* Re: [PATCH 0/2] [PPC 4xx] L2-cache synchronization for ppc44x
From: Eugene Surovegin @ 2008-01-11 17:41 UTC (permalink / raw)
To: Yuri Tikhonov; +Cc: linuxppc-dev, sr, dzu
In-Reply-To: <200801111824.46920.yur@emcraft.com>
On Fri, Jan 11, 2008 at 06:24:46PM +0300, Yuri Tikhonov wrote:
>
> Hello, Eugene,
>
> The h/w snooping mechanism you are talking about is limited to the Low
> Latency (LL) segment of the PLB bus in ppc440sp and ppc440spe chips (see
> section "7.2.7 L2 Cache Coherency" of the ppc440spe spec), whereas DMA and
> XOR engines use the High Bandwidth (HB) segment of PLB bus (see
> section "1.1.2 Internal Buses" of the ppc440spe spec).
>
> Thus, the h/w snooping mechanism is not able to trace the results of
> operations performed by DMA and XOR engines and keep L2-cache coherent with
> SDRAM, because the data flow through the HB PLB segment. This leads to, for
> example, incorrect results of RAID-parity calculations if one uses the h/w
> accelerated ppc440spe ADMA driver with L2-cache enabled.
>
> The s/w synchronization algorithms proposed in my patches has no LL PLB
> limitations as opposed to h/w snooping, but, probably, this is not the best
> way of how it might be implemented. Even though with these patches the h/w
> accelerated RAID starts to operate correctly (with L2-cache enabled) there is
> a performance degradation (induced by loops in the L2-cache synchronization
> routines) observed in the most cases. So, as a result, there is no benefit
> from using L2-cache for these, RAID, cases at all.
Thanks a lot for explanation, Yuri. I'd never imagine they were so
stupid to make new chips with such behaviour.
--
Eugene
^ permalink raw reply
* Re: [PATCH] ibm_newemac: Increase number of default rx-/tx-buffers
From: Eugene Surovegin @ 2008-01-11 17:48 UTC (permalink / raw)
To: Stefan Roese; +Cc: linuxppc-dev, netdev
In-Reply-To: <200801051338.17957.sr@denx.de>
On Sat, Jan 05, 2008 at 01:38:17PM +0100, Stefan Roese wrote:
> On Saturday 05 January 2008, Benjamin Herrenschmidt wrote:
> > On Sat, 2008-01-05 at 10:50 +0100, Stefan Roese wrote:
> > > Performance tests done by AMCC have shown that 256 buffer increase the
> > > performance of the Linux EMAC driver. So let's update the default
> > > values to match this setup.
> > >
> > > Signed-off-by: Stefan Roese <sr@denx.de>
> > > ---
> >
> > Do we have the numbers ? Did they also measure latency ?
>
> I hoped this question would not come. ;) No, unfortunately I don't have any
> numbers. Just the recommendation from AMCC to always use 256 buffers.
This cannot be true for all chips. Default numbers I selected weren't
random. In particular, 256 for Tx doesn't make a lot of sense for 405.
You just gonna waste memory.
I'd be quite reluctant to follow such advices from AMCC without actual
details.
--
Eugene
^ permalink raw reply
* Re: 2.6.22-ppc8xx fec.c bugs
From: Scott Wood @ 2008-01-11 17:27 UTC (permalink / raw)
To: raul.moreno; +Cc: linuxppc-embedded
In-Reply-To: <OF431AD1B5.8B0529FD-ONC12573CD.002EB1C5-C12573CD.0030B37E@abengoa.com>
raul.moreno@telvent.abengoa.com wrote:
> I didn't know ppc was deprecated. Why is it so?
Having separate 32 and 64 bit powerpc trees was a maintenance headache.
The way arch/ppc is structured was a maintenance headache.
> I have been using this architecture since I started with Linux (not too
> long ago) because I manage a mpc866 processor.
mpc866 chips are supported in arch/powerpc.
> Do you mean that I should port to arch/powerpc?
Yes.
-Scott
^ permalink raw reply
* Re: Help with device tree binding for SMC serial
From: Scott Wood @ 2008-01-11 17:24 UTC (permalink / raw)
To: Rune Torgersen; +Cc: linuxppc-dev
In-Reply-To: <DCEAAC0833DD314AB0B58112AD99B93B03D9A092@ismail.innsys.innovsys.com>
Rune Torgersen wrote:
> Ok, needed a valid console= line on the command line. WHen we tried
> that, we had a typo, so it was not recognized.
> Our old 2.6.18 arch/ppc kernel didn't need a console parameter, it got
> the baudrate from u-boot somehow.
> Anyway of doing that here too?
