Systems using the MIPS Coherence Manager (CM) cannot support multi-core
SMP with dcache aliasing. This is because CPU caches are VIPT, but
interventions in CM-based systems provide only the physical address to
remote caches. This means that interventions may behave incorrectly in
the presence of an aliasing dcache, since the physical address used
when handling an intervention may lead to operation on an aliased cache
line rather than the correct line.
Prevent us from running into this issue by refusing to boot secondary
cores in systems where dcache aliasing may occur.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/16196/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 5570ba2ee9)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Prior to MIPSr6 multithreading is only supported if CONFIG_MIPS_MT_SMP
is enabled, so CONFIG_MIPS_MT_SMP selects CONFIG_SYS_SUPPORTS_SCHED_SMT.
With MIPSr6 the CONFIG_MIPS_CPS SMP implementation always supports
multithreading, so have it select CONFIG_SYS_SUPPORTS_SCHED_SMT in order
to allow the scheduler to make better informed decisions on
multithreaded MIPSr6 systems (for example those using I6400 or I6500
CPUs).
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/16195/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit c8b7712c34)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
CM3 provides a GCR_CL_OTHER register per VP, rather than only per core.
This means that we don't need to prevent other VPs within a core from
racing with code that makes use of the core-other register region.
Reduce locking overhead by demoting the per-core spinlock providing
protection for CM2.5 & lower to a per-CPU/per-VP spinlock for CM3 &
higher.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/16193/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 516db1c61f)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
When CONFIG_MIPS_CPS_NS16550 is enabled, some register state is dumped
to the UART when an exception is taken via the BEV on secondary cores.
EJTAG exceptions are architecturally expected to be handled by the BEV
even when Status.BEV is 0. This effectively means that if userland
executes an sdbbp instruction on a secondary core then the kernel dumps
register state to the UART even though the exception is perfectly normal
& expected. Prevent this by simply not dumping information to the UART
for EJTAG exceptions.
Fixes: 609cf6f229 ("MIPS: CPS: Early debug using an ns16550-compatible UART")
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/12341/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 6609ccdc85)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
In preparation for supporting MIPSr6 multithreading (ie. VPs) which will
begin execution from the core reset vector, skip core level setup if the
core is already coherent. This is never the case when a core is first
started, since boot_core explicitly clears the cores GCR_Cx_COH_EN
register, and always the case when secondary VPs start since the first
VP to start will have enabled coherence after initialising the core &
its caches.
One notable side effect of this patch is that eva_init gets called
slightly earlier, prior to mips_cps_core_init rather than after it.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/12338/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 87a70bcdb4)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Fix mips_cm_max_vp_width for UP kernels where it previously referenced
smp_num_siblings, which is not declared for UP kernels. This led to
build errors such as the following:
drivers/built-in.o: In function `$L446':
irq-mips-gic.c:(.text+0x1994): undefined reference to `smp_num_siblings'
drivers/built-in.o:irq-mips-gic.c:(.text+0x199c): more undefined references to `smp_num_siblings' follow
On UP kernels simply return 1, leaving the reference to smp_num_siblings
in place only for SMP kernels.
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Matt Redfearn <matt.redfearn@imgtec.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/12332/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit a60ae81e5e)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
This is similar to commit 934c79231c ("MIPS: asm: r4kcache: Add MIPS
R6 cache unroll functions"). The CACHE instruction has been redefined
for MIPSr6 and it reduced its offset field to 8 bits. This leads to
micro-assembler field overflow warnings when booting SMP MIPSr6 cores
like the following one:
Call Trace:
[<ffffffff8010af88>] show_stack+0x68/0x88
[<ffffffff8056ddf0>] dump_stack+0x68/0x88
[<ffffffff801305bc>] warn_slowpath_common+0x8c/0xc8
[<ffffffff80130630>] warn_slowpath_fmt+0x38/0x48
[<ffffffff80125814>] build_insn+0x514/0x5c0
[<ffffffff806ee134>] cps_gen_cache_routine.isra.3+0xe0/0x1b8
[<ffffffff806ee570>] cps_pm_init+0x364/0x9ec
[<ffffffff80100538>] do_one_initcall+0x90/0x1a8
[<ffffffff806e8c14>] kernel_init_freeable+0x160/0x21c
[<ffffffff8056b6a0>] kernel_init+0x10/0xf8
[<ffffffff801059f8>] ret_from_kernel_thread+0x14/0x1c
We fix this by incrementing the base register on every loop.
