jzbydniewski

Results 25 comments of jzbydniewski

Looking at stack traces, I can see that it's always one of registers (r0 usually, might depend on debug changes introduced) that contains invalid value. Could it be that there...

I added the following change to the tcp_check_space to catch the issue earlier: ``` diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index fd9067062f2a..0152a30bfc35 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -5463,6 +5463,12 @@ void...

Here is yet another reproduction, SmaRT: [smart-log-core-1.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15307580/smart-log-core-1.txt) [smart-log-core-2.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15307581/smart-log-core-2.txt) [smart-log-core-3.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15307582/smart-log-core-3.txt) ![image](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/assets/73646606/2e982806-2096-4710-a55d-4c89264ed5bf)

Another reproduction, now I can see SmaRT looks matching with call stack, here is my analysis of the situation ![image](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/assets/73646606/7604658d-9574-42b2-a13c-be1077e10e8e) From what I see content of registers r16..r24 looks matching...

@xxkent I tried with this patch but it's the same ![image](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/assets/73646606/7f4737bb-e462-43b9-995c-3cdd60ded729) Do I understand correctly that SmaRT is showing full trace of instructions executed by a particular CPU (even if...

Contents of dcache for each core + L2 cache: [dcache-core1.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15321449/dcache-core1.txt) [dcache-core2.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15321450/dcache-core2.txt) [dcache-core3.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15321451/dcache-core3.txt) [l2-cache.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15321453/l2-cache.txt) ![image](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/assets/73646606/5131d9f5-c594-4209-9767-913409087c31) According to the contents of data cache for core 1 where the issue triggered, the ldd.ab...

I'm testing now with kernel option CONFIG_ARC_HAS_LL64=n. No reproduction within 3 days. Buildroot didn't require change as it's not passing "mll64" to gcc, and from what I see by default...

This is what I see with -off=flush_dcache ![image](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/assets/73646606/5dfabd2f-5517-416f-8efc-e18dc57a8734) it looks like the value that should have been loaded to r14 (at address 0x81c35cbc) is in L1 data cache line marked...

The good news is that with patch https://github.com/foss-for-synopsys-dwc-arc-processors/linux/commit/28e6344920c628041a9447472dbf1ce3c2c5834f I no longer can see boot-time freeze, tried over 2k times.

> Could you provide call stacks for each core from MDB? Yes, here new capture of SmaRT + call stacks for each core: [callstack-core0.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15259297/callstack-core0.txt) [callstack-core1.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15259299/callstack-core1.txt) [callstack-core2.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15259300/callstack-core2.txt) [smart-core0.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15259302/smart-core0.txt) [smart-core1.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15259303/smart-core1.txt) [smart-core2.txt](https://github.com/foss-for-synopsys-dwc-arc-processors/linux/files/15259304/smart-core2.txt) This...