You could add something to the cuboot code to fill in current-speed
based on the value in the bd_t.
-Scott
^ permalink raw reply
* Re: [PATCH 1/8] pseries: phyp dump: Docmentation
From: Linas Vepstas @ 2008-01-11 16:57 UTC (permalink / raw)
To: Nathan Lynch; +Cc: linuxppc-dev, lkessler, mahuja, Olof Johansson, strosake
In-Reply-To: <20080111012641.GX14201@localdomain>
On 10/01/2008, Nathan Lynch <ntl@pobox.com> wrote:
> Mike Strosaker wrote:
> >
> > At the risk of repeating what others have already said, the PHYP-assistance
> > method provides some advantages that the kexec method cannot:
> > - Availability of the system for production use before the dump data is
> > collected. As was mentioned before, some production systems may choose not
> > to operate with the limited memory initially available after the reboot,
> > but it sure is nice to provide the option.
>
> I'm more concerned that this design encourages the user to resume a
> workload *which is almost certainly known to result in a system crash*
> before collection of crash data is complete. Maybe the gamble will
> pay off most of the time, but I wouldn't want to be working support
> when it doesn't.
Workloads that cause crashes within hours of startup tend to be
weeded-out/discovered during pre-production test of the system
to be deployed. Since its pre-production test, dumps can be
taken in a leisurely manner. Heck, even a session at the
xmon prompt can be contemplated.
The problem is when the crash only reproduces after days or
weeks of uptime, on a production machine. Since the machine
is in production, its got to be brought back up ASAP. Since
its crashing only after days/weeks, the dump should have
plenty of time to complete. (And if it crashes quickly after
that reboot ... well, support people always welcome ways
in which a bug can be reproduced more quickly/easily).
--linas
^ permalink raw reply
* Re: PCI Failed to allocate mem for PCI ROM
From: Jiri Slaby @ 2008-01-11 16:52 UTC (permalink / raw)
To: Kumar Gala; +Cc: Greg KH, linux-pci, LKML, linuxppc-dev list
In-Reply-To: <FB6DCE73-5D8A-42AC-AE5C-0C1AD095B50B@kernel.crashing.org>
Kumar Gala napsal(a):
> I saw that patch, but if you notice that its just x86 specific and I'm
> having the issue on a powerpc 32-bit system.
My bad, sorry.
^ permalink raw reply
* RFC: Only deal with apple fix if res is memory
From: Kumar Gala @ 2008-01-11 16:38 UTC (permalink / raw)
To: Benjamin Herrenschmidt; +Cc: linuxppc-dev
Ben,
Will this impact the apple fix? I'm getting some _IO cases hitting this.
- k
diff --git a/arch/powerpc/kernel/pci-common.c b/arch/powerpc/kernel/pci-common.c
index d394d41..42a9dab 100644
--- a/arch/powerpc/kernel/pci-common.c
+++ b/arch/powerpc/kernel/pci-common.c
@@ -807,6 +807,7 @@ static void __devinit __pcibios_fixup_bus(struct pci_bus *bus)
* their size is smaller than 1M.
*/
if (res->start == hose->pci_mem_offset &&
+ res->flags & IORESOURCE_MEM &&
res->end < 0x100000) {
printk(KERN_INFO
"PCI: Closing bogus Apple Firmware"
^ permalink raw reply related
* Re: [PATCH 1/5] Warp Base Platform
From: Sean MacLennan @ 2008-01-11 16:26 UTC (permalink / raw)
To: Stephen Rothwell; +Cc: linuxppc-dev
In-Reply-To: <20080111210237.5636d0c3.sfr@canb.auug.org.au>
Stephen Rothwell wrote:
> I don't where that function is actually defined - I assume it is in one
> of the other recent patch sets.
>
> /me searches ...
> /me reads "ad7414 driver" email ...
> /me notes spaces missing there as well :-)
>
>
I didn't write the ad7414 driver and wanted to change the original code
as little as possible.
> Tricky. You have something that can only be built in (pika_dtm_thread in
> warp.c) calling something that might be a module ... I think you need to
> think of a new way of organizing these pieces.
>
The DTM is a very small, but important part of the taco. Basically it
controls the fan. I could have wrapped it in a CONFIG_AD7414, but I
would rather have a compile time error and ship people the ad7414.c.
That said, I could just slam a CONFIG_AD7414 around the DTM code. It
would just mean that the fan would run at high speed all the time if you
didn't have it.
I really don't want to have to write a driver just to get one register
from the temperature chip. And we want the DTM running ASAP on startup.
I think forcing taco users to build in the i2c and ad7414 drivers is not
a hardship. And we do ship a working kernel with the taco.
