Wolfgang Denk wrote on 2010/06/29 14:55:44:
>
> Dear Joakim Tjernlund,
>
> In message 0045d...@transmode.se> you wrote:
> >
> > > I could not come up with a better name... What is "align32wrap"
> > > supposed to mean?
> >
> > mpc5200_
>
> Dear Detlev Zundel,
>
> In message you wrote:
> >
> > Acked-by: Detlev Zundel
>
> thanks.
>
> > The only thing I wonder is the filename 'memcpy_mpc5200.c' as the code
> > doesn't really have any 5200 specifics in it. What about
>
> As far as I understand this behaviour is specific to the MPC
> Hi All,
>
> Just a really quick message to say that I now have Linux Kernel 2.6.33 and
> a JFFS2 Root File System up and running on my AMD SC520 board.
>
> I big thanks to everyone for helping out on this 'little' project of mine.
>
> There is no way I could have possibly gained so much personal
"Peter Vollmer" wrote on 2010/04/20 14:12:41:
>
> On Tue, 20 Apr 2010 11:57:09 +0200, Joakim Tjernlund
> wrote:
> > Probably a BDI2000 issue. If memory serves right BDI2000 flushes the
> > cache when it hits a BP.
> >
> > Try this addin this to
>
> On Tue, 20 Apr 2010 10:36:08 +0200, Peter Vollmer
> wrote:
>
>
> > I then inserted 8 pairs of lwzu/stwu inside the loop to see when exactly
> > the problem occurs:
> >
> > 1: lwzur0,4(r8)
> > stwur0,4(r7)
> > lwzur0,4(r8)
> > stwur0,4(r7)
> > ...
> > lwzu
>
> Michael Zaidman wrote on 2010/04/16 18:44:12:
> >
> > On Fri, Apr 16, 2010 at 1:25 AM, Kim Phillips
> > wrote:
> > > On Thu, 8 Apr 2010 10:37:08 +0200
> > > Joakim Tjernlund wrote:
> > >
> > >> Kim Phillips wrote on 2010-04-08
Michael Zaidman wrote on 2010/04/16 18:44:12:
>
> On Fri, Apr 16, 2010 at 1:25 AM, Kim Phillips
> wrote:
> > On Thu, 8 Apr 2010 10:37:08 +0200
> > Joakim Tjernlund wrote:
> >
> >> Kim Phillips wrote on 2010-04-08 10:27:03:
> Although this is good for m
>
>
> Hi Detlev,
>
> I used "iminfo 0xffdc" and got following log
> Image Name: Linux-2.6.31-LE
> Image Type: PowerPC Linux Kernel Image (gzip compressed)
> Data Size: 936145 Bytes = 914.2 kB
> Load Address:
> Entry Point:
> Bad Data CRC
>
> Here i am
Detlev Zundel wrote on 2010-04-08 18:06:40:
>
> Hi Jocke,
Hi again :)
>
> [...]
>
> > Since an escape char appear to exist, one should be able to use it much
> > like you
> > did above so I don't think that any use case disappears. Instead the common
> > usage
> > becomes simpler and the so fa
Detlev Zundel wrote on 2010-04-08 12:00:05:
>
> Hi Jocke,
Hi Detlev :)
>
> >> > > To me it looks like the new code would indeed do a "greedy"
> >> > > substitution
> >> > > only stopping when no more substitutions can be done. This is very
> >> > > un-unixy and thus not something I'd like to s
Kim Phillips wrote on 2010-04-08 10:27:03:
>
> On Thu, 8 Apr 2010 09:49:08 +0300
> Michael Zaidman wrote:
>
> > On Wed, Apr 7, 2010 at 7:01 PM, Joakim Tjernlund
> > wrote:
> > >>
> > >> On Wed, Apr 7, 2010 at 6:10 PM, Timur Tabi wrote:
> >
>
> On Wed, Apr 7, 2010 at 6:10 PM, Timur Tabi wrote:
> > On Wed, Apr 7, 2010 at 9:47 AM, Michael Zaidman
> > wrote:
> >
> >> The version #2 of the patch is located here
> >> http://lists.denx.de/pipermail/u-boot/2010-April/069579.html
> >
> > What is ppcDWstore(), and why do you use it sometimes
Just started to look at reserving some RAM from u-boot and the kernel
and got a bit confused about what config options to use.
