Re: [PATCH] powerpc: process.c: fix Kconfig typo

Cyril Bur [off-list ref] writes:

On Wed, 2016-10-05 at 07:57 +0200, Valentin Rothberg wrote:

quoted

s/ALIVEC/ALTIVEC/

Oops, nice catch

quoted

Signed-off-by: Valentin Rothberg <redacted>

Reviewed-by: Cyril Bur <redacted>

How did we not notice? Sounds like we need a new selftest.

Looks like this should have:

Fixes: dc16b553c949 ("powerpc: Always restore FPU/VEC/VSX if hardware transactional memory in use")


And I guess I need to start running checkkconfigsymbols.py on every
commit.

cheers

On Wed, Oct 26, 2016 at 7:52 AM, Michael Ellerman [off-list ref] wrote:

Cyril Bur [off-list ref] writes:

quoted

On Wed, 2016-10-05 at 07:57 +0200, Valentin Rothberg wrote:

quoted

s/ALIVEC/ALTIVEC/

Oops, nice catch

quoted

Signed-off-by: Valentin Rothberg <redacted>

Reviewed-by: Cyril Bur <redacted>

How did we not notice? Sounds like we need a new selftest.

Looks like this should have:

Fixes: dc16b553c949 ("powerpc: Always restore FPU/VEC/VSX if hardware transactional memory in use")


And I guess I need to start running checkkconfigsymbols.py on every
commit.

Happy to find a new user :-)  You can also run the script on a range
of commits via '--diff commitA..commitB', which can safe some time.

Best regards,
 Valentin

cheers

Valentin Rothberg [off-list ref] writes:

On Wed, Oct 26, 2016 at 7:52 AM, Michael Ellerman [off-list ref] wrote:

quoted

Cyril Bur [off-list ref] writes:

quoted

On Wed, 2016-10-05 at 07:57 +0200, Valentin Rothberg wrote:

quoted

s/ALIVEC/ALTIVEC/

Oops, nice catch

quoted

Signed-off-by: Valentin Rothberg <redacted>

Reviewed-by: Cyril Bur <redacted>

How did we not notice? Sounds like we need a new selftest.

Looks like this should have:

Fixes: dc16b553c949 ("powerpc: Always restore FPU/VEC/VSX if hardware transactional memory in use")


And I guess I need to start running checkkconfigsymbols.py on every
commit.

Happy to find a new user :-)  You can also run the script on a range
of commits via '--diff commitA..commitB', which can safe some time.

Yeah I just discovered that, very useful.

But my instinct was to do:

$ ./scripts/checkkconfigsymbols.py --diff HEAD^^^..HEAD

Which didn't work, the regexp is a bit tight, fixed by:

diff --git a/scripts/checkkconfigsymbols.py b/scripts/checkkconfigsymbols.py
index a32e4da4c117..3820f00b066a 100755
--- a/scripts/checkkconfigsymbols.py
+++ b/scripts/checkkconfigsymbols.py

@@ -88,7 +88,7 @@ def parse_options():
     if args.commit and args.diff:
         sys.exit("Please specify only one option at once.")
 
-    if args.diff and not re.match(r"^[\w\-\.]+\.\.[\w\-\.]+$", args.diff):
+    if args.diff and not re.match(r"^[\w\-\.\^]+\.\.[\w\-\.\^]+$", args.diff):
         sys.exit("Please specify valid input in the following format: "
                  "\'commit1..commit2\'")
 

cheers

On Wed, Oct 26, 2016 at 12:40 PM, Michael Ellerman [off-list ref] wrote:

Looks great.  If you agree, I'll pick this up and send the patch to
Greg (who takes patches for this script)?

Thanks,
 Valentin

Valentin Rothberg [off-list ref] writes:

On Wed, Oct 26, 2016 at 12:40 PM, Michael Ellerman [off-list ref] wrote:

quoted

diff --git a/scripts/checkkconfigsymbols.py b/scripts/checkkconfigsymbols.py
index a32e4da4c117..3820f00b066a 100755
--- a/scripts/checkkconfigsymbols.py
+++ b/scripts/checkkconfigsymbols.py

@@ -88,7 +88,7 @@ def parse_options():
     if args.commit and args.diff:
         sys.exit("Please specify only one option at once.")

-    if args.diff and not re.match(r"^[\w\-\.]+\.\.[\w\-\.]+$", args.diff):
+    if args.diff and not re.match(r"^[\w\-\.\^]+\.\.[\w\-\.\^]+$", args.diff):
         sys.exit("Please specify valid input in the following format: "
                  "\'commit1..commit2\'")

Looks great.  If you agree, I'll pick this up and send the patch to
Greg (who takes patches for this script)?

That'd be great, thanks.

cheers

On Wed, 2016-10-26 at 16:52 +1100, Michael Ellerman wrote:

Cyril Bur [off-list ref] writes:

quoted

On Wed, 2016-10-05 at 07:57 +0200, Valentin Rothberg wrote:

quoted

s/ALIVEC/ALTIVEC/

Oops, nice catch

quoted

Signed-off-by: Valentin Rothberg <redacted>

Reviewed-by: Cyril Bur <redacted>

How did we not notice? Sounds like we need a new selftest.

Indeed...

Here is probably a good place to say why we didn't catch it and under
what circumstances this will have a negative effect.

If a thread performs and transaction with altivec and then gets
preempted for whatever reason, this bug may cause the kernel to not
reenable altivec when that thread runs again. This will result in an
altivec unavailable fault, when these faults happen inside a user
transaction the kernel has no choice but enable altivec and doom the
transaction.

The result is that transactions using altivec may get aborted more than
they should.

The difficulty in catching this with a selftest is my deliberate use of
the word may above. Optimisations to avoid FPU/altivec/VSX faults mean
that the kernel will always leave them on for 255 switches, this code
prevents the kernel turning it off if it got to the 256th switch (and
userspace was transactional)...


Cyril

Looks like this should have:

Fixes: dc16b553c949 ("powerpc: Always restore FPU/VEC/VSX if hardware
transactional memory in use")


And I guess I need to start running checkkconfigsymbols.py on every
commit.

cheers