| Age | Commit message (Collapse) | Author |
|---|
|
This patch has been reworked since RFC version. In the RFC, this patch
was preceded by a patch clearing MSR RI for all PPC32 at all time at
exception prologs. Now MSR RI clearing is done only when this 8xx perf
events functionality is compiled in, it is therefore limited to 8xx
and merged inside this patch.
Other main changes have been to take into account detailed review from
Peter Zijlstra. The instructions counter has been reworked to behave
as a free running counter like the three other counters.
The 8xx has no PMU, however some events can be emulated by other means.
This patch implements the following events (as reported by 'perf list'):
cpu-cycles OR cycles [Hardware event]
instructions [Hardware event]
dTLB-load-misses [Hardware cache event]
iTLB-load-misses [Hardware cache event]
'cycles' event is implemented using the timebase clock. Timebase clock
corresponds to CPU clock divided by 16, so number of cycles is
approximatly 16 times the number of TB ticks
On the 8xx, TLB misses are handled by software. It is therefore
easy to count all TLB misses each time the TLB miss exception is
called.
'instructions' is calculated by using instruction watchpoint counter.
This patch sets counter A to count instructions at address greater
than 0, hence we count all instructions executed while MSR RI bit is
set. The counter is set to the maximum which is 0xffff. Every 65535
instructions, debug instruction breakpoint exception fires. The
exception handler increments a counter in memory which then
represent the upper part of the instruction counter. We therefore
end up with a 48 bits counter. In order to avoid unnecessary overhead
while no perf event is active, this counter is started when the first
event referring to this counter is added, and the counter is stopped
when the last event referring to it is deleted. In order to properly
support breakpoint exceptions, MSR RI bit has to be unset in exception
epilogs in order to avoid breakpoint exceptions during critical
sections during changes to SRR0 and SRR1 would be problematic.
All counters are handled as free running counters.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
This patch implements HW breakpoint on the 8xx. The 8xx has
capability to manage HW breakpoints, which is slightly different
than BOOK3S:
1/ The breakpoint match doesn't trigger a DSI exception but a
dedicated data breakpoint exception.
2/ The breakpoint happens after the instruction has completed,
no need to single step or emulate the instruction,
3/ Matched address is not set in DAR but in BAR,
4/ DABR register doesn't exist, instead we have registers
LCTRL1, LCTRL2 and CMPx registers,
5/ The match on one comparator is not on a double word but
on a single word.
The patch does:
1/ Prepare the dedicated registers in call to __set_dabr(). In order
to emulate the double word handling of BOOK3S, comparator E is set to
DABR address value and comparator F to address + 4. Then breakpoint 1
is set to match comparator E or F,
2/ Skip the singlestepping stage when compiled for CONFIG_PPC_8xx,
3/ Implement the exception. In that exception, the matched address
is taken from SPRN_BAR and manage as if it was from SPRN_DAR.
4/ I/D TLB error exception routines perform a tlbie on bad TLBs. That
tlbie triggers the breakpoint exception when performed on the
breakpoint address. For this reason, the routine returns if the match
is from one of those two tlbie.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
8xx uses a two level page table with two different linux page size
support (4k and 16k). 8xx also support two different hugepage sizes
512k and 8M. In order to support them on linux we define two different
page table layout.
The size of pages is in the PGD entry, using PS field (bits 28-29):
00 : Small pages (4k or 16k)
01 : 512k pages
10 : reserved
11 : 8M pages
For 512K hugepage size a pgd entry have the below format
[<hugepte address >0101] . The hugepte table allocated will contain 8
entries pointing to 512K huge pte in 4k pages mode and 64 entries in
16k pages mode.
For 8M in 16k mode, a pgd entry have the below format
[<hugepte address >1101] . The hugepte table allocated will contain 8
entries pointing to 8M huge pte.
For 8M in 4k mode, multiple pgd entries point to the same hugepte
address and pgd entry will have the below format
[<hugepte address>1101]. The hugepte table allocated will only have one
entry.
For the time being, we do not support CPU15 ERRATA when HUGETLB is
selected
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Reviewed-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> (v3, for the generic bits)
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild
Pull kbuild updates from Michal Marek:
- EXPORT_SYMBOL for asm source by Al Viro.
This does bring a regression, because genksyms no longer generates
checksums for these symbols (CONFIG_MODVERSIONS). Nick Piggin is
working on a patch to fix this.
Plus, we are talking about functions like strcpy(), which rarely
change prototypes.
- Fixes for PPC fallout of the above by Stephen Rothwell and Nick
Piggin
- fixdep speedup by Alexey Dobriyan.
