1994624 - Build failure in JIT on OpenBSD/riscv64 since bug 1982266

Status: RESOLVED FIXED

(Core :: JavaScript Engine: JIT, defect)

Description

[Landry Breuil (:gaston)]

I'm just the messenger as i don't have such hardware, but the build fails in the JIT engine since its been enabled by default in bug 1982266:

/local/pobj/firefox-143.0.3/firefox-143.0.3/js/src/jit/riscv64/Architecture-riscv64.cpp:89:4: error: "Unsupported platform" 89
 | #  error "Unsupported platform"

i suppose it might affect the other BSDs.

a simple fix (from jca, in cc) is inlined below, which i can push if deemed acceptable :

--- js/src/jit/riscv64/Architecture-riscv64.cpp.orig
+++ js/src/jit/riscv64/Architecture-riscv64.cpp
@@ -76,7 +76,7 @@ void FlushICache(void* code, size_t size) {
 #if defined(JS_SIMULATOR)
   js::jit::SimulatorProcess::FlushICache(code, size);
 
-#elif defined(__linux__)
+#elif defined(__linux__) || defined(__OpenBSD__)
 #  if defined(__GNUC__)
   intptr_t end = reinterpret_cast<intptr_t>(code) + size;
   __builtin___clear_cache(reinterpret_cast<char*>(code),
--- js/src/wasm/WasmSignalHandlers.cpp.orig
+++ js/src/wasm/WasmSignalHandlers.cpp
@@ -100,6 +100,12 @@ using namespace js::wasm;
 #      define RLR_sig(p) ((p)->sc_lr)
 #      define R31_sig(p) ((p)->sc_sp)
 #    endif
+#    if defined(__riscv)
+#      define RPC_sig(p) ((p)->sc_sepc)
+#      define RRA_sig(p) ((p)->sc_ra)
+#      define RFP_sig(p) ((p)->sc_s[0])
+#      define R02_sig(p) ((p)->sc_sp)
+#    endif
 #    if defined(__mips__)
 #      define EPC_sig(p) ((p)->sc_pc)
 #      define RFP_sig(p) ((p)->sc_regs[30])

given the full list of riscv64 bugs (https://bugzilla.mozilla.org/buglist.cgi?quicksearch=riscv64&list_id=17728043) we might disable the jit on OpenBSD/riscv64 until it's been more battle tested - or at least someone (tm) has run the full jit testsuite and compared the results with other platforms.

Comment 1

[BugBot [:suhaib / :marco/ :calixte]]

Set release status flags based on info from the regressing bug 1982266

:csmantle, since you are the author of the regressor, bug 1982266, could you take a look?

For more information, please visit BugBot documentation.

Comment 2

[Rong Bao [:csmantle]]

(In reply to Landry Breuil (:gaston) from comment #0)

given the full list of riscv64 bugs (https://bugzilla.mozilla.org/buglist.cgi?quicksearch=riscv64&list_id=17728043) we might disable the jit on OpenBSD/riscv64 until it's been more battle tested - or at least someone (tm) has run the full jit testsuite and compared the results with other platforms.

Agreed.

Comment 3

[Landry Breuil (:gaston)]

Attachment: Bug 1994624: fix JIT build on OpenBSD/riscv64. r=#spidermonkey-reviewers

Comment 4

[Saiteja]

Hi Landry,

I tried the patch that you mentioned in #comment 3 in my firefox-v143.0.3 sources and built the firefox.
It was built successfully but still i am facing the crashes when i am trying to launch the new tabs.
can you look into this issue?

Regards,
Saiteja.

Comment 5

[Rong Bao [:csmantle]]

(In reply to Saiteja from comment #4)

I tried the patch that you mentioned in #comment 3 in my firefox-v143.0.3 sources and built the firefox.
It was built successfully but still i am facing the crashes when i am trying to launch the new tabs.

Hi Saiteja, are you building for OpenBSD on RV64? If so, I think we can try to look at if it's also caused by the same (seemingly) graphics issues you encountered earlier.

(Replying since I'm on the CC list.)

Comment 6

[Saiteja]

Hi Rong Bao,

I built firefox for riscv64 on ubuntu-22.04, i observed the similar crash on ubuntu-22.04. so,i patched the #comment 3 in my firefox and built the firefox still observing the tabs getting crashed, not on OpenBSD.

Regards,
Saiteja

Comment 7

[Rong Bao [:csmantle]]

(In reply to Saiteja from comment #6)

I built firefox for riscv64 on ubuntu-22.04, i observed the similar crash on ubuntu-22.04. so,i patched the #comment 3 in my firefox and built the firefox still observing the tabs getting crashed, not on OpenBSD.

I'm afraid this patch is a noop for non-OpenBSD builds, so this is consistent with what you have observed.

Comment 8

[Pulsebot]

Pushed by [e-mail]:
https://github.com/mozilla-firefox/firefox/commit/eaa05d03ac16
https://hg.mozilla.org/integration/autoland/rev/36f9e80ba3d2
fix JIT build on OpenBSD/riscv64. r=spidermonkey-reviewers,iain

Comment 9

[Narcis Beleuzu [:NarcisB]]

https://hg.mozilla.org/mozilla-central/rev/36f9e80ba3d2

Comment 10

[BugBot [:suhaib / :marco/ :calixte]]

The patch landed in nightly and beta is affected.
:gaston, is this bug important enough to require an uplift?

• If yes, please nominate the patch for beta approval.
  • See https://wiki.mozilla.org/Release_Management/Requesting_an_Uplift for documentation on how to request an uplift.
• If no, please set status-firefox145 to wontfix.

For more information, please visit BugBot documentation.

Comment 11

[Landry Breuil (:gaston)]

Comment on attachment 9520384 [details]
Bug 1994624: fix JIT build on OpenBSD/riscv64. r=#spidermonkey-reviewers

Beta/Release Uplift Approval Request

• User impact if declined/Reason for urgency: FTBFS on OpenBSD/riscv64
• Is this code covered by automated tests?: No
• Has the fix been verified in Nightly?: Yes
• Needs manual test from QE?: No
• If yes, steps to reproduce:
• List of other uplifts needed: None
• Risk to taking this patch: Low
• Why is the change risky/not risky? (and alternatives if risky): Tier3
• String changes made/needed:
• Is Android affected?: No

Comment 12

[Pulsebot]

https://github.com/mozilla-firefox/firefox/commit/a3bff277a037
https://hg.mozilla.org/releases/mozilla-beta/rev/6bb27ca75983