From: bugzilla-daemon@freedesktop.org
To: dri-devel@lists.freedesktop.org
Subject: [Bug 91828] R600 LLVM Assertion in Instructions.cpp:1499
Date: Mon, 31 Aug 2015 20:23:15 +0000 [thread overview]
Message-ID: <bug-91828-502@http.bugs.freedesktop.org/> (raw)
[-- Attachment #1.1: Type: text/plain, Size: 2264 bytes --]
https://bugs.freedesktop.org/show_bug.cgi?id=91828
Bug ID: 91828
Summary: R600 LLVM Assertion in Instructions.cpp:1499
Product: Mesa
Version: git
Hardware: Other
OS: All
Status: NEW
Severity: normal
Priority: medium
Component: Drivers/Gallium/r600
Assignee: dri-devel@lists.freedesktop.org
Reporter: ikalvachev@gmail.com
QA Contact: dri-devel@lists.freedesktop.org
Created attachment 118020
--> https://bugs.freedesktop.org/attachment.cgi?id=118020&action=edit
Log of running Xonotic with R600_DEBUG=llvm,ps,vs. The last shader causes
assertion.
This bug happen with wide variety of games. I managed to reproduce it with the
free Xonotic v0.8.1. The exact command is
`R600_DEBUG=llvm,ps,vs ./xonotic-linux32-glx`.
(Then start the first single player map. Crash happens at first in-game frame.)
The error I get is:
---
xonotic-linux32-glx: Instructions.cpp:1499:
llvm::InsertElementInst::InsertElementInst(llvm::Value *, llvm::Value *,
llvm::Value *, const llvm::Twine &, llvm::Instruction *): Assertion
`isValidOperands(Vec, Elt, Index) && "Invalid insertelement instruction
operands!"' failed.
---
I've obtained that error by compiling debug versions of LLVM 3.6.2 and current
git of Mesa. Unfortunately the backtrace doesn't produce any relevant info.
I'm reporting the bug here, because the error happens during the phase where
TGSI code is translated into the intrinsics that LLVM understand. I do not see
llvm intrinsics code for the last TGSI shader, so I cannot reproduce the crash
with `llc`. I think it is quite likely that the problem is caused by some LLVM
API change, that have not been reflected in Mesa.
I know that R600 LLVM is not under active development. However at the moment
that code is crashing on regular basis. Mesa should be able to work with any
recent LLVM release and 3.6.2 is the latest current release. If you do not
intend to maintain it, then you might consider removing it.
I would prefer the bug(s) to be fixed. Nine exposes quite a number of SB bugs
and having other options is always good for testing.
--
You are receiving this mail because:
You are the assignee for the bug.
[-- Attachment #1.2: Type: text/html, Size: 3843 bytes --]
[-- Attachment #2: Type: text/plain, Size: 159 bytes --]
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
http://lists.freedesktop.org/mailman/listinfo/dri-devel
next reply other threads:[~2015-08-31 20:23 UTC|newest]
Thread overview: 2+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-08-31 20:23 bugzilla-daemon [this message]
2015-09-09 8:52 ` [Bug 91828] R600 LLVM Assertion in Instructions.cpp:1499 bugzilla-daemon
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=bug-91828-502@http.bugs.freedesktop.org/ \
--to=bugzilla-daemon@freedesktop.org \
--cc=dri-devel@lists.freedesktop.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.