- preparatory work by Nick Piggin to allow architectures to build with
-ffunction-sections, -fdata-sections and --gc-sections
- CONFIG_THIN_ARCHIVES support by Stephen Rothwell
- fix for filenames with colons in the initramfs source by me.
* 'kbuild' of git://git.kernel.org/pub/scm/linux/kernel/git/mmarek/kbuild: (22 commits)
initramfs: Escape colons in depfile
ppc: there is no clear_pages to export
powerpc/64: whitelist unresolved modversions CRCs
kbuild: -ffunction-sections fix for archs with conflicting sections
kbuild: add arch specific post-link Makefile
kbuild: allow archs to select link dead code/data elimination
kbuild: allow architectures to use thin archives instead of ld -r
kbuild: Regenerate genksyms lexer
kbuild: genksyms fix for typeof handling
fixdep: faster CONFIG_ search
ia64: move exports to definitions
sparc32: debride memcpy.S a bit
[sparc] unify 32bit and 64bit string.h
sparc: move exports to definitions
ppc: move exports to definitions
arm: move exports to definitions
s390: move exports to definitions
m68k: move exports to definitions
alpha: move exports to actual definitions
x86: move exports to actual definitions
...
|
|
User space DTLB miss represent approximatly 90% of TLB misses
so make it the shortest path.
Also remove an unneccessary double jump in FixupDAR
Before this patch, we spend 3.3 TB ticks in the handler for each
user address miss and 3.4 TB ticks for each kernel address miss
After this patch, we send 3.0 TB ticks in the handler for each
user address miss and 3.9 TB ticks for each kernel address miss
Taking into account that user misses represent 90% of the total,
this patch provides an improvement of approx. 9%
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
When all options are activated, there is not enough space for the
DTLBMiss handlers that handles IMMR area and linear RAM pages in
the exception area once we have added hugepage handling.
So lets move them after .0x2000
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
When the watchdog is in NMI mode, the system reset interrupt is
generated when the watchdog counter expires.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
CLR_TOP32() is defined as blank. Last useful instance of CLR_TOP32()
was removed by commit 40ef8cbc6d360 ("powerpc: Get 64-bit configs to
compile with ARCH=powerpc") in 2005.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
|
|
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
|
|
CONFIG_PIN_TLB maps IMMR area and the first 24 Mbytes of memory.
In some circunstances it might be more interesting to not map
IMMR but map 32 Mbytes of memory instead.
Therefore we add config option CONFIG_PIN_TLB_IMMR to select if
IMMR shall be pinned or not, hence whether we pin 24 or 32 Mbytes of RAM
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
On recent kernels, with some debug options like for instance
CONFIG_LOCKDEP, the BSS requires more than 8M memory, allthough
the kernel code fits in the first 8M.
Today, it is necessary to activate CONFIG_PIN_TLB to get more than 8M
at startup, allthough pinning TLB is not necessary for that.
We could have inconditionaly mapped 16 or 24M bytes at startup
but some old hardware only have 8M and mapping non-existing RAM
would be an issue due to speculative accesses.
With the preceding patch however, the TLB entries are populated on
demand. By setting up the TLB miss handler to handle up to 24M until
the handler is patched for the entire memory space, it is possible
to allow access up to more memory without mapping non-existing RAM.
It is therefore not needed anymore to map memory data at all
at startup. It will be handled by the TLB miss handler.
One might still want to PIN the IMMR and the first 24M of RAM.
It is now possible to do it in the C memory initialisation
functions. In addition, we now know how much memory we have
when we do it, so we are able to adapt the pining to the
real amount of memory available. So boards with less than 24M
can now also benefit from PIN_TLB.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Instead of using the first level page table to define mappings for
the linear memory space, we can use direct mapping from the TLB
handling routines. This has several advantages:
* No need to read the tables at each TLB miss
* No issue in 16k pages mode where the 1st level table maps 64 Mbytes
The size of the available linear space is known at system startup.
In order to avoid data access at each TLB miss to know the memory
size, the TLB routine is patched at startup with the proper size
This patch provides a 10%-15% improvment of TLB miss handling for
kernel addresses
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Bootloader may have pinned some TLB entries so the kernel must
unpin them before flushing TLBs with tlbia otherwise pinned TLB
entries won't get flushed
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Once the linear memory space has been mapped with 8Mb pages, as
seen in the related commit, we get 11 millions DTLB missed during
the reference 600s period. 77% of the misses are on user addresses
and 23% are on kernel addresses (1 fourth for linear address space
and 3 fourth for virtual address space)
Traditionaly, each driver manages one computer board which has its
own components with its own memory maps.
But on embedded chips like the MPC8xx, the SOC has all registers
located in the same IO area.
