* Re: [PATCH] ppc: add support for new powerbooks
From: Olof Johansson @ 2005-11-10 16:44 UTC (permalink / raw)
To: Andrew Morton; +Cc: linuxppc-dev
In-Reply-To: <20051109213837.47c8dce7.akpm@osdl.org>
On Wed, Nov 09, 2005 at 09:38:37PM -0800, Andrew Morton wrote:
> Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:
> >
> > Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>
> I don't think I'm adding much value handling ppc/ppc64 patches, really.
> I'd prefer to get out of that business.
Main benefit for me, as a patch submitter, is that you consistently,
always, ACK when you pick up or drop a patch. In other words, I don't
have to go polling a tree to see if it made it in or not.
> It would be better to have an arch maintainer who runs a git tree, same as
> ia64, arm, etc.
>
> And, given the amount of shared infrastructure, I suspect it would have to
> be a single git tree for both architectures.
>
> I'd still sweep up random ppc patches, but those will go into mainline via the
> originator->mm->git-powerpc->linus route.
>
> Possible?
I'm fine with all that; to be honest I haven't been certain where the
smaller ppc/ppc64 patches should go during this whole merge business (or
after, for that matter).
If Paulus wants to start an arch tree, that's fine with me. I do share
Ben's opinion about less review though, I guess we'll just have to deal
with that.
Also, Paul, I kindly request some sort of ACK mechanism similar to
Andrew's when patches are picked up. :-)
Thanks,
-Olof
^ permalink raw reply
* Re: [PATCH] ppc: add support for new powerbooks
From: Kumar Gala @ 2005-11-10 17:12 UTC (permalink / raw)
To: Olof Johansson; +Cc: Andrew Morton, linuxppc-dev
In-Reply-To: <20051110164404.GA7303@pb15.lixom.net>
On Nov 10, 2005, at 10:44 AM, Olof Johansson wrote:
> On Wed, Nov 09, 2005 at 09:38:37PM -0800, Andrew Morton wrote:
>> Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:
>>>
>>> Acked-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
>>
>> I don't think I'm adding much value handling ppc/ppc64 patches,
>> really.
>> I'd prefer to get out of that business.
>
> Main benefit for me, as a patch submitter, is that you consistently,
> always, ACK when you pick up or drop a patch. In other words, I don't
> have to go polling a tree to see if it made it in or not.
>
>> It would be better to have an arch maintainer who runs a git tree,
>> same as
>> ia64, arm, etc.
>>
>> And, given the amount of shared infrastructure, I suspect it would
>> have to
>> be a single git tree for both architectures.
>>
>> I'd still sweep up random ppc patches, but those will go into
>> mainline via the
>> originator->mm->git-powerpc->linus route.
>>
>> Possible?
>
> I'm fine with all that; to be honest I haven't been certain where the
> smaller ppc/ppc64 patches should go during this whole merge
> business (or
> after, for that matter).
>
> If Paulus wants to start an arch tree, that's fine with me. I do share
> Ben's opinion about less review though, I guess we'll just have to
> deal
> with that.
>
> Also, Paul, I kindly request some sort of ACK mechanism similar to
> Andrew's when patches are picked up. :-)
I agree with Olof's comments. Additionally, I think one reason we
had been sending things to Andrew for PPC32 embedded was that the sub-
arch maintainers (Matt, Marcelo, Tom, Dan, Sylvain, myself, etc..)
needed a responsive path to get things to Linus which Andrew was
providing. If Paul's going to be able to handle accepting and
dealing with all the embedded patches I'm fine with having a PPC git
tree. I'm even willing to setup some GIT tree's for sub-arch's I
deal with for Paul to pull from if he wants.
Just my two cents.
- kumar
^ permalink raw reply
* Re: "Now booting the kernel"
From: Nitesh Guinde @ 2005-11-10 18:26 UTC (permalink / raw)
To: linuxppc-embedded
Hello Nathael,
I tried the stuff but I am still getting the same result.
I did not find the section "/* 2. Invalidate all entries except the
entry we're executing in */" So I commented the TLBWE line in the
section which says "invalidate all TLB entries " . Is it correct? The
ASM code for rebooting will have to be changed for my board... right?
Nitesh
^ permalink raw reply
* 82xxHCI driver for MPC8280 Specifically
From: Russell McGuire @ 2005-11-10 21:16 UTC (permalink / raw)
To: linuxppc-embedded
In-Reply-To: <20051110145136.3E4EC68715@ozlabs.org>
All,
I have been trying to get the 82xxHCI driver to work on our protoboard and
after a lot of correspondence with Freescale FAE's there are some real
differences in code implementation for some CPU parts.
Our board is using a MPC8280 Rev 0.1 part, the Rev number of course is the
last bit of the IMMR register. So 0xXXXX0A01 is our CPU, the 0A is all 8280
parts.
I was curious if anyone else has gotten this driver to operation
specifically on an 8280 CPU. Freescale says that ALL 8280's do NOT NEED any
microcode patch only some Rev's only need hardware workarounds to generate
the SOF packets, mainly wiring a BRG or TIMER to the DREQ1 / 4 lines at
least for Rev 0.0 and Rev 0.1. Rev A parts do not need this, as it is all
working internally, guess they finally fixed it.
This documentation is not conclusive as to how to wire and setup the DREQ
lines, and is no longer available on their websites, I have the PDF file
entitled MPC8280UMAD_D.pdf if anyone is curious about Rev 0.0 and Rev 0.1
parts.
So help, has anyone got this working with the Rev 0.0 and Rev 0.1 parts, and
if so, how? I have been spending way to much time reading and hacking code
to this barely functional. I would be happy to send anyone the code I have
currently, if they want to start in on this.
Russell McGuire
-----Original Message-----
From: linuxppc-embedded-bounces@ozlabs.org
[mailto:linuxppc-embedded-bounces@ozlabs.org] On Behalf Of
linuxppc-embedded-request@ozlabs.org
Sent: Thursday, November 10, 2005 6:52 AM
To: linuxppc-embedded@ozlabs.org
Subject: Linuxppc-embedded Digest, Vol 15, Issue 21
Send Linuxppc-embedded mailing list submissions to
linuxppc-embedded@ozlabs.org
To subscribe or unsubscribe via the World Wide Web, visit
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
or, via email, send a message with subject or body 'help' to
linuxppc-embedded-request@ozlabs.org
You can reach the person managing the list at
linuxppc-embedded-owner@ozlabs.org
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Linuxppc-embedded digest..."
Today's Topics:
1. Re: [PATCH 2.6.14] mm: 8xx MM fix for (David Jander)
2. PowerPC reservations (Kalle Pokki)
3. Re: "Now booting the kernel" (Nathael PAJANI)
4. floating point + kernel (PPC) (srideep.devireddy@wipro.com)
5. SCC QMC driver for 8247 (Robin Mathew)
6. Re: "Now booting the kernel" (Nathael PAJANI)
7. Re: [PATCH] ppc32 8xx: MPC8xx PCMCIA update (Dominik Brodowski)
----------------------------------------------------------------------
Message: 1
Date: Thu, 10 Nov 2005 10:18:07 +0200
From: David Jander <david.jander@protonic.nl>
Subject: Re: [PATCH 2.6.14] mm: 8xx MM fix for
To: linuxppc-embedded@ozlabs.org
Message-ID: <200511100918.08060.david.jander@protonic.nl>
Content-Type: text/plain; charset="iso-8859-1"
On Thursday 10 November 2005 08:48, David Jander wrote:
>[...]
> Hmmm. This is a lot in the line of the tests I did with (the more generic
> benchmark) nbench. After looking at those results (see my other post in
> this thread) I already suspected something like this.
Sorry, I obviously did not mean this thread, but the following post on
another
thread:
http://ozlabs.org/pipermail/linuxppc-embedded/2005-November/020775.html
Regards,
--
David Jander
------------------------------
Message: 2
Date: Thu, 10 Nov 2005 10:20:30 +0200
From: Kalle Pokki <kalle.pokki@iki.fi>
Subject: PowerPC reservations
To: linuxppc-embedded <linuxppc-embedded@ozlabs.org>
Message-ID: <437302CE.8070309@iki.fi>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Hi,
Can someone please help me understand how the memory reservations in
PowerPC actually work. Let's just assume uniprocessor with a pre-emptive
scheduler, and take a text-book example of an atomic increment case,
which is also frequently used in e.g. the Linux kernel. With two atomic
operations, everything seems to be just fine. But how about with three
concurrent threads of execution?
From the following code, assume r3 contains the same address for each
incrementing operation. If the first atomic increment is pre-empted in
the middle, execution then jumps to the second increment (by the
scheduler). The second increment runs through and succeeds, and
continues straight to the third increment. Then it is again pre-empted
in the middle, execution returning to the first increment. Now the
processor has the reservation with the correct address, and the first
increment succeeds when still holding the original input value. The
first and the second increment thus write the same value in memory.
After the first increment, the scheduler again continues the third
increment, which doesn't succeed a first, but the second round succeeds.
However, the value in the address pointed by r3 was not increased by
three, but by two.
Am I just not getting how this is really supposed to work? Are there
still some other constructs in use to prevent this, e.g. extra stwcx.
instructions when changing the thread of execution?
I'm also wondering why the architecture specifically defines the stwcx.
instruction to have, well, undefined behavior in case the reservation
address differs from the address of the previous lwarx...
1: lwarx r6, r0, r3
addi r6, r6, 1
stwcx. r6, r0, r3
bne- 1b
.....
2: lwarx r7, r0, r3
addi r7, r7, 1
stwcx. r7, r0, r3
bne- 2b
3: lwarx r8, r0, r3
addi r8, r8, 1
stwcx. r8, r0, r3
bne- 3b
------------------------------
Message: 3
Date: Thu, 10 Nov 2005 11:40:14 +0100
From: Nathael PAJANI <nathael.pajani@cpe.fr>
Subject: Re: "Now booting the kernel"
To: linuxppc-embedded@ozlabs.org
Message-ID: <4373238E.8000405@cpe.fr>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
(Sorry, I first replyed with the wrong mail address)
>Linux/PPC load: console=ttyS0,9600 root=/dev/xsysace/disc0/part2 rw
>Uncompressing Linux...done.
>Now booting the kernel
Hi!
At this state the bootloader stops executing and the Linux Kernel
starts. The problem is that the Linux kernel does not know yet how to
use the serial.
You won't have any message before it is set up (in early-console if my
memory is allright)
So what you can do to check what's going on, is put "breakpoints" in the
boot sequence. This means in the file arch/ppc/kernel/head*.S used for
your board you should try to comment the line with the "tlbwe"
instruction in the section "/* 2. Invalidate all entries except the
entry we're executing in */"
This will allow you to keep access to your board registers.
Then you go step by step, putting some code which will reboot the board
when executed, so you know you're going up to that point, and then move
the "breakpoint" further.
This code does the reboot (for the booke I can reboot the board by
writting '4' at address 0xfa001001):
ASM:
lis r4,0xfa00
li r5,4
stb r5,0x1001(r4)
msync
C:
*((volatile unsigned char*)0xfa001001 = 4;
This way, instead of hanging up, the board reboots and you know where
you are.
If you're going up to this:
bl machine_init /* arch/ppc/kernel/setup.c */
bl MMU_init /* arch/ppc/mm/init.c */
It's quite good, these are C functions, but they are processor specific,
once again, check that the ones used (compiled) are those you need.
And next you've got the "start_kernel" call, which leads you to C code
definitely. It's in init/main.c.
I hope I did not tell anything wrong, and that this will help.
Have fun.
Nathael.
------------------------------
Message: 4
Date: Thu, 10 Nov 2005 18:07:31 +0530
From: <srideep.devireddy@wipro.com>
Subject: floating point + kernel (PPC)
To: <linuxppc-embedded@ozlabs.org>
Message-ID:
<6AD9F6A5F6E096408F0B703773355A07605C40@CHN-SNR-MBX01.wipro.com>
Content-Type: text/plain; charset="us-ascii"
Thanks for your help friends ...
I am able to bring my kernel up ...... right now I have a problem
after some time at the console .... I get floating point used in kernel
(task=C0188480,pc=1100)
10.0.0.2 login: root
Last login: Tnhu Jan 1 00:00sole
Linux 10.0.0.2 2.4.22 #1 Tue Nov 8 12:03:26 UTC 2005 ppc unknown
MontaVista(R) Linux(R) Professional Edition 3.1
root@10.0.0.2:~# <mailto:root@10.0.0.2:~>
root@10.0.0.2:/home# cd ..
root@10.0.0.2:/# ls
bin boot dev etc home lib mnt opt proc root sbin tmp usr
var
root@10.0.0.2:/# floating point used in kernel (task=c0188480, pc=1100)
What does this mean ...? This makes the console hang .....Can any 1 help
me on in this?
Best Regards
srideep
Confidentiality Notice
The information contained in this electronic message and any attachments to
this message are intended
for the exclusive use of the addressee(s) and may contain confidential or
privileged information. If
you are not the intended recipient, please notify the sender at Wipro or
Mailadmin@wipro.com immediately
and destroy all copies of this message and any attachments.
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://ozlabs.org/pipermail/linuxppc-embedded/attachments/20051110/30c43a42/
attachment-0001.htm
------------------------------
Message: 5
Date: Thu, 10 Nov 2005 20:18:48 +0530
From: Robin Mathew <robin@india.tejasnetworks.com>
Subject: SCC QMC driver for 8247
To: Linuxppc-embedded <linuxppc-embedded@ozlabs.org>
Cc: Nicholas Basker <nbasker@india.tejasnetworks.com>
Message-ID: <43735DD0.9030205@india.tejasnetworks.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Hello:
We are working on implementing driver for SCC operating in QMC mode that
supports multiple HDLC channels. The details of the system we are using
are the following.
CPU Version = 8247 based on 82xx family of processor(PVR 80822014)
The Linux OS Version = DENKS Linux version 2.4.20.
The Peripherals used are = FCC1 in 10/100 ethernet
FCC2 in 10/100 ethernet
SCC1 currently not used, but
proposed to be used in HDLC mode.
SCC3 is not used.
SCC4 operating in QMC mode
with super -channel capability. Trying to operate 192kbps HDLC channel
(using 3x64kbps QMC channels).
SMC2 in UART mode (as a
simple debug port without any modem signals).
SPI interface currently not
used but shall be used in future.
I2C is being used.
01. Can the channel specific parameters of the SCC4 in QMC mode be
relocated from DPRAM base address?? There is no mention in the
8272RM.pdf about configuring the base address of channel specific
parameters. From the QMC memory structure diagram, it looks like the
channel specific parameters will be taken from the the DPRAM base
address. Can you please confirm it?
02. We have changed the function m8260_cpm_dpalloc() used for DPRAM
allocation in the linux source code. It was allocating memory from
128Byte location to 8KByte location in DPRAM. But since QMC requires the
lower 4KByte for channel specific parameters, we changed the function to
allocate DPRAM from 4KByte location. Will this change lead to any
problem for proper linux operation?
03. We are encountering a strange problem with scc4 parameter RAM. When
the driver is coming up, its trying to initialize the SCC4 parameter RAM
for QMC. In the location (immr + 0x8318), we are writing the value
0x8320 but the value magically changes to 0x8300. We tried to write to
the location using BDI and again the problem is seen. Is this a known
problem with the processor?? This problem doesnot appear everytime. Is
there any known cause for this problem?
Regards,
Robin
------------------------------
Message: 6
Date: Thu, 10 Nov 2005 09:46:30 +0100
From: Nathael PAJANI <nathael.pajani@ecrin.com>
Subject: Re: "Now booting the kernel"
To: linuxppc-embedded@ozlabs.org
Message-ID: <437308E6.2070804@ecrin.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>Linux/PPC load: console=ttyS0,9600 root=/dev/xsysace/disc0/part2 rw
>Uncompressing Linux...done.
>Now booting the kernel
Hi!
At this state the bootloader stops executing and the Linux Kernel
starts. The problem is that the Linux kernel does not know yet how to
use the serial.
You won't have any message before it is set up (in early-console if my
memory is allright)
So what you can do to check what's going on, is put "breakpoints" in the
boot sequence. This means in the file arch/ppc/kernel/head*.S used for
your board you should try to comment the line with the "tlbwe"
instruction in the section "/* 2. Invalidate all entries except the
entry we're executing in */"
This will allow you to keep access to your board registers.