And as a special bonus to readers of this thread, an exclusive, never
before seen, picture of tigger, my development taco. I have to point out
it is a development prototype and not the final product.
ftp://ftp.seanm.ca/stuff/tigger.jpg or
http://ftp.seanm.ca/stuff/tigger.jpg .
Cheers,
Sean
^ permalink raw reply
* Re: [PATCH 0/2] [PPC 4xx] L2-cache synchronization for ppc44x
From: Yuri Tikhonov @ 2008-01-11 15:24 UTC (permalink / raw)
To: Eugene Surovegin; +Cc: linuxppc-dev, sr, dzu
In-Reply-To: <20071128195037.GB22325@gate.ebshome.net>
Hello, Eugene,
The h/w snooping mechanism you are talking about is limited to the Low
Latency (LL) segment of the PLB bus in ppc440sp and ppc440spe chips (see
section "7.2.7 L2 Cache Coherency" of the ppc440spe spec), whereas DMA and
XOR engines use the High Bandwidth (HB) segment of PLB bus (see
section "1.1.2 Internal Buses" of the ppc440spe spec).
Thus, the h/w snooping mechanism is not able to trace the results of
operations performed by DMA and XOR engines and keep L2-cache coherent with
SDRAM, because the data flow through the HB PLB segment. This leads to, for
example, incorrect results of RAID-parity calculations if one uses the h/w
accelerated ppc440spe ADMA driver with L2-cache enabled.
The s/w synchronization algorithms proposed in my patches has no LL PLB
limitations as opposed to h/w snooping, but, probably, this is not the best
way of how it might be implemented. Even though with these patches the h/w
accelerated RAID starts to operate correctly (with L2-cache enabled) there is
a performance degradation (induced by loops in the L2-cache synchronization
routines) observed in the most cases. So, as a result, there is no benefit
from using L2-cache for these, RAID, cases at all.
Regards, Yuri
On Wednesday 28 November 2007 22:50, Eugene Surovegin wrote:
> On Wed, Nov 07, 2007 at 01:40:10AM +0300, Yuri Tikhonov wrote:
> >
> > Hello all,
> >
> > Here is a patch-set for support L2-cache synchronization routines for
> > the ppc44x processors family. I know that the "ppc" branch is for
bug-fixing only, thus
> > the patch-set is just FYI [though enabled but non-coherent L2-cache may
appear as a bug for
> > someone who uses one of the boards listed below :)].
> >
> > [PATCH 1/2] [PPC 4xx] invalidate_l2cache_range() implementation for
ppc44x;
> > [PATCH 2/2] [PPC 44x] enable L2-cache for the following ppc44x-based
boards: ALPR,
> > Katmai, Ocotea, and Taishan.
>
> Why is this all needed?
>
> IIRC ibm440gx_l2c_enable() configures 64G snoop region for L2C.
>
> Did AMCC made non-only-coherent L2C chips recently?
>
> --
> Eugene
>
>
--
Yuri Tikhonov, Senior Software Engineer
Emcraft Systems, www.emcraft.com
^ permalink raw reply
* Re: [i2c] [PATCH 0/5] Version 17, series to add device tree naming to i2c
From: Jochen Friedrich @ 2008-01-11 16:05 UTC (permalink / raw)
To: Jon Smirl; +Cc: Jean Delvare, linuxppc-dev, i2c, linux-kernel
In-Reply-To: <9e4733910801110752k57f1fd7crd5f143900fc64f0b@mail.gmail.com>
Hi Jon,
>>> The following series implements standard linux module aliasing for i2c modules on arch=powerpc. It then converts the mpc i2c driver from being a platform driver to an open firmware one. I2C device names are picked up from the device tree. Module aliasing is used to translate from device tree names into to linux kernel names. Several i2c drivers are updated to use the new aliasing.
>> Now that I have read all the previous versions of this patch series
>> and, more importantly, all objections that were raised on the way, I
>> can start reviewing the latest iteration of your patches. I'll also do
>> some testing, although I have no powerpc stuff here, but at least I
>> want to make sure that there are no regressions introduced by your
>> patches on x86.
> Various people were worried about x86. Around version 15 I altered the
> patches so that they only impacted PowerPC. If they impact x86 in
> current form that is a bug.
I can only second this. The latest version of i2c-cpm
(http://patchwork.ozlabs.org/linuxppc/patch?person=1023&id=15902) makes
use of this patch, as well. On the dbox2, loading and unloading of
modules in any order just works fine.