When looking at ioremaps done during startup, we see that
many drivers are re-mapping small parts of the IMMR for their own use
and all those small pieces gets their own 4k page, amplifying the
number of TLB misses: in our system we get 0xff000000 mapped 31 times
and 0xff003000 mapped 9 times.
Even if each part of IMMR was mapped only once with 4k pages, it would
still be several small mappings towards linear area.
This patch maps the IMMR with a single 512k page.
With this patch applied, the number of DTLB misses during the 10 min
period is reduced to 11.8 millions for a duration of 5.8s, which
represents 2% of the non-idle time hence yet another 10% reduction.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
Memory: 124428K/131072K available (3748K kernel code, 188K rwdata,
648K rodata, 508K init, 290K bss, 6644K reserved)
Kernel virtual memory layout:
* 0xfffdf000..0xfffff000 : fixmap
* 0xfde00000..0xfe000000 : consistent mem
* 0xfddf6000..0xfde00000 : early ioremap
* 0xc9000000..0xfddf6000 : vmalloc & ioremap
SLUB: HWalign=16, Order=0-3, MinObjects=0, CPUs=1, Nodes=1
Today, IMMR is mapped 1:1 at startup
Mapping IMMR 1:1 is just wrong because it may overlap with another
area. On most mpc8xx boards it is OK as IMMR is set to 0xff000000
but for instance on EP88xC board, IMMR is at 0xfa200000 which
overlaps with VM ioremap area
This patch fixes the virtual address for remapping IMMR with the fixmap
regardless of the value of IMMR.
The size of IMMR area is 256kbytes (CPM at offset 0, security engine
at offset 128k) so a 512k page is enough
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
There is no real need to have set_context() in assembly.
Now that we have mtspr() handling CPU6 ERRATA directly, we
can rewrite set_context() in C language for easier maintenance.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
CPU6 ERRATA is now handled directly in mtspr(), so we can use the
standard set_dec() fonction in all cases.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
On a live running system (VoIP gateway for Air Trafic Control), over
a 10 minutes period (with 277s idle), we get 87 millions DTLB misses
and approximatly 35 secondes are spent in DTLB handler.
This represents 5.8% of the overall time and even 10.8% of the
non-idle time.
Among those 87 millions DTLB misses, 15% are on user addresses and
85% are on kernel addresses. And within the kernel addresses, 93%
are on addresses from the linear address space and only 7% are on
addresses from the virtual address space.
MPC8xx has no BATs but it has 8Mb page size. This patch implements
mapping of kernel RAM using 8Mb pages, on the same model as what is
done on the 40x.
In 4k pages mode, each PGD entry maps a 4Mb area: we map every two
entries to the same 8Mb physical page. In each second entry, we add
4Mb to the page physical address to ease life of the FixupDAR
routine. This is just ignored by HW.
In 16k pages mode, each PGD entry maps a 64Mb area: each PGD entry
will point to the first page of the area. The DTLB handler adds
the 3 bits from EPN to map the correct page.
With this patch applied, we now get only 13 millions TLB misses
during the 10 minutes period. The idle time has increased to 313s
and the overall time spent in DTLB miss handler is 6.3s, which
represents 1% of the overall time and 2.2% of non-idle time.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
We are spending between 40 and 160 cycles with a mean of 65 cycles in
the DTLB handling routine (measured with mftbl) so make it more
simple althought it adds one instruction.
With this modification, we get three registers available at all time,
which will help with following patch.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
When CONFIG_DEBUG_PAGEALLOC is activated, the initial TLB mapping gets
flushed to track accesses to wrong areas. Therefore, kernel addresses
will also generate ITLB misses.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <oss@buserror.net>
|
|
This patch implements PAGE_EXEC capability on the 8xx.
All pages PP exec bits are set to 000, which means Execute for
Supervisor and no Execute for User.
Then we use the APG to say whether accesses are according to Page
rules, "all Supervisor" rules (Exec for all) and
"all User" rules (Exec for noone)
Therefore, we define 4 APG groups. msb is _PAGE_EXEC,
lsb is _PAGE_USER. MI_AP is initialised as follows:
GP0 (00) => Not User, no exec => 11 (all accesses performed as user)
GP1 (01) => User but no exec => 11 (all accesses performed as user)
GP2 (10) => Not User, exec => 01 (rights according to page definition)
GP3 (11) => User, exec => 00 (all accesses performed as supervisor)
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
[scottwood: comments: s/exec/data/ on data side, and s/pages/pages'/]
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Use of APG for handling PAGE_USER.
All pages PP exec bits are set to either 000 or 011, which means
respectively RW for Supervisor and no access for User, or RO for
Supervisor and no access for user.
Then we use the APG to say whether accesses are according to
Page rules or "all Supervisor" rules (Access to all)
Therefore, we define 2 APG groups corresponding to _PAGE_USER.