Then you go step by step, putting some code which will reboot the board
when executed, so you know you're going up to that point, and then move
the "breakpoint" further.
This code does the reboot (for the booke I can reboot the board by
writting '4' at address 0xfa001001):
ASM:
lis r4,0xfa00
li r5,4
stb r5,0x1001(r4)
msync
C:
*((volatile unsigned char*)0xfa001001 = 4;
This way, instead of hanging up, the board reboots and you know where
you are.
If you're going up to this:
bl machine_init /* arch/ppc/kernel/setup.c */
bl MMU_init /* arch/ppc/mm/init.c */
It's quite good, these are C functions, but they are processor specific,
once again, check that the ones used (compiled) are those you need.
And next you've got the "start_kernel" call, which leads you to C code
definitely. It's in init/main.c.
I hope I did not tell anything wrong, and that this will help.
Have fun.
Nathael.
------------------------------
Message: 7
Date: Thu, 10 Nov 2005 11:03:04 +0100
From: Dominik Brodowski <linux@dominikbrodowski.net>
Subject: Re: [PATCH] ppc32 8xx: MPC8xx PCMCIA update
To: Paul Mackerras <paulus@samba.org>
Cc: linux-ppc-embedded <linuxppc-embedded@ozlabs.org>
Message-ID: <20051110100304.GA21855@dominikbrodowski.de>
Content-Type: text/plain; charset=us-ascii
Hi,
On Mon, Nov 07, 2005 at 12:03:45PM +1100, Paul Mackerras wrote:
> Marcelo Tosatti writes:
>
> > The following patch updates the MPC8xx PCMCIA driver:
>
> This and the following patch look OK as far as I can tell - Dominik,
> will you take care of sending them to Linus?
Yes, will do.
Thanks,
Dominik
------------------------------
_______________________________________________
Linuxppc-embedded mailing list
Linuxppc-embedded@ozlabs.org
https://ozlabs.org/mailman/listinfo/linuxppc-embedded
End of Linuxppc-embedded Digest, Vol 15, Issue 21
*************************************************
^ permalink raw reply
* Re: 2.6.14 USB vs. sleep issues
From: Wolfgang Pfeiffer @ 2005-11-11 0:24 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc-dev list, debian-powerpc@lists.debian.org
In-Reply-To: <1130999620.4680.28.camel@gaston>
Hi Ben
Thanks for your work, for the patches: on a Titanium IV a kernel-org
based, self-made (via "make-kpkg") 2.6.14.1 patched kernel works. USB
seems being fine here: I put the machine to sleep with a camera and a
Logitech mouse plugged in to USB connectors: After resuming the system
from sleep the mouse is still working, and the camera is usable and
fine. I'm running pbbuttons here.
It seems I have problems because of a - I still can't believe that -
missing pcmciautils package on Debian. But that's unrelated to what
we're discussing here, I believe. I'll start another thread perhaps for
this latter issue ...
On Thu, Nov 03, 2005 at 05:33:39PM +1100, Benjamin Herrenschmidt wrote:
> For those who experience crashes on sleep and/or wakeup (typically due
> to USB) with 2.6.14, I made a test patch that might help. Please let me
> know if it makes things more reliable.
>
> http://gate.crashing.org/~benh/fix-ohci-sleep.diff
[ ... ]
I didn't use the one above, but just the one you posted here Nov.8
(this thread), IINM:
$ patch -p1 < /home/shorty/downloads/ben-patches/benh.nov.8.sleep.diff
patching file drivers/usb/core/hcd-pci.c
patching file drivers/usb/core/hcd.c
patching file drivers/usb/core/hcd.h
patching file drivers/usb/host/ehci-hcd.c
patching file drivers/usb/host/ehci-q.c
patching file drivers/usb/host/ehci-sched.c
patching file drivers/usb/host/ohci-hcd.c
patching file drivers/usb/host/ohci-hub.c
patching file drivers/usb/host/ohci-pci.c
patching file drivers/usb/host/uhci-hcd.c
Following that, provided my .bash_history is well organised enough to
be trusted ( ... :) I applied the 2.6.14.1 patch
("patch-2.6.14.1.bz2")
So: As to sleep/USB: Everything seems being fine here on 2.6.14.1.
FYI:
Kismet still does not seem work for hermes/orinoco on 2.6.14.1.
Thanks again for the patches, Ben.
Best Regards
Wolfgang
--
Wolfgang Pfeiffer
http://profiles.yahoo.com/wolfgangpfeiffer
Key ID: E3037113
Key fingerprint = A8CA 9D8C 54C4 4CC1 0B26 AA3C 9108 FB42 E303 7113
^ permalink raw reply
* 11-11-05 2.6.14 on AMCC Yosemite board(boot error)
From: KylongMu @ 2005-11-11 2:04 UTC (permalink / raw)
To: wd; +Cc: Linuxppc-embedded
In-Reply-To: <20051102164644.6D167353A2C@atlas.denx.de>
[-- Attachment #1: Type: text/plain, Size: 204 bytes --]
Dear Denk:
With today's update, I can make the kernel by Yosemite_defconfig,
but
When boot my board with this new kernel, it halt at boot time, the
attachment
Is the boot message.
Cordially,
Kylong Mu
[-- Attachment #2: bootmessage.txt --]
[-- Type: text/plain, Size: 5911 bytes --]
U-Boot 1.1.3 (Sep 7 2005 - 09:03:02)
IBM PowerPC 440 EP Rev. B
Board: AMCC YOSEMITE
VCO: 1066 MHz
CPU: 533 MHz
PLB: 133 MHz
OPB: 66 MHz
PER: 66 MHz
PCI: 66 MHz
I2C: ready
DRAM: 256 MB
FLASH: 32 MB
PCI: Bus Dev VenId DevId Class Int
In: serial
Out: serial
Err: serial
Net: ppc_440x_eth0, ppc_440x_eth1
Hit any key to stop autoboot: 0
=> tftp 0x400000 uImage
Waiting for PHY auto negotiation to complete.. done
Using ppc_440x_eth0 device
TFTP from server 192.168.0.93; our IP address is 192.168.0.66
Filename 'uImage'.
Load address: 0x400000
Loading: #################################################################
#################################################################
#################################################################
##########################################
done
Bytes transferred = 1212626 (1280d2 hex)
=> bootm 0x400000 0xfed40000
## Booting image at 00400000 ...
Image Name: Linux-2.6.14
Created: 2005-11-11 1:51:20 UTC
Image Type: PowerPC Linux Kernel Image (gzip compressed)
Data Size: 1212562 Bytes = 1.2 MB
Load Address: 00000000
Entry Point: 00000000
Verifying Checksum ... OK
Uncompressing Kernel Image ... OK
## Loading RAMDisk Image at fed40000 ...
Image Name: Yosemite Ramdisk
Created: 2005-08-30 2:13:49 UTC
Image Type: PowerPC Linux RAMDisk Image (gzip compressed)
Data Size: 17770706 Bytes = 16.9 MB
Load Address: 00000000
Entry Point: 00000000
Verifying Checksum ... OK
Loading Ramdisk to 0ee9b000, end 0ff8d8d2 ... OK
Linux version 2.6.14 (root@dxp) (gcc version 4.0.0) #1 Fri Nov 11 09:51:04 CST 2005
AMCC PowerPC 440EP Yosemite Platform
Built 1 zonelists
Kernel command line: ramdisk=49152 root=/dev/ram ip=192.168.1.62:::::eth0:off
PID hash table entries: 2048 (order: 11, 32768 bytes)
Dentry cache hash table entries: 65536 (order: 6, 262144 bytes)
Inode-cache hash table entries: 32768 (order: 5, 131072 bytes)
Memory: 239488k available (1852k kernel code, 604k data, 136k init, 0k highmem)
Mount-cache hash table entries: 512
checking if image is initramfs...it isn't (no cpio magic); looks like an initrd
softlockup thread 0 started up.
Freeing initrd memory: 17354k freed
NET: Registered protocol family 16
PCI: Probing PCI hardware
usbcore: registered new driver usbfs
usbcore: registered new driver hub
SCSI subsystem initialized
io scheduler noop registered
io scheduler anticipatory registered
io scheduler deadline registered
io scheduler cfq registered
usbmon: debugfs is not available
ppc-soc-ohci ppc-soc-ohci.0: USB Host Controller
ppc-soc-ohci ppc-soc-ohci.0: new USB bus registered, assigned bus number 1
ppc-soc-ohci ppc-soc-ohci.0: irq 40, io mem 0xef601000
usb usb1: Product: USB Host Controller
usb usb1: Manufacturer: Linux 2.6.14 ohci_hcd
usb usb1: SerialNumber: PPC-SOC USB
hub 1-0:1.0: USB hub found
hub 1-0:1.0: 2 ports detected
Initializing USB Mass Storage driver...
usbcore: registered new driver usb-storage
USB Mass Storage support registered.
pegasus: v0.6.12 (2005/01/13), Pegasus/Pegasus II USB Ethernet driver
usbcore: registered new driver pegasus
Serial: 8250/16550 driver $Revision: 1.90 $ 4 ports, IRQ sharing enabled
serial8250: ttyS0 at MMIO 0x0 (irq = 0) is a 16550A
serial8250: ttyS1 at MMIO 0x0 (irq = 1) is a 16550A
RAMDISK driver initialized: 16 RAM disks of 49152K size 1024 blocksize
PPC 4xx OCP EMAC driver, version 3.53
mal0: initialized, 4 TX channels, 2 RX channels
zmii0: bridge in RMII mode
eth0: emac0, MAC 00:10:ec:00:8f:aa
eth0: found Generic MII PHY (0x01)
eth1: emac1, MAC 00:10:ec:80:8f:aa
eth1: found Generic MII PHY (0x03)
Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 50MHz system bus speed for PIO modes; override with idebus=xx
Data machine check in kernel mode.
PLB0: BEAR=0x0000000000000000 ACR= 0x00000000 BESR= 0xfdfff7ff
POB0: BEAR=0xc27e3194ffffffff BESR0=0x00000000 BESR1=0x00000000
OPB0: BEAR=0x0000000000000151 BSTAT=0x00000000
Oops: machine check, sig: 7 [#1]
NIP: 00000000 LR: C0157E9C SP: CFFA1EE0 REGS: c0270f50 TRAP: 0202 Not tainted
MSR: 00000000 EE: 0 PR: 0 FP: 0 ME: 0 IR/DR: 00
TASK = cee98b10[1] 'swapper' THREAD: cffa0000
Last syscall: 120
GPR00: 00000002 CFFA1EE0 CEE98B10 C022673C 00000000 00000000 CFFA1F18 00000000
GPR08: 00000002 00000002 00000002 00000001 000F1FFF 60000428 0FFE6E00 00000001
GPR16: FFFFFFFF 007FFF00 0FFE1AD4 00000000 00000003 007FFE70 00800000 00000001
GPR24: C01D0000 C01D0000 C0270000 00000000 C022673C CFFA1F18 00000000 C022673C
NIP [00000000] 0x0
LR [c0157e9c] mtd_do_chip_probe+0xdc/0x334
Call trace:
[c0157e9c] mtd_do_chip_probe+0xdc/0x334
[c0152214] cfi_probe+0x18/0x28
[c0152160] do_map_probe+0x40/0xdc
[c0265568] init_yosemite_flash+0xb4/0x110
[c00013ac] init+0xb4/0x254
[c00046c0] kernel_thread+0x48/0x64
Data machine check in kernel mode.
PLB0: BEAR=0x0000000000000000 ACR= 0x00000000 BESR= 0xfdfff7ff
POB0: BEAR=0xc27e3194ffffffff BESR0=0x00000000 BESR1=0x00000000
OPB0: BEAR=0x0000000000000151 BSTAT=0x00000000
Oops: machine check, sig: 7 [#2]
NIP: 00000000 LR: C000235C SP: C0270E70 REGS: c0270f50 TRAP: 0202 Not tainted
MSR: 00000000 EE: 0 PR: 0 FP: 0 ME: 0 IR/DR: 00
TASK = cee98b10[1] 'swapper' THREAD: cffa0000
Last syscall: 120
GPR00: 08000000 C0270E70 CEE98B10 C0270E80 00001529 FFFFFFFF C0270000 00004000
GPR08: C0270000 C000235C 00021002 C00039E8 CEE98CD8 60000428 0FFE6E00 00000001
GPR16: FFFFFFFF 007FFF00 0FFE1AD4 00000000 00000003 007FFE70 00800000 00000001
GPR24: C01D0000 C01D0000 C0270000 00000000 C022673C CFFA1F18 00000007 C0270F50
NIP [00000000] 0x0
LR [c000235c] ret_from_except+0x0/0x18
Kernel panic - not syncing: Attempted to kill init!
<0>Rebooting in 180 seconds..
^ permalink raw reply
* Re: 11-11-05 2.6.14 on AMCC Yosemite board(boot error)
From: Wolfgang Denk @ 2005-11-11 8:39 UTC (permalink / raw)
To: KylongMu; +Cc: Linuxppc-embedded
In-Reply-To: <20051111020420.9644C1073@smtp.263.net>
In message <20051111020420.9644C1073@smtp.263.net> you wrote:
>
> With today's update, I can make the kernel by Yosemite_defconfig,
> but
> When boot my board with this new kernel, it halt at boot time, the
> attachment
> Is the boot message.
What you show is NOT the current code.
> U-Boot 1.1.3 (Sep 7 2005 - 09:03:02)
Please update U-Boot as well. You need this commit included:
commit f190c11b1f22ff766e046588e5a7bb55f28ae305
tree 1c41a7beffe8a293174dcb7c28b394349204a047
parent 4d03a21f2329bb13a9661a23742ac0b77a7a76d5
author Stefan Roese <sr@denx.de> Sat, 05 Nov 2005 15:50:16 +0100
committer Stefan Roese <sr@denx.de> Sat, 05 Nov 2005 15:50:16 +0100
Adjust bd->bi_flashstart on Yellowstone & Yosemite to correct size
Patch by Stefan Roese, 05 Nov 2005
I recommend to use the top-of-tree from the CVS or GIT repositories.
> Linux version 2.6.14 (root@dxp) (gcc version 4.0.0) #1 Fri Nov 11
> 09:51:04 CST 2005
This is not the current code base. You should use top-of-tree from
GIT repository (commit ff1df84b3c3154ffdb646941e1c70d17554e3042); the
kernel will display something like this:
Linux version 2.6.14-gff1df84b ...
Best regards,
Wolfgang Denk
--
Software Engineering: Embedded and Realtime Systems, Embedded Linux
Phone: (+49)-8142-66989-10 Fax: (+49)-8142-66989-80 Email: wd@denx.de
Nail here --X-- for new monitor.
^ permalink raw reply
* floating point operation of ppc
From: lily @ 2005-11-11 9:13 UTC (permalink / raw)
To: linuxppc-embedded
[-- Attachment #1: Type: text/plain, Size: 543 bytes --]
hi,all:
i build an embedded linux on powerpc 405ep,i need execute floating point operation,and i enable "math emulate" option in kernel, i run a while(1)loop ,in the loop there was floating point expression ,after test ,i found the while (1) exit because of the floating point computation,that is the floating point can't execute on ppc stablely.Is there someone meets the same problems ,please give me some suggestion,thanks!!!
lily
[-- Attachment #2: Type: text/html, Size: 1466 bytes --]
^ permalink raw reply
* packet name created by rpmbuild
From: debora liu @ 2005-11-11 9:58 UTC (permalink / raw)
To: linuxppc-embedde
SGVsbG8gRGVueA0KDQpJJ20gdXNpbmcgcnBtYnVpbGQsIGJ1dCBpdCBjcmVhdCBycG0gcGFja2V0
IG5hbWUgaXMgIk5BTUUtcHBjXzh4eC1WRVJTSU9OLVJFTEVBU0UucHBjLnJwbSIsDQpJIHRoaW5r
IGl0IHNob3VsZCBpcyAiTkFNRS1WRVJTSU9OLVJFTEVBU0UucHBjLnJwbSIsIHRydWU/DQpJIHVz
ZSBuYXRpdmUgcnBtYnVpbGQgb24gbXkgYm9hcmQgd2l0aCBlbGRrLTMuMS4xIFBQQzh4eCBORlMg
cm9vdC4NCg0KoaGhoaGhoaGhoaGhoaGhoWRlYm9yYSBsaXUNCqGhoaGhoaGhoaGhoaGhoaFkZWJv
cmFsaEBzaW5vdmVlLmNvbQ0KoaGhoaGhoaGhoaGhoaGhoaGhoaEyMDA1LTExLTExDQo=
^ permalink raw reply
* [PATCH] powerpc: Merge vdso's and add vdso support to 32 bits kernel
From: Benjamin Herrenschmidt @ 2005-11-11 10:15 UTC (permalink / raw)
To: Paul Mackerras
Cc: linuxppc64-dev, linuxppc-dev list, Steve Munroe, tom_gall@mac.com
This patch moves the vdso's to arch/powerpc, adds support for the 32
bits vdso to the 32 bits kernel, rename systemcfg (finally !), and adds
some new (still untested) routines to both vdso's: clock_gettime() with
support for CLOCK_REALTIME and CLOCK_MONOTONIC, clock_getres() (same
clocks) and get_tbfreq() for glibc to retreive the timebase frequency.
