9 Nov
2017
9 Nov
'17
1:18 p.m.
Hi Waldemar,
I personally made this tst-cancel test start to pass in the past,
related to unwinding issues we had in our tools.
However, I have noticed that some of the tst-cancel tests would have
some non-deterministic behavior.
This was related to some pipes not blocking, when writing some data, but
not blocking.
If I increased the written elements to the pipe, those would block
always, and the test would pass.
It is possible that there was some change in the kernel that might have
fixed this.
Vineet: Do you know anything about this ?
Also be sure you are using our latest tools, as like I said I have
addressed these problems.
Indeed, clobbering blink was an issue when millicode option was enabled.
Regarding your patch I would advise not to apply it until we find the
real reason for blink not to be restored.
Best regards,
Cupertino
On 11/09/2017 08:18 AM, Waldemar Brodkorb wrote:
Hi Cupertino,
Cupertino Miranda wrote,
Hi Waldemar,
Can you share in which context does this happens, for example by sharing
a .i file for the compilation and point out for the actual location of
the macro expansion?
The reason to request this is to understand if the compiler is not doing
the right thing with blink, as semantically the patched and unpatched
code should be the same, as long as the surrounding code is correct.
It could also be that the problem is not necessarily the blink not being
clobbered but by some other reason it not expecting it to be needed.
I tried to
generate the files for you, but had updated the Linux
kernel before your mail. It seems now that the same test suite tests
are executing fine without the blink removal patch.
Is there any related bugfix made between Linux 4.9.50 and 4.9.60
someone remembers? Strange, I need to verify with 4.9.50 again.
best regards
Waldemar