Re: [linux-yocto] [PATCH 00/48] Intel Axxia updates to linux-yocto-4.9

2017-12-13 Thread Dragomir Daniel


On 12.12.2017 21:12, Bruce Ashfield wrote:

On 2017-12-11 8:13 AM, Daniel Dragomir wrote:

Hello Bruce!

Please review and merge this series of patches in Axxia branches from
linux-yocto-4.9.

If all the patches are ok, please pull them from
https://github.com/axxia/axxia_yocto_linux_4.9_pull.git
into
git://git.yoctoproject.org/linux-yocto-4.9
Pull
standard/axxia/base-1.15 -> standard/axxia/base
standard/preempt-rt/axxia/base-1.15 -> standard/preempt-rt/axxia/base



I'm not seeing these branches on the pull repository. Can you
double check that they are in place ?

Bruce


Hi Bruce!

I cloned again the repo to check the branches, and I was able to find them:
https://github.com/axxia/axxia_yocto_linux_4.9_pull.git
standard/axxia/base-1.15
standard/preempt-rt/axxia/base-1.15

Maybe after a fresh clone instead just a pull, you'll see the branches.
Regards,
Daniel




NOTE: Preempt-rt branch has 4 patches in addition to the standard one.

If you add more patches to standard/(preempt-rt)/axxia/base beforehand
please notify me. I will rebase our changes so you can do a clean,
fast-forward pull.

Changes included in this series (axxia tag 1.7 to 1.15):

Linux 4.9, Axxia 1.15 Changes
 Support the Axxia reset driver and, optionally, DDR retention 
resets.

 Backport the commits necessary to boot reliably with the Preempt-RT
 patches and debug configuration.

Linux 4.9, Axxia 1.14 Changes
 Update the PCI LOS work around to make sure the configuration 
gets put

 back to normal when the speed changes from gen1 to gen2/gen3.
 Support additional SPI flash devices (Winbond W25Q128JVSIM and 
Cypress S25FL128L).

 EDAC updates:
 - Change default log settings -- nothing on the console.
 - Add 6700 support.
 - Fix issue with DDR retention reset and the use of the 
scratch register.

 Only update coresight on 6700 hardware, not simulation.
 Create a custom gpio driver for Axxia as the interrupt structure 
is different.

 Include a device tree for the axm5612.
 Fix compile errors when CONFIG_PCI_MSI is not defined.
 Axxia network interface cleanup:
 - Remove unused reference, CONFIG_NET_VENDOR_LSI.
 - Include dependecy on Axxia MDIO.

Linux 4.9, Axxia 1.13 Changes
 Update from 4.9.49 to 4.9.61.

Linux 4.9, Axxia 1.12 Changes
 Update coresight hardware tracing support to include all 
clusters on 6700.
 Only include coresight in the hardware device trees, not 
simulation.

 Update the GPIO driver for 4.9.
 Include a 12 core device tree for 5600.
 Fix compile errors when MSI is not enabled.
 Remove unused reference to CONFIG_NET_VENDOR_LSI.
 Include dependency on Axxia MDIO when enabling Axxia NEMAC.

Linux 4.9, Axxia 1.11 Changes
 Enable EDAC CMEM support when CMEM gets initialized.
 Reduce the PCIe LOS work around timeout to 1 second.
 Update the debug configuration for 5600.

Linux 4.9, Axxia 1.10 Changes
 Rebase from Linux 4.9.36 to 4.9.49.

Linux 4.9, Axxia 1.9 Changes
 Support hardware tracing using coresight and tools/perf (cluster 
0 only, 6700 only).

 Add CCN504 interrupt driven EDAC driver.

Linux 4.9, Axxia 1.8 Changes
 Remove NOSMP configurations.
 Add support for MSI interruts on 5600 and 6700.
 APC3000 support.