Tom,Steve: The implementation of get_tbfreq() I've done for 32 bits
returns a long long (r3, r4) not a long. This is such that if we ever
add support for >4Ghz timebases on ppc32, the userland interface won't
have to change.
I have tested gettimeofday() using some glibc patches in both ppc32 and
ppc64 kernels using 32 bits userland (I haven't had a chance to test a
64 bits userland yet, but the implementation didn't change and was
tested earlier). I haven't tested yet the new functions.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Due to the size of the patch, I haven't posted it to the lists, it's
available at http://gate.crashing.org/~benh/ppc64-vdso-update.diff
Ben.
^ permalink raw reply
* Re: "Now booting the kernel"
From: drizzt @ 2005-11-11 12:51 UTC (permalink / raw)
To: linuxppc-embedded
In-Reply-To: <20051111010005.016A868717@ozlabs.org>
Hi,
> I tried the stuff but I am still getting the same result.
This step can be quite hard to go over... do not deseparate :)
> I did not find the section "/* 2. Invalidate all entries except the
> entry we're executing in */" So I commented the TLBWE line in the
> section which says "invalidate all TLB entries " . Is it correct?
I think it is.
Do not forget to uncomment it when you start having outputs.
> The ASM code for rebooting will have to be changed for my board... right?
Yes
On my board, this is a feature of the VME bus, which you should'nt have,
so check for an instruction or register to perform a reset.
Have fun :) :)
Nathael.
^ permalink raw reply
* Re: [PATCH] gianfar mii needs to zero out the mii_bus structure
From: Jeff Garzik @ 2005-11-11 13:27 UTC (permalink / raw)
To: Kumar Gala; +Cc: netdev, linuxppc-embedded
In-Reply-To: <Pine.LNX.4.44.0511091211030.26349-100000@gate.crashing.org>
applied
^ permalink raw reply
* Re: floating point operation of ppc
From: Greg Weeks @ 2005-11-11 12:51 UTC (permalink / raw)
To: lily; +Cc: linuxppc-embedded
In-Reply-To: <331697990.25627@st.lzu.edu.cn>
lily wrote:
> hi,all:
> i build an embedded linux on powerpc 405ep,i need execute floating
> point operation,and i enable "math emulate" option in kernel, i run a
> while(1)loop ,in the loop there was floating point expression ,after
> test ,i found the while (1) exit because of the floating point
> computation,that is the floating point can't execute on ppc
> stablely.Is there someone meets the same problems ,please give me some
> suggestion,thanks!!!
The 405ep has some errata on floating point use. I didn't work on the
processor so I don't know exactly what they are, but Timesys has a
toolchain just for the 405ep to work around the problem. The other
workaround we've used is much slower and it involved disabling the
hardware floating point unit in the head.s for the 405ep and turning on
the emulator. Just turning on the emulator is insufficient. The Timesys
toolchain isn't the only one out there specifically for the 405ep.
Greg Weeks
^ permalink raw reply
* Re: floating point operation of ppc
From: Eugene Surovegin @ 2005-11-11 19:35 UTC (permalink / raw)
To: Greg Weeks; +Cc: linuxppc-embedded
In-Reply-To: <437493D3.2070301@timesys.com>
On Fri, Nov 11, 2005 at 07:51:31AM -0500, Greg Weeks wrote:
> lily wrote:
>
> > hi,all:
> > i build an embedded linux on powerpc 405ep,i need execute floating
> > point operation,and i enable "math emulate" option in kernel, i run a
> > while(1)loop ,in the loop there was floating point expression ,after
> > test ,i found the while (1) exit because of the floating point
> > computation,that is the floating point can't execute on ppc
> > stablely.Is there someone meets the same problems ,please give me some
> > suggestion,thanks!!!
>
> The 405ep has some errata on floating point use.
Are you sure it was 405ep? There is no floating point errata for this
chip due to the simple reason, it doesn't support floating point in
hardware at all.
You probably confusing it with 440ep.
I'm not aware of any 405ep specific toolchain. In fact, all what is
needed is simple 4xx specific toolchain with soft-float enabled by
default. This is how floating point is usually done for 4xx - just soft
emulation in user-space by the compiler without in-kernel emulation,
which is slow.
--
Eugene
^ permalink raw reply
* Re: floating point operation of ppc
From: Greg Weeks @ 2005-11-11 19:47 UTC (permalink / raw)
To: Eugene Surovegin; +Cc: linuxppc-embedded
In-Reply-To: <20051111193516.GA31237@gate.ebshome.net>
Eugene Surovegin wrote:
>On Fri, Nov 11, 2005 at 07:51:31AM -0500, Greg Weeks wrote:
>
>
>>lily wrote:
>>
>>
>>
>>>hi,all:
>>>i build an embedded linux on powerpc 405ep,i need execute floating
>>>point operation,and i enable "math emulate" option in kernel, i run a
>>>while(1)loop ,in the loop there was floating point expression ,after
>>>test ,i found the while (1) exit because of the floating point
>>>computation,that is the floating point can't execute on ppc
>>>stablely.Is there someone meets the same problems ,please give me some
>>>suggestion,thanks!!!
>>>
>>>
>>The 405ep has some errata on floating point use.
>>
>>
>
>Are you sure it was 405ep? There is no floating point errata for this
>chip due to the simple reason, it doesn't support floating point in
>hardware at all.
>
>You probably confusing it with 440ep.
>
>
>
You're right. It was the 440ep. The 405ep BSP is using the 4xx
toolchain. I don't think I've ever tried the math emulator on a 405.
Greg Weeks
^ permalink raw reply
* MPC8272 + kernel
From: srideep.devireddy @ 2005-11-11 20:05 UTC (permalink / raw)
To: linuxppc-embedded
[-- Attachment #1: Type: text/plain, Size: 1374 bytes --]
Hi Sir ,
With the help of mailing list I am able to get my u-boot 1.1.3
version ... right held up with kernel
....
I am able to boot my MPC8272ADS with u-boot 1.1.3 version . how can
build my uImaze(compressed linux image so that I can boot it ) ,what are
the options .. I got the source from Montavista BSP for MPC 8272 .. so
can you guide me in making a compressed image so that I can boot my
Linux kernel .
As of now I am getting and error of
## Booting image at 01000000 ...
Image Name: Linux-2.4.22
Image Type: PowerPC Linux Kernel Image (gzip compressed)
Data Size: 782360 Bytes = 764 kB
Load Address: 00000000
Entry Point: 00000000
Verifying Checksum ... OK
Uncompressing Kernel Image ... OK
That's it ... kernel hangs .. why is this ? can you kindly help me
resolving in the issue
Best Regards
Confidentiality Notice
The information contained in this electronic message and any attachments to this message are intended
for the exclusive use of the addressee(s) and may contain confidential or privileged information. If
you are not the intended recipient, please notify the sender at Wipro or Mailadmin@wipro.com immediately
and destroy all copies of this message and any attachments.
[-- Attachment #2: Type: text/html, Size: 7652 bytes --]
^ permalink raw reply
* [RFC] Attempt to clean up sigsuspend et al
From: David Woodhouse @ 2005-11-11 21:37 UTC (permalink / raw)
To: paulus; +Cc: linuxppc-dev
We currently play tricks with assembly wrappers for various syscalls, in
case a signal is delivered. We need to make sure the registers are all
saved for the benefit of ptrace, and we need to return to userspace
without clobbering r4 and r5, as the normal syscall exit path would.
As usual, these tricks differ subtly between ppc32 and ppc64 too.
This is an attempt to get rid of it all, motivated mostly because I want
to add two _new_ functions which would need such treatment (pselect,
ppoll).
We define a new thread flag TIF_RESTOREALL which is set by the syscall
which needs this treatment, and make the syscall return path honour it.
This seems to work as far as I've tested it (i.e. the system boots) for
ppc64, but I haven't even attempted to compile the 32-bit version yet.
I've also removed the explicit calls to ptrace_notify() which were in
#ifdef CONFIG_PPC64 -- if we're now stopping on the way back to
userspace, we don't need to do that any more. I'm going to need to go
through some of the sigsuspend/signal cases with a fine-tooth comb and
make sure we're still getting a ptrace stop (and only _one_ ptrace stop)
in all cases where that's appropriate.
I also tried to clean up entry_64.S so it only checks the thread flags
in the fast path _once_. This means that for ptrace we'll disable
interrupts, then check the flags and enable interrupts again before
calling do_syscall_trace_leave(). That makes the slow path match the
PPC32 code and makes the fast path smaller.
Comments?
diff --git a/arch/powerpc/kernel/entry_32.S b/arch/powerpc/kernel/entry_32.S
index 2e99ae4..59dc2e7 100644
--- a/arch/powerpc/kernel/entry_32.S
+++ b/arch/powerpc/kernel/entry_32.S
@@ -222,25 +222,25 @@ ret_from_syscall:
bl do_show_syscall_exit
#endif
mr r6,r3
- li r11,-_LAST_ERRNO
- cmplw 0,r3,r11
+ lwz r11,_CCR(r1) /* Load CR */
+ li r10,-_LAST_ERRNO
+ cmplw 0,r3,r10
rlwinm r12,r1,0,0,(31-THREAD_SHIFT) /* current_thread_info() */
blt+ 30f
- lbz r11,TI_SC_NOERR(r12)
- cmpwi r11,0
+ lbz r10,TI_SC_NOERR(r12)
+ cmpwi r10,0
bne 30f
neg r3,r3
- lwz r10,_CCR(r1) /* Set SO bit in CR */
- oris r10,r10,0x1000
- stw r10,_CCR(r1)
+ oris r11,r11,0x1000 /* Set SO bit in CR */
/* disable interrupts so current_thread_info()->flags can't change */
30: LOAD_MSR_KERNEL(r10,MSR_KERNEL) /* doesn't include MSR_EE */
SYNC
MTMSRD(r10)
lwz r9,TI_FLAGS(r12)
- andi. r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SIGPENDING|_TIF_NEED_RESCHED)
+ andi. r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SIGPENDING|_TIF_NEED_RESCHED|_TIF_RESTOREALL)
bne- syscall_exit_work
+ stw r11,_CCR(r1)
syscall_exit_cont:
#if defined(CONFIG_4xx) || defined(CONFIG_BOOKE)
/* If the process has its own DBCR0 value, load it up. The single
@@ -292,9 +292,13 @@ syscall_dotrace:
b syscall_dotrace_cont
syscall_exit_work:
+ andi. r0,r9,_TIF_RESTOREALL
+ bne- 3f
+
stw r6,RESULT(r1) /* Save result */
stw r3,GPR3(r1) /* Update return value */
- andi. r0,r9,_TIF_SYSCALL_T_OR_A
+ stw r11,_CCR(r1)
+3: andi. r0,r9,_TIF_SYSCALL_T_OR_A
beq 5f
ori r10,r10,MSR_EE
SYNC
@@ -407,22 +411,6 @@ show_syscalls_task:
* to save all the nonvolatile registers (r13 - r31) before calling
* the C code.
*/
- .globl ppc_sigsuspend
-ppc_sigsuspend:
- SAVE_NVGPRS(r1)
- lwz r0,_TRAP(r1)
- rlwinm r0,r0,0,0,30 /* clear LSB to indicate full */
- stw r0,_TRAP(r1) /* register set saved */
- b sys_sigsuspend
-
- .globl ppc_rt_sigsuspend
-ppc_rt_sigsuspend:
- SAVE_NVGPRS(r1)
- lwz r0,_TRAP(r1)
- rlwinm r0,r0,0,0,30
- stw r0,_TRAP(r1)
- b sys_rt_sigsuspend
-
.globl ppc_fork
ppc_fork:
SAVE_NVGPRS(r1)
@@ -447,14 +435,6 @@ ppc_clone:
stw r0,_TRAP(r1) /* register set saved */
b sys_clone
- .globl ppc_swapcontext
-ppc_swapcontext:
- SAVE_NVGPRS(r1)
- lwz r0,_TRAP(r1)
- rlwinm r0,r0,0,0,30 /* clear LSB to indicate full */
- stw r0,_TRAP(r1) /* register set saved */
- b sys_swapcontext
-
/*
* Top-level page fault handling.
* This is in assembler because if do_page_fault tells us that
@@ -626,16 +606,6 @@ END_FTR_SECTION_IFSET(CPU_FTR_601)
.long ret_from_except
#endif
- .globl sigreturn_exit
-sigreturn_exit:
- subi r1,r3,STACK_FRAME_OVERHEAD
- rlwinm r12,r1,0,0,(31-THREAD_SHIFT) /* current_thread_info() */
- lwz r9,TI_FLAGS(r12)
- andi. r0,r9,_TIF_SYSCALL_T_OR_A
- beq+ ret_from_except_full
- bl do_syscall_trace_leave
- /* fall through */
-
.globl ret_from_except_full
ret_from_except_full:
REST_NVGPRS(r1)
@@ -658,7 +628,7 @@ user_exc_return: /* r10 contains MSR_KE
/* Check current_thread_info()->flags */
rlwinm r9,r1,0,0,(31-THREAD_SHIFT)
lwz r9,TI_FLAGS(r9)
- andi. r0,r9,(_TIF_SIGPENDING|_TIF_NEED_RESCHED)
+ andi. r0,r9,(_TIF_SIGPENDING|_TIF_NEED_RESCHED|_TIF_RESTOREALL)
bne do_work
restore_user:
@@ -876,6 +846,8 @@ load_dbcr0:
#endif /* !(CONFIG_4xx || CONFIG_BOOKE) */
do_work: /* r10 contains MSR_KERNEL here */
+ andi. r0,r9,_TIF_RESTOREALL
+ bne- do_clear_restoreall
andi. r0,r9,_TIF_NEED_RESCHED
beq do_user_signal
@@ -890,6 +862,8 @@ recheck:
MTMSRD(r10) /* disable interrupts */
rlwinm r9,r1,0,0,(31-THREAD_SHIFT)
lwz r9,TI_FLAGS(r9)
+ andi. r0,r9,_TIF_RESTOREALL
+ bne- do_clear_restoreall
andi. r0,r9,_TIF_NEED_RESCHED
bne- do_resched
andi. r0,r9,_TIF_SIGPENDING
@@ -911,6 +885,9 @@ do_user_signal: /* r10 contains MSR_KE
REST_NVGPRS(r1)
b recheck
+do_clear_restoreall:
+ bl do_clear_tif_restoreall
+ b recheck
/*
* We come here when we are at the end of handling an exception
* that occurred at a place where taking an exception will lose
diff --git a/arch/powerpc/kernel/entry_64.S b/arch/powerpc/kernel/entry_64.S
index 2d22bf0..756663e 100644
--- a/arch/powerpc/kernel/entry_64.S
+++ b/arch/powerpc/kernel/entry_64.S
@@ -157,12 +157,6 @@ syscall_exit:
bge- syscall_error
syscall_error_cont:
- /* check for syscall tracing or audit */
- ld r9,TI_FLAGS(r12)
- andi. r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
- bne- syscall_exit_trace
-syscall_exit_trace_cont:
-
/* disable interrupts so current_thread_info()->flags can't change,
and so that we don't get interrupted after loading SRR0/1. */
ld r8,_MSR(r1)
@@ -173,8 +167,9 @@ syscall_exit_trace_cont:
rotldi r10,r10,16
mtmsrd r10,1
ld r9,TI_FLAGS(r12)
- andi. r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SIGPENDING|_TIF_NEED_RESCHED)
+ andi. r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP|_TIF_SIGPENDING|_TIF_NEED_RESCHED|_TIF_RESTOREALL)
bne- syscall_exit_work
+ std r5,_CCR(r1)
ld r7,_NIP(r1)
stdcx. r0,0,r1 /* to clear the reservation */
andi. r6,r8,MSR_PR
@@ -205,7 +200,6 @@ syscall_error:
bne- syscall_error_cont
neg r3,r3
oris r5,r5,0x1000 /* Set SO bit in CR */
- std r5,_CCR(r1)
b syscall_error_cont
/* Traced system call support */
@@ -225,8 +219,22 @@ syscall_dotrace:
ld r10,TI_FLAGS(r10)
b syscall_dotrace_cont
-syscall_exit_trace:
+syscall_exit_work:
+ /* Avoid saving r3 and ccr if TIF_RESTOREALL flag is set */
+ andi. r0,r9,_TIF_RESTOREALL
+ bne- 1f
+
std r3,GPR3(r1)
+ std r5,_CCR(r1)
+
+ /* If tracing, re-enable interrupts and do it */
+1: andi. r0,r9,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
+ beq .ret_from_except_lite
+
+ mfmsr r10
+ ori r10,r10,MSR_EE
+ mtmsrd r10,1
+
bl .save_nvgprs
addi r3,r1,STACK_FRAME_OVERHEAD
bl .do_syscall_trace_leave
@@ -234,12 +242,14 @@ syscall_exit_trace:
ld r3,GPR3(r1)
ld r5,_CCR(r1)
clrrdi r12,r1,THREAD_SHIFT
- b syscall_exit_trace_cont
-/* Stuff to do on exit from a system call. */
-syscall_exit_work:
- std r3,GPR3(r1)
- std r5,_CCR(r1)
+ /* Disable interrupts again and handle other work if any */
+ mfmsr r10
+ rldicl r10,r10,48,1
+ rotldi r10,r10,16
+ mtmsrd r10,1
+
+ ld r9,TI_FLAGS(r12)
b .ret_from_except_lite
/* Save non-volatile GPRs, if not already saved. */
@@ -260,35 +270,6 @@ _GLOBAL(save_nvgprs)
* the C code. Similarly, fork, vfork and clone need the full
* register state on the stack so that it can be copied to the child.