Thanks,
Jochen
^ permalink raw reply
* Re: [PATCH for 2.6.24][NET] fs_enet: check for phydev existence in the ethtool handlers
From: Sergej Stepanov @ 2008-01-11 16:01 UTC (permalink / raw)
To: hs; +Cc: Scott Wood, linuxppc-dev, Jeff Garzik
In-Reply-To: <478666C6.2030104@denx.de>
Am Donnerstag, den 10.01.2008, 19:41 +0100 schrieb Heiko Schocher:
> Hello Scott,
>
> Scott Wood wrote:
> > Heiko Schocher wrote:
> >> diff --git a/drivers/net/fs_enet/fs_enet-main.c
> >> b/drivers/net/fs_enet/fs_enet-main.c
> >> index f2a4d39..f432a18 100644
> >> --- a/drivers/net/fs_enet/fs_enet-main.c
> >> +++ b/drivers/net/fs_enet/fs_enet-main.c
> >> @@ -810,6 +810,10 @@ static int fs_enet_open(struct net_device *dev)
> >> if (fep->fpi->use_napi)
> >> napi_enable(&fep->napi);
> >>
> >> + /* to initialize the fep->cur_rx,... */
> >> + /* not doing this, will cause a crash in fs_enet_rx_napi */
> >> + fs_init_bds(fep->ndev);
> >
> > We should do this just before napi_enable() rather than just after, to
> > eliminate any chance of a race window.
>
> Yes, you are right.
> Here is the new patch:
>
It works fine.
Sergej.
^ permalink raw reply
* Re: [i2c] [PATCH 0/5] Version 17, series to add device tree naming to i2c
From: Jon Smirl @ 2008-01-11 15:52 UTC (permalink / raw)
To: Jean Delvare; +Cc: linuxppc-dev, i2c, linux-kernel
In-Reply-To: <20080111095638.775670ed@hyperion.delvare>
On 1/11/08, Jean Delvare <khali@linux-fr.org> wrote:
> Hi Jon,
>
> On Wed, 19 Dec 2007 23:41:36 -0500, Jon Smirl wrote:
> > Since copying i2c-mpc.c to maintain support for the ppc architecture seems to be an issue; instead rework i2c-mpc.c to use CONFIG_PPC_MERGE #ifdefs to support both the ppc and powerpc architecture. When ppc is deleted in six months these #ifdefs will need to be removed.
> >
> > Another rework of the i2c for powerpc device tree patch. This version implements standard alias naming only on the powerpc platform and only for the device tree names. The old naming mechanism of i2c_client.name,driver_name is left in place and not changed for non-powerpc platforms. This patch is fully capable of dynamically loading the i2c modules. You can modprobe in the i2c-mpc driver and the i2c modules described in the device tree will be automatically loaded. Modules also work if compiled in.
> >
> > The follow on patch to module-init-tools is also needed since the i2c subsystem has never implemented dynamic loading.
> >
> > The following series implements standard linux module aliasing for i2c modules on arch=powerpc. It then converts the mpc i2c driver from being a platform driver to an open firmware one. I2C device names are picked up from the device tree. Module aliasing is used to translate from device tree names into to linux kernel names. Several i2c drivers are updated to use the new aliasing.
>
> Now that I have read all the previous versions of this patch series
> and, more importantly, all objections that were raised on the way, I
> can start reviewing the latest iteration of your patches. I'll also do
> some testing, although I have no powerpc stuff here, but at least I
> want to make sure that there are no regressions introduced by your
> patches on x86.
Various people were worried about x86. Around version 15 I altered the
patches so that they only impacted PowerPC. If they impact x86 in
current form that is a bug.
When x86 is ready for it I do think dynamic module loading should be
implemented there also.
>
> --
> Jean Delvare
>
--
Jon Smirl
jonsmirl@gmail.com
^ permalink raw reply
* Re: PCI Failed to allocate mem for PCI ROM
From: Kumar Gala @ 2008-01-11 15:41 UTC (permalink / raw)
To: Jiri Slaby; +Cc: Greg KH, linux-pci, LKML, linuxppc-dev list
In-Reply-To: <47873333.5020202@gmail.com>
On Jan 11, 2008, at 3:13 AM, Jiri Slaby wrote:
> On 01/11/2008 10:07 AM, Kumar Gala wrote:
>> On Jan 11, 2008, at 2:41 AM, Jiri Slaby wrote:
>>> On 01/11/2008 09:29 AM, Kumar Gala wrote:
>>>> Greg,
>>>> I'm getting the following message from the kernel on an embedded
>>>> ppc32 system:
>>>> PCI: Failed to allocate mem resource #9:100000@e0000000 for
>>>> 0000:00:00.0
>>>> The HW setup is a PCIe host controller and an e1000 NIC card. It
>>>> appears that pci_bus_assign_resources() is trying to call
>>>> pci_assign_resource() for the ROM and the resource for the ROM is
>>>> [100000:1fffff] where the PHB is [c0000000:dfffffff].