Linux 4.9, Axxia 1.7 Changes
 Rebase to the latest Yocto (4.9.21 to 4.9.36).
 Update USB to use the original reset sequence instead of 
asserting HCRST:

 - The HCRST reset breaks USB2 and USB3 in some cases.
 Fixes for PEI configuration 5.
 Fix a boot problem on 5500. In the new Axxia fault code, 
exceptions need to be masked

 on 5500 by default.
 The final change to the PCI work around (If the link is not 
established at the requested
 number of lanes...) was not included. As the hardware does 
not have the correct default
 class code, software needs to set it. With the width change 
work around, the class code
 needs to be set to the correct value each time the width is 
changed.
 Update the default configurations to get rid of some warnings 
when using systemd instead of

 sysv init.
 Add tracing of NCR, PEI, and PCI accesses. Tracing can be 
enabled using the bootargs or after

 booting using /sys:
 - For NCR, use trace for all accesses, and trace_value_read 
to display the value read

   after a read.
 - For PEI and PCI, use trace.
 PCI Updates:
 - Add a work around for a hardware problem. If the link is 
not established at the requested
   number of lanes, reduce and try again. If there is no link 
at x1, fail.
 - Update the LOS work around to work the the lane reduction 
work around mentioned above.
 - Add a new configuration for 5600, unsupported on the Axxia 
development board, configuration 5.
 - In configuration 2, 5600, set sw_port_1 to 0x2 instead of 
0x4.

 - Always set

[linux-yocto] New branches for AXXIA in linux-yocto-4.12

2017-12-13 Thread Dragomir, Daniel
Hello Bruce!

Can you please create new branches for AXXIA in linux-yocto-4.12 repository?
We'll start submitting patches soon.

standard/preempt-rt/axxia/base from standard/preempt-rt/base
standard/preempt-rt/axxia/base from standard/base

Thank you!
Daniel
-- 
___
linux-yocto mailing list
linux-yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/linux-yocto


Re: [linux-yocto] New branches for AXXIA in linux-yocto-4.12

2017-12-13 Thread Dragomir Daniel


On 13.12.2017 19:31, Dragomir, Daniel wrote:

Hello Bruce!

Can you please create new branches for AXXIA in linux-yocto-4.12 
repository?

We'll start submitting patches soon.

standard/preempt-rt/axxia/base from standard/preempt-rt/base
standard/preempt-rt/axxia/base from standard/base


standard/preempt-rt/axxia/base from standard/preempt-rt/base
standard/axxia/base from standard/base

My mistake. Sorry.



Thank you!
Daniel




-- 
___
linux-yocto mailing list
linux-yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/linux-yocto


Re: [linux-yocto] [PATCH 00/48] Intel Axxia updates to linux-yocto-4.9

2017-12-13 Thread Bruce Ashfield

On 2017-12-11 8:13 AM, Daniel Dragomir wrote:

Hello Bruce!

Please review and merge this series of patches in Axxia branches from
linux-yocto-4.9.



The branches showed up for me, and have now been merged.

Bruce


If all the patches are ok, please pull them from
https://github.com/axxia/axxia_yocto_linux_4.9_pull.git
into
git://git.yoctoproject.org/linux-yocto-4.9
Pull
standard/axxia/base-1.15 -> standard/axxia/base
standard/preempt-rt/axxia/base-1.15 -> standard/preempt-rt/axxia/base

NOTE: Preempt-rt branch has 4 patches in addition to the standard one.

If you add more patches to standard/(preempt-rt)/axxia/base beforehand
please notify me. I will rebase our changes so you can do a clean,
fast-forward pull.

Changes included in this series (axxia tag 1.7 to 1.15):

Linux 4.9, Axxia 1.15 Changes
 Support the Axxia reset driver and, optionally, DDR retention resets.
 Backport the commits necessary to boot reliably with the Preempt-RT
 patches and debug configuration.

Linux 4.9, Axxia 1.14 Changes
 Update the PCI LOS work around to make sure the configuration gets put
 back to normal when the speed changes from gen1 to gen2/gen3.
 Support additional SPI flash devices (Winbond W25Q128JVSIM and Cypress 
S25FL128L).
 EDAC updates:
 - Change default log settings -- nothing on the console.
 - Add 6700 support.
 - Fix issue with DDR retention reset and the use of the scratch 
register.
 Only update coresight on 6700 hardware, not simulation.
 Create a custom gpio driver for Axxia as the interrupt structure is 
different.
 Include a device tree for the axm5612.
 Fix compile errors when CONFIG_PCI_MSI is not defined.
 Axxia network interface cleanup:
 - Remove unused reference, CONFIG_NET_VENDOR_LSI.
 - Include dependecy on Axxia MDIO.