*/
-_GLOBAL(ppc32_sigsuspend)
- bl .save_nvgprs
- bl .compat_sys_sigsuspend
- b 70f
-
-_GLOBAL(ppc64_rt_sigsuspend)
- bl .save_nvgprs
- bl .sys_rt_sigsuspend
- b 70f
-
-_GLOBAL(ppc32_rt_sigsuspend)
- bl .save_nvgprs
- bl .compat_sys_rt_sigsuspend
-70: cmpdi 0,r3,0
- /* If it returned an error, we need to return via syscall_exit to set
- the SO bit in cr0 and potentially stop for ptrace. */
- bne syscall_exit
- /* If sigsuspend() returns zero, we are going into a signal handler. We
- may need to call audit_syscall_exit() to mark the exit from sigsuspend() */
-#ifdef CONFIG_AUDITSYSCALL
- ld r3,PACACURRENT(r13)
- ld r4,AUDITCONTEXT(r3)
- cmpdi 0,r4,0
- beq .ret_from_except /* No audit_context: Leave immediately. */
- li r4, 2 /* AUDITSC_FAILURE */
- li r5,-4 /* It's always -EINTR */
- bl .audit_syscall_exit
-#endif
- b .ret_from_except
_GLOBAL(ppc_fork)
bl .save_nvgprs
@@ -305,37 +286,6 @@ _GLOBAL(ppc_clone)
bl .sys_clone
b syscall_exit
-_GLOBAL(ppc32_swapcontext)
- bl .save_nvgprs
- bl .compat_sys_swapcontext
- b 80f
-
-_GLOBAL(ppc64_swapcontext)
- bl .save_nvgprs
- bl .sys_swapcontext
- b 80f
-
-_GLOBAL(ppc32_sigreturn)
- bl .compat_sys_sigreturn
- b 80f
-
-_GLOBAL(ppc32_rt_sigreturn)
- bl .compat_sys_rt_sigreturn
- b 80f
-
-_GLOBAL(ppc64_rt_sigreturn)
- bl .sys_rt_sigreturn
-
-80: cmpdi 0,r3,0
- blt syscall_exit
- clrrdi r4,r1,THREAD_SHIFT
- ld r4,TI_FLAGS(r4)
- andi. r4,r4,(_TIF_SYSCALL_T_OR_A|_TIF_SINGLESTEP)
- beq+ 81f
- addi r3,r1,STACK_FRAME_OVERHEAD
- bl .do_syscall_trace_leave
-81: b .ret_from_except
-
_GLOBAL(ret_from_fork)
bl .schedule_tail
REST_NVGPRS(r1)
@@ -602,7 +552,12 @@ user_work:
bl .schedule
b .ret_from_except_lite
-1: bl .save_nvgprs
+1: andi. r0,r4,_TIF_RESTOREALL
+ beq 2f
+ bl .do_clear_tif_restoreall
+ b .ret_from_except_lite
+
+2: bl .save_nvgprs
li r3,0
addi r4,r1,STACK_FRAME_OVERHEAD
bl .do_signal
diff --git a/arch/powerpc/kernel/signal_32.c b/arch/powerpc/kernel/signal_32.c
index 081d931..d18aebd 100644
--- a/arch/powerpc/kernel/signal_32.c
+++ b/arch/powerpc/kernel/signal_32.c
@@ -75,7 +75,6 @@
* registers from *regs. This is what we need
* to do when a signal has been delivered.
*/
-#define sigreturn_exit(regs) return 0
#define GP_REGS_SIZE min(sizeof(elf_gregset_t32), sizeof(struct pt_regs32))
#undef __SIGNAL_FRAMESIZE
@@ -178,8 +177,6 @@ static inline int restore_general_regs(s
#else /* CONFIG_PPC64 */
-extern void sigreturn_exit(struct pt_regs *);
-
#define GP_REGS_SIZE min(sizeof(elf_gregset_t), sizeof(struct pt_regs))
static inline int put_sigset_t(sigset_t __user *uset, sigset_t *set)
@@ -255,8 +252,10 @@ long sys_sigsuspend(old_sigset_t mask, i
while (1) {
current->state = TASK_INTERRUPTIBLE;
schedule();
- if (do_signal(&saveset, regs))
- sigreturn_exit(regs);
+ if (do_signal(&saveset, regs)) {
+ set_thread_flag(TIF_RESTOREALL);
+ return 0;
+ }
}
}
@@ -291,8 +290,10 @@ long sys_rt_sigsuspend(
while (1) {
current->state = TASK_INTERRUPTIBLE;
schedule();
- if (do_signal(&saveset, regs))
- sigreturn_exit(regs);
+ if (do_signal(&saveset, regs)) {
+ set_thread_flag(TIF_RESTOREALL);
+ return 0;
+ }
}
}
@@ -829,12 +830,6 @@ static int handle_rt_signal(unsigned lon
regs->gpr[6] = (unsigned long) rt_sf;
regs->nip = (unsigned long) ka->sa.sa_handler;
regs->trap = 0;
-#ifdef CONFIG_PPC64
- regs->result = 0;
-
- if (test_thread_flag(TIF_SINGLESTEP))
- ptrace_notify(SIGTRAP);
-#endif
return 1;
badframe:
@@ -912,8 +907,8 @@ long sys_swapcontext(struct ucontext __u
*/
if (do_setcontext(new_ctx, regs, 0))
do_exit(SIGSEGV);
- sigreturn_exit(regs);
- /* doesn't actually return back to here */
+
+ set_thread_flag(TIF_RESTOREALL);
return 0;
}
@@ -946,12 +941,11 @@ long sys_rt_sigreturn(int r3, int r4, in
* nobody does any...
*/
compat_sys_sigaltstack((u32)(u64)&rt_sf->uc.uc_stack, 0, 0, 0, 0, 0, regs);
- return (int)regs->result;
#else
do_sigaltstack(&rt_sf->uc.uc_stack, NULL, regs->gpr[1]);
- sigreturn_exit(regs); /* doesn't return here */
- return 0;
#endif
+ set_thread_flag(TIF_RESTOREALL);
+ return 0;
bad:
force_sig(SIGSEGV, current);
@@ -1042,9 +1036,7 @@ int sys_debug_setcontext(struct ucontext
*/
do_sigaltstack(&ctx->uc_stack, NULL, regs->gpr[1]);
- sigreturn_exit(regs);
- /* doesn't actually return back to here */
-
+ set_thread_flag(TIF_RESTOREALL);
out:
return 0;
}
@@ -1109,12 +1101,6 @@ static int handle_signal(unsigned long s
regs->gpr[4] = (unsigned long) sc;
regs->nip = (unsigned long) ka->sa.sa_handler;
regs->trap = 0;
-#ifdef CONFIG_PPC64
- regs->result = 0;
-
- if (test_thread_flag(TIF_SINGLESTEP))
- ptrace_notify(SIGTRAP);
-#endif
return 1;
@@ -1162,12 +1148,8 @@ long sys_sigreturn(int r3, int r4, int r
|| restore_user_regs(regs, sr, 1))
goto badframe;
-#ifdef CONFIG_PPC64
- return (int)regs->result;
-#else
- sigreturn_exit(regs); /* doesn't return */
+ set_thread_flag(TIF_RESTOREALL);
return 0;
-#endif
badframe:
force_sig(SIGSEGV, current);
diff --git a/arch/powerpc/kernel/signal_64.c b/arch/powerpc/kernel/signal_64.c
index 58194e1..7520ad4 100644
--- a/arch/powerpc/kernel/signal_64.c
+++ b/arch/powerpc/kernel/signal_64.c
@@ -96,8 +96,10 @@ long sys_rt_sigsuspend(sigset_t __user *
while (1) {
current->state = TASK_INTERRUPTIBLE;
schedule();
- if (do_signal(&saveset, regs))
+ if (do_signal(&saveset, regs)) {
+ set_thread_flag(TIF_RESTOREALL);
return 0;
+ }
}
}
@@ -343,6 +345,7 @@ int sys_swapcontext(struct ucontext __us
do_exit(SIGSEGV);
/* This returns like rt_sigreturn */
+ set_thread_flag(TIF_RESTOREALL);
return 0;
}
@@ -375,7 +378,8 @@ int sys_rt_sigreturn(unsigned long r3, u
*/
do_sigaltstack(&uc->uc_stack, NULL, regs->gpr[1]);
- return regs->result;
+ set_thread_flag(TIF_RESTOREALL);
+ return 0;
badframe:
#if DEBUG_SIG
@@ -454,9 +458,6 @@ static int setup_rt_frame(int signr, str
if (err)
goto badframe;
- if (test_thread_flag(TIF_SINGLESTEP))
- ptrace_notify(SIGTRAP);
-
return 1;
badframe:
diff --git a/arch/powerpc/kernel/syscalls.c b/arch/powerpc/kernel/syscalls.c
index f72ced1..f8ec816 100644
--- a/arch/powerpc/kernel/syscalls.c
+++ b/arch/powerpc/kernel/syscalls.c
@@ -356,3 +356,8 @@ void do_show_syscall_exit(unsigned long
{
printk(" -> %lx, current=%p cpu=%d\n", r3, current, smp_processor_id());
}
+
+void do_clear_tif_restoreall(void)
+{
+ clear_thread_flag(TIF_RESTOREALL);
+}
diff --git a/arch/powerpc/kernel/systbl.S b/arch/powerpc/kernel/systbl.S
index 65eaea9..4bb3650 100644
--- a/arch/powerpc/kernel/systbl.S
+++ b/arch/powerpc/kernel/systbl.S
@@ -113,7 +113,7 @@ SYSCALL(sgetmask)
COMPAT_SYS(ssetmask)
SYSCALL(setreuid)
SYSCALL(setregid)
-SYSX(sys_ni_syscall,ppc32_sigsuspend,ppc_sigsuspend)
+SYS32ONLY(sigsuspend)
COMPAT_SYS(sigpending)
COMPAT_SYS(sethostname)
COMPAT_SYS(setrlimit)
@@ -160,7 +160,7 @@ SYSCALL(swapoff)
COMPAT_SYS(sysinfo)
COMPAT_SYS(ipc)
SYSCALL(fsync)
-SYSX(sys_ni_syscall,ppc32_sigreturn,sys_sigreturn)
+SYS32ONLY(sigreturn)
PPC_SYS(clone)
COMPAT_SYS(setdomainname)
PPC_SYS(newuname)
@@ -213,13 +213,13 @@ COMPAT_SYS(nfsservctl)
SYSCALL(setresgid)
SYSCALL(getresgid)
COMPAT_SYS(prctl)
-SYSX(ppc64_rt_sigreturn,ppc32_rt_sigreturn,sys_rt_sigreturn)
+COMPAT_SYS(rt_sigreturn)
COMPAT_SYS(rt_sigaction)
COMPAT_SYS(rt_sigprocmask)
COMPAT_SYS(rt_sigpending)
COMPAT_SYS(rt_sigtimedwait)
COMPAT_SYS(rt_sigqueueinfo)
-SYSX(ppc64_rt_sigsuspend,ppc32_rt_sigsuspend,ppc_rt_sigsuspend)
+COMPAT_SYS(rt_sigsuspend)
COMPAT_SYS(pread64)
COMPAT_SYS(pwrite64)
SYSCALL(chown)
@@ -290,7 +290,7 @@ COMPAT_SYS(clock_settime)
COMPAT_SYS(clock_gettime)
COMPAT_SYS(clock_getres)
COMPAT_SYS(clock_nanosleep)
-SYSX(ppc64_swapcontext,ppc32_swapcontext,ppc_swapcontext)
+COMPAT_SYS(swapcontext)
COMPAT_SYS(tgkill)
COMPAT_SYS(utimes)
COMPAT_SYS(statfs64)
diff --git a/include/asm-powerpc/thread_info.h b/include/asm-powerpc/thread_info.h
index e525f49..d10ec17 100644
--- a/include/asm-powerpc/thread_info.h
+++ b/include/asm-powerpc/thread_info.h
@@ -123,6 +123,7 @@ static inline struct thread_info *curren
#define TIF_SINGLESTEP 9 /* singlestepping active */
#define TIF_MEMDIE 10
#define TIF_SECCOMP 11 /* secure computing */
+#define TIF_RESTOREALL 12 /* signal handler invoked */
/* as above, but as bit values */
#define _TIF_SYSCALL_TRACE (1<<TIF_SYSCALL_TRACE)
@@ -136,10 +137,11 @@ static inline struct thread_info *curren
#define _TIF_SYSCALL_AUDIT (1<<TIF_SYSCALL_AUDIT)
#define _TIF_SINGLESTEP (1<<TIF_SINGLESTEP)
#define _TIF_SECCOMP (1<<TIF_SECCOMP)
+#define _TIF_RESTOREALL (1<<TIF_RESTOREALL)
#define _TIF_SYSCALL_T_OR_A (_TIF_SYSCALL_TRACE|_TIF_SYSCALL_AUDIT|_TIF_SECCOMP)
#define _TIF_USER_WORK_MASK (_TIF_NOTIFY_RESUME | _TIF_SIGPENDING | \
- _TIF_NEED_RESCHED)
+ _TIF_NEED_RESCHED | _TIF_RESTOREALL)
#endif /* __KERNEL__ */
--
dwmw2
^ permalink raw reply related
* Re: 2.6.14 USB vs. sleep issues
From: Eddy Petrisor @ 2005-11-11 21:47 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc-dev list, debian-powerpc@lists.debian.org
In-Reply-To: <1131229041.5229.29.camel@gaston>
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain; charset=x-user-defined; format=flowed, Size: 1075 bytes --]
Benjamin Herrenschmidt wrote:
> On Sat, 2005-11-05 at 22:58 +0100, Bin Zhang wrote:
>
>>On 11/3/05, Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote:
>>
>>>For those who experience crashes on sleep and/or wakeup (typically due
>>>to USB) with 2.6.14, I made a test patch that might help. Please let me
>>>know if it makes things more reliable.