>>>> It seems like the resno that pci_assign_resource is getting
>>>> called with is wrong and thus pci_update_resource() doesn't get
>>>> called.
>>>> any ideas?
>>>
>>> Kernel version, please.
>> Sorry, its 2.6.24-rc7 + some ppc patches queued for 2.6.25
>
> Could you try this patch?
> http://git.kernel.org/?p=linux/kernel/git/gregkh/patches.git;a=blob_plain;f=pci/pci-remove-default-pci-expansion-rom-memory-allocation.patch
>
> Greg: is this 2.6.25 material, please? We need this for SP2.
I saw that patch, but if you notice that its just x86 specific and I'm
having the issue on a powerpc 32-bit system.
- k
^ permalink raw reply
* Re: [alsa-devel] [PATCH v2] [ALSA] Add ASoC drivers for the Freescale MPC8610 SoC
From: Timur Tabi @ 2008-01-11 15:08 UTC (permalink / raw)
To: Olof Johansson; +Cc: linuxppc-dev, alsa-devel, david
In-Reply-To: <20080111002450.GA13291@lixom.net>
Olof Johansson wrote:
> Having been in a similar situation myself (needing to share resources
> between DMA, ethernet and function offload), I recommend creating a
> separate small library that all those drivers use, instead of making
> some sort of dependency between drivers in completely different parts
> of the kernel.
Well, the DMA driver should be in soon. Actually, it should be in now, because
I saw a blurb in this month's Linux Journal about it. As soon as I find it :-),
I'll post a new patch that adds arbitration.
--
Timur Tabi
Linux kernel developer at Freescale
^ permalink raw reply
* Re: PPC vs POWERPC
From: Grant Likely @ 2008-01-11 14:52 UTC (permalink / raw)
To: samppa@sundmangroup.com; +Cc: linuxppc-embedded
In-Reply-To: <23145.147.11.3.128.1200054808.squirrel@www.sundmangroup.com>
On 1/11/08, samppa@sundmangroup.com <samppa@sundmangroup.com> wrote:
> Hi,
> I want to port a board ( WRS SBC8265 ) from the PPC branch to the POWERPC
> branch in the Linux Kernel -- do you have any good starting points that
> describes what I need to pay attention to?
>
> I've already made a port of U-boot-1.3.1 and enabled CONFIG_OF_LIBFDT.
>
> So I need to create a DTS file, compile it with the DTC compiler and port
> the current PPC branch to the POWERPC branch to my understanding.
Yes, you're exactly right. Use one of the dts files in
arch/powerpc/boot/dts as a starting point and read
Documentation/powerpc/booting-without-of.txt to find out what is
expected in the device tree.
Cheers,
g.
--
Grant Likely, B.Sc., P.Eng.
Secret Lab Technologies Ltd.
^ permalink raw reply
* Re: libfdt: Add fdt_set_name() function
From: Jon Loeliger @ 2008-01-11 14:38 UTC (permalink / raw)
To: Josh Boyer; +Cc: linuxppc-dev, David Gibson
In-Reply-To: <20080111083003.38b556d6@zod.rchland.ibm.com>
So, like, the other day Josh Boyer mumbled:
> On Fri, 11 Jan 2008 07:44:33 -0600
> Jon Loeliger <jdl@jdl.com> wrote:
>
> > So, like, the other day David Gibson mumbled:
> > > This patch adds an fdt_set_name() function to libfdt, mirroring
> > > fdt_get_name(). This is a r/w function which alters the name of a
> > > given device tree node.
> > >
> > > Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
> >
> > Applied.
>
> Awesome. Now if we could just get that into the kernel and usable in
> the wrappers... :)
I've pushed it out now.
I guess we'll try to get the real v1.1.0 into the kernel then?
Maybe at 2.6.25?
jdl
^ permalink raw reply
* Re: libfdt: Add fdt_set_name() function
From: Josh Boyer @ 2008-01-11 14:30 UTC (permalink / raw)
To: Jon Loeliger; +Cc: linuxppc-dev, David Gibson
In-Reply-To: <E1JDKBV-0007Hq-Dm@jdl.com>
On Fri, 11 Jan 2008 07:44:33 -0600
Jon Loeliger <jdl@jdl.com> wrote:
> So, like, the other day David Gibson mumbled:
> > This patch adds an fdt_set_name() function to libfdt, mirroring
> > fdt_get_name(). This is a r/w function which alters the name of a
> > given device tree node.
> >
> > Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
>
> Applied.
