make use of 64 bit FP & MSA.
Hybrid FPRs are implemented by setting both the FR & FRE bits, trapping
& emulating single precision FP instructions (via Reserved Instruction
exceptions) whilst allowing others to execute natively. It therefore has
a penalty in terms of execution speed, and should only be used when no
fully native mode can be. As more binaries are recompiled to use either
the FPXX or FP64(A) ABIs, the need for hybrid FPRs should diminish.
However in the short to mid term it allows for a gradual transition
towards that world, rather than a complete ABI break which is not
feasible for some users & not desirable for many.
A task will be executed using the hybrid FPR scheme when its
TIF_HYBRID_FPREGS flag is set & TIF_32BIT_FPREGS is clear. A further
patch will set the flags as necessary, this patch simply adds the
infrastructure necessary for the hybrid FPR mode to work.
Signed-off-by: Paul Burton <paul.burton@imgtec.com> Cc: linux-mips@linux-mips.org Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: linux-fsdevel@vger.kernel.org Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/7683/ Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
projects / linux-beck.git / log