>>>
>>
>>I've tried your patch with usb wifi dlink dwl-g122 (my eth1). It works.
>>There are some differences in /var/log/syslog :
>
>
> I have another patch tho:
With this patch the kernel panics on my PowerBook 5,2
I made a few pictures with my phone and I can put them smewhere, if needed.
I will try tonight the third patch you proposed and the kernel is
compiling now...
BTW, is there any difference between the resulted .ko files when made with:
make SUBDIRS=drivers/usb/core
make SUBDIRS=drivers/usb/core modules
and what the make command ran from the root of the kernel source?
--
Regards,
EddyP
=============================================
"Imagination is more important than knowledge" A.Einstein
^ permalink raw reply
* Re: 2.6.14 USB vs. sleep issues
From: Benjamin Herrenschmidt @ 2005-11-11 21:52 UTC (permalink / raw)
To: Eddy Petrisor; +Cc: linuxppc-dev list, debian-powerpc@lists.debian.org
In-Reply-To: <43751182.1020009@gmail.com>
The latest patch that is candidate for upstream is:
Index: 2.6.14-ben/drivers/usb/core/hcd-pci.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/core/hcd-pci.c
+++ 2.6.14-ben/drivers/usb/core/hcd-pci.c
@@ -32,6 +32,13 @@
#include <linux/usb.h>
#include "hcd.h"
+#ifdef CONFIG_PPC_PMAC
+#include <asm/machdep.h>
+#include <asm/pmac_feature.h>
+#include <asm/pci-bridge.h>
+#include <asm/prom.h>
+#endif
+
/* PCI-based HCs are common, but plenty of non-PCI HCs are used too */
@@ -278,6 +285,18 @@ int usb_hcd_pci_suspend (struct pci_dev
break;
}
+#ifdef CONFIG_PPC_PMAC
+ if (retval == 0 && _machine == _MACH_Pmac) {
+ struct device_node *of_node;
+
+ /* Disable USB PAD & cell clock */
+ of_node = pci_device_to_OF_node (to_pci_dev(hcd->self.
+ controller));
+ if (of_node)
+ pmac_call_feature(PMAC_FTR_USB_ENABLE, of_node, 0, 0);
+ }
+#endif /* CONFIG_PPC_PMAC */
+
/* update power_state **ONLY** to make sysfs happier */
if (retval == 0)
dev->dev.power.power_state = message;
@@ -303,6 +322,18 @@ int usb_hcd_pci_resume (struct pci_dev *
return 0;
}
+#ifdef CONFIG_PPC_PMAC
+ if (_machine == _MACH_Pmac) {
+ struct device_node *of_node;
+
+ /* Re-enable USB PAD & cell clock */
+ of_node = pci_device_to_OF_node (to_pci_dev(hcd->self.
+ controller));
+ if (of_node)
+ pmac_call_feature(PMAC_FTR_USB_ENABLE, of_node, 0, 1);
+ }
+#endif /* CONFIG_PPC_PMAC */
+
/* NOTE: chip docs cover clean "real suspend" cases (what Linux
* calls "standby", "suspend to RAM", and so on). There are also
* dirty cases when swsusp fakes a suspend in "shutdown" mode.
Index: 2.6.14-ben/drivers/usb/core/hcd.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/core/hcd.c
+++ 2.6.14-ben/drivers/usb/core/hcd.c
@@ -1600,7 +1600,8 @@ irqreturn_t usb_hcd_irq (int irq, void *
struct usb_hcd *hcd = __hcd;
int start = hcd->state;
- if (start == HC_STATE_HALT)
+ if (start == HC_STATE_HALT ||
+ !test_bit(HC_FLAG_IRQ_ON, &hcd->bitflags))
return IRQ_NONE;
if (hcd->driver->irq (hcd, r) == IRQ_NONE)
return IRQ_NONE;
@@ -1736,6 +1737,9 @@ int usb_add_hcd(struct usb_hcd *hcd,
if (hcd->driver->irq) {
char buf[8], *bufp = buf;
+ set_bit(HC_FLAG_IRQ_ON, &hcd->bitflags);
+ mb();
+
#ifdef __sparc__
bufp = __irq_itoa(irqnum);
#else
Index: 2.6.14-ben/drivers/usb/core/hcd.h
===================================================================
--- 2.6.14-ben.orig/drivers/usb/core/hcd.h
+++ 2.6.14-ben/drivers/usb/core/hcd.h
@@ -71,6 +71,9 @@ struct usb_hcd { /* usb_bus.hcpriv point
/*
* hardware info/state
*/
+ unsigned long bitflags; /* various single-bit flags */
+#define HC_FLAG_IRQ_ON 0
+
const struct hc_driver *driver; /* hw-specific hooks */
unsigned saw_irq : 1;
unsigned can_wakeup:1; /* hw supports wakeup? */
Index: 2.6.14-ben/drivers/usb/host/ehci-hcd.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/host/ehci-hcd.c
+++ 2.6.14-ben/drivers/usb/host/ehci-hcd.c
@@ -750,6 +750,12 @@ static int ehci_suspend (struct usb_hcd
if (time_before (jiffies, ehci->next_statechange))
msleep (100);
+ /* Disable emission of interrupts during suspend */
+ writel(0, &ehci->regs->intr_enable);
+ mb();
+ clear_bit(HC_FLAG_IRQ_ON, &hcd->bitflags);
+ synchronize_irq(to_pci_dev(hcd->self.controller)->irq);
+
#ifdef CONFIG_USB_SUSPEND
(void) usb_suspend_device (hcd->self.root_hub, message);
#else
@@ -776,6 +782,8 @@ static int ehci_resume (struct usb_hcd *
if (time_before (jiffies, ehci->next_statechange))
msleep (100);
+ set_bit(HC_FLAG_IRQ_ON, &hcd->bitflags);
+
/* If any port is suspended (or owned by the companion),
* we know we can/must resume the HC (and mustn't reset it).
*/
Index: 2.6.14-ben/drivers/usb/host/ehci-q.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/host/ehci-q.c
+++ 2.6.14-ben/drivers/usb/host/ehci-q.c
@@ -926,6 +926,11 @@ submit_async (
#endif
spin_lock_irqsave (&ehci->lock, flags);
+ if (HC_IS_SUSPENDED(ehci_to_hcd(ehci)->state)) {
+ spin_unlock_irqrestore (&ehci->lock, flags);
+ return -ESHUTDOWN;
+ }
+
qh = qh_append_tds (ehci, urb, qtd_list, epnum, &ep->hcpriv);
/* Control/bulk operations through TTs don't need scheduling,
Index: 2.6.14-ben/drivers/usb/host/ehci-sched.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/host/ehci-sched.c
+++ 2.6.14-ben/drivers/usb/host/ehci-sched.c
@@ -602,6 +602,11 @@ static int intr_submit (
spin_lock_irqsave (&ehci->lock, flags);
+ if (HC_IS_SUSPENDED(ehci_to_hcd(ehci)->state)) {
+ spin_unlock_irqrestore (&ehci->lock, flags);
+ return -ESHUTDOWN;
+ }
+
/* get qh and force any scheduling errors */
INIT_LIST_HEAD (&empty);
qh = qh_append_tds (ehci, urb, &empty, epnum, &ep->hcpriv);
@@ -1456,6 +1461,11 @@ static int itd_submit (struct ehci_hcd *
/* schedule ... need to lock */
spin_lock_irqsave (&ehci->lock, flags);
+ if (HC_IS_SUSPENDED(ehci_to_hcd(ehci)->state)) {
+ spin_unlock_irqrestore (&ehci->lock, flags);
+ status = -ESHUTDOWN;
+ goto done;
+ }
status = iso_stream_schedule (ehci, urb, stream);
if (likely (status == 0))
itd_link_urb (ehci, urb, ehci->periodic_size << 3, stream);
@@ -1815,6 +1825,11 @@ static int sitd_submit (struct ehci_hcd
/* schedule ... need to lock */
spin_lock_irqsave (&ehci->lock, flags);
+ if (HC_IS_SUSPENDED(ehci_to_hcd(ehci)->state)) {
+ spin_unlock_irqrestore (&ehci->lock, flags);
+ status = -ESHUTDOWN;
+ goto done;
+ }
status = iso_stream_schedule (ehci, urb, stream);
if (status == 0)
sitd_link_urb (ehci, urb, ehci->periodic_size << 3, stream);
Index: 2.6.14-ben/drivers/usb/host/ohci-hcd.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/host/ohci-hcd.c
+++ 2.6.14-ben/drivers/usb/host/ohci-hcd.c
@@ -252,6 +252,10 @@ static int ohci_urb_enqueue (
spin_lock_irqsave (&ohci->lock, flags);
+ if (HC_IS_SUSPENDED(hcd->state)) {
+ retval = -ESHUTDOWN;
+ goto fail;
+ }
/* don't submit to a dead HC */
if (!HC_IS_RUNNING(hcd->state)) {
retval = -ENODEV;
Index: 2.6.14-ben/drivers/usb/host/ohci-hub.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/host/ohci-hub.c
+++ 2.6.14-ben/drivers/usb/host/ohci-hub.c
@@ -219,13 +219,6 @@ static int ohci_hub_resume (struct usb_h
/* Sometimes PCI D3 suspend trashes frame timings ... */
periodic_reinit (ohci);
- /* interrupts might have been disabled */
- ohci_writel (ohci, OHCI_INTR_INIT, &ohci->regs->intrenable);
- if (ohci->ed_rm_list)
- ohci_writel (ohci, OHCI_INTR_SF, &ohci->regs->intrenable);
- ohci_writel (ohci, ohci_readl (ohci, &ohci->regs->intrstatus),
- &ohci->regs->intrstatus);
-
/* Then re-enable operations */
ohci_writel (ohci, OHCI_USB_OPER, &ohci->regs->control);
(void) ohci_readl (ohci, &ohci->regs->control);
@@ -241,6 +234,13 @@ static int ohci_hub_resume (struct usb_h
/* TRSMRCY */
msleep (10);
+ /* interrupts might have been disabled */
+ ohci_writel (ohci, OHCI_INTR_INIT, &ohci->regs->intrenable);
+ if (ohci->ed_rm_list)
+ ohci_writel (ohci, OHCI_INTR_SF, &ohci->regs->intrenable);
+ ohci_writel (ohci, ohci_readl (ohci, &ohci->regs->intrstatus),
+ &ohci->regs->intrstatus);
+
/* keep it alive for ~5x suspend + resume costs */
ohci->next_statechange = jiffies + msecs_to_jiffies (250);
Index: 2.6.14-ben/drivers/usb/host/ohci-pci.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/host/ohci-pci.c
+++ 2.6.14-ben/drivers/usb/host/ohci-pci.c
@@ -14,13 +14,6 @@
* This file is licenced under the GPL.
*/
-#ifdef CONFIG_PPC_PMAC
-#include <asm/machdep.h>
-#include <asm/pmac_feature.h>
-#include <asm/pci-bridge.h>
-#include <asm/prom.h>
-#endif
-
#ifndef CONFIG_PCI
#error "This file is PCI bus glue. CONFIG_PCI must be defined."
#endif
@@ -118,6 +111,12 @@ static int ohci_pci_suspend (struct usb_
if (time_before (jiffies, ohci->next_statechange))
msleep (100);
+ /* Disable emission of interrupts during suspend */
+ ohci_writel(ohci, OHCI_INTR_MIE, &ohci->regs->intrdisable);
+ mb();
+ clear_bit(HC_FLAG_IRQ_ON, &hcd->bitflags);
+ synchronize_irq(to_pci_dev(hcd->self.controller)->irq);
+
#ifdef CONFIG_USB_SUSPEND
(void) usb_suspend_device (hcd->self.root_hub, message);
#else
@@ -129,16 +128,6 @@ static int ohci_pci_suspend (struct usb_
/* let things settle down a bit */
msleep (100);
-#ifdef CONFIG_PPC_PMAC
- if (_machine == _MACH_Pmac) {
- struct device_node *of_node;
-
- /* Disable USB PAD & cell clock */
- of_node = pci_device_to_OF_node (to_pci_dev(hcd->self.controller));
- if (of_node)
- pmac_call_feature(PMAC_FTR_USB_ENABLE, of_node, 0, 0);
- }
-#endif /* CONFIG_PPC_PMAC */
return 0;
}
@@ -148,20 +137,11 @@ static int ohci_pci_resume (struct usb_h
struct ohci_hcd *ohci = hcd_to_ohci (hcd);
int retval = 0;
-#ifdef CONFIG_PPC_PMAC
- if (_machine == _MACH_Pmac) {
- struct device_node *of_node;
-
- /* Re-enable USB PAD & cell clock */
- of_node = pci_device_to_OF_node (to_pci_dev(hcd->self.controller));
- if (of_node)
- pmac_call_feature (PMAC_FTR_USB_ENABLE, of_node, 0, 1);
- }
-#endif /* CONFIG_PPC_PMAC */
-
/* resume root hub */
if (time_before (jiffies, ohci->next_statechange))
msleep (100);
+ set_bit(HC_FLAG_IRQ_ON, &hcd->bitflags);
+
#ifdef CONFIG_USB_SUSPEND
/* get extra cleanup even if remote wakeup isn't in use */
retval = usb_resume_device (hcd->self.root_hub);
Index: 2.6.14-ben/drivers/usb/host/uhci-hcd.c
===================================================================
--- 2.6.14-ben.orig/drivers/usb/host/uhci-hcd.c
+++ 2.6.14-ben/drivers/usb/host/uhci-hcd.c
@@ -797,6 +797,11 @@ static int uhci_suspend(struct usb_hcd *
done:
spin_unlock_irq(&uhci->lock);
+ if (rc == 0) {
+ mb();
+ clear_bit(HC_FLAG_IRQ_ON, &hcd->bitflags);
+ synchronize_irq(hcd->irq);
+ }
return rc;
}
@@ -818,6 +823,8 @@ static int uhci_resume(struct usb_hcd *h
* system wakeup. Check it and reconfigure to avoid problems.
*/
check_and_reset_hc(uhci);
+ set_bit(HC_FLAG_IRQ_ON, &hcd->bitflags);
+ mb();
configure_hc(uhci);
#ifndef CONFIG_USB_SUSPEND
^ permalink raw reply
* [PATCH] glibc vDSO bits for ppc/ppc64 update
From: Tom Gall @ 2005-11-11 22:11 UTC (permalink / raw)
To: libc-alpha; +Cc: linuxppc-dev, linuxppc64-dev, munroesj
Greetings,
Enclosed is the next revision of the glibc vDSO implementation for
ppc/ppc64 that goes hand in hand with benh's vDSO kernel work currently
available at http://gate.crashing.org/~benh/ppc64-vdso-update.diff.
Implemented are:
__vdso_gettimeofdate
__vdso_clock_gettime
__vdso_clock_getres
__vdso_get_tbfreq
New to this patch as compared to the last posting is a rework for
symbol lookup to use _dl_vdso_weakref which as a nice side effect
when vdso symbols aren't found anoying error messages are no longer
printed.
Additional hp_timer_t is now declared as unsigned long long int
to match up with benh's get_tbfreq on 32 bit ppc. (ppc64 is already
64 bit)
Appreciate comments, questions and suggestions!
2005-11-11 Steven Munroe <sjmunroe@us.ibm.com>
Tom Gall <tom_gall@vnet.ibm.com>
* elf/rtld.c (dl_main): Initialize l_local_scope for sysinfo_map.
* sysdeps/powerpc/elf/libc-start.c: Move this.
* sysdeps/unix/sysv/linux/powerpc/libc-start.c: To here.
* sysdeps/powerpc/Versions: add __vdso_get_tbfreq symbol
* sysdeps/unix/sysv/linux/powerpc/bits/libc-vdso.h: New file.
* sysdeps/unix/sysv/linux/powerpc/clock_getres.c: New file.
* sysdeps/unix/sysv/linux/powerpc/clock_gettime.c: New file.
* sysdeps/unix/sysv/linux/powerpc/dl-vdso.c: New file.