Signed-off-by: Markos Chandras <markos.chandras@imgtec.com>
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/12329/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 0f2a148448)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
GCC-7 complains about a boolean value being used with an arithmetic
AND:
arch/mips/math-emu/cp1emu.c: In function 'cop1Emulate':
arch/mips/math-emu/cp1emu.c:838:14: warning: '~' on a boolean expression [-Wbool-operation]
fpr = (x) & ~(cop1_64bit(xcp) == 0); \
^
arch/mips/math-emu/cp1emu.c:1068:3: note: in expansion of macro 'DITOREG'
DITOREG(dval, MIPSInst_RT(ir));
^~~~~~~
arch/mips/math-emu/cp1emu.c:838:14: note: did you mean to use logical not?
fpr = (x) & ~(cop1_64bit(xcp) == 0); \
Since cop1_64bit() returns and int, just flip the LSB.
Suggested-by: Maciej W. Rozycki <macro@imgtec.com>
Signed-off-by: Manuel Lauss <manuel.lauss@gmail.com>
Reviewed-by: Maciej W. Rozycki <macro@imgtec.com>
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/17058/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 8535f2ba0a)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Add early console functionality to the Goldfish tty driver.
When 'earlycon' kernel command line parameter is used with no options,
the early console is determined by the 'stdout-path' property in device
tree's 'chosen' node. This is illustrated in the following device tree
source example:
Device tree example:
chosen {
stdout-path = "/goldfish_tty@1f004000";
};
goldfish_tty@1f004000 {
interrupts = <0xc>;
reg = <0x1f004000 0x0 0x1000>;
compatible = "google,goldfish-tty";
};
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 3840ed9548)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Implement tty r/w operations using streaming DMA.
Goldfish tty for Ranchu platforms has been modified to use
streaming DMA mappings for read/write operations. This change
eliminates the need for snooping through the TLB in QEMU using
cpu_get_phys_page_debug() which does not guarantee that it will
return the valid va -> pa mapping.
The streaming DMA mapping is implemented using dma_map_single() per
transfer, while dma_unmap_single() is used for unmapping right after
the DMA transfer.
Using DMA API is the proper way for handling r/w transfers and
makes this driver more portable, thus effectively eliminating
the need for virt_to_page() and page_to_phys() conversions.
This change does not affect the old style Goldfish tty behaviour
which is still used by the Goldfish emulator. Version register has
been added and probed to see which platform is running this driver.
Reading from the new register GOLDFISH_TTY_REG_VERSION using the
Goldfish emulator will return 0 and driver will work with virtual
addresses. Whereas if run on Ranchu it returns 1, and thus DMA is
used.
(Goldfish and Ranchu are code names for the first and the second
generation of virtual boards used by Android emulator.)
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 7157d2be23)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Classify constants GOLDFISH_TTY_xxx into two groups: command ids and
register offsets. Apply different naming for register offsets (add
'REG_' after 'GOLDFISH_TTY_' in constant names). Change implementation
to use preprocessor's '#define' statements instead of 'enum'
declaration (as this is more common way of implementation in such
cases).
This makes the driver code easier to follow and hopefully prevents
future bugs.
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
(cherry picked from commit 2296eee704)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Add FP emulation debugfs statistics for individual instructions. The
debugfs files that contain counter values are placed in a separate
directory called "instructions". This means that the default path for
these new stat is "/sys/kernel/debug/mips/fpuemustats/instructions".
Each instruction counter is mapped to the debugfs file that has the
same name as instruction name. The lowercase is choosen as more
commonly used case for instruction names.
One example of usage:
mips_host::/sys/kernel/debug/mips/fpuemustats/instructions # grep "" *
The shortened output of this command is:
abs.d:34
abs.s:5711
add.d:10401
add.s:399307
bc1eqz:3199
...
...
...
sub.s:167211
trunc.l.d:375
trunc.l.s:8054
trunc.w.d:421
trunc.w.s:27032
The limitation of this patch is that it handles R6 FP emulation
instructions only. There are altogether 114 handled instructions.
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Cc: Douglas Leung <douglas.leung@imgtec.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: Maciej W. Rozycki <macro@imgtec.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Petar Jovanovic <petar.jovanovic@imgtec.com>
Cc: Raghu Gandham <raghu.gandham@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/17145/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 454854ace2)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Reimplement RINT.<D|S> kernel emulation so that all RINT.<D|S>
specifications are met.
For the sake of simplicity, let's analyze RINT.S only. Prior to
this patch, RINT.S emulation was essentially implemented as (in
pseudocode) <output> = ieee754sp_flong(ieee754sp_tlong(<input>)),
where ieee754sp_tlong() and ieee754sp_flong() are functions
providing conversion from double to integer, and from integer
to double, respectively. On surface, this implementation looks
correct, but actually fails in many cases. Following problems
were detected:
1. NaN and infinity cases will not be handled properly. The
function ieee754sp_flong() never returns NaN nor infinity.
2. For RINT.S, for all inputs larger than LONG_MAX, and smaller
than FLT_MAX, the result will be wrong, and the overflow
exception will be erroneously set. A similar problem for
negative inputs exists as well.
3. For some rounding modes, for some negative inputs close to zero,
the return value will be zero, and should be -zero. This is
because ieee754sp_flong() never returns -zero.
This patch removes the problems above by implementing dedicated
functions for RINT.<D|S> emulation.
The core of the new function functionality is adapted version of
the core of the function ieee754sp_tlong(). However, there are many
details that are implemented to match RINT.<D|S> specification. It
should be said that the functionality of ieee754sp_tlong() actually
closely corresponds to CVT.L.S instruction, and it is used while
emulating CVT.L.S. However, RINT.S and CVT.L.S instructions differ
in many aspects. This patch fulfills missing support for RINT.<D|S>.
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Douglas Leung <douglas.leung@imgtec.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Hans Verkuil <hans.verkuil@cisco.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: Maciej W. Rozycki <macro@imgtec.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Petar Jovanovic <petar.jovanovic@imgtec.com>
Cc: Raghu Gandham <raghu.gandham@imgtec.com>
Cc: Randy Dunlap <rdunlap@infradead.org>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/17141/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 3ec404d88c)
Conflicts:
MAINTAINERS
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Fix CMP.Sxxx.<D|S> SIGILL crashes by fixing main switch/case statement
in fpu_emul() function so that inadvertent fall-troughs are prevented.
Consider, let's say, CMP.SAF.S instruction when one of inputs is zero
and another input is a signaling NaN. The desired output is zero, and
the exception flag "invalid operation" set. For such case, the main
portion of the implementation is within "d_fmt" case of the main
"switch/case" statement in fpu_emul() function. The execution will
follow one of "if-else" branches that doesn't contain "goto cop1scr;"
statement, and will therefore reach the end of "d_fmt" case. It will
subsequently fall through to the next case, "l_fmt". After following
similar pattern, the execution will fall through to the succeeding
case, which is "default". The "default" case contains "return SIGILL;"
statement only. This means that the caller application will crash
with "illegal instruction" message.
It is obvious that above described fall-throughs are unnecessary and
harmful. This patch rectifies that behavior by providing "break;"
statements at the end of cases "d_fmt" and "l_fmt".
There are 22 instructions affected by this problem:
CMP.<SAF|SEQ|SLE|SLT|SNE|SOR|SUEQ|SULE|SULT|SUN|SUNE>.<D|S>.
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Cc: Douglas Leung <douglas.leung@imgtec.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: Maciej W. Rozycki <macro@imgtec.com>
Cc: Masahiro Yamada <yamada.masahiro@socionext.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Petar Jovanovic <petar.jovanovic@imgtec.com>
Cc: Raghu Gandham <raghu.gandham@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/17140/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit 1ff8560ac9)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
Fix the cases of <MADDF|MSUBF>.<D|S> when any of three inputs is any
NaN. Correct behavior of <MADDF|MSUBF>.<D|S> fd, fs, ft is following:
- if any of inputs is sNaN, return a sNaN using following rules: if
only one input is sNaN, return that one; if more than one input is
sNaN, order of precedence for return value is fd, fs, ft
- if no input is sNaN, but at least one of inputs is qNaN, return a
qNaN using following rules: if only one input is qNaN, return that
one; if more than one input is qNaN, order of precedence for
return value is fd, fs, ft
The previous code contained correct handling of some above cases, but
not all. Also, such handling was scattered into various cases of
"switch (CLPAIR(xc, yc))" statement, and elsewhere. With this patch,
this logic is placed in one place, and "switch (CLPAIR(xc, yc))" is
significantly simplified.
A relevant example:
MADDF.S fd,fs,ft:
If fs contains qNaN1, ft contains qNaN2, and fd contains qNaN3, fd
is going to contain qNaN3 (without this patch, it used to contain
qNaN1).
Fixes: e24c3bec3e ("MIPS: math-emu: Add support for the MIPS R6 MADDF FPU instruction")
Fixes: 83d43305a1 ("MIPS: math-emu: Add support for the MIPS R6 MSUBF FPU instruction")
Signed-off-by: Miodrag Dinic <miodrag.dinic@imgtec.com>
Signed-off-by: Goran Ferenc <goran.ferenc@imgtec.com>
Signed-off-by: Aleksandar Markovic <aleksandar.markovic@imgtec.com>
Reviewed-by: James Hogan <james.hogan@imgtec.com>
Cc: Bo Hu <bohu@google.com>
Cc: Douglas Leung <douglas.leung@imgtec.com>
Cc: Jin Qian <jinqian@google.com>
Cc: Paul Burton <paul.burton@imgtec.com>
Cc: Petar Jovanovic <petar.jovanovic@imgtec.com>
Cc: Raghu Gandham <raghu.gandham@imgtec.com>
Cc: <stable@vger.kernel.org> # 4.7+
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/16886/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit e840be6e70)
Signed-off-by: Greg Kroah-Hartman <gregkh@google.com>