Awesome. Now if we could just get that into the kernel and usable in
the wrappers... :)
josh
^ permalink raw reply
* RE: Help with device tree binding for SMC serial
From: Rune Torgersen @ 2008-01-11 14:13 UTC (permalink / raw)
To: Scott Wood; +Cc: linuxppc-dev
In-Reply-To: <DCEAAC0833DD314AB0B58112AD99B93B03D9A042@ismail.innsys.innovsys.com>
> From: Rune Torgersen
> Finally got it (sort-of) working.
> Turned out that for some reason the console init is setting=20
> the baudrate
> to 9600
> the options string passed in to the console init fuunction is NULL.
>=20
> Any idea oon how this should be passed in from u-boot?
Ok, needed a valid console=3D line on the command line. WHen we tried
that, we had a typo, so it was not recognized.
Our old 2.6.18 arch/ppc kernel didn't need a console parameter, it got
the baudrate from u-boot somehow.
Anyway of doing that here too?
^ permalink raw reply
* Re: libfdt: Add fdt_set_name() function
From: Jon Loeliger @ 2008-01-11 13:44 UTC (permalink / raw)
To: David Gibson; +Cc: linuxppc-dev
In-Reply-To: <20080111035505.GB25055@localhost.localdomain>
So, like, the other day David Gibson mumbled:
> This patch adds an fdt_set_name() function to libfdt, mirroring
> fdt_get_name(). This is a r/w function which alters the name of a
> given device tree node.
>
> Signed-off-by: David Gibson <david@gibson.dropbear.id.au>
Applied.
jdl
^ permalink raw reply
* [PATCH 2/2] ps3fb: fix deadlock on kexec()
From: Geert Uytterhoeven @ 2008-01-11 13:28 UTC (permalink / raw)
To: Linus Torvalds, Andrew Morton
Cc: Linux Frame Buffer Device Development, Linux Kernel Development,
Linux/PPC Development, Jeremy Kerr,
Cell Broadband Engine OSS Development
In-Reply-To: <Pine.LNX.4.64.0801111417450.12038@vixen.sonytel.be>
[-- Attachment #1: Type: TEXT/PLAIN, Size: 3570 bytes --]
From: Jeremy Kerr <jk@ozlabs.org>
Since the introduction of the acquire_console_sem calls in
0333d83509c7d8496c8965b5ba9bc0c98e83c259, kexecing can cause the
kernel to deadlock:
ps3fb_shutdown()
-> unregister_framebuffer()
-> fb_notifier_call_chain(FB_EVENT_FB_UNBIND)
-> fbcon_fb_unbind()
-> unbind_con_driver()
-> bind_con_driver()
[ acquires console_sem ]
-> fbcon_deinit()
-> fbops->fb_release(newinfo, 0)
-> ps3fb_release()
-> ps3fb_sync()
[ acquires console_sem ]
This change avoids the deadlock by moving the acquire_console_sem()
out of ps3fb_sync(), and puts it into the two other callsites, leaving
ps3fb_release() to call ps3fb_sync() without the console semaphore.
[Geert]
- Corrected call sequence above
- ps3fb_release() may be called with and without console_sem held. This is an
inconsistency that should be fixed at the fb level, but for now, try to
acquire console_sem in ps3fb_release().
I think it's safer to let ps3fb_release() try to acquire console_sem and
not refresh the screen if it fails, than to call ps3fb_sync() without
holding console_sem, as ps3fb_par may be modified at the same time, causing
crashes or lockups.
Besides, ps3fb_release() only calls ps3fb_sync() to refresh the screen
when display flipping is disabled, which is an uncommon case (except during
shutdown/kexec).