I found CONFIG_PRAM and CONFIG_SYS_MEM_TOP_HIDE and from what I can tell
one needs to set both to hide some memory from both u-boot and linux
but CONFIG_SYS_MEM_TOP_HIDE i
Wolfgang Denk wrote on 2010-04-01 20:27:47:
>
> Dear Joakim Tjernlund,
>
> In message 005e9...@transmode.se> you wrote:
> >
> > > To me it looks like the new code would indeed do a "greedy" substitution
> > > only stopping when no more substituti
Detlev Zundel wrote on 2010-04-01 16:56:40:
>
> Hi Jocke,
>
> >> > > Hm... will this not make escaping impossible? Assume you want to pass
> >> > >
> >> > >arg=${name}
> >> > >
> >> > > to Linux. How would you escape this so it does NOT get expanded if
> >> > > you run process_macros() arbitra
Wolfgang Denk wrote on 2010-04-01 15:05:29:
>
> Dear Joakim Tjernlund,
>
> In message 00471...@transmode.se> you wrote:
> >
> > > I don't think you ever posted this before. I cannot find any trace of
> > > such a patch - not in the public archives
Wolfgang Denk wrote on 2010-04-01 14:47:45:
>
> Dear Joakim Tjernlund,
>
> In message 0044d...@transmode.se> you wrote:
> >
> > linuxip=ip=$(ipaddr)::$(gatewayip):$(netmask):$(hostname):$(linuxif):off
> > tboot=setenv bootargs $(linuxroot) $(linuxip) $(extr
>
> Hi Nitin,
>
> > Hi!
> >
> > I am doing env settings some thing like this,
> >
> > ROOT1=/dev/mmcblk0p1
> > ROOT2=/dev/mmcblk0p2
> > ROOT=${ROOT1}
> > bootargs1=console=ttyS0,115200n8 mem=256M noinitrd rw rootdelay=1 ${ROOT}
> >
> > when I say 'setenv bootargs ${bootargs1}', ${ROOT} gets resolve
Graeme Russ wrote on 2010/03/28 09:38:47:
>
> Joakim Tjernlund wrote:
> >
> > Graeme Russ wrote on 2010/03/27 12:54:03:
> >> Joakim Tjernlund wrote:
> >>> I did a few months ago for MPC83xx and sent it to the list. There was some
> >>> discuss
Graeme Russ wrote on 2010/03/27 12:54:03:
>
> Joakim Tjernlund wrote:
> >> In the near future, I intend to make it 100% relocatable (i.e. does not
> >> need to be loaded at TEXT_BASE) and maybe even add the ability to specify
> >> the relocation address.
&
>
> In the near future, I intend to make it 100% relocatable (i.e. does not
> need to be loaded at TEXT_BASE) and maybe even add the ability to specify
> the relocation address.
I did a few months ago for MPC83xx and sent it to the list. There was some
discussion but in the end it wasn't applied b
Jeff Angielski wrote on 2010/03/18 02:14:19:
> From: Jeff Angielski
> To: u-boot@lists.denx.de
> Cc: Joakim Tjernlund
> Date: 2010/03/18 02:14
> Subject: [PATCH] env: fix endian ordering in crc table
>
>
> The crc table was being built as little endian for big endia
Jeff Angielski wrote on 2010/03/17 16:10:50:
>
> Joakim Tjernlund wrote:
> >
> > Wolfgang Denk wrote on 2010/03/17 12:57:31:
> >> Dear Joakim Tjernlund,
> >>
> >> In message >> 003b4...@transmode.se> you wrote:
> >>> hmm,
Wolfgang Denk wrote on 2010/03/17 12:57:31:
>
> Dear Joakim Tjernlund,
>
> In message 003b4...@transmode.se> you wrote:
> >
> > hmm, I recently discovered that normal user space headers always define
> > both __LITTLE_ENDIAN and __BIG_ENDIAN so theref
>
> Hi Jeff,
>
> > I seem to be having a problem with fw_printenv on my PowerPC 85xx target
> > whereby I constantly get CRC32 errors:
> >
> > # fw_printenv
> > Warning: Bad CRC, using default environment
> >
> > I tracked this down to the fact that u-boot/lib_generic/crc32.c is
> > getting compile
jrjr wrote on 2010/03/17 10:49:15:
>
> Because I want to use LED controlled by OR3 to debug the code,so I don't want
> to clear OR0.