Linux 4.9, Axxia 1.13 Changes
 Update from 4.9.49 to 4.9.61.

Linux 4.9, Axxia 1.12 Changes
 Update coresight hardware tracing support to include all clusters on 6700.
 Only include coresight in the hardware device trees, not simulation.
 Update the GPIO driver for 4.9.
 Include a 12 core device tree for 5600.
 Fix compile errors when MSI is not enabled.
 Remove unused reference to CONFIG_NET_VENDOR_LSI.
 Include dependency on Axxia MDIO when enabling Axxia NEMAC.

Linux 4.9, Axxia 1.11 Changes
 Enable EDAC CMEM support when CMEM gets initialized.
 Reduce the PCIe LOS work around timeout to 1 second.
 Update the debug configuration for 5600.

Linux 4.9, Axxia 1.10 Changes
 Rebase from Linux 4.9.36 to 4.9.49.

Linux 4.9, Axxia 1.9 Changes
 Support hardware tracing using coresight and tools/perf (cluster 0 only, 
6700 only).
 Add CCN504 interrupt driven EDAC driver.

Linux 4.9, Axxia 1.8 Changes
 Remove NOSMP configurations.
 Add support for MSI interruts on 5600 and 6700.
 APC3000 support.

Linux 4.9, Axxia 1.7 Changes
 Rebase to the latest Yocto (4.9.21 to 4.9.36).
 Update USB to use the original reset sequence instead of asserting HCRST:
 - The HCRST reset breaks USB2 and USB3 in some cases.
 Fixes for PEI configuration 5.
 Fix a boot problem on 5500. In the new Axxia fault code, exceptions need 
to be masked
 on 5500 by default.
 The final change to the PCI work around (If the link is not established at 
the requested
 number of lanes...) was not included. As the hardware does not have 
the correct default
 class code, software needs to set it. With the width change work 
around, the class code
 needs to be set to the correct value each time the width is changed.
 Update the default configurations to get rid of some warnings when using 
systemd instead of
 sysv init.
 Add tracing of NCR, PEI, and PCI accesses. Tracing can be enabled using 
the bootargs or after
 booting using /sys:
 - For NCR, use trace for all accesses, and trace_value_read to display 
the value read
   after a read.
 - For PEI and PCI, use trace.
 PCI Updates:
 - Add a work around for a hardware problem. If the link is not 
established at the requested
   number of lanes, reduce and try again. If there is no link at x1, 
fail.
 - Update the LOS work around to work the the lane reduction work 
around mentioned above.
 - Add a new configuration for 5600, unsupported on the Axxia 
development board, configuration 5.
 - In configuration 2, 5600, set sw_port_1 to 0x2 instead of 0x4.
 - Always set the number of phys to 4 for 5600 and 1 for 6700.
 - Always set the link width speed change register to x1 and use the 
link control register to
   set the desired width.
 USB Updates and USB support for 5600 B* parts:
 - Use HCRST for soft resets in the driver and don't reset the PHY 
seperately sinc HCRST resets
   the PHY.
 - Move all changes to constant v

Re: [linux-yocto] New branches for AXXIA in linux-yocto-4.12

2017-12-13 Thread Bruce Ashfield

On 2017-12-13 12:33 PM, Dragomir Daniel wrote:


On 13.12.2017 19:31, Dragomir, Daniel wrote:

Hello Bruce!

Can you please create new branches for AXXIA in linux-yocto-4.12 
repository?

We'll start submitting patches soon.

standard/preempt-rt/axxia/base from standard/preempt-rt/base
standard/preempt-rt/axxia/base from standard/base


standard/preempt-rt/axxia/base from standard/preempt-rt/base
standard/axxia/base from standard/base



To ssh://g...@git.yoctoproject.org/linux-yocto-4.12.git
 * [new branch]standard/axxia/base -> standard/axxia/base
 * [new branch]standard/preempt-rt/axxia/base -> 
standard/preempt-rt/axxia/base



Bruce


My mistake. Sorry.



