oweals/u-boot.git
14 years agoppc/85xx: Add tracking of TLB CAM usage
Kumar Gala [Thu, 12 Nov 2009 16:26:16 +0000 (10:26 -0600)]
ppc/85xx: Add tracking of TLB CAM usage

We need to track which TLB CAM entries are used to allow us to
"dynamically" allocate entries later in the code.  For example the SPD
DDR code today hard codes which TLB entries it uses.  We can now make
that pick entries that are free.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/8xxx: Remove is_fsl_pci_agent
Kumar Gala [Wed, 4 Nov 2009 19:00:55 +0000 (13:00 -0600)]
ppc/8xxx: Remove is_fsl_pci_agent

All users of is_fsl_pci_agent have been converted to fsl_is_pci_agent
that uses the standard PCI programming model to determine host vs
agent/end-point.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Move to using fsl_setup_hose on TQM 85xx
Kumar Gala [Wed, 4 Nov 2009 17:39:55 +0000 (11:39 -0600)]
ppc/85xx: Move to using fsl_setup_hose on TQM 85xx

We can use fsl_setup_hose to determine if we are a agent/end-point or
a host.  Rather than using some SoC specific register we can just look
at the PCI cfg space of the host controller to determine this.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Move to using fsl_setup_hose on P2020 DS
Kumar Gala [Wed, 4 Nov 2009 19:01:51 +0000 (13:01 -0600)]
ppc/85xx: Move to using fsl_setup_hose on P2020 DS

We can use fsl_setup_hose to determine if we are a agent/end-point or
a host.  Rather than using some SoC specific register we can just look
at the PCI cfg space of the host controller to determine this.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Move to using fsl_setup_hose on P1/P2 RDB
Kumar Gala [Wed, 4 Nov 2009 19:01:38 +0000 (13:01 -0600)]
ppc/85xx: Move to using fsl_setup_hose on P1/P2 RDB

We can use fsl_setup_hose to determine if we are a agent/end-point or
a host.  Rather than using some SoC specific register we can just look
at the PCI cfg space of the host controller to determine this.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Move to using fsl_setup_hose on MPC8572 DS
Kumar Gala [Wed, 4 Nov 2009 19:01:17 +0000 (13:01 -0600)]
ppc/85xx: Move to using fsl_setup_hose on MPC8572 DS

We can use fsl_setup_hose to determine if we are a agent/end-point or
a host.  Rather than using some SoC specific register we can just look
at the PCI cfg space of the host controller to determine this.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/86xx: Clean up MPC8610 HPCD PCI setup code
Kumar Gala [Wed, 4 Nov 2009 18:51:10 +0000 (12:51 -0600)]
ppc/86xx: Clean up MPC8610 HPCD PCI setup code

Use new fsl_pci_init_port() that reduces amount of duplicated code in the
board ports, use IO accessors and clean up printing of status info.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Clean up MPC8548 CDS PCI setup code
Kumar Gala [Wed, 4 Nov 2009 17:15:29 +0000 (11:15 -0600)]
ppc/85xx: Clean up MPC8548 CDS PCI setup code

Use new fsl_pci_init_port() that reduces amount of duplicated code in the
board ports, use IO accessors and clean up printing of status info.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Clean up ATUM8548 PCI setup code
Kumar Gala [Wed, 4 Nov 2009 17:05:02 +0000 (11:05 -0600)]
ppc/85xx: Clean up ATUM8548 PCI setup code

Use new fsl_pci_init_port() that reduces amount of duplicated code in the
board ports, use IO accessors and clean up printing of status info.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Clean up MPC8568 MDS PCI setup code
Kumar Gala [Wed, 4 Nov 2009 16:31:53 +0000 (10:31 -0600)]
ppc/85xx: Clean up MPC8568 MDS PCI setup code

Use new fsl_pci_init_port() that reduces amount of duplicated code in the
board ports, use IO accessors and clean up printing of status info.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Clean up MPC8569 MDS PCI setup code
Kumar Gala [Wed, 4 Nov 2009 16:26:30 +0000 (10:26 -0600)]
ppc/85xx: Clean up MPC8569 MDS PCI setup code

Use new fsl_pci_init_port() that reduces amount of duplicated code in the
board ports, use IO accessors and clean up printing of status info.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Clean up MPC8544 DS PCI setup code
Kumar Gala [Wed, 4 Nov 2009 16:22:26 +0000 (10:22 -0600)]
ppc/85xx: Clean up MPC8544 DS PCI setup code

Use new fsl_pci_init_port() that reduces amount of duplicated code in the
board ports, use IO accessors and clean up printing of status info.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/85xx: Clean up MPC8536 DS PCI setup code
Mingkai Hu [Wed, 28 Oct 2009 02:49:31 +0000 (10:49 +0800)]
ppc/85xx: Clean up MPC8536 DS PCI setup code

Use new fsl_pci_init_port() that reduces amount of duplicated code in the
board ports, use IO accessors and clean up printing of status info.

Signed-off-by: Mingkai Hu <Mingkai.hu@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoNET: Base support for etsec2.0
Kumar Gala [Sat, 31 Oct 2009 16:23:41 +0000 (11:23 -0500)]
NET: Base support for etsec2.0

1. Modified the tsec_mdio structure to include the new regs
2. Modified the MDIO_BASE_ADDR so that it will handle both
older version and new version of etsec.

Signed-off-by: Sandeep Gopalpet <sandeep.kumar@freescale.com>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoNET: Move MDIO regs out of TSEC Space
Sandeep Gopalpet [Fri, 30 Oct 2009 19:05:04 +0000 (00:35 +0530)]
NET: Move MDIO regs out of TSEC Space

Moved the mdio regs out of the tsec structure,and
provided different offsets for tsec base and mdio
base so that provision for etsec2.0 can be provided.

This patch helps in providing the support for etsec2.0
In etsec2.0, the MDIO register space and the etsec reg
space are different.

Also, moved the TSEC_BASE_ADDR and MDIO_BASE_ADDR definitons into
platform specific files.

Signed-off-by: Sandeep Gopalpet <sandeep.kumar@freescale.com>
Acked-by: Kim Phillips <kim.phillips@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years ago85xx: Add support to set DPAA (data path) devices clock frequencies
Kumar Gala [Thu, 19 Mar 2009 07:46:28 +0000 (02:46 -0500)]
85xx: Add support to set DPAA (data path) devices clock frequencies

Set clock-frequency for Frame Manager 0/1 and Patter Match Engine on p4080.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/8xxx: Don't use pci_cfg on FSL_CORENET platforms
Kumar Gala [Thu, 17 Sep 2009 05:01:14 +0000 (00:01 -0500)]
ppc/8xxx: Don't use pci_cfg on FSL_CORENET platforms

The FSL_CORENET platforms use a completely different means to determine
which PCIe port is enabled as well as if its a host or agent/end-point.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoppc/p4080: Added p4080 SERDES registers & USB offset
Kumar Gala [Fri, 16 Oct 2009 04:22:10 +0000 (23:22 -0500)]
ppc/p4080: Added p4080 SERDES registers & USB offset

Added immap definition for SERDES registers on p4080, the USB offset
(since it was missing) and a GPL header.

Signed-off-by: Li Yang <leoli@freescale.com>
Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years ago85xx: Add support for e500mc cache stashing
Kumar Gala [Thu, 19 Mar 2009 07:53:01 +0000 (02:53 -0500)]
85xx: Add support for e500mc cache stashing

The e500mc core supports the ability to stash into the L1 or L2 cache,
however we need to uniquely identify the caches with an id.

We use the following equation to set the various stash-ids:

32 + coreID*2 + 0(L1) or 1(L2)

The 0 (for L1) or 1 (for L2) matches the CT field used be various cache
control instructions.

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoMerge branch 'master' of git://git.denx.de/u-boot-ppc4xx
Wolfgang Denk [Sun, 3 Jan 2010 20:07:04 +0000 (21:07 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-ppc4xx

14 years agoppc4xx: alpr: Remove some not needed commands to make image fit again
Stefan Roese [Wed, 16 Dec 2009 08:27:31 +0000 (09:27 +0100)]
ppc4xx: alpr: Remove some not needed commands to make image fit again

The latest changes increased the size of the alpr image a bit more.
Now it doesn't fit into the 256k reserved for it. This patch now removes
the commands "ping" and "diag" which are not needed in the production
systems.

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Pieter Voorthuijsen <pieter.voorthuijsen@prodrive.nl>
14 years agoi2c, ppc4xx: fix compiling KAREF and METROBOX boards.
Heiko Schocher [Thu, 3 Dec 2009 10:23:17 +0000 (11:23 +0100)]
i2c, ppc4xx: fix compiling KAREF and METROBOX boards.

commit eb5eb2b0f744f0cba405160c5d01335c40f09acf

ppc4xx: Cleanup PPC4xx I2C infrastructure

This patch cleans up the PPC4xx I2C intrastructure:

- Use C struct to describe the I2C registers instead of defines
- Coding style cleanup (braces, whitespace, comments, line length)
- Extract common code from i2c_read() and i2c_write()
- Remove unneeded IIC defines from ppc405.h & ppc440.h

breaks comiling for the KAREF and METROBOX boards.

This patch fixes this issue.

Signed-off-by: Heiko Schocher <hs@denx.de>
Signed-off-by: Stefan Roese <sr@denx.de>
14 years agomove prototypes for gunzip() and zunzip() to common.h
Wolfgang Wegner [Wed, 9 Dec 2009 14:16:47 +0000 (15:16 +0100)]
move prototypes for gunzip() and zunzip() to common.h

Prototype for gunzip/zunzip was only in lib_generic/gunzip.c and thus
repeated in every file using it. This patch moves the prototypes to
common.h and removes all prototypes distributed anywhere else.

Signed-off-by: Wolfgang Wegner <w.wegner@astro-kom.de>
14 years agoMerge branch 'master' of git://git.denx.de/u-boot-usb
Wolfgang Denk [Mon, 21 Dec 2009 20:01:56 +0000 (21:01 +0100)]
Merge branch 'master' of git://git.denx.de/u-boot-usb

14 years agousb: write command for RAW partition.
Mahavir Jain [Tue, 3 Nov 2009 06:52:10 +0000 (12:22 +0530)]
usb: write command for RAW partition.

This patch implements write support to usb device with raw partition.
It will be useful for filesystem write support to usb device from
u-boot in future.

Tested with writing kernel image to raw usb disk & booting with usb
read command into ram.

[Note:  run usb part to get info about start sector & number of
sectors on a partition for usb write operation.]

Signed-off-by: Mahavir Jain <mjain@marvell.com>
14 years agoomap3evm: musb: add USB config
Ajay Kumar Gupta [Wed, 4 Nov 2009 21:58:23 +0000 (15:58 -0600)]
omap3evm: musb: add USB config

Added USB host and device config for host (MSC, Keyboard) and
device (ACM) functionalities.

Signed-off-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
14 years agoOMAP3 USB Initialize twl4030 only if required
Tom Rix [Sat, 31 Oct 2009 17:37:46 +0000 (12:37 -0500)]
OMAP3 USB Initialize twl4030 only if required

OMAP3EVM uses ISP1504 phy and so twl4030 related init is not required.

Submitted-by: Ajay Kumar Gupta <ajay.gupta@ti.com>
Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoOMAP3 zoom2 Use usbtty if the debug board is not connected.
Tom Rix [Sat, 31 Oct 2009 17:37:45 +0000 (12:37 -0500)]
OMAP3 zoom2 Use usbtty if the debug board is not connected.

The preferred serial output comes from the debug board.
When the debug board is disconnected, fall back on using
usbtty from the usb connector on the Zoom2 board.

