1525630 - choose the correct frame pointer enablement options on aarch64 windows

Status: RESOLVED FIXED

(Firefox Build System :: General, enhancement)

Description

[Nathan Froyd [:froydnj]]

clang-cl doesn't currently support -Oy correctly, so we need to resort
to the GNU-ish based spelling.

Comment 1

[Nathan Froyd [:froydnj]]

Attachment: choose the correct frame pointer enablement options on aarch64 windows

Comment 2

[Nathan Froyd [:froydnj]]

Comment on attachment 9041837 [details] [diff] [review]
choose the correct frame pointer enablement options on aarch64 windows

Review of attachment 9041837 [details] [diff] [review]:
-----------------------------------------------------------------

Ugh, I copypasta'd...but -Xclang -fomit-frame-pointer throws errors, so clearly I'm doing something else wrong.

Comment 3

[(Away)]

Yes - I was going to ask you to show me the resulting .sym files from a try build. I'm not convinced that this is going to work...

Comment 4

[Nathan Froyd [:froydnj]]

Attachment: choose the correct frame pointer enablement options on aarch64 windows

This appears to work better, insofar as the compiler accepts it.  I'm not
completely convinced that we're actually reading minidumps correctly, though;
I'll try to write out findings on that tomorrow.

Comment 5

[(Away)]

Comment on attachment 9041946 [details] [diff] [review]
choose the correct frame pointer enablement options on aarch64 windows

The code itself seems fine, but let's figure out why about:crashparent is unhappy before landing this.

Comment 6

[Nathan Froyd [:froydnj]]

Comment on attachment 9041946 [details] [diff] [review]
choose the correct frame pointer enablement options on aarch64 windows

I'm not able to figure out why about:crash{parent,content} isn't working, because I can't convince windbg to give me a reasonable debug experience with symbols.  Can we land this patch so that we can see if we get reasonable crash reports with it, and leave the about:crash* issues to another bug?

Comment 7

[(Away)]

Try one more thing: comment out this line, it will make your browser quite crashy. https://searchfox.org/mozilla-central/source/js/src/jit/arm64/vixl/MozCpu-vixl.cpp#52

Comment 8

[(Away)]

Comment on attachment 9041946 [details] [diff] [review]
choose the correct frame pointer enablement options on aarch64 windows

This can be a straight revert the next time we pick up clang, right?

Comment 9

[Nathan Froyd [:froydnj]]

(In reply to David Major [:dmajor] from comment #8)

Comment on attachment 9041946 [details] [diff] [review]
choose the correct frame pointer enablement options on aarch64 windows

This can be a straight revert the next time we pick up clang, right?

Assuming we upgrade to 8.0, yes.

Comment 10

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/mozilla-inbound/rev/6f86eefc83a5
choose the correct frame pointer enablement options on aarch64 windows; r=dmajor

Comment 11

[Razvan Maries]

https://hg.mozilla.org/mozilla-central/rev/6f86eefc83a5

Comment 12

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

https://hg.mozilla.org/mozilla-central/rev/6f86eefc83a5