Thank you!
Daniel






--
___
linux-yocto mailing list
linux-yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/linux-yocto


[linux-yocto] [yocto-4.12] [PATCH] x86/debug: Handle warnings before the notifier chain, to fix KGDB crash

2017-12-13 Thread Hongzhi.Song
From: Alexander Shishkin 

commit b8347c2196492f4e1cccde3d92fda1cc2cc7de7e upstream

Commit:

  9a93848fe787 ("x86/debug: Implement __WARN() using UD0")

turned warnings into UD0, but the fixup code only runs after the
notify_die() chain. This is a problem, in particular, with kgdb,
which kicks in as if it was a BUG().

Fix this by running the fixup code before the notifier chain in
the invalid op handler path.

Signed-off-by: Alexander Shishkin 
Tested-by: Ilya Dryomov 
Acked-by: Daniel Thompson 
Acked-by: Thomas Gleixner 
Cc: Jason Wessel 
Cc: Arjan van de Ven 
Cc: Borislav Petkov 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Richard Weinberger 
Cc:  # v4.12+
Link: 
http://lkml.kernel.org/r/20170724100428.19173-1-alexander.shish...@linux.intel.com
Signed-off-by: Ingo Molnar 

This patch is to resolve the problem shown as follows, when testing KGDB with 
'echo kgdbts=V1 > /sys/module/kgdbts/parameters/kgdbts':

[  196.275931] Kernel panic - not syncing: Recursive entry to debugger
[  196.276658] CPU: 0 PID: 1495 Comm: sh Not tainted 4.12.16-yocto-standard #2
[  196.277648] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 
rel-1.10.2-0-g5f4c7b1-prebuilt.qemu-project.org 04/01/2014
[  196.278634] Call Trace:
[  196.278891]  
[  196.279124]  dump_stack+0x65/0x8d
[  196.279477]  ? kgdb_breakpoint+0x13/0x20
[  196.279968]  kgdb_handle_exception+0x16e/0x1a0
[  196.280405]  __kgdb_notify+0x6d/0xd0
[  196.280761]  kgdb_notify+0x23/0x40
[  196.281189]  notifier_call_chain+0x4a/0x70
[  196.281589]  atomic_notifier_call_chain+0x33/0x50
[  196.282033]  notify_die+0x2e/0x30
[  196.282469]  do_int3+0x70/0x100
[  196.282792]  int3+0x35/0x70
[  196.283080] RIP: 0010:kgdb_breakpoint+0x14/0x20
[  196.283596] RSP: :9f2a0fc01d00 EFLAGS: 0002
[  196.284078] RAX:  RBX:  RCX: 
[  196.284751] RDX: a90cd700 RSI:  RDI: a8e4ca10
[  196.285356] RBP: 9f2a0fc01d00 R08:  R09: a8e4ca10
[  196.286002] R10: 9f2a0fc01e00 R11: a90d40ed R12: a8e4d2e0
[  196.286614] R13:  R14: a90cd700 R15: 

Signed-off-by: Hongzhi.Song 
---
 arch/x86/kernel/traps.c | 10 +++---
 1 file changed, 7 insertions(+), 3 deletions(-)

diff --git a/arch/x86/kernel/traps.c b/arch/x86/kernel/traps.c
index 67db4f4..5a6b8f8 100644
--- a/arch/x86/kernel/traps.c
+++ b/arch/x86/kernel/traps.c
@@ -209,9 +209,6 @@ do_trap_no_signal(struct task_struct *tsk, int trapnr, char 
*str,
if (fixup_exception(regs, trapnr))
return 0;
 
-   if (fixup_bug(regs, trapnr))
-   return 0;
-
tsk->thread.error_code = error_code;
tsk->thread.trap_nr = trapnr;
die(str, regs, error_code);
@@ -292,6 +289,13 @@ static void do_error_trap(struct pt_regs *regs, long 
error_code, char *str,
 
RCU_LOCKDEP_WARN(!rcu_is_watching(), "entry code didn't wake RCU");
 
+   /*
+* WARN*()s end up here; fix them up before we call the
+* notifier chain.
+*/
+   if (!user_mode(regs) && fixup_bug(regs, trapnr))
+   return;
+
if (notify_die(DIE_TRAP, str, regs, error_code, trapnr, signr) !=
NOTIFY_STOP) {
cond_local_irq_enable(regs);
-- 
2.8.1

-- 
___
linux-yocto mailing list
linux-yocto@yoctoproject.org
https://lists.yoctoproject.org/listinfo/linux-yocto