* sysdeps/unix/sysv/linux/powerpc/dl-vdso.h: New file.
* sysdeps/unix/sysv/linux/powerpc/get_clockfreq.c: use vDSO / format
* sysdeps/unix/sysv/linux/powerpc/gettimeofday.c: New file.
* sysdeps/unix/sysv/linux/powerpc/Makefile: Add routines += dl-vdso.
* sysdeps/powerpc/powerpc32/hp-timing.h: New file.
diff -uNr libc/elf/rtld.c libc-41-32/elf/rtld.c
--- libc/elf/rtld.c 2005-11-11 11:18:09.389605816 -0500
+++ libc-41-32/elf/rtld.c 2005-11-11 09:31:48.000000000 -0500
@@ -1296,6 +1296,13 @@
elf_get_dynamic_info (l, dyn_temp);
_dl_setup_hash (l);
l->l_relocated = 1;
+ /* Initialize l_local_scope to contain just this map. This allows
+ the use of dl_lookup_symbol_x to resolve symbols within the vdso.
+ So we create a single entry list pointing to l_real as its only
+ element */
+
+ l->l_local_scope[0]->r_nlist = 1;
+ l->l_local_scope[0]->r_list = &l->l_real;
/* Now that we have the info handy, use the DSO image's soname
so this object can be looked up by name. Note that we do not
diff -uNr libc/sysdeps/powerpc/elf/libc-start.c libc-41-32/sysdeps/powerpc/elf/libc-start.c
--- libc/sysdeps/powerpc/elf/libc-start.c 2005-11-11 11:18:20.366667224 -0500
+++ libc-41-32/sysdeps/powerpc/elf/libc-start.c 1969-12-31 19:00:00.000000000 -0500
@@ -1,99 +0,0 @@
-/* Copyright (C) 1998,2000,2001,2002,2003,2004 Free Software Foundation, Inc.
- This file is part of the GNU C Library.
-
- The GNU C Library is free software; you can redistribute it and/or
- modify it under the terms of the GNU Lesser General Public
- License as published by the Free Software Foundation; either
- version 2.1 of the License, or (at your option) any later version.
-
- The GNU C Library is distributed in the hope that it will be useful,
- but WITHOUT ANY WARRANTY; without even the implied warranty of
- MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
- Lesser General Public License for more details.
-
- You should have received a copy of the GNU Lesser General Public
- License along with the GNU C Library; if not, write to the Free
- Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
- 02111-1307 USA. */
-
-#include <stdlib.h>
-#include <unistd.h>
-#include <ldsodefs.h>
-#include <bp-start.h>
-#include <bp-sym.h>
-
-extern int __cache_line_size;
-weak_extern (__cache_line_size)
-
-/* The main work is done in the generic function. */
-#define LIBC_START_MAIN generic_start_main
-#define LIBC_START_DISABLE_INLINE
-#define LIBC_START_MAIN_AUXVEC_ARG
-#define MAIN_AUXVEC_ARG
-#include <sysdeps/generic/libc-start.c>
-
-
-struct startup_info
-{
- void *__unbounded sda_base;
- int (*main) (int, char **, char **, void *);
- int (*init) (int, char **, char **, void *);
- void (*fini) (void);
-};
-
-
-int
-/* GKM FIXME: GCC: this should get __BP_ prefix by virtue of the
- BPs in the arglist of startup_info.main and startup_info.init. */
-BP_SYM (__libc_start_main) (int argc, char *__unbounded *__unbounded ubp_av,
- char *__unbounded *__unbounded ubp_ev,
- ElfW(auxv_t) *__unbounded auxvec,
- void (*rtld_fini) (void),
- struct startup_info *__unbounded stinfo,
- char *__unbounded *__unbounded stack_on_entry)
-{
-#if __BOUNDED_POINTERS__
- char **argv;
-#else
-# define argv ubp_av
-#endif
-
- /* the PPC SVR4 ABI says that the top thing on the stack will
- be a NULL pointer, so if not we assume that we're being called
- as a statically-linked program by Linux... */
- if (*stack_on_entry != NULL)
- {
- char *__unbounded *__unbounded temp;
- /* ...in which case, we have argc as the top thing on the
- stack, followed by argv (NULL-terminated), envp (likewise),
- and the auxilary vector. */
- /* 32/64-bit agnostic load from stack */
- argc = *(long int *__unbounded) stack_on_entry;
- ubp_av = stack_on_entry + 1;
- ubp_ev = ubp_av + argc + 1;
-#ifdef HAVE_AUX_VECTOR
- temp = ubp_ev;
- while (*temp != NULL)
- ++temp;
- auxvec = (ElfW(auxv_t) *)++temp;
-#endif
- rtld_fini = NULL;
- }
-
- /* Initialize the __cache_line_size variable from the aux vector. */
- for (ElfW(auxv_t) *av = auxvec; av->a_type != AT_NULL; ++av)
- switch (av->a_type)
- {
- case AT_DCACHEBSIZE:
- {
- int *cls = & __cache_line_size;
- if (cls != NULL)
- *cls = av->a_un.a_val;
- }
- break;
- }
-
- return generic_start_main (stinfo->main, argc, ubp_av, auxvec,
- stinfo->init, stinfo->fini, rtld_fini,
- stack_on_entry);
-}
diff -uNr libc/sysdeps/powerpc/Versions libc-41-32/sysdeps/powerpc/Versions
--- libc/sysdeps/powerpc/Versions 2005-11-11 11:18:20.367667072 -0500
+++ libc-41-32/sysdeps/powerpc/Versions 2005-11-11 09:31:48.000000000 -0500
@@ -13,5 +13,6 @@
GLIBC_PRIVATE {
__novmx__libc_longjmp; __novmx__libc_siglongjmp;
__vmx__libc_longjmp; __vmx__libc_siglongjmp;
+ __vdso_get_tbfreq;
}
}
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/bits/libc-vdso.h libc-41-32/sysdeps/unix/sysv/linux/powerpc/bits/libc-vdso.h
--- libc/sysdeps/unix/sysv/linux/powerpc/bits/libc-vdso.h 1969-12-31 19:00:00.000000000 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/bits/libc-vdso.h 2005-11-11 09:31:48.000000000 -0500
@@ -0,0 +1,36 @@
+/* Resolved function pointers to VDSO functions.
+ Copyright (C) 2005 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+
+#ifndef _LIBC_VDSO_H
+#define _LIBC_VDSO_H
+
+#ifdef SHARED
+
+typedef int (* __vdso_gettimeofday_t)(void *, void *);
+
+typedef int (* __vdso_clock_gettime_t)(clockid_t, struct timespec *);
+
+typedef int (* __vdso_clock_getres_t)(struct timeval*, void*);
+
+typedef hp_timing_t (* __vdso_get_tbfreq_t)(void);
+
+#endif
+
+#endif /* _LIBC_VDSO_H */
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/clock_getres.c libc-41-32/sysdeps/unix/sysv/linux/powerpc/clock_getres.c
--- libc/sysdeps/unix/sysv/linux/powerpc/clock_getres.c 1969-12-31 19:00:00.000000000 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/clock_getres.c 2005-11-11 09:31:48.000000000 -0500
@@ -0,0 +1,197 @@
+/* clock_getres -- Get the resolution of a POSIX clockid_t. Linux version.
+ Copyright (C) 2003, 2004 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <sysdep.h>
+#include <errno.h>
+#include <time.h>
+#include <sysdeps/unix/sysv/linux/kernel-posix-cpu-timers.h>
+#include <sysdeps/unix/sysv/linux/kernel-features.h>
+
+#undef __clock_getres
+#include <bits/libc-vdso.h>
+
+
+#define SYSCALL_GETRES_MONOTONIC \
+#ifdef SHARED \
+ if (__vdso_clock_getres== NULL)\
+ retval = INLINE_SYSCALL (clock_getres, 2, clock_id, res); \
+ else \
+ return (__vdso_clock_getres) (clock_id, tp); \
+ break
+
+#define SYSCALL_GETRES_REALTIME \
+ retval = INLINE_SYSCALL (clock_getres, 2, clock_id, res); \
+ break
+
+#ifdef __ASSUME_POSIX_TIMERS
+
+/* This means the REALTIME and MONOTONIC clock are definitely
+ supported in the kernel. */
+# define SYSDEP_GETRES \
+ SYSDEP_GETRES_CPUTIME \
+ case CLOCK_REALTIME: \
+ SYSCALL_GETRES_REALTIME \
+ case CLOCK_MONOTONIC: \
+ SYSCALL_GETRES_MONOTONIC
+
+# define __libc_missing_posix_timers 0
+#elif defined __NR_clock_getres
+/* Is the syscall known to exist? */
+extern int __libc_missing_posix_timers attribute_hidden;
+
+static inline int
+maybe_syscall_getres (clockid_t clock_id, struct timespec *res)
+{
+ int e = EINVAL;
+
+ if (!__libc_missing_posix_timers)
+ {
+ INTERNAL_SYSCALL_DECL (err);
+ int r = INTERNAL_SYSCALL (clock_getres, err, 2, clock_id, res);
+ if (!INTERNAL_SYSCALL_ERROR_P (r, err))
+ return 0;
+
+ e = INTERNAL_SYSCALL_ERRNO (r, err);
+ if (e == ENOSYS)
+ {
+ __libc_missing_posix_timers = 1;
+ e = EINVAL;
+ }
+ }
+
+ return e;
+}
+
+/* The REALTIME and MONOTONIC clock might be available. Try the
+ syscall first. */
+# define SYSDEP_GETRES \
+ SYSDEP_GETRES_CPUTIME \
+ case CLOCK_REALTIME: \
+ case CLOCK_MONOTONIC: \
+ retval = maybe_syscall_getres (clock_id, res); \
+ if (retval == 0) \
+ break; \
+ /* Fallback code. */ \
+ if (retval == EINVAL && clock_id == CLOCK_REALTIME) \
+ retval = realtime_getres (res); \
+ else \
+ { \
+ __set_errno (retval); \
+ retval = -1; \
+ } \
+ break;
+#endif
+
+#ifdef __NR_clock_getres
+/* We handled the REALTIME clock here. */
+# define HANDLED_REALTIME 1
+# define HANDLED_CPUTIME 1
+
+# if __ASSUME_POSIX_CPU_TIMERS > 0
+
+# define SYSDEP_GETRES_CPU SYSCALL_GETRES
+# define SYSDEP_GETRES_CPUTIME /* Default catches them too. */
+
+# else
+
+extern int __libc_missing_posix_cpu_timers attribute_hidden;
+
+static int
+maybe_syscall_getres_cpu (clockid_t clock_id, struct timespec *res)
+{
+ int e = EINVAL;
+
+ if (!__libc_missing_posix_cpu_timers)
+ {
+ INTERNAL_SYSCALL_DECL (err);
+ int r = INTERNAL_SYSCALL (clock_getres, err, 2, clock_id, res);
+ if (!INTERNAL_SYSCALL_ERROR_P (r, err))
+ return 0;
+
+ e = INTERNAL_SYSCALL_ERRNO (r, err);
+# ifndef __ASSUME_POSIX_TIMERS
+ if (e == ENOSYS)
+ {
+ __libc_missing_posix_timers = 1;
+ __libc_missing_posix_cpu_timers = 1;
+ e = EINVAL;
+ }
+ else
+# endif
+ {
+ if (e == EINVAL)
+ {
+ /* Check whether the kernel supports CPU clocks at all.
+ If not, record it for the future. */
+ r = INTERNAL_SYSCALL (clock_getres, err, 2,
+ MAKE_PROCESS_CPUCLOCK (0, CPUCLOCK_SCHED),
+ NULL);
+ if (INTERNAL_SYSCALL_ERROR_P (r, err))
+ __libc_missing_posix_cpu_timers = 1;
+ }
+ }
+ }
+
+ return e;
+}
+
+# define SYSDEP_GETRES_CPU \
+ retval = maybe_syscall_getres_cpu (clock_id, res); \
+ if (retval == 0) \
+ break; \
+ if (retval != EINVAL || !__libc_missing_posix_cpu_timers) \
+ { \
+ __set_errno (retval); \
+ retval = -1; \
+ break; \
+ } \
+ retval = -1 /* Otherwise continue on to the HP_TIMING version. */;
+
+static inline int
+maybe_syscall_getres_cputime (clockid_t clock_id, struct timespec *res)
+{
+ return maybe_syscall_getres_cpu
+ (clock_id == CLOCK_THREAD_CPUTIME_ID
+ ? MAKE_THREAD_CPUCLOCK (0, CPUCLOCK_SCHED)
+ : MAKE_PROCESS_CPUCLOCK (0, CPUCLOCK_SCHED),
+ res);
+}
+
+# define SYSDEP_GETRES_CPUTIME \
+ case CLOCK_PROCESS_CPUTIME_ID: \
+ case CLOCK_THREAD_CPUTIME_ID: \
+ retval = maybe_syscall_getres_cputime (clock_id, res); \
+ if (retval == 0) \
+ break; \
+ if (retval != EINVAL || !__libc_missing_posix_cpu_timers) \
+ { \
+ __set_errno (retval); \
+ retval = -1; \
+ break; \
+ } \
+ retval = hp_timing_getres (res); \
+ break;
+# if !HP_TIMING_AVAIL
+# define hp_timing_getres(res) (__set_errno (EINVAL), -1)
+# endif
+
+# endif
+#endif
+
+#include <sysdeps/posix/clock_getres.c>
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/clock_gettime.c libc-41-32/sysdeps/unix/sysv/linux/powerpc/clock_gettime.c
--- libc/sysdeps/unix/sysv/linux/powerpc/clock_gettime.c 1969-12-31 19:00:00.000000000 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/clock_gettime.c 2005-11-11 09:31:48.000000000 -0500
@@ -0,0 +1,196 @@
+/* clock_gettime -- Get current time from a POSIX clockid_t. Linux version.
+ Copyright (C) 2003, 2004 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <sysdep.h>
+#include <errno.h>
+#include <time.h>
+#include <sysdeps/unix/sysv/linux/kernel-posix-cpu-timers.h>
+#include <sysdeps/unix/sysv/linux/kernel-features.h>
+
+#undef __clock_gettime
+#include <bits/libc-vdso.h>
+
+#define SYSCALL_GETTIME_CLOCK_MONOTONIC \
+#ifdef SHARED \
+ if (__vdso_clock_gettime== NULL) \
+ retval = INLINE_SYSCALL (clock_gettime, 2, clock_id, tp); \
+ else \
+ return (__vdso_clock_gettime) (clock_id, tp); \
+ break
+
+#define SYSCALL_GETTIME_CLOCK_REALTIME \
+ retval = INLINE_SYSCALL (clock_gettime, 2, clock_id, tp); \
+ break
+
+#ifdef __ASSUME_POSIX_TIMERS
+
+/* This means the REALTIME and MONOTONIC clock are definitely
+ supported in the kernel. */
+# define SYSDEP_GETTIME \
+ SYSDEP_GETTIME_CPUTIME \
+ case CLOCK_REALTIME: \
+ SYSCALL_GETTIME_CLOCK_REALTIME \
+ case CLOCK_MONOTONIC: \
+ SYSCALL_GETTIME_CLOCK_MONOTONIC
+
+# define __libc_missing_posix_timers 0
+#elif defined __NR_clock_gettime
+/* Is the syscall known to exist? */
+int __libc_missing_posix_timers attribute_hidden;
+
+static inline int
+maybe_syscall_gettime (clockid_t clock_id, struct timespec *tp)
+{
+ int e = EINVAL;
+
+ if (!__libc_missing_posix_timers)
+ {
+ INTERNAL_SYSCALL_DECL (err);
+ int r = INTERNAL_SYSCALL (clock_gettime, err, 2, clock_id, tp);
+ if (!INTERNAL_SYSCALL_ERROR_P (r, err))
+ return 0;
+
+ e = INTERNAL_SYSCALL_ERRNO (r, err);
+ if (e == ENOSYS)
+ {
+ __libc_missing_posix_timers = 1;
+ e = EINVAL;
+ }
+ }
+
+ return e;
+}
+
+/* The REALTIME and MONOTONIC clock might be available. Try the
+ syscall first. */
+# define SYSDEP_GETTIME \
+ SYSDEP_GETTIME_CPUTIME \
+ case CLOCK_REALTIME: \
+ case CLOCK_MONOTONIC: \
+ retval = maybe_syscall_gettime (clock_id, tp); \
+ if (retval == 0) \
+ break; \
+ /* Fallback code. */ \
+ if (retval == EINVAL && clock_id == CLOCK_REALTIME) \
+ retval = realtime_gettime (tp); \
+ else \
+ { \
+ __set_errno (retval); \
+ retval = -1; \
+ } \
+ break;
+#endif
+
+#ifdef __NR_clock_gettime
+/* We handled the REALTIME clock here. */
+# define HANDLED_REALTIME 1
+# define HANDLED_CPUTIME 1
+
+# if __ASSUME_POSIX_CPU_TIMERS > 0
+
+# define SYSDEP_GETTIME_CPU SYSCALL_GETTIME
+# define SYSDEP_GETTIME_CPUTIME /* Default catches them too. */
+
+# else
+
+int __libc_missing_posix_cpu_timers attribute_hidden;
+
+static int
+maybe_syscall_gettime_cpu (clockid_t clock_id, struct timespec *tp)
+{
+ int e = EINVAL;
+
+ if (!__libc_missing_posix_cpu_timers)
+ {
+ INTERNAL_SYSCALL_DECL (err);
+ int r = INTERNAL_SYSCALL (clock_gettime, err, 2, clock_id, tp);
+ if (!INTERNAL_SYSCALL_ERROR_P (r, err))
+ return 0;
+
+ e = INTERNAL_SYSCALL_ERRNO (r, err);
+# ifndef __ASSUME_POSIX_TIMERS
+ if (e == ENOSYS)
+ {
+ __libc_missing_posix_timers = 1;
+ __libc_missing_posix_cpu_timers = 1;
+ e = EINVAL;
+ }
+ else
+# endif
+ {
+ if (e == EINVAL)
+ {
+ /* Check whether the kernel supports CPU clocks at all.