Mx_AP are initialised as follows:
GP0 => No user => 01 (all accesses performed according
to page definition)
GP1 => User => 00 (all accesses performed as supervisor
according to page definition)
This removes the special 8xx handling in pte_update()
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
By default, TASK_SIZE is set to 0x80000000 for PPC_8xx, which is most
likely sufficient for most cases. However, kernel configuration allows
to set TASK_SIZE to another value, so the 8xx shall handle it.
This patch also takes into account the case of PAGE_OFFSET lower than
0x80000000, allthought most of the time it is equal to 0xC0000000
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
We now have SPRG2 available as in it not used anymore for saving CR, so we don't
need to crash DAR anymore for saving r3 for CPU6 ERRATA handling.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
CR only needs to be preserved when checking if we are handling a kernel address.
So we can preserve CR in a register:
- In ITLBMiss, check is done only when CONFIG_MODULES is defined. Otherwise we
don't need to do anything at all with CR.
- We use r10, then we reload SRR0/MD_EPN into r10 when CR is restored
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
In order to be able to reduce scope during which CR is saved, we take
CR saving/restoring out of exception PROLOG and EPILOG
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Having a macro will help keep clear code.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
All accessed to PGD entries are done via 0(r11).
By using lower part of swapper_pg_dir as load index to r11, we can remove the
ori instruction.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
L1 base address is now aligned so we can insert L1 index into r11 directly and
then preserve r10
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Kernel MMU handling code handles validity of entries via _PMD_PRESENT which
corresponds to V bit in MD_TWC and MI_TWC. When the V bit is not set, MPC8xx
triggers TLBError exception. So we don't have to check that and branch ourself
to TLBError. We can set TLB entries with non present entries, remove all those
tests and let the 8xx handle it. This reduce the number of cycle when the
entries are valid which is the case most of the time, and doesn't significantly
increase the time for handling invalid entries.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Since commit 33fb845a6f01 ("powerpc/8xx: Don't use MD_TWC for walk"), MD_EPN and
MD_TWC are not writen anymore in FixupDAR so saving r3 has become useless.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
On powerpc 8xx, in TLB entries, 0x400 bit is set to 1 for read-only pages
and is set to 0 for RW pages. So we should use _PAGE_RO instead of _PAGE_RW
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
8xx sometimes need to load a invalid/non-present TLBs in
it DTLB asm handler.
These must be invalidated separaly as linux mm doesn't.
Commit 5efab4a02c89c252fb4cce097aafde5f8208dbfe was invalidating them in
arch/powerpc/mm/fault.c.
This patch does the invalidation earlier in order to free the TLB as soon as
possible. This also has the advantage of removing some 8xx specific code from
fault.c
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
As we are not using anymore DAR to save registers, it is now available for
saving the r3 register used for CPU6 ERRATA handling. Therefore we can
remove the major hack which was to use memory location 0 to save r3.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
There is not need to restore r10, r11 and cr registers at this end of ITLBmiss
handler as they are saved again to the same place in ITLBError handler we are
jumping to.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
When a PMD entry is valid, _PMD_PRESENT is set. Therefore, forcing that bit
during TLB loading is useless.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
No need to re-set this bit at each TLB miss. Let's set it in the PTE.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
This patch hiddes that SPR address needed for CPU6 ERRATA handling in the macro.
Then we don't have to worry about this address directly in the code.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
This patch activates the handling of 16k pages on the MPC8xx.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Value 0x00f0 is used to force bits in TLB level 2 entry. This value is linked
to the page size and will vary when we change the page size. Lets define a const
for it in order to have it at only one place.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
For PAGE size related operations, use PAGE size consts in order to be able to
use different page size in the futur.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
MD_TWC can only be used properly with 4k pages.
So lets calculate level 2 table index by ourselves.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Use M_TW instead of M_TWB for storing Level 1 table address as M_TWB requires
4k aligned tables, which is only the case with 4k pages.
Consequently, we have to calculate the level 1 table index by ourselves.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
In DTLBError handler there is not need to restore r10, r11 and cr registers
after fixing DAR as they are saved again to the same place just after.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
DataAccess exception is never generated by MPC8xx so do the job directly where
it is used to avoid an unnecessary branching.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Exception InstructionAccess does not exist on MPC8xx. No need to branch there from somewhere else.
Handling can be done directly in InstructionTLBError Exception.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
Branching takes two cycles on MPC8xx. Lets duplicate the two instructions
and avoid the branching.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
|
By XORing the upper part of the instruction code, we get a value that can
directly be verified with the second test and we can remove the first test.
Signed-off-by: Christophe Leroy <christophe.leroy@c-s.fr>
Signed-off-by: Scott Wood <scottwood@freescale.com>
|