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
---
drivers/video/ps3fb.c | 12 ++++++++----
1 file changed, 8 insertions(+), 4 deletions(-)
--- a/drivers/video/ps3fb.c
+++ b/drivers/video/ps3fb.c
@@ -443,8 +443,6 @@ static int ps3fb_sync(struct fb_info *in
u32 ddr_line_length, xdr_line_length;
u64 ddr_base, xdr_base;
- acquire_console_sem();
-
if (frame > par->num_frames - 1) {
dev_dbg(info->device, "%s: invalid frame number (%u)\n",
__func__, frame);
@@ -464,7 +462,6 @@ static int ps3fb_sync(struct fb_info *in
xdr_line_length);
out:
- release_console_sem();
return error;
}
@@ -479,7 +476,10 @@ static int ps3fb_release(struct fb_info
if (atomic_dec_and_test(&ps3fb.f_count)) {
if (atomic_read(&ps3fb.ext_flip)) {
atomic_set(&ps3fb.ext_flip, 0);
- ps3fb_sync(info, 0); /* single buffer */
+ if (!try_acquire_console_sem()) {
+ ps3fb_sync(info, 0); /* single buffer */
+ release_console_sem();
+ }
}
}
return 0;
@@ -865,7 +865,9 @@ static int ps3fb_ioctl(struct fb_info *i
break;
dev_dbg(info->device, "PS3FB_IOCTL_FSEL:%d\n", val);
+ acquire_console_sem();
retval = ps3fb_sync(info, val);
+ release_console_sem();
break;
default:
@@ -885,7 +887,9 @@ static int ps3fbd(void *arg)
set_current_state(TASK_INTERRUPTIBLE);
if (ps3fb.is_kicked) {
ps3fb.is_kicked = 0;
+ acquire_console_sem();
ps3fb_sync(info, 0); /* single buffer */
+ release_console_sem();
}
schedule();
}
With kind regards,
Geert Uytterhoeven
Software Architect
Sony Network and Software Technology Center Europe
The Corporate Village · Da Vincilaan 7-D1 · B-1935 Zaventem · Belgium
Phone: +32 (0)2 700 8453
Fax: +32 (0)2 700 8622
E-mail: Geert.Uytterhoeven@sonycom.com
Internet: http://www.sony-europe.com/
Sony Network and Software Technology Center Europe
A division of Sony Service Centre (Europe) N.V.
Registered office: Technologielaan 7 · B-1840 Londerzeel · Belgium
VAT BE 0413.825.160 · RPR Brussels
Fortis Bank Zaventem · Swift GEBABEBB08A · IBAN BE39001382358619
^ permalink raw reply
* [PATCH 1/2] ps3fb: prevent use after free of fb_info
From: Geert Uytterhoeven @ 2008-01-11 13:27 UTC (permalink / raw)
To: Linus Torvalds, Andrew Morton
Cc: Linux/PPC Development, Linux Frame Buffer Device Development,
Jeremy Kerr, Cell Broadband Engine OSS Development
In-Reply-To: <Pine.LNX.4.64.0801111417450.12038@vixen.sonytel.be>
[-- Attachment #1: Type: TEXT/PLAIN, Size: 1869 bytes --]
From: Jeremy Kerr <jk@ozlabs.org>
In ps3fb_shutdown, freeing the framebuffer will cause fb_info (in
dev->core.driver_data) to be free()ed, which we potentially access
from the ps3fbd kthread.
This change frees the framebuffer after stopping the ps3fbd kthread.
Signed-off-by: Jeremy Kerr <jk@ozlabs.org>
Signed-off-by: Geert Uytterhoeven <Geert.Uytterhoeven@sonycom.com>
---
drivers/video/ps3fb.c | 12 ++++++------
1 file changed, 6 insertions(+), 6 deletions(-)
--- a/drivers/video/ps3fb.c
+++ b/drivers/video/ps3fb.c
@@ -1234,12 +1234,6 @@ static int ps3fb_shutdown(struct ps3_sys
ps3fb_flip_ctl(0, &ps3fb); /* flip off */
ps3fb.dinfo->irq.mask = 0;
- if (info) {
- unregister_framebuffer(info);
- fb_dealloc_cmap(&info->cmap);
- framebuffer_release(info);
- }
-
ps3av_register_flip_ctl(NULL, NULL);
if (ps3fb.task) {
struct task_struct *task = ps3fb.task;
@@ -1250,6 +1244,12 @@ static int ps3fb_shutdown(struct ps3_sys
free_irq(ps3fb.irq_no, &dev->core);
ps3_irq_plug_destroy(ps3fb.irq_no);
}
+ if (info) {
+ unregister_framebuffer(info);
+ fb_dealloc_cmap(&info->cmap);
+ framebuffer_release(info);
+ info = dev->core.driver_data = NULL;
+ }
iounmap((u8 __iomem *)ps3fb.dinfo);
status = lv1_gpu_context_free(ps3fb.context_handle);
With kind regards,
Geert Uytterhoeven
Software Architect
Sony Network and Software Technology Center Europe
The Corporate Village · Da Vincilaan 7-D1 · B-1935 Zaventem · Belgium
Phone: +32 (0)2 700 8453
Fax: +32 (0)2 700 8622
E-mail: Geert.Uytterhoeven@sonycom.com
Internet: http://www.sony-europe.com/
Sony Network and Software Technology Center Europe
A division of Sony Service Centre (Europe) N.V.
Registered office: Technologielaan 7 · B-1840 Londerzeel · Belgium
VAT BE 0413.825.160 · RPR Brussels
Fortis Bank Zaventem · Swift GEBABEBB08A · IBAN BE39001382358619
^ permalink raw reply
* [PATCH 0/2] ps3fb fixes for 2.6.24
From: Geert Uytterhoeven @ 2008-01-11 13:26 UTC (permalink / raw)
To: Linus Torvalds, Andrew Morton
Cc: Linux/PPC Development, Linux Frame Buffer Device Development,
Jeremy Kerr, Cell Broadband Engine OSS Development
[-- Attachment #1: Type: TEXT/PLAIN, Size: 989 bytes --]
Hi Linus, Andrew,
Here are 2 more fixes for ps3fb:
[1] ps3fb: prevent use after free of fb_info
[2] ps3fb: fix deadlock on kexec()
The first one fixes a potential use after free.
The second one fixes a lockup when using kexec or shutdown while /dev/fb0 is
open (e.g. when using the petitboot graphical bootloader to load the second
stage kernel).
Can we please get these in 2.6.24? Thanks!
With kind regards,
Geert Uytterhoeven
Software Architect
Sony Network and Software Technology Center Europe
The Corporate Village · Da Vincilaan 7-D1 · B-1935 Zaventem · Belgium
Phone: +32 (0)2 700 8453
Fax: +32 (0)2 700 8622
E-mail: Geert.Uytterhoeven@sonycom.com
Internet: http://www.sony-europe.com/
Sony Network and Software Technology Center Europe
A division of Sony Service Centre (Europe) N.V.
Registered office: Technologielaan 7 · B-1840 Londerzeel · Belgium
VAT BE 0413.825.160 · RPR Brussels
Fortis Bank Zaventem · Swift GEBABEBB08A · IBAN BE39001382358619
^ permalink raw reply
* PPC vs POWERPC
From: samppa @ 2008-01-11 12:33 UTC (permalink / raw)
To: linuxppc-embedded
Hi,
I want to port a board ( WRS SBC8265 ) from the PPC branch to the POWERPC
branch in the Linux Kernel -- do you have any good starting points that
describes what I need to pay attention to?
I've already made a port of U-boot-1.3.1 and enabled CONFIG_OF_LIBFDT.
So I need to create a DTS file, compile it with the DTC compiler and port
the current PPC branch to the POWERPC branch to my understanding.
Cheers // Matias
^ permalink raw reply
* Re: Please pull linux-2.6-virtex.git
From: Josh Boyer @ 2008-01-11 13:09 UTC (permalink / raw)
To: Grant Likely; +Cc: linuxppc-dev
In-Reply-To: <fa686aa40801090703m5f411795j890bfa714a128b1e@mail.gmail.com>
On Wed, 9 Jan 2008 08:03:31 -0700
"Grant Likely" <grant.likely@secretlab.ca> wrote:
> Gah; teach me to pick up patches right before bed. I shouldn't have
> picked up the ulite console changes patch just yet. I've dropped it
> from the series until I have a chance to rework it. Sorry.
>
> Here's the new pull request
>
> The following changes since commit 4f43143f9fbbb679c38d2ff99e44d3aaa00d0fe1:
> Paul Mackerras (1):
> Merge branch 'for-2.6.25' of git://git.kernel.org/.../olof/pasemi
>
> are available in the git repository at:
>
> git://git.secretlab.ca/git/linux-2.6-virtex.git virtex-for-2.6.25
>
> Stephen Neuendorffer (5):
> [POWERPC] Xilinx: update compatible list for interrupt controller
> [POWERPC] Xilinx: Add correct compatible list for device tree
> bus bindings.
> [POWERPC] Xilinx: Update booting-without-of.
> [POWERPC] Xilinx: updated device tree compatibility to match
> uboot bsp generator.
> [POWERPC] Xilinx uartlite: Section type fixups
Pulled and pushed out. I'll ask Paul to pull after I finish chasing
the Warp patches Sean is working on, and the RTC patch from David.
josh
^ permalink raw reply
* Re: Linux for ml310
From: Enno Lübbers @ 2008-01-11 11:40 UTC (permalink / raw)
To: linuxppc-embedded
Hi Joachim,
I haven't tried the XUP or the ML310 yet, but I just managed to get
Linux 2.4.26 from Bitkeeper running on a ML403 board (a Virtex-4FX12).
However, I'm seriously considering moving to the 2.6 kernel, for
various reasons.
Your problems seem to be related to the BSP configuration. Have you
set the corerct options in the Software Platform Settings of the EDK?
In particular, you need to set the clock frequency and the attached
peripherals in the "OS and Libraries" section.
Regards
- Enno
--
Dipl.-Ing. Enno Luebbers
Computer Science Department
University of Paderborn
Warburger Str. 100
33098 Paderborn
http://wwwcs.upb.de/cs/ag-platzner
phone: 05251 / 60-5397
fax: 05251 / 60-5377
^ 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