+ If not, record it for the future. */
+ r = INTERNAL_SYSCALL (clock_getres, err, 2,
+ MAKE_PROCESS_CPUCLOCK (0, CPUCLOCK_SCHED),
+ NULL);
+ if (INTERNAL_SYSCALL_ERROR_P (r, err))
+ __libc_missing_posix_cpu_timers = 1;
+ }
+ }
+ }
+
+ return e;
+}
+
+# define SYSDEP_GETTIME_CPU \
+ retval = maybe_syscall_gettime_cpu (clock_id, tp); \
+ if (retval == 0) \
+ break; \
+ if (retval != EINVAL || !__libc_missing_posix_cpu_timers) \
+ { \
+ __set_errno (retval); \
+ retval = -1; \
+ break; \
+ } \
+ retval = -1 /* Otherwise continue on to the HP_TIMING version. */;
+
+static inline int
+maybe_syscall_gettime_cputime (clockid_t clock_id, struct timespec *tp)
+{
+ return maybe_syscall_gettime_cpu
+ (clock_id == CLOCK_THREAD_CPUTIME_ID
+ ? MAKE_THREAD_CPUCLOCK (0, CPUCLOCK_SCHED)
+ : MAKE_PROCESS_CPUCLOCK (0, CPUCLOCK_SCHED),
+ tp);
+}
+
+# define SYSDEP_GETTIME_CPUTIME \
+ case CLOCK_PROCESS_CPUTIME_ID: \
+ case CLOCK_THREAD_CPUTIME_ID: \
+ retval = maybe_syscall_gettime_cputime (clock_id, tp); \
+ if (retval == 0) \
+ break; \
+ if (retval != EINVAL || !__libc_missing_posix_cpu_timers) \
+ { \
+ __set_errno (retval); \
+ retval = -1; \
+ break; \
+ } \
+ retval = hp_timing_gettime (clock_id, tp); \
+ break;
+# if !HP_TIMING_AVAIL
+# define hp_timing_gettime(clock_id, tp) (__set_errno (EINVAL), -1)
+# endif
+
+# endif
+#endif
+
+#include <sysdeps/unix/clock_gettime.c>
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/dl-vdso.c libc-41-32/sysdeps/unix/sysv/linux/powerpc/dl-vdso.c
--- libc/sysdeps/unix/sysv/linux/powerpc/dl-vdso.c 1969-12-31 19:00:00.000000000 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/dl-vdso.c 2005-11-11 11:09:33.054571056 -0500
@@ -0,0 +1,128 @@
+/* VDSO symbol handling in the ELF dynamic linker.
+ Copyright (C) 2005 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include "config.h"
+#include <dl-hash.h>
+#include <ldsodefs.h>
+
+/* This function attempts to resolve a symbol in the VDSO map but
+ simply returns NULL if it is not there. This is intended for the
+ syscall replacement case where the runtime can fail-safe back to
+ the syscall. For example if we are running on an older kernel
+ that does not implement the specific function yet. */
+
+void *internal_function
+_dl_vdso_weakref (const char *name)
+{
+ const ElfW (Sym) * ref = NULL;
+ struct link_map *map = GLRO (dl_sysinfo_map);
+ void *value = NULL;
+ lookup_t result;
+
+ if (map != NULL)
+ {
+ /* This is a tricky because by default _dl_lookup_symbol_x will fail
+ and exit if the symbol is not found. Specifying a "skip_map" is
+ one case where _dl_loopup_symbol_x will do what we want, but the
+ l_local_scope in the dl_sysinfo_map will not to give the desired
+ result. So we create a dummy scope and link-map struct to be
+ the "skip_map". As the dummy.l_real points back to the real VDSO
+ map, do_lookup_x can find the symbol if it is there. */
+ struct r_scope_elem *symbol_scope[2];
+ struct link_map *scope_list[2];
+ struct r_scope_elem dummy_scope;
+ struct link_map dummy;
+
+ symbol_scope[0] = &dummy_scope;
+ symbol_scope[1] = NULL;
+ symbol_scope[0]->r_nlist = 2;
+ symbol_scope[0]->r_list = &scope_list[0];
+ scope_list[0] = map->l_local_scope[0]->r_list[0];
+ scope_list[1] = &dummy;
+ dummy.l_real = map->l_real;
+
+ /* Search the scope of the given vdso map. */
+ result = GLRO (dl_lookup_symbol_x) (name, map, &ref,
+ symbol_scope, NULL, 0, 0, &dummy);
+
+ if (ref != NULL)
+ {
+ value = DL_SYMBOL_ADDRESS (result, ref);
+ }
+ }
+
+ return value;
+}
+
+
+void *internal_function
+_dl_vdso_sym (const char *name)
+{
+ const ElfW (Sym) * ref = NULL;
+ struct link_map *map = GLRO (dl_sysinfo_map);
+ void *value = NULL;
+ lookup_t result;
+
+ if (map != NULL)
+ {
+ /* Search the scope of the given vdso map. */
+ result = GLRO (dl_lookup_symbol_x) (name, map, &ref,
+ map->l_local_scope,
+ NULL, 0, 0, NULL);
+
+ if (ref != NULL)
+ {
+ value = DL_SYMBOL_ADDRESS (result, ref);
+ }
+ }
+
+ return value;
+}
+
+void *internal_function
+_dl_vdso_vsym (const char *name, const char *version)
+{
+ const ElfW (Sym) * ref = NULL;
+ struct link_map *map = GLRO (dl_sysinfo_map);
+ void *value = NULL;
+ struct r_found_version vers;
+ lookup_t result;
+
+ if (map != NULL)
+ {
+ /* Compute hash value to the version string. */
+ vers.name = version;
+ vers.hidden = 1;
+ vers.hash = _dl_elf_hash (version);
+ /* We don't have a specific file where the symbol can be found. */
+ vers.filename = NULL;
+
+ /* Search the scope of the vdso map. */
+ result = GLRO (dl_lookup_symbol_x) (name, map, &ref,
+ map->l_local_scope,
+ &vers, 0, 0, NULL);
+
+ if (ref != NULL)
+ {
+ value = DL_SYMBOL_ADDRESS (result, ref);
+
+ }
+ }
+ return value;
+}
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/dl-vdso.h libc-41-32/sysdeps/unix/sysv/linux/powerpc/dl-vdso.h
--- libc/sysdeps/unix/sysv/linux/powerpc/dl-vdso.h 1969-12-31 19:00:00.000000000 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/dl-vdso.h 2005-11-11 09:31:48.000000000 -0500
@@ -0,0 +1,33 @@
+/* Run-time dynamic linker data structures for loaded ELF shared objects.
+ Copyright (C) 2005 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#ifndef _DL_VDSO_H
+#define _DL_VDSO_H
+
+/* Functions for resolving symbols in the VDSO link map. */
+ extern void *_dl_vdso_weakref (const char *name)
+ internal_function attribute_hidden;
+
+ extern void *_dl_vdso_sym (const char *name)
+ internal_function attribute_hidden;
+
+ extern void *_dl_vdso_vsym (const char *name, const char *version)
+ internal_function attribute_hidden;
+
+#endif /* ldsodefs.h */
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/get_clockfreq.c libc-41-32/sysdeps/unix/sysv/linux/powerpc/get_clockfreq.c
--- libc/sysdeps/unix/sysv/linux/powerpc/get_clockfreq.c 2005-11-11 11:18:22.804563648 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/get_clockfreq.c 2005-11-11 09:31:48.000000000 -0500
@@ -22,7 +22,11 @@
#include <string.h>
#include <unistd.h>
#include <libc-internal.h>
+#include <bits/libc-vdso.h>
+#ifdef SHARED
+extern __vdso_get_tbfreq_t __vdso_get_tbfreq;
+#endif
hp_timing_t
__get_clockfreq (void)
@@ -38,68 +42,78 @@
if (timebase_freq != 0)
return timebase_freq;
- int fd = open ("/proc/cpuinfo", O_RDONLY);
- if (__builtin_expect (fd != -1, 1))
+ /* if we can use the vDSO to obtain the timebase even better */
+#ifdef SHARED
+ if (__vdso_get_tbfreq != NULL)
{
- /* The timebase will be in the 1st 1024 bytes for systems with up
- to 8 processors. If the first read returns less then 1024
- bytes read, we have the whole cpuinfo and can start the scan.
- Otherwise we will have to read more to insure we have the
- timebase value in the scan. */
- char buf[1024];
- ssize_t n;
-
- n = read (fd, buf, sizeof (buf));
- if (n == sizeof (buf))
- {
- /* We are here because the 1st read returned exactly sizeof
- (buf) bytes. This implies that we are not at EOF and may
- not have read the timebase value yet. So we need to read
- more bytes until we know we have EOF. We copy the lower
- half of buf to the upper half and read sizeof (buf)/2
- bytes into the lower half of buf and repeat until we
- reach EOF. We can assume that the timebase will be in
- the last 512 bytes of cpuinfo, so two 512 byte half_bufs
- will be sufficient to contain the timebase and will
- handle the case where the timebase spans the half_buf
- boundry. */
- const ssize_t half_buf = sizeof (buf) / 2;
- while (n >= half_buf)
- {
- memcpy (buf, buf + half_buf, half_buf);
- n = read (fd, buf + half_buf, half_buf);
- }
- if (n >= 0)
- n += half_buf;
- }
-
- if (__builtin_expect (n, 1) > 0)
- {
- char *mhz = memmem (buf, n, "timebase", 7);
-
- if (__builtin_expect (mhz != NULL, 1))
- {
- char *endp = buf + n;
-
- /* Search for the beginning of the string. */
- while (mhz < endp && (*mhz < '0' || *mhz > '9') && *mhz != '\n')
- ++mhz;
-
- while (mhz < endp && *mhz != '\n')
- {
- if (*mhz >= '0' && *mhz <= '9')
- {
- result *= 10;
- result += *mhz - '0';
- }
+ timebase_freq=(__vdso_get_tbfreq)();
+ }
+ else
+#endif
+ {
+ int fd = open ("/proc/cpuinfo", O_RDONLY);
+ if (__builtin_expect (fd != -1, 1))
+ {
+ /* The timebase will be in the 1st 1024 bytes for systems with up
+ to 8 processors. If the first read returns less then 1024
+ bytes read, we have the whole cpuinfo and can start the scan.
+ Otherwise we will have to read more to insure we have the
+ timebase value in the scan. */
+ char buf[1024];
+ ssize_t n;
+
+ n = read (fd, buf, sizeof (buf));
+ if (n == sizeof (buf))
+ {
+ /* We are here because the 1st read returned exactly sizeof
+ (buf) bytes. This implies that we are not at EOF and may
+ not have read the timebase value yet. So we need to read
+ more bytes until we know we have EOF. We copy the lower
+ half of buf to the upper half and read sizeof (buf)/2
+ bytes into the lower half of buf and repeat until we
+ reach EOF. We can assume that the timebase will be in
+ the last 512 bytes of cpuinfo, so two 512 byte half_bufs
+ will be sufficient to contain the timebase and will
+ handle the case where the timebase spans the half_buf
+ boundry. */
+ const ssize_t half_buf = sizeof (buf) / 2;
+ while (n >= half_buf)
+ {
+ memcpy (buf, buf + half_buf, half_buf);
+ n = read (fd, buf + half_buf, half_buf);
+ }
+ if (n >= 0)
+ n += half_buf;
+ }
+
+ if (__builtin_expect (n, 1) > 0)
+ {
+ char *mhz = memmem (buf, n, "timebase", 7);
+
+ if (__builtin_expect (mhz != NULL, 1))
+ {
+ char *endp = buf + n;
+
+ /* Search for the beginning of the string. */
+ while (mhz < endp && (*mhz < '0' || *mhz > '9') && *mhz != '\n')
++mhz;
- }
- }
- timebase_freq = result;
- }
- close (fd);
- }
+ while (mhz < endp && *mhz != '\n')
+ {
+ if (*mhz >= '0' && *mhz <= '9')
+ {
+ result *= 10;
+ result += *mhz - '0';
+ }
+
+ ++mhz;
+ }
+ }
+ timebase_freq = result;
+ }
+ close (fd);
+ }
+ }
return timebase_freq;
}
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/gettimeofday.c libc-41-32/sysdeps/unix/sysv/linux/powerpc/gettimeofday.c
--- libc/sysdeps/unix/sysv/linux/powerpc/gettimeofday.c 1969-12-31 19:00:00.000000000 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/gettimeofday.c 2005-11-11 09:31:48.000000000 -0500
@@ -0,0 +1,54 @@
+/* Copyright (C) 2005 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <sysdep.h>
+#include <bp-checks.h>
+#include <stddef.h>
+#include <sys/time.h>
+#include <time.h>
+#include <hp-timing.h>
+
+#undef __gettimeofday
+#include <bits/libc-vdso.h>
+
+/* Get the current time of day and timezone information,
+ putting it into *TV and *TZ. If TZ is NULL, *TZ is not filled.