This shows up as /dev/ttyACM0 in a linux host.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoUSBTTY make some function declarations easier to use.
Tom Rix [Sat, 31 Oct 2009 17:37:44 +0000 (12:37 -0500)]
USBTTY make some function declarations easier to use.

Zoom2 needs to use these declarations and the include directory is a
better place from them than in the middle of the driver directory.
It did not make sense to create a new file for just a couple of
lines so they were appended to the serial.h

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoOMAP3 beagle Add usbtty configuration
Tom Rix [Sat, 31 Oct 2009 17:37:43 +0000 (12:37 -0500)]
OMAP3 beagle Add usbtty configuration

The primary console of beagle is the serial header.

A secondary console is to use the usbtty.  The user can set this
manually by doing

setenv stdout usbtty; setenv stdin usbtty; setenv stderr usbtty
saveenv

usbtty will be usable by accessing the /dev/ttyACM0 on a linux host.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoOMAP3 zoom1 Add usbtty configuration
Tom Rix [Sat, 31 Oct 2009 17:37:42 +0000 (12:37 -0500)]
OMAP3 zoom1 Add usbtty configuration

The primary console of zoom1 is the serial out from the jumpers
accessed by removing the back panel.

A secondary console is to use the usbtty.  The user can set this
manually by doing

setenv stdout usbtty; setenv stdin usbtty; setenv stderr usbtty
saveenv

usbtty will be usable by accessing the /dev/ttyACM0 on a linux host.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoOMAP3 Add usb device support
Tom Rix [Sat, 31 Oct 2009 17:37:41 +0000 (12:37 -0500)]
OMAP3 Add usb device support

This change adds the usb device support for musb.

Omap3 platform support added at the same level as davinci.

The interface for usbtty to use the musb device support was added.

Verified on omap3 beagle, zoom1 and zoom2.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoTWL4030 Add usb PHY support
Tom Rix [Sat, 31 Oct 2009 17:37:40 +0000 (12:37 -0500)]
TWL4030 Add usb PHY support

The twl4030 provides a PHY device for connecting a link device,
like musb, to physical connection.

This change adds the twl4030 usb registers and functions for
initializing the PHY as required by omap3.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoUSB add macros for debugging usb device setup.
Tom Rix [Sat, 31 Oct 2009 17:37:39 +0000 (12:37 -0500)]
USB add macros for debugging usb device setup.

When developing usb device features, it is useful to print out
common usb structures.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoUSB Consolidate descriptor definitions
Tom Rix [Sat, 31 Oct 2009 17:37:38 +0000 (12:37 -0500)]
USB Consolidate descriptor definitions

The header files usb.h and usbdescriptors.h have the same nameed
structure definitions for

usb_config_descriptor
usb_interface_descriptor
usb_endpoint_descriptor
usb_device_descriptor
usb_string_descriptor

These are out right duplicates in usb.h

usb_device_descriptor
usb_string_descriptor

This one has extra unused elements

usb_endpoint_descriptor

unsigned char bRefresh
unsigned char bSynchAddress;

These in usb.h have extra elements at the end of the usb 2.0
specified descriptor and are used.

usb_config_descriptor
usb_interface_descriptor

The change is to consolidate the definition of the descriptors
to usbdescriptors.h.  The dublicates in usb.h are removed.
The extra element structure will have their name shorted by
removing the '_descriptor' suffix.

So

usb_config_descriptor -> usb_config
usb_interface_descriptor -> usb_interface

For these, the common descriptor elements are accessed now
by an element 'desc'.

As an example

- if (iface->bInterfaceClass != USB_CLASS_HUB)
+ if (iface->desc.bInterfaceClass != USB_CLASS_HUB)

This has been compile tested on MAKEALL arm, ppc and mips.

Signed-off-by: Tom Rix <Tom.Rix@windriver.com>
14 years agoMerge branch 'next' of git://git.denx.de/u-boot-net
Wolfgang Denk [Thu, 17 Dec 2009 21:59:27 +0000 (22:59 +0100)]
Merge branch 'next' of git://git.denx.de/u-boot-net

14 years agoUpdate Makefile for tag generating
Li Yang [Wed, 9 Dec 2009 10:13:26 +0000 (18:13 +0800)]
Update Makefile for tag generating

Get tag directories from the $(__LIB) and also generate tag for .S files.

Signed-off-by: Li Yang <leoli@freescale.com>
14 years agodrivers/bios_emulator: Fix compile error in .depend not being generated
Kumar Gala [Wed, 16 Dec 2009 20:12:11 +0000 (14:12 -0600)]
drivers/bios_emulator: Fix compile error in .depend not being generated

make -C drivers/bios_emulator/
make[2]: Entering directory
`drivers/bios_emulator'
In file included from atibios.c:49:
biosemui.h:47:21: error: biosemu.h: No such file or directory
...
x86emu/decode.c:40:28: error: x86emu/x86emui.h: No such file or directory
...

Due to lack of proper CPPFLAGS being passed to .depend generation rule

Signed-off-by: Kumar Gala <galak@kernel.crashing.org>
14 years agoMerge branch 'next' of ../next
Wolfgang Denk [Tue, 15 Dec 2009 22:38:34 +0000 (23:38 +0100)]
Merge branch 'next' of ../next

14 years agoUpdate CHANGELOG; prepare Prepare v2009.11 v2009.11
Wolfgang Denk [Tue, 15 Dec 2009 22:20:54 +0000 (23:20 +0100)]
Update CHANGELOG; prepare Prepare v2009.11

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agompc85xx, mpc86xx: Fix gd->cpu pointer after relocation
Peter Tyser [Tue, 15 Dec 2009 18:10:47 +0000 (12:10 -0600)]
mpc85xx, mpc86xx: Fix gd->cpu pointer after relocation

The gd->cpu pointer is set to an address located in flash when the
probecpu() function is called while U-Boot is executing from flash.
This pointer needs to be updated to point to an address in RAM after
relocation has occurred otherwise Linux may not be able to boot due to
"fdt board" crashing if flash has been erased or changed.

This bug was introduced in commit
a0e2066f392782730f0398095e583c87812d97f2.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Reported-by: Ed Swarthout <Ed.Swarthout@freescale.com>
Tested-by: Kumar Gala <galak@kernel.crashing.org>
Tested on MPC8527DS.
Tested by: Ed Swarthout <Ed.Swarthout@freescale.com>

14 years agoNet: Clean up LAN91C96 Support
Ben Warren [Tue, 15 Dec 2009 00:30:39 +0000 (16:30 -0800)]
Net: Clean up LAN91C96 Support

A previous Commit converted the LAN91C96 Ethernet driver to using the
CONFIG_NET_MULTI API, but did not include full board support.  This patch
finishes the job.

Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agoMVBLUE: Remove CONFIG_CMD_IRQ
Peter Tyser [Sun, 13 Dec 2009 23:58:34 +0000 (17:58 -0600)]
MVBLUE: Remove CONFIG_CMD_IRQ

Neither the MVBLUE nor its underlying architecture implement the
do_irqinfo() function which is required when CONFIG_CMD_IRQ is defined.
This change fixes the following MVBLUE compiler error:

-> ./MAKEALL MVBLUE
Configuring for MVBLUE board...
common/libcommon.a(cmd_irq.o):(.u_boot_cmd+0x24): undefined reference to `do_irqinfo'
make: *** [u-boot] Error 1

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Acked-by: Andre Schwarz <andre.schwarz@matrix-vision.de>
14 years agoimx27lite: Reenable MTD support on NOR flash.
Detlev Zundel [Mon, 14 Dec 2009 16:54:40 +0000 (17:54 +0100)]
imx27lite: Reenable MTD support on NOR flash.

The support for this was silently dropped by a configuration
split during the merge of the imx27lite board support in commit
864aa034f3a0e10ce710e8bbda171df3cab59414 (cmd_mtdparts: Move to common
handling of FLASH devices via MTD layer).

Signed-off-by: Detlev Zundel <dzu@denx.de>
14 years agonet: dm9000x: fix debug output
Thomas Weber [Wed, 9 Dec 2009 08:38:04 +0000 (09:38 +0100)]
net: dm9000x: fix debug output

commit 60f61e6d7655400bb785a2ef637581679941f6d1 breaks compile with gcc by introducing __func__
instead of constant string "func" in the macro call but missed to change the macro.

Signed-off-by: Thomas Weber <weber@corscience.de>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agofec_mxc: incomplete error handling
John Ogness [Fri, 11 Dec 2009 08:47:28 +0000 (09:47 +0100)]
fec_mxc: incomplete error handling

fec_init() will only allocate fec->base_ptr if it is non-NULL. But
the cleanup routine on error will free the pointer without setting
it to NULL. This means that a later call to fec_init() would result
in using an invalid pointer.

Signed-off-by: John Ogness <john.ogness@linutronix.de>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agonet: e1000: Add support for the Intel 82546GB controller
Reinhard Arlt [Fri, 4 Dec 2009 08:52:17 +0000 (09:52 +0100)]
net: e1000: Add support for the Intel 82546GB controller

This chip is equipped for example on the esd PMC-ETH2-GB board. So let's
add it to the list of supported chips to the e1000 driver.

Signed-off-by: Reinhard Arlt <reinhard.arlt@esd.eu>
Signed-off-by: Stefan Roese <sr@denx.de>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agoAdd error codes/handling for TFTP-server
Remy Bohmer [Wed, 28 Oct 2009 21:13:40 +0000 (22:13 +0100)]
Add error codes/handling for TFTP-server

Signed-off-by: Remy Bohmer <linux@bohmer.net>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agoLAN91C96: Enable NET_MULTI LAN driver
Nishanth Menon [Fri, 16 Oct 2009 05:06:37 +0000 (00:06 -0500)]
LAN91C96: Enable NET_MULTI LAN driver

This modification is NOT tested on any of the
platforms modified as I dont have them. please
help by testing+building+fixing

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agoTI OMAP3: SDP3430 FIX NET_MULTI Warning
Nishanth Menon [Fri, 16 Oct 2009 05:06:36 +0000 (00:06 -0500)]
TI OMAP3: SDP3430 FIX NET_MULTI Warning

Enable the NET MULTI option and remove build warning

Tested: SDP3430

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agoNET: LAN91C96 CONFIG_NET_MULTIify
Nishanth Menon [Fri, 16 Oct 2009 05:06:35 +0000 (00:06 -0500)]
NET: LAN91C96 CONFIG_NET_MULTIify

Make the lan91c96 driver capable of CONFIG_NET_MULTI
to be clean for the new arch, add a a lil detect function
Most of the formatting change was done to keep checkpatch
silent, but a few functions and #if 0ed code which
does not make sense for NET_MULTI have been removed

Now, use the lan91c96_initialize() function to init the driver

Signed-off-by: Nishanth Menon <nm@ti.com>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agonet: pull CONFIG checks out of source and into makefile
Mike Frysinger [Tue, 3 Nov 2009 16:35:42 +0000 (11:35 -0500)]
net: pull CONFIG checks out of source and into makefile

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
Signed-off-by: Ben Warren <biggerbadderben@gmail.com>
14 years agonand: Fix access to last block in NAND devices
Stefan Roese [Wed, 9 Dec 2009 08:01:43 +0000 (09:01 +0100)]
nand: Fix access to last block in NAND devices

Currently, the last block of NAND devices can't be accessed. This patch
fixes this issue by correcting the boundary checking (off-by-one error).

Signed-off-by: Stefan Roese <sr@denx.de>
Cc: Scott Wood <scottwood@freescale.com>
Cc: Wolfgang Denk <wd@denx.de>
14 years agompc83xx: boot time regression, move LCRR setup back to cpu_init_f
Peter Korsgaard [Tue, 8 Dec 2009 21:20:34 +0000 (22:20 +0100)]
mpc83xx: boot time regression, move LCRR setup back to cpu_init_f

Commit c7190f02 (retain POR values of non-configured ACR, SPCR, SCCR,
and LCRR bitfields) moved the LCRR assignment to after relocation
to RAM because of the potential problem with changing the local bus
clock while executing from flash.

This change unfortunately adversely affects the boot time, as running
all code up to cpu_init_r can cause significant slowdown.

E.G. on a 8347 board a bootup time increase of ~600ms has been observed:

   0.020 CPU:   e300c1, MPC8347_PBGA_EA, Rev: 3.0 at 400 MHz, CSB: 266.667 MHz
   0.168 RS:    232
   0.172 I2C:   ready
   0.176 DRAM:  64 MB
   1.236 FLASH: 32 MB

Versus:

   0.016 CPU:   e300c1, MPC8347_PBGA_EA, Rev: 3.0 at 400 MHz, CSB: 266.667 MHz
   0.092 RS:    232
   0.092 I2C:   ready
   0.096 DRAM:  64 MB
   0.644 FLASH: 32 MB

So far no boards have needed the late LCRR setup, so simply revert it
for now - If it is needed at a later time, those boards can either do
their own final LCRR setup in board code (E.G. in board_early_init_r),
or we can introduce a CONFIG_SYS_LCRR_LATE config option to only do
the setup in cpu_init_r.

Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
Signed-off-by: Kim Phillips <kim.phillips@freescale.com>
14 years agocommon: delete CONFIG_SYS_64BIT_VSPRINTF and CONFIG_SYS_64BIT_STRTOUL
Heiko Schocher [Thu, 3 Dec 2009 10:21:21 +0000 (11:21 +0100)]
common: delete CONFIG_SYS_64BIT_VSPRINTF and CONFIG_SYS_64BIT_STRTOUL

There is more and more usage of printing 64bit values,
so enable this feature generally, and delete the
CONFIG_SYS_64BIT_VSPRINTF and CONFIG_SYS_64BIT_STRTOUL
defines.

Signed-off-by: Heiko Schocher <hs@denx.de>
14 years ago5xxx, fdt: move fdt_fixup_memory() to cpu.c file
Heiko Schocher [Thu, 3 Dec 2009 10:20:06 +0000 (11:20 +0100)]
5xxx, fdt: move fdt_fixup_memory() to cpu.c file

u-boot updates, before starting Linux, the memory node in the
DTS. As this is a "standard" feature, move this functionality
to the cpu.c file for mpc5xxx and mpc512x processors.

Signed-off-by: Heiko Schocher <hs@denx.de>
14 years agomicroblaze: Correct ffs regression for Microblaze
Michal Simek [Tue, 8 Dec 2009 08:12:49 +0000 (09:12 +0100)]
microblaze: Correct ffs regression for Microblaze

We are using generic implementation of ffs. This should
be part of Simon's commit 0413cfecea350000eab5e591a0965c3e3ee0ff00

Here is warning message which this patch removes.

In file included from /tmp/u-boot-microblaze/include/common.h:38,
                 from cmd_mtdparts.c:87:
/tmp/u-boot-microblaze/include/linux/bitops.h:123:1: warning: "ffs" redefined
In file included from /tmp/u-boot-microblaze/include/linux/bitops.h:110,
                 from /tmp/u-boot-microblaze/include/common.h:38,
                 from cmd_mtdparts.c:87:
/tmp/u-boot-microblaze/include/asm/bitops.h:269:1:
warning: this is the location of the previous definition

Signed-off-by: Michal Simek <monstr@monstr.eu>
14 years agomicroblaze: Stop stack clobbering in microblaze-generic.
Graeme Smecher [Mon, 7 Dec 2009 16:09:57 +0000 (08:09 -0800)]
microblaze: Stop stack clobbering in microblaze-generic.

A typo caused the stack and malloc regions to overlap, which prevented
mem_malloc_init() from returning. This commit makes the memory layout match
the example described in include/configs/microblaze-generic.h

Signed-off-by: Graeme Smecher <graeme.smecher@mail.mcgill.ca>
Signed-off-by: Michal Simek <monstr@monstr.eu>
14 years agoMerge branch 'master' of ../work into next
Wolfgang Denk [Mon, 7 Dec 2009 22:42:33 +0000 (23:42 +0100)]
Merge branch 'master' of ../work into next

14 years agoUpdate CHANGELOG, prepare -rc2 v2009.11-rc2
Wolfgang Denk [Mon, 7 Dec 2009 22:14:13 +0000 (23:14 +0100)]
Update CHANGELOG, prepare -rc2

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agoMAKEALL: Fix return value
Peter Tyser [Mon, 7 Dec 2009 05:58:28 +0000 (23:58 -0600)]
MAKEALL: Fix return value

Previously MAKEALL would always return a value of 0, even if 1 or more
boards did not compile.  This change causes MAKEALL to return 0 if all
boards were able to build, otherwise 1.

This change also requires changing the script interpreter from sh to
bash to support bash's PIPESTATUS variable.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
14 years agotools/mkimage: Remove duplicate line of code
Peter Tyser [Sun, 6 Dec 2009 07:33:24 +0000 (01:33 -0600)]
tools/mkimage: Remove duplicate line of code

Recent commits 1a99de2cb4d08eb3bf9fb3f60a9d533150de8c0e and
6a590c5f5fd12cdd27f3153522acfac3854590e7 both fixed the same bug in the
same manner.  Unfortunately git was "smart" enough to merge both changes
which resulted in some duplicate code.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Reordered code and comment a bit.

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agoi2c: fix dangling comment in do_i2c_mw()
Heiko Schocher [Sat, 5 Dec 2009 10:59:51 +0000 (11:59 +0100)]
i2c: fix dangling comment in do_i2c_mw()

commit bd3784df94bfeca43fbf34094df9cb1bd3ecca3b deleted some unused
code in do_i2c_mw(), but missed to also remove the respective
commment. This patch fixes this.

Signed-off-by: Heiko Schocher <hs@denx.de>
14 years agompc52xx, manroland: add some commands
Heiko Schocher [Thu, 3 Dec 2009 10:20:42 +0000 (11:20 +0100)]
mpc52xx, manroland: add some commands

add the following commands for the manroland boards:

CONFIG_CMDLINE_EDITING
CONFIG_COMMAND_HISTORY
CONFIG_AUTO_COMPLETE

Signed-off-by: Heiko Schocher <hs@denx.de>
14 years agoMerge branch 'master' into next
Wolfgang Denk [Mon, 7 Dec 2009 21:47:17 +0000 (22:47 +0100)]
Merge branch 'master' into next

Conflicts:
lib_generic/zlib.c

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agoPOST: Remove duplicated post_hotkey_pressed() functions
Stefan Roese [Thu, 3 Dec 2009 05:24:30 +0000 (06:24 +0100)]
POST: Remove duplicated post_hotkey_pressed() functions

This patch introduces a weak default function for post_hotkey_pressed(),
returning 0, for boards without hotkey support. The long-running tests
won't be started on those boards. This default function was implemented
in many board directories. By implementing this weak default we can
remove all those duplicate versions.

Boards with hotkey support, can override this weak default function
by defining one in their board specific code.

Signed-off-by: Stefan Roese <sr@denx.de>
14 years ago52xx, manroland: add fdt_fixup_memory() in ft_board_setup()
Heiko Schocher [Tue, 1 Dec 2009 18:30:47 +0000 (19:30 +0100)]
52xx, manroland: add fdt_fixup_memory() in ft_board_setup()

To update the real memory size in the memory node on the
uc101 and mucmc52 boards call fdt_fixup_memory() in
ft_board_setup().

Signed-off-by: Heiko Schocher <hs@denx.de>
14 years agoFix computation in nand_util.c:get_len_incl_bad
Daniel Hobi [Tue, 1 Dec 2009 13:05:55 +0000 (14:05 +0100)]
Fix computation in nand_util.c:get_len_incl_bad

Depending on offset, flash size and the number of bad blocks,
get_len_incl_bad may return a too small value which may lead to:

1) If there are no bad blocks, nand_{read,write}_skip_bad chooses the
bad block aware read/write code. This may hurt performance, but does
not have any adverse effects.

2) If there are bad blocks, the nand_{read,write}_skip_bad may choose
the bad block unaware read/write code (if len_incl_bad == *length)
which leads to corrupted data.

Signed-off-by: Daniel Hobi <daniel.hobi@schmid-telecom.ch>
14 years agonfs: NfsTimeout() updates
Evan Samanas [Tue, 10 Nov 2009 02:08:36 +0000 (20:08 -0600)]
nfs: NfsTimeout() updates

- NfsTimeout() does not correctly update the NFS timeout value which
  results in NfsTimeout() only being called once in certain situations.
  This can result in the 'nfs' command hanging indefinetly.  For
  example, the command:

    nfs 192.168.0.1:/home/user/file

  will not exit until ctrl-c is pressed if 192.168.0.1 does not have an
  NFS server running.

  This issue is resolved by reinitializting the NFS timeout value inside
  NfsTimeout() when a timeout occurs.

- Make the 'nfs' command print the 'T' character when a timeout occurs.
  Previously there was no indication that timeouts were occuring.

- Mimic the 'tftpboot' command and when a download fails print "Retry
  count exceeded; starting again", and restart the download taking the
  'netretry' environment variable into account.

Signed-off-by: Evan Samanas <esamanas@xes-inc.com>
Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
Tested on TQM8xxL.

Tested by: Wolfgang Denk <wd@denx.de>

Tested on MPC8527DS.

Tested by: Ed Swarthout <Ed.Swarthout@freescale.com>

14 years agobootm: Fix help message's sub-command ordering
Peter Tyser [Thu, 19 Nov 2009 01:08:59 +0000 (19:08 -0600)]
bootm: Fix help message's sub-command ordering

The help message for the 'bootm' command listed the 'cmdline' and 'bdt'
sub-commands in the wrong order which resulted in the error below when
following the 'help' command's instructions:

  "Trying to execute a command out of order"

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
14 years agohelp: Correct syntax of nandecc help output.
Robert P. J. Day [Tue, 17 Nov 2009 12:30:23 +0000 (07:30 -0500)]
help: Correct syntax of nandecc help output.

"nandecc" help output should not reproduce the command name, nor have
a trailing newline.

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
14 years agosmc911x: update SMC911X related configuration description
Mike Rapoport [Wed, 11 Nov 2009 08:03:03 +0000 (10:03 +0200)]
smc911x: update SMC911X related configuration description

Since commit 736fead8fdbf8a8407048bebc373cd551d01ec98 "Convert SMC911X
Ethernet driver to CONFIG_NET_MULTI API" SMC911X configration options
are called CONFIG_SMC911X rather than CONFIG_DRIVER_SMC911X. Update
README to reflect that change.

Signed-off-by: Mike Rapoport <mike@compulab.co.il>
14 years agosmc911x: fix typo in smc911x_handle_mac_address name
Mike Rapoport [Wed, 11 Nov 2009 08:03:09 +0000 (10:03 +0200)]
smc911x: fix typo in smc911x_handle_mac_address name

Signed-off-by: Mike Rapoport <mike@compulab.co.il>
14 years agofix nfs symlink name corruption
Ed Swarthout [Thu, 19 Nov 2009 08:47:28 +0000 (02:47 -0600)]
fix nfs symlink name corruption

An off by one error may cause nfs readlink lookup fail if
nfs_path_buff has non-zero data from a previous use.

Loading: *** ERROR: File lookup fail

Signed-off-by: Ed Swarthout <Ed.Swarthout@freescale.com>
14 years agoat91sam9261ek.c: fix minor coding style issue.
Wolfgang Denk [Mon, 7 Dec 2009 20:06:40 +0000 (21:06 +0100)]
at91sam9261ek.c: fix minor coding style issue.

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agotrab: fix warning: implicit declaration of function 'disable_vfd'
Wolfgang Denk [Sun, 6 Dec 2009 00:21:28 +0000 (01:21 +0100)]
trab: fix warning: implicit declaration of function 'disable_vfd'

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agozlib.c: avoid build conflicts for cradle board
Wolfgang Denk [Sat, 5 Dec 2009 23:53:18 +0000 (00:53 +0100)]
zlib.c: avoid build conflicts for cradle board

Commit dce3d79710 updated the zlib code to v0.95; this caused
conflicts when building for the "cradle" board, because the (pretty
generic) preprocessor variable "OFF" was used in multiple files.
Make sure to avoid further conflicts by #undef'ing it in zlib.c
before redefining it.

Signed-off-by: Wolfgang Denk <wd@denx.de>
cc: Giuseppe Condorelli <giuseppe.condorelli@st.com>
cc: Angelo Castello <angelo.castello@st.com>
cc: Alessandro Rubini <rubini-list@gnudd.com>

14 years agoFix out-of-tree building of "apollon" board.
Wolfgang Denk [Sat, 5 Dec 2009 23:26:19 +0000 (00:26 +0100)]
Fix out-of-tree building of "apollon" board.

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agoMerge branch 'master' into next
Wolfgang Denk [Sat, 5 Dec 2009 01:11:59 +0000 (02:11 +0100)]
Merge branch 'master' into next

Conflicts:
board/esd/plu405/plu405.c
drivers/rtc/ftrtc010.c

Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agolzma: ignore unset filesizes
Mike Frysinger [Fri, 4 Dec 2009 10:35:15 +0000 (05:35 -0500)]
lzma: ignore unset filesizes

The Linux kernel build system changed how it compresses things with LZMA
such that the header no longer contains the filesize (it is instead set to
all F's).  So if we get a LZMA image that has -1 for the 64bit field,
let's just assume that the decompressed size is unknown and continue on.

Signed-off-by: Mike Frysinger <vapier@gentoo.org>
14 years agoREADME: Rearrange paragraphs to regain linear arrangement.
Detlev Zundel [Tue, 1 Dec 2009 16:16:19 +0000 (17:16 +0100)]
README: Rearrange paragraphs to regain linear arrangement.

Two later additions to the Configuration Option section unfortunately
split the description of Show boot progress and the list of its call outs.

Signed-off-by: Detlev Zundel <dzu@denx.de>
14 years agozlib: Optimize decompression
Joakim Tjernlund [Thu, 19 Nov 2009 12:22:44 +0000 (13:22 +0100)]
zlib: Optimize decompression

This patch optimizes the direct copy procedure.
Uses get_unaligned() but only in one place.
The copy loop just above this one can also use this
optimization, but I havn't done so as I have not tested if it
is a win there too.
On my MPC8321 this is about 17% faster on my JFFS2 root FS
than the original. No speed test has been performed in u-boot.

Size increase on ppc: 484 bytes

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>
Acked-by: Peter Korsgaard <jacmet@sunsite.dk>
14 years agoadd lzop decompression support
Peter Korsgaard [Thu, 19 Nov 2009 10:37:51 +0000 (11:37 +0100)]
add lzop decompression support

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 lzop decompression
code is based on Albin Tonnerre's recent ARM Linux lzo support patch.

Cc: albin.tonnerre@free-electrons.com
Signed-off-by: Peter Korsgaard <jacmet@sunsite.dk>
14 years agotools/mkimage: Print FIT image contents after creation
Peter Tyser [Tue, 24 Nov 2009 22:42:10 +0000 (16:42 -0600)]
tools/mkimage: Print FIT image contents after creation

Previously, there was no indication to the user that a FIT image was
successfully created after executing mkimage.  For example:

  $ mkimage -f uImage.its uImage.itb
  DTC: dts->dtb  on file "uImage.its"

Adding some additional output after creating a FIT image lets the user
know exactly what is contained in their image, eg:

  $ mkimage -f uImage.its uImage.itb
  DTC: dts->dtb  on file "uImage.its"
  FIT description: Linux kernel 2.6.32-rc7-00201-g7550d6f-dirty
  Created:         Tue Nov 24 15:43:01 2009
   Image 0 (kernel@1)
    Description:  Linux Kernel 2.6.32-rc7-00201-g7550d6f-dirty
    Type:         Kernel Image
    Compression:  gzip compressed
    Data Size:    2707311 Bytes = 2643.86 kB = 2.58 MB
    Architecture: PowerPC
    OS:           Linux
    Load Address: 0x00000000
    Entry Point:  0x00000000
    Hash algo:    crc32
    Hash value:   efe0798b
    Hash algo:    sha1
    Hash value:   ecafba8c95684f2c8fec67e33c41ec88df1534d7
   Image 1 (fdt@1)
    Description:  Flattened Device Tree blob
    Type:         Flat Device Tree
    Compression:  uncompressed
    Data Size:    12288 Bytes = 12.00 kB = 0.01 MB
    Architecture: PowerPC
    Hash algo:    crc32
    Hash value:   a5cab676
    Hash algo:    sha1
    Hash value:   168722b13e305283cfd6603dfe8248cc329adea6
   Default Configuration: 'config@1'
   Configuration 0 (config@1)
    Description:  Default Linux kernel
    Kernel:       kernel@1
    FDT:          fdt@1

This brings the behavior of creating a FIT image in line with creating a
standard uImage, which also prints out the uImage contents after
creation.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
14 years agotools/fit_image.c: Remove unused fit_set_header()
Peter Tyser [Tue, 24 Nov 2009 22:42:09 +0000 (16:42 -0600)]
tools/fit_image.c: Remove unused fit_set_header()

The FIT fit_set_header() function was copied from the standard uImage's
image_set_header() function during mkimage reorganization.  However, the
fit_set_header() function is not used since FIT images use a standard
device tree blob header.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
14 years agotools/mkimage: Assume FDT image type for FIT images
Peter Tyser [Tue, 24 Nov 2009 22:42:08 +0000 (16:42 -0600)]
tools/mkimage: Assume FDT image type for FIT images

When building a Flattened Image Tree (FIT) the image type needs to be
"flat_dt".  Commit 89a4d6b12fd6394898b8a454cbabeaf1cd59bae5 introduced a
regression which caused the user to need to specify the "-T flat_dt"
parameter on the command line when building a FIT image.  The "-T
flat_dt" parameter should not be needed and is at odds with the current
FIT image documentation.

Signed-off-by: Peter Tyser <ptyser@xes-inc.com>
14 years agoGeneric udelay() with watchdog support
Ingo van Lil [Tue, 24 Nov 2009 13:09:21 +0000 (14:09 +0100)]
Generic udelay() with watchdog support

According to the PPC reference implementation the udelay() function is
responsible for resetting the watchdog timer as frequently as needed.
Most other architectures do not meet that requirement, so long-running
operations might result in a watchdog reset.

This patch adds a generic udelay() function which takes care of
resetting the watchdog before calling an architecture-specific
__udelay().

Signed-off-by: Ingo van Lil <inguin@gmx.de>
14 years agoi386: Final Relocation
Graeme Russ [Tue, 24 Nov 2009 09:04:21 +0000 (20:04 +1100)]
i386: Final Relocation

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Move references to link script exports
Graeme Russ [Tue, 24 Nov 2009 09:04:20 +0000 (20:04 +1100)]
i386: Move references to link script exports

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Remove inline asm symbols from .dynsym
Graeme Russ [Tue, 24 Nov 2009 09:04:19 +0000 (20:04 +1100)]
i386: Remove inline asm symbols from .dynsym

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Rearrange Interupt Handling
Graeme Russ [Tue, 24 Nov 2009 09:04:18 +0000 (20:04 +1100)]
i386: Rearrange Interupt Handling

In preperation for full relocation

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Fix race condition when using SC520 timers
Graeme Russ [Tue, 24 Nov 2009 09:04:17 +0000 (20:04 +1100)]
i386: Fix race condition when using SC520 timers

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Fix global label in inline asm compile error
Graeme Russ [Tue, 24 Nov 2009 09:04:16 +0000 (20:04 +1100)]
i386: Fix global label in inline asm compile error

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Reorder source objects in lib_i386 Makefile
Graeme Russ [Tue, 24 Nov 2009 09:04:15 +0000 (20:04 +1100)]
i386: Reorder source objects in lib_i386 Makefile

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Fix link collisions resulting from gcc4.4.1 upgrade
Graeme Russ [Tue, 24 Nov 2009 09:04:14 +0000 (20:04 +1100)]
i386: Fix link collisions resulting from gcc4.4.1 upgrade

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Fix malloc initialization
Graeme Russ [Tue, 24 Nov 2009 09:04:13 +0000 (20:04 +1100)]
i386: Fix malloc initialization

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoi386: Fix dlmalloc compile warning
Graeme Russ [Tue, 24 Nov 2009 09:04:12 +0000 (20:04 +1100)]
i386: Fix dlmalloc compile warning

Signed-off-by: Graeme Russ <graeme.russ@gmail.com>
14 years agoEXT2FS: fix inode size for ext2fs rev#0
Michael Brandt [Sun, 22 Nov 2009 13:13:27 +0000 (14:13 +0100)]
EXT2FS: fix inode size for ext2fs rev#0

extfs.c assumes that there is always a valid inode_size field in the
superblock. But this is not true for ext2fs rev 0. Such ext2fs images
are for instance generated by genext2fs. Symptoms on ARM machines are
messages like: "raise: Signal # 8 caught"; on PowerPC "ext2ls" will
print nothing.
This fix checks for rev 0 and uses then 128 bytes as inode size.

Signed-off-by: Michael Brandt <Michael.Brandt@emsyso.de>
Tested on: TQM5200S
Tested-by: Wolfgang Denk <wd@denx.de>
Signed-off-by: Wolfgang Denk <wd@denx.de>
14 years agoRemove superfluous preprocessor tests from some cmd_*.c files.
Robert P. J. Day [Thu, 19 Nov 2009 16:00:28 +0000 (11:00 -0500)]
Remove superfluous preprocessor tests from some cmd_*.c files.

A small number of common/cmd_*.c files contain preprocessor tests that
are apparently superfluous since those same tests are used in the
Makefile to control the compilation of those files.  Those tests are
clearly redundant as long as they surround the entirety of the source
in those files.

Signed-off-by: Robert P. J. Day <rpjday@crashcourse.ca>
14 years agocrc32: Impl. linux optimized crc32()
Joakim Tjernlund [Thu, 19 Nov 2009 12:44:16 +0000 (13:44 +0100)]
crc32: Impl. linux optimized crc32()

Ported over the more efficient linux crc32() function.
A quick comparsion on ppc:
After changing the old crc32 to do 4 bytes in the
inner loop to be able to compare with new version one can note:
- old inner loop has 61 insn, new has 19 insn.
- new crc32 does one 32 bit load of data to crc while
  the old does four 8 bits loads.
- size is bit bigger for the new crc32:
  1392(old) 1416(new) of text. The is because the new version
  shares code with crc32_no_comp() instead of duplicating code.
- about 33% faster on ppc:
  New > crc 0 0xfffffff -> 39 secs
  Old > crc 0 0xfffffff -> 60 secs

Signed-off-by: Joakim Tjernlund <Joakim.Tjernlund@transmode.se>