>
> You remind me, so I enable the "Calculate absolute address in FLASH and jump
> there " part,
>
> It seems that I get close to the point: jump to "in_flash:" fai
jrjr wrote on 2010/03/17 09:45:08:
>
> And I have made some change in cpu/mpc8260/start.S, I removed following lines:
>
> #ifndef CONFIG_SYS_RAMBOOT
> /* When booting from ROM (Flash or EPROM), clear the*/
> /* Address Mask in OR0 so ROM appears everywhere*/
> /*---
>
> Hi, everyone.
>
> I have got a problem when I porting U-BOOT to my board
>
> with MPC8247. My configuration file is based on the MPC8260ADS_config.
>
> The problem is : The instruction in TEXT section can not load data in
>
> other (DATA) sections. It seems there is a relocation to the data se
"Martin Krause" wrote on 2010/03/10 17:52:40:
>
> Hi Jocke,
>
> Joakim Tjernlund wrote on Wednesday, March 10, 2010 5:36 PM:
> >> Andrew Dyer wrote on Wednesday, March 10, 2010 4:50 PM: [SNIP]
> >>
> >> Hm, on the wikipedia article for the
>
> Andrew Dyer wrote on Wednesday, March 10, 2010 4:50 PM:
[SNIP]
>
> Hm, on the wikipedia article for the 'variable length arrays' (VLA) I
> read, that the GNU C compiler always uses the stack for this type of
> variables. So I think, if the stack is working for all other local
> variables, it sh
is is the key
>information I need)
> 2) date
> 3) Subject
> 4) Link to gmane mailing list archive
>
[SNIP]
> 6862 01/19 Joakim Tjernlund [PATCH 1/4] ppc: Loose GOT access in IRQ
> http://article.gmane.org/gmane.comp.boot-loaders.u-boot/74001
> 686
>
> Hello Michael,
>
> Thanks for posting your patches in plain text.
>
> Michael Durrant wrote:
> > drivers_i2c_fsl_i2c.patch
> >- need to set I2C to be idle according to the MCF5282 user's manual
> >
> >If I2SR[IBB] is set when the I2C bus module is enabled,
> >execute the following c
r14 is no longer used as non volatile GOT ptr. Instead
the volatile r12 is used so be sure to do GET_GOT in
asm code when you need to access global data.
Signed-off-by: Joakim Tjernlund
---
README |4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)
diff --git a/README b/README
index
This is no loger needed, free up r14 for general usage.
Signed-off-by: Joakim Tjernlund
---
cpu/74xx_7xx/config.mk |2 +-
cpu/mpc512x/config.mk |2 +-
cpu/mpc5xx/config.mk |2 +-
cpu/mpc5xxx/config.mk |2 +-
cpu/mpc8220/config.mk |2 +-
cpu/mpc824x/config.mk |2
r14 is not supposed to be clobbered by functions. Switch
to r12 and call GET_GOT when needed. This will allow u-boot
to loose the -ffixed-r14 gcc option.
Signed-off-by: Joakim Tjernlund
---
cpu/74xx_7xx/start.S | 11 ++-
cpu/mpc512x/start.S | 11 ++-
cpu/mpc5xx/start.S
Using the GOT in IRQ handlers requires r14 to be -ffixed-r14.
Avoid this by relocatate transfer_to_handler too.
This will allow to free up r14 later on.
Signed-off-by: Joakim Tjernlund
---
cpu/74xx_7xx/start.S | 36 +++-
cpu/mpc512x/start.S | 35
>
>
> Not sure what is going on but my build broke. I suspect commit
> d984fed068b3bec8a7edaf7a3de71479abde080a
> makefiles: fixes for building build tools
Answering myself, the symlinks under tools isn't removed, not even a by a
distclean.
Kinda makes sense since now they aren't created anymore
Not sure what is going on but my build broke. I suspect commit
d984fed068b3bec8a7edaf7a3de71479abde080a
makefiles: fixes for building build tools
Currently, some of the tools instead set CC to be HOSTCC in order to re-use
some pattern rules -- but this fails when the user overrides CC on
>
> Hi Richard,
>
> > Detlev Zundel wrote:
> >
> >>
> >> Please excuse my ignorance, but why not simply remove the #ifdef
> >> CONFIG_AMIGAONEG3SE in board_init_f? Actually I was hoping to remove
> >> the Amigaone special case.
> >>
> >> Cheers
> >> Detlev
> >>
> > I prefer getting the data fro
Wolfgang Denk wrote on 17/01/2010 20:12:31:
>
> Dear Joakim Tjernlund,
>
> In message 00510...@transmode.se> you wrote:
> > Has this been forgotten(again)?
> >
> > Jocke
> > >
> > > This series removes the needs to access the GOT in
> >
seems like the series was lost
>
> Joakim Tjernlund (3):
> ppc: Loose GOT access in IRQ
> ppc: Use r12 instead of r14 as GOT pointer.
> ppc: remove -ffixed-r14 gcc option.
> ppc: Update README about the new GOT ptr.
>
> cpu/74xx_7xx/config.
u-boot-boun...@lists.denx.de wrote on 05/01/2010 21:20:32:
> From: Scott Wood
> To: Albert ARIBAUD
> Cc: u-boot@lists.denx.de
> Date: 05/01/2010 21:22
> Subject: Re: [U-Boot] [PATCH 2/4] Use LINK_OFF to access global data
> Sent by: u-boot-boun...@lists.denx.de
>
> On Sun, Jan 03, 2010 at 09:0
Wolfgang Denk wrote on 03/01/2010 20:51:53:
>
> Dear Joakim Tjernlund,
>
> In message 0039f...@transmode.se> you wrote:
> >
> > > Will such changes be needed to all functions that work on (constant?)
> > > strings? Or why here?
> >
> > Only th
Wolfgang Denk wrote on 02/01/2010 19:17:52:
>
> Dear Joakim Tjernlund,
>
> in message 00091...@transmode.se> you wrote:
> >
> > Happy new year :)
>
> Thanks, and same to you and everybody else.
>
> > The bigger question is if the LINN_OFF changes in gene
Wolfgang Denk wrote on 02/01/2010 19:13:29:
>
> Dear Joakim Tjernlund,
>
> In message <1262185712-11890-3-git-send-email-joakim.tjernl...@transmode.se>
> you wrote:
> > Accessing global data before relocation needs
> > special handling if link address !=
Mike Frysinger wrote on 01/01/2010 07:18:44:
> From: Mike Frysinger
> To: Joakim Tjernlund
> Cc: u-boot@lists.denx.de
> Date: 01/01/2010 07:19
> Subject: Re: [U-Boot] [PATCH 2/4] Use LINK_OFF to access global data
>
> On Thursday 31 December 2009 20:39:10 Joakim Tjer
Mike Frysinger wrote on 31/12/2009 19:44:40:
Happy new year :)
>
> On Wednesday 30 December 2009 10:08:30 Joakim Tjernlund wrote:
> > --- a/common/cmd_nvedit.c
> > +++ b/common/cmd_nvedit.c
> > @@ -512,6 +512,7 @@ char *getenv (char *name)
> > {
> > int
To use a different link address than load address,
start.S must not make any absolute accesses. This
makes it so. Use link_off(), if defined, to calculate
the difference in load and link address.
---
cpu/mpc83xx/start.S | 35 ---
1 files changed, 28 insertions(+),
This is the most complex change. Keep this
one as a separate commit for now.
---
common/env_flash.c | 65 +++
1 files changed, 39 insertions(+), 26 deletions(-)
diff --git a/common/env_flash.c b/common/env_flash.c
index b860c48..64882d2 100644
---
Accessing global data before relocation needs
special handling if link address != load address.
Use LINK_OFF to calculate the difference.
---
common/cmd_nvedit.c |2 ++
common/console.c | 12 +---
common/env_common.c |2 +-
cpu/mpc83xx/cpu.c
Calculates the offset between global data link address
and where the data is actually loaded. Add this offset
to 'addr' arg and return the result.
Useful for true PIC and when relocating code to RAM.
---
include/common.h |7 +++
lib_ppc/reloc.S | 21 +
2 files change
.
Jocke
Joakim Tjernlund (4):
ppc: Add const void *link_off(const void *addr)
Use LINK_OFF to access global data
Use LINK_OFF in enviroment too
ppc: Make mpc83xx start.S relative.
common/cmd_nvedit.c |2 +
common/console.c | 12 ++--
common/env_common.c
This is no loger needed, free up r14 for general usage.
---
cpu/74xx_7xx/config.mk |2 +-
cpu/mpc512x/config.mk |2 +-
cpu/mpc5xx/config.mk |2 +-
cpu/mpc5xxx/config.mk |2 +-
cpu/mpc8220/config.mk |2 +-
cpu/mpc824x/config.mk |2 +-
cpu/mpc8260/config.mk |2 +-
r14 is no longer used as non volatile GOT ptr. Instead
the volatile r12 is used so be sure to do GET_GOT in
asm code when you need to access global data.
Signed-off-by: Joakim Tjernlund
---
README |4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)
diff --git a/README b/README
index
r14 is not supposed to be clobbered by functions. Switch
to r12 and call GET_GOT when needed. This will allow u-boot
to loose the -ffixed-r14 gcc option.
---
cpu/74xx_7xx/start.S | 11 ++-
cpu/mpc512x/start.S | 11 ++-
cpu/mpc5xx/start.S | 11 ++-
cpu/mpc5xxx/start
Using the GOT in IRQ handlers requires r14 to be -ffixed-r14.
Avoid this by relocatate transfer_to_handler too.
This will allow to free up r14 later on.
---
cpu/74xx_7xx/start.S | 36 +++-
cpu/mpc512x/start.S | 35 +++
cpu/mpc5xx
This series removes the needs to access the GOT in
IRQ handlers, then switches GOT PTR in start.S asm to
use r12 instead of r14. This leads up to the removal of
-ffixed-r14 gcc option for ppc
This is a resend since the it seems like the series was lost
Joakim Tjernlund (3):
ppc: Loose GOT
>Dear Joakim Tjernlund,
>
>In message <1257180242-1299-1-git-send-email-joakim.tjernl...@transmode.se> you
>wrote:
>> This series removes the needs to access the GOT in
>> IRQ handlers, then switches GOT PTR in start.S asm to
>> use r12 instead of
Wolfgang Denk wrote on 14/12/2009 21:11:17:
>
> Dear Joakim Tjernlund,
>
> In message 004f9...@transmode.se> you wrote:
> > Any interest in these patches I sent some time ago?
>
> Lots of ionterest actually. But so little time :-(
OK, good to know. I w
Wolfgang Denk wrote on 03/11/2009 20:40:50:
> From: Wolfgang Denk
> To: Joakim Tjernlund
> Cc: u-boot@lists.denx.de
> Date: 03/11/2009 20:40
> Subject: Re: [U-Boot] [PATCH 0/3] No GOT in IRQ and use r12 as GOT ptr
>
> Dear Joakim Tjernlund,
>
> In message <12
Any interest in these patches I sent some time ago?
Jocke
Joakim Tjernlund wrote on 02/11/2009 19:01:16:
> From: Joakim Tjernlund
> To: u-boot@lists.denx.de
> Cc: Joakim Tjernlund
> Date: 02/11/2009 19:01
> Subject: [PATCH 0/4] Make u-boot true PIC for ppc
>
>
> Ah. I just realize that the endianess can be supported as simple as:
>
> In stead of this:
>xsum += (*p & 0xff00);
> Use this:
>xsum += (*p & ntohs(0xff00));
Did you look at the suggestion I sent you?
I know it works because I use in ospf.
Jocke
_
Wolfgang Denk wrote on 03/12/2009 15:08:24:
> Dear Joakim Tjernlund,
>
> In message 00448...@transmode.se> you wrote:
> >
> > > > + if (len == 1) {
> > > > + xsum += (*p & 0xff00);
> > >
> > > I doubt that this
> Dear "Greg Ren",
>
> In message you
> wrote:
> >
> > I am new to u-boot and got assignment to debug some networking issue. I
> > traced the checksum failure and was able to fix it with the patch below.
>
> It would be important to know on which system(s) you have actually
> tested your patch -
Wolfgang Denk wrote on 30/11/2009 22:02:44:
>
>
> Dear Joakim Tjernlund,
>
> In message 00453...@transmode.se> you wrote:
> > Wolfgang Denk wrote on 27/11/2009 21:18:28:
> > > I am not convinced yet that the new code is actually an improvement.
> > >
Wolfgang Denk wrote on 27/11/2009 21:18:28:
> I am not convinced yet that the new code is actually an improvement.
> Changing the array of pointers into a list of function calls does not
> solve any of the real issues we have with the init seuqnece - like
> that some board need the PCI initializat
Wolfgang Denk wrote on 27/11/2009 21:18:28:
> From: Wolfgang Denk
> To: Joakim Tjernlund
> Cc: u-boot@lists.denx.de
> Date: 27/11/2009 21:18
> Subject: Re: [U-Boot] [PATCH] ppc: transform init_sequence into a function.
>
> Dear Joakim Tjernlund,
>
> In message 00508
rub...@gnudd.com wrote on 27/11/2009 15:54:57:
>
> > I think most of the ugliness comes from the #ifdef hell in this list.
> > replacing the #ifdefs is another matter so looking behind
> > the #ifdef mess, I don't think it looks too bad.
>
> I think the approach I took with lib_arm/board.c may appl
Wolfgang Denk wrote on 27/11/2009 15:06:34:
>
> Dear Joakim Tjernlund,
>
> In message <1259317926-9820-1-git-send-email-joakim.tjernl...@transmode.se>
> you wrote:
> > init_sequence is an array with function pointers.
> > It produces lots of relocation data
Tom wrote on 26/11/2009 20:18:22:
>
> Joakim Tjernlund wrote:
> >> From: Alessandro Rubini
> >>
> >> While it's a matter of personal taste, I prefer to avoid ifdef when
> >> possible. For example, I don't like to add BOARD_LATE_INIT i
/board.org
1224 92 01316 524 lib_ppc/board.new
Signed-off-by: Joakim Tjernlund
---
lib_ppc/board.c | 123 +-
1 files changed, 75 insertions(+), 48 deletions(-)
diff --git a/lib_ppc/board.c b/lib_ppc/board.c
index 765f97a
>
> From: Alessandro Rubini
>
> While it's a matter of personal taste, I prefer to avoid ifdef when
> possible. For example, I don't like to add BOARD_LATE_INIT in the
> config file just to add a board_late_init() function.
> Also, I think the file is more readable without the ifdef stuff.
> This
Mike Frysinger wrote on 19/11/2009 15:49:55:
>
> On Thursday 19 November 2009 07:31:18 Joakim Tjernlund wrote:
> > Perhaps you can take a look at getting X_unaligned definitions
> > into compiler.h too?
>
> that would be much trickier i think ...
>
> > I use ge
Thiago A. Corrêa wrote on 19/11/2009 16:41:05:
>
> Hi,
>
> On Thu, Nov 19, 2009 at 10:44 AM, Joakim Tjernlund
> wrote:
> >
> > +#define tole(x) cpu_to_le32(x)
> > +
>
> This define is only used once and at the constants table, after that,
> cpu_to_le32
Peter Korsgaard wrote on 19/11/2009 15:07:12:
>
> >>>>> "Joakim" == Joakim Tjernlund writes:
>
> Joakim> This patch optimizes the direct copy procedure.
> Joakim> Uses get_unaligned() but only in one place.
> Joakim> The copy loop just abov
> 60 secs
Signed-off-by: Joakim Tjernlund
---
V3: squashed "crc32: Make crc32 build on non Linux too"
into this patch to form one clean patch.
I think this patch is ready for general use.
lib_generic/crc32.c | 200 --
1 files c
Mike Frysinger wrote on 17/11/2009 05:06:02:
>
> On Monday 16 November 2009 23:01:59 Mike Frysinger wrote:
> > On Friday 13 November 2009 03:52:05 Joakim Tjernlund wrote:
> > > --- a/lib_generic/crc32.c
> > > +++ b/lib_generic/crc32.c
> > > @@ -13,
>
> Add lzop decompression support to the existing lzo bitstream handling
> (think gzip versus zlib), and support it for uImage decompression if
> CONFIG_LZO is enabled.
>
> Lzop doesn't compress as good as gzip (~10% worse), but decompression
> is very fast (~0.7s faster here on a slow ppc). The l
n the original. No speed test has been performed in u-boot.
Size increase on ppc: 484 bytes
Signed-off-by: Joakim Tjernlund
---
I have the same optimization queued for linux. Figured it
would be useful for u-boot too.
Testing and feedback welcome.
lib_generic/zlib.c |
Mike Frysinger wrote on 19/11/2009 00:45:16:
>
> On Wednesday 18 November 2009 17:59:59 Joakim Tjernlund wrote:
> > Wolfgang Denk wrote on 18/11/2009 23:45:56:
> > > Joakim Tjernlund wrote:
> > > > > And PLEASE keep the In-reply-to: and References: headers so
Wolfgang Denk wrote on 18/11/2009 23:45:56:
>
> Dear Joakim Tjernlund,
>
> In message 003a0...@transmode.se> you wrote:
> >
> > This was a "reply" for Mikes comment on
> > [PATCH v2] crc32: Impl. linux optimized crc32()
>
> If it was a rep
Wolfgang Denk wrote on 17/11/2009 22:06:02:
>
> Dear Joakim Tjernlund,
>
> In message <1258446374-2919-1-git-send-email-joakim.tjernl...@transmode.se>
> you wrote:
> > The #include is Linux specific. Use
> > compiler.h and adjust the code to cope.
> &
The #include is Linux specific. Use
compiler.h and adjust the code to cope.
Signed-off-by: Joakim Tjernlund
---
Mike, how does this work for you?
lib_generic/crc32.c | 14 ++
1 files changed, 6 insertions(+), 8 deletions(-)
diff --git a/lib_generic/crc32.c b/lib_generic/crc32
> 60 secs
Signed-off-by: Joakim Tjernlund
---
Resending a new patch with all the improvents included.
Forget the previous versions.
lib_generic/crc32.c | 196 +--
1 files changed, 111 insertions(+), 85 deletions(-)
diff --git a/lib_gener
patch "crc32: more optimizations" introduced a bug.
This fixes it.
Signed-off-by: Joakim Tjernlund
---
Perhaps I should resend as one whole patch with
everything instead?
lib_generic/crc32.c | 16
1 files changed, 8 insertions(+), 8 deletions(-)
diff --git a/l
Shave off yet 4 bytes and make if faster for unaligned and/or
len & 3 != 0.
Signed-off-by: Joakim Tjernlund
---
Now I am done with crc32 optimizations.
Someone should look at porting over crc32_be from linux
since CRC32 BE is used by bzip and it could probably
benefit from a faster
Don't optimize for len < 4. This reduces size with
8 bytes too.
This crc32 impl is smaller that the orginal if crc32
and crc32_no_comp is impl. two separate functions(like the orginal).
That would, however, be a waste of space when both are defined.
Signed-off-by: Joakim T
r14 is no longer used as non volatile GOT ptr. Instead
the volatile r12 is used so be sure to do GET_GOT in
asm code when you need to access global data.
Signed-off-by: Joakim Tjernlund
---
README |4 +++-
1 files changed, 3 insertions(+), 1 deletions(-)
diff --git a/README b/README
index
Wolfgang Denk wrote on 03/11/2009 20:40:50:
>
> Dear Joakim Tjernlund,
>
> In message <1257180242-1299-1-git-send-email-joakim.tjernl...@transmode.se>
> you wrote:
> > This series removes the needs to access the GOT in
> > IRQ handlers, then switches GOT PTR in s
Wolfgang Denk wrote on 03/11/2009 17:02:27:
>
> Dear Joakim Tjernlund,
>
> In message <1257263550-21504-1-git-send-email-joakim.tjernl...@transmode.se>
> you wrote:
> > Ported over the more efficient linux crc32() function.
> > A quick comparsion on ppc:
> &g
Wolfgang Denk wrote on 03/11/2009 14:21:08:
>
> Dear Joakim Tjernlund,
>
> In message 003bf...@transmode.se> you wrote:
> > >
> > > HeLei wrote on 03/11/2009 09:21:04:
> > > >
> > > > Thank you, Jocke
> >
> > > > &g
the old does four 8 bits loads.
- size is bit bigger for the new crc32:
1392(old) 1428(new) of text. Can probably be shrunk
somewhat by inlining crc32().
Signed-off-by: Joakim Tjernlund
---
Only tested on PPC. little endian testing would be nice too.
lib_generic/crc32.c | 202
Wolfgang Denk wrote on 03/11/2009 14:21:08:
>
> Dear Joakim Tjernlund,
>
> In message 003bf...@transmode.se> you wrote:
> > >
> > > HeLei wrote on 03/11/2009 09:21:04:
> > > >
> > > > Thank you, Jocke
> >
> > > > &g
HeLei wrote on 03/11/2009 12:08:39:
>
> > Subject: Re: [U-Boot] Does uboot EBS(erase block summary) to reduce JFFS2
> scaning time?
> > CC: leon...@msn.com; u-boot@lists.denx.de
> > From: joakim.tjernl...@transmode.se
> > Date: Tue, 3 Nov 2009 11:55:18 +0100
> >
> > >
> > > HeLei wrote on 03/11/2
little or big endian target?
Jocke
>From 7b98aab2aa940b47b81d3a548c8d786016cd2ee8 Mon Sep 17 00:00:00 2001
From: Joakim Tjernlund
Date: Tue, 3 Nov 2009 11:39:43 +0100
Subject: [PATCH] crc32: Impl. linux optimized crc32()
Ported over the more efficient linux crc32() function.
---
lib_generic/crc32.c | 208
HeLei wrote on 03/11/2009 10:39:48:
>
>
> > Subject: RE: [U-Boot] Does uboot EBS(erase block summary) to reduce JFFS2
> scaning time?
> > To: leon...@msn.com
> > CC: u-boot@lists.denx.de
> > From: joakim.tjernl...@transmode.se
> > Date: Tue, 3 Nov 2009 10:29:47 +0100
> >
> > HeLei wrote on 03/11/
HeLei wrote on 03/11/2009 09:21:04:
> > - Don't scan more than you have to, that is, ls/read /some/file
> > should only scan and keep records to do the ls/read of that
> > particular file.
>So we have to have an index, or something like that, to tell which file in
> which EBs. This index is al
HeLei wrote on 03/11/2009 10:26:44:
> Jocke, you are so kind.
> Thank you very much:)
NP, it was easy considering I did the first impl. for linux :)
It is probably easier to paste the missing bits into the current crc32 impl.
though.
Jocke
___
U-B
HeLei wrote on 03/11/2009 10:21:13:
>
> Ok, thank you, jocke.
> Can you tell me how much time can be saved according to your idea, by your
> experience?
Can't remember any figures, but is was lots by not scanning empty space. It all
depends on how full your fs is.
Jocke
_
HeLei wrote on 03/11/2009 09:21:04:
>
> Thank you, Jocke
>
> > Subject: Re: [U-Boot] Does uboot EBS(erase block summary) to reduce JFFS2
> scaning time?
> > To: leon...@msn.com
> > CC: u-boot@lists.denx.de
> > From: joakim.tjernl...@transmode.se
> > Date: Tue, 3 Nov 2009 08:41:08 +0100
> >
> >
> >
501 - 600 of 747 matches
Mail list logo