+ Returns 0 on success, -1 on errors. */
+
+#ifdef SHARED
+extern __vdso_gettimeofday_t __vdso_gettimeofday;
+#endif
+
+int
+__gettimeofday (tv, tz)
+ struct timeval *tv;
+ struct timezone *tz;
+{
+#ifdef SHARED
+ if (__vdso_gettimeofday == NULL)
+#endif
+ return INLINE_SYSCALL (gettimeofday, 2, CHECK_1 (tv), CHECK_1 (tz));
+#ifdef SHARED
+ else
+ {
+ return (*__vdso_gettimeofday) (tv, tz);
+ }
+#endif
+}
+
+INTDEF (__gettimeofday) weak_alias (__gettimeofday, gettimeofday)
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/libc-start.c libc-41-32/sysdeps/unix/sysv/linux/powerpc/libc-start.c
--- libc/sysdeps/unix/sysv/linux/powerpc/libc-start.c 1969-12-31 19:00:00.000000000 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/libc-start.c 2005-11-11 09:31:48.000000000 -0500
@@ -0,0 +1,201 @@
+/* Copyright (C) 1998,2000,2001,2002,2003,2004,2005 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#include <stdlib.h>
+#include <unistd.h>
+#include <ldsodefs.h>
+#include <bp-start.h>
+#include <bp-sym.h>
+
+extern int __cache_line_size;
+weak_extern (__cache_line_size)
+/* The main work is done in the generic function. */
+#define LIBC_START_MAIN generic_start_main
+#define LIBC_START_DISABLE_INLINE
+#define LIBC_START_MAIN_AUXVEC_ARG
+#define MAIN_AUXVEC_ARG
+#define INIT_MAIN_ARGS
+#include <sysdeps/generic/libc-start.c>
+ struct startup_info
+ {
+ void *__unbounded sda_base;
+ int (*main) (int, char **, char **, void *);
+ int (*init) (int, char **, char **, void *);
+ void (*fini) (void);
+ };
+
+
+#ifdef SHARED
+#include <sys/time.h>
+#include <dl-vdso.h>
+#undef __gettimeofday
+#undef __clock_gettime
+#undef __clock_getres
+#include <bits/libc-vdso.h>
+
+ __vdso_gettimeofday_t __vdso_gettimeofday;
+ __vdso_clock_gettime_t __vdso_clock_gettime;
+ __vdso_clock_getres_t __vdso_clock_getres;
+ __vdso_get_tbfreq_t __vdso_get_tbfreq;
+
+#if __WORDSIZE == 64
+ typedef struct
+ {
+ void *func;
+ void *toc_ptr;
+ void *extra;
+ } _ppc_func_desciptor;
+
+ static struct
+ {
+ _ppc_func_desciptor fd_gettimeofday;
+ _ppc_func_desciptor fd_clock_gettime;
+ _ppc_func_desciptor fd_clock_getres;
+ _ppc_func_desciptor fd_get_tbfreq;
+ } __vdso_descriptor;
+#endif
+
+ static inline void _libc_vdso_platform_setup (void)
+{
+ void *vdso_ref = NULL;
+
+ __vdso_gettimeofday = NULL;
+ __vdso_clock_gettime = NULL;
+ __vdso_clock_getres = NULL;
+ __vdso_get_tbfreq = NULL;
+
+ vdso_ref = _dl_vdso_weakref ("__kernel_gettimeofday");
+
+#if __WORDSIZE == 64
+ if (vdso_ref != NULL)
+ {
+ __vdso_descriptor.fd_gettimeofday.func = vdso_ref;
+ __vdso_descriptor.fd_gettimeofday.toc_ptr = NULL;
+ __vdso_descriptor.fd_gettimeofday.extra = NULL;
+ __vdso_gettimeofday = (__vdso_gettimeofday_t)
+ & __vdso_descriptor.fd_gettimeofday;
+ }
+#else
+ __vdso_gettimeofday = (__vdso_gettimeofday_t) vdso_ref;
+#endif
+
+ vdso_ref = _dl_vdso_weakref ("__kernel_clock_gettime");
+
+#if __WORDSIZE == 64
+ if (vdso_ref != NULL)
+ {
+ __vdso_descriptor.fd_clock_gettime.func = vdso_ref;
+ __vdso_descriptor.fd_clock_gettime.toc_ptr = NULL;
+ __vdso_descriptor.fd_clock_gettime.extra = NULL;
+ __vdso_clock_gettime = (__vdso_clock_gettime_t)
+ & __vdso_descriptor.fd_clock_gettime;
+ }
+#else
+ __vdso_clock_gettime = (__vdso_clock_gettime_t) vdso_ref;
+#endif
+
+ vdso_ref = _dl_vdso_weakref ("__kernel_clock_getres");
+
+#if __WORDSIZE == 64
+ if (vdso_ref != NULL)
+ {
+ __vdso_descriptor.fd_clock_getres.func = vdso_ref;
+ __vdso_descriptor.fd_clock_getres.toc_ptr = NULL;
+ __vdso_descriptor.fd_clock_getres.extra = NULL;
+ __vdso_clock_getres = (__vdso_clock_getres_t)
+ & __vdso_descriptor.fd_clock_getres;
+ }
+#else
+ __vdso_clock_getres = (__vdso_clock_getres_t) vdso_ref;
+#endif
+
+ vdso_ref = _dl_vdso_weakref ("__kernel_vdso_get_tbfreq");
+
+#if __WORDSIZE == 64
+ if (vdso_ref != NULL)
+ {
+ __vdso_descriptor.fd_clock_getres.func = vdso_ref;
+ __vdso_descriptor.fd_clock_getres.toc_ptr = NULL;
+ __vdso_descriptor.fd_clock_getres.extra = NULL;
+ __vdso_get_tbfreq = (__vdso_get_tbfreq_t)
+ & __vdso_descriptor.fd_get_tbfreq;
+ }
+#else
+ __vdso_get_tbfreq = (__vdso_get_tbfreq_t) vdso_ref;
+#endif
+}
+#endif
+
+int
+/* GKM FIXME: GCC: this should get __BP_ prefix by virtue of the
+ BPs in the arglist of startup_info.main and startup_info.init. */
+ BP_SYM (__libc_start_main) (int argc, char *__unbounded * __unbounded ubp_av,
+ char *__unbounded * __unbounded ubp_ev,
+ ElfW (auxv_t) * __unbounded auxvec,
+ void (*rtld_fini) (void),
+ struct startup_info * __unbounded stinfo,
+ char *__unbounded * __unbounded stack_on_entry)
+{
+#if __BOUNDED_POINTERS__
+ char **argv;
+#else
+# define argv ubp_av
+#endif
+
+ /* the PPC SVR4 ABI says that the top thing on the stack will
+ be a NULL pointer, so if not we assume that we're being called
+ as a statically-linked program by Linux... */
+ if (*stack_on_entry != NULL)
+ {
+ char *__unbounded * __unbounded temp;
+ /* ...in which case, we have argc as the top thing on the
+ stack, followed by argv (NULL-terminated), envp (likewise),
+ and the auxilary vector. */
+ /* 32/64-bit agnostic load from stack */
+ argc = *(long int *__unbounded) stack_on_entry;
+ ubp_av = stack_on_entry + 1;
+ ubp_ev = ubp_av + argc + 1;
+#ifdef HAVE_AUX_VECTOR
+ temp = ubp_ev;
+ while (*temp != NULL)
+ ++temp;
+ auxvec = (ElfW (auxv_t) *)++ temp;
+#endif
+ rtld_fini = NULL;
+ }
+
+ /* Initialize the __cache_line_size variable from the aux vector. */
+ for (ElfW (auxv_t) * av = auxvec; av->a_type != AT_NULL; ++av)
+ switch (av->a_type)
+ {
+ case AT_DCACHEBSIZE:
+ {
+ int *cls = &__cache_line_size;
+ if (cls != NULL)
+ *cls = av->a_un.a_val;
+ }
+ break;
+ }
+#ifdef SHARED
+ /* Resolve and initialize function pointers for VDSO functions. */
+ _libc_vdso_platform_setup ();
+#endif
+ return generic_start_main (stinfo->main, argc, ubp_av, auxvec,
+ stinfo->init, stinfo->fini, rtld_fini,
+ stack_on_entry);
+}
diff -uNr libc/sysdeps/unix/sysv/linux/powerpc/Makefile libc-41-32/sysdeps/unix/sysv/linux/powerpc/Makefile
--- libc/sysdeps/unix/sysv/linux/powerpc/Makefile 2005-11-11 11:18:22.776567904 -0500
+++ libc-41-32/sysdeps/unix/sysv/linux/powerpc/Makefile 2005-11-11 09:31:48.000000000 -0500
@@ -2,3 +2,8 @@
ifeq ($(subdir),rt)
librt-routines += rt-sysdep
endif
+
+ifeq ($(subdir),misc)
+routines += dl-vdso
+endif
+
--- /dev/null 2005-08-19 11:20:54.270914304 -0400
+++ libc-34-32/sysdeps/powerpc/powerpc32/hp-timing.h 2005-11-11 15:44:16.817619528 -0500
@@ -0,0 +1,83 @@
+/* High precision, low overhead timing functions. Generic version.
+ Copyright (C) 1998, 2000 Free Software Foundation, Inc.
+ This file is part of the GNU C Library.
+ Contributed by Ulrich Drepper <drepper@cygnus.com>, 1998.
+
+ The GNU C Library is free software; you can redistribute it and/or
+ modify it under the terms of the GNU Lesser General Public
+ License as published by the Free Software Foundation; either
+ version 2.1 of the License, or (at your option) any later version.
+
+ The GNU C Library is distributed in the hope that it will be useful,
+ but WITHOUT ANY WARRANTY; without even the implied warranty of
+ MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+ Lesser General Public License for more details.
+
+ You should have received a copy of the GNU Lesser General Public
+ License along with the GNU C Library; if not, write to the Free
+ Software Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
+ 02111-1307 USA. */
+
+#ifndef _HP_TIMING_H
+#define _HP_TIMING_H 1
+
+
+/* There are no generic definitions for the times. We could write something
+ using the `gettimeofday' system call where available but the overhead of
+ the system call might be too high.
+
+ In case a platform supports timers in the hardware the following macros
+ and types must be defined:
+
+ - HP_TIMING_AVAIL: test for availability.
+
+ - HP_TIMING_INLINE: this macro is non-zero if the functionality is not
+ implemented using function calls but instead uses some inlined code
+ which might simply consist of a few assembler instructions. We have to
+ know this since we might want to use the macros here in places where we
+ cannot make function calls.
+
+ - hp_timing_t: This is the type for variables used to store the time
+ values.
+
+ - HP_TIMING_ZERO: clear `hp_timing_t' object.
+
+ - HP_TIMING_NOW: place timestamp for current time in variable given as
+ parameter.
+
+ - HP_TIMING_DIFF_INIT: do whatever is necessary to be able to use the
+ HP_TIMING_DIFF macro.
+
+ - HP_TIMING_DIFF: compute difference between two times and store it
+ in a third. Source and destination might overlap.
+
+ - HP_TIMING_ACCUM: add time difference to another variable. This might
+ be a bit more complicated to implement for some platforms as the
+ operation should be thread-safe and 64bit arithmetic on 32bit platforms
+ is not.
+
+ - HP_TIMING_ACCUM_NT: this is the variant for situations where we know
+ there are no threads involved.
+
+ - HP_TIMING_PRINT: write decimal representation of the timing value into
+ the given string. This operation need not be inline even though
+ HP_TIMING_INLINE is specified.
+
+*/
+
+/* Provide dummy definitions. */
+#define HP_TIMING_AVAIL (0)
+#define HP_TIMING_INLINE (0)
+typedef unsigned long long int hp_timing_t;
+#define HP_TIMING_ZERO(Var)
+#define HP_TIMING_NOW(var)
+#define HP_TIMING_DIFF_INIT()
+#define HP_TIMING_DIFF(Diff, Start, End)
+#define HP_TIMING_ACCUM(Sum, Diff)
+#define HP_TIMING_ACCUM_NT(Sum, Diff)
+#define HP_TIMING_PRINT(Buf, Len, Val)
+
+/* Since this implementation is not available we tell the user about it. */
+#define HP_TIMING_NONAVAIL 1
+
+#endif /* hp-timing.h */
^ permalink raw reply
* Re: [PATCH] glibc vDSO bits for ppc/ppc64 update
From: Benjamin Herrenschmidt @ 2005-11-11 22:13 UTC (permalink / raw)
To: Tom Gall; +Cc: linuxppc64-dev, munroesj, libc-alpha, linuxppc-dev
In-Reply-To: <Pine.LNX.4.64.0511111555580.6838@triton>
> Additional hp_timer_t is now declared as unsigned long long int
> to match up with benh's get_tbfreq on 32 bit ppc. (ppc64 is already
> 64 bit)
I'm not sure having hp_timer_t be 64 bits on 32 bits libraries is the
right thing to do. I suppose it's a matter of performances tradeoff. The
fact that the vDSO __kernel_get_tbfreq() returns a long long doesn't
force you to use a long long hp_timer_t, the vDSO doesn't have to return
a value of type hp_timer_t :)
I did that on the vDSO to avoid having to change the vDSO interface if
we ever have to deal with >4Ghz time sources.
Ben.
^ permalink raw reply
* Re: [PATCH] glibc vDSO bits for ppc/ppc64 update
From: Steve Munroe @ 2005-11-11 22:32 UTC (permalink / raw)
To: Benjamin Herrenschmidt
Cc: linuxppc-dev, Tom Gall, munroesj, libc-alpha, linuxppc64-dev
In-Reply-To: <1131747209.24637.255.camel@gaston>
Benjamin Herrenschmidt <benh@kernel.crashing.org> wrote on 11/11/2005
04:13:28 PM:
>
> > Additional hp_timer_t is now declared as unsigned long long int
> > to match up with benh's get_tbfreq on 32 bit ppc. (ppc64 is already
> > 64 bit)
>
> I'm not sure having hp_timer_t be 64 bits on 32 bits libraries is the
> right thing to do. I suppose it's a matter of performances tradeoff. The
> fact that the vDSO __kernel_get_tbfreq() returns a long long doesn't
> force you to use a long long hp_timer_t, the vDSO doesn't have to return
> a value of type hp_timer_t :)
>
> I did that on the vDSO to avoid having to change the vDSO interface if
> we ever have to deal with >4Ghz time sources.
>
A 64-bit timebase is required to the use hp-timing for clock_gettime(
CLOCK_PROCESS_CPUTIME_ID, ) and clock_gettime(CLOCK_THREAD_CPUTIME_ID, ).
So 64-bit it is.
Steven J. Munroe
Linux on Power Toolchain Architect
IBM Corporation, Linux Technology Center
^ permalink raw reply
* Re: [RFC] Attempt to clean up sigsuspend et al
From: Paul Mackerras @ 2005-11-12 0:28 UTC (permalink / raw)
To: David Woodhouse; +Cc: linuxppc-dev
In-Reply-To: <1131745047.21212.194.camel@hades.cambridge.redhat.com>
David Woodhouse writes:
> This is an attempt to get rid of it all, motivated mostly because I want
> to add two _new_ functions which would need such treatment (pselect,
> ppoll).
>
> We define a new thread flag TIF_RESTOREALL which is set by the syscall
> which needs this treatment, and make the syscall return path honour it.
> This seems to work as far as I've tested it (i.e. the system boots) for
> ppc64, but I haven't even attempted to compile the 32-bit version yet.
Ummm, I don't see where you save the non-volatile GPRs to the
exception frame _before_ do_signal (called from within *_sigsuspend)
copies them to the user stack...
Paul.
^ permalink raw reply
* Re: [RFC] Attempt to clean up sigsuspend et al
From: David Woodhouse @ 2005-11-12 0:48 UTC (permalink / raw)
To: Paul Mackerras; +Cc: linuxppc-dev
In-Reply-To: <17269.14127.936147.778781@cargo.ozlabs.ibm.com>
On Sat, 2005-11-12 at 11:28 +1100, Paul Mackerras wrote:
> Ummm, I don't see where you save the non-volatile GPRs to the
> exception frame _before_ do_signal (called from within *_sigsuspend)
> copies them to the user stack...
Hm, yeah.
I suppose I could make a TIF_xxx flag make entry.S do that. It's not
massively pretty, but then it wasn't pretty beforehand either. At least
I'd have made the ppc32 and ppc64 code match, removed the assembly
wrappers for a bunch of syscalls, and shortened the syscall exit fast
path on ppc64.
I'd probably still be removing more lines of code than I add, but I'm
not entirely convinced by the whole plan. What do you think?
Failing that, I'll just switch them both to do either what ppc32
currently does with sigreturn_exit() or what ppc64 does with returning
zero to take the ret_from_except path. Probably the former, since
pselect() and ppoll() will sometimes want to return zero without
invoking a signal handler. And we can probably shorten the ppc64
syscall exit path anyway, either way.
--
dwmw2
^ permalink raw reply
* Re: 2.6.14 USB vs. sleep issues
From: Eddy Petrisor @ 2005-11-12 8:53 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc-dev list, debian-powerpc@lists.debian.org
In-Reply-To: <1131745946.24637.251.camel@gaston>
Benjamin Herrenschmidt wrote:
> The latest patch that is candidate for upstream is:
>
> Index: 2.6.14-ben/drivers/usb/core/hcd-pci.c
This version also panics :(
--
Regards,
EddyP
=============================================
"Imagination is more important than knowledge" A.Einstein
^ permalink raw reply
page: next (older) | prev (newer) | latest
- recent:[subjects (threaded)|topics (new)|topics (active)]
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox