similar to: [RFC] Enable thread specific cl::opt values for multi-threaded support

So... looking at OptBisect, I have a few thoughts: 1) what's the purpose of the virtual interface/OptPassGate? I'm guessing maybe that worked around the circular referencing in these APIs? hmm, no, I suppose that wouldn't work/be relevant here. 2) Why is OptBisector a ManagedStatic? That seems pretty antithetical to the role of LLVMContext. When/why would a user be bisecting over

On 03/30/2018 12:05 AM, David Blaikie via llvm-dev wrote: > & now looking back at the patch-in-progress, I see it allows setting the OptBisector/OptPassGate as suggested in (2). Well, the patch currently discussed does not attempt to solve the passgate object management issue. It is left for the discretion of passgate object provider. > > If that becomes the /only/ option (ie:

& now looking back at the patch-in-progress, I see it allows setting the OptBisector/OptPassGate as suggested in (2). If that becomes the /only/ option (ie: LLVMContext has no default OptPassGate) then the virtual interface could be kept down in IR (though it's still a bit questionable to have those Analysis types (Loop, Region, CallGraphSCC) even declared in IR). Then the implementation

Andrew, I would not make the caller pass the description of the IR unit. That is because it would result in the description generated every time even if OptBisect is disabled. Description generation is not very chip. Thinking on the OptBisect extension, I believe passing the units are the right choice because OptPassGates may use them to make pass skipping decisions. -Yevgeny Rouban

> Pass Manager (PassManager.h) itself, does not - it's only templates, none of it depends on Region, Loop, etc. Well, true but the problem happens when you try to instantiate the thing. And for generic features like opt-bisection, ir-print-after-all etc that want:   - to have a say before/on/after every execution of every pass   - have a shared implementation of the main logic you have

On Mon, Apr 2, 2018 at 11:32 PM Fedor Sergeev via llvm-dev < llvm-dev at lists.llvm.org> wrote: > On 03/30/2018 12:05 AM, David Blaikie via llvm-dev wrote: > > & now looking back at the patch-in-progress, I see it allows setting > the OptBisector/OptPassGate as suggested in (2). > Well, the patch currently discussed does not attempt to solve the > passgate object

Ping on this - any chance we can look at fixing the OptBisect layering here/now? Could we move the implementation into Analysis & require users to set it, rather than having it as a default value in IR? On Tue, Apr 3, 2018 at 9:25 AM David Blaikie <dblaikie at gmail.com> wrote: > On Tue, Apr 3, 2018 at 8:50 AM Fedor Sergeev <fedor.sergeev at azul.com> > wrote: >

On Tue, Apr 3, 2018 at 8:50 AM Fedor Sergeev <fedor.sergeev at azul.com> wrote: > > Pass Manager (PassManager.h) itself, does not - it's only templates, > none of it depends on Region, Loop, etc. > Well, true but the problem happens when you try to instantiate the thing. > And for generic features like opt-bisection, ir-print-after-all etc that > want: > - to

David, we definitely need to address this issue and I did not forget about it. I'm still fleshing out a proposal on a generic solution for "pass execution control points", which was inspired by the new pass manager needs but then it appears we can largerly reuse the implementation for both managers. Since the implementation should be based on a special Analysis, it definitely

Lately I've been using some utilities to increase the number of logic errors caught at compile time. I thought they might be useful to the LLVM project. I'd appreciate feedback on the below proposal. Would the community find these useful? -David RFC: Strong typedef utilities for LLVM -------------------------------------- Abstract -------- This proposal

Hello, I am not finding how to read SSL keys/certificates from database. Is this possible? Or only file reads allowed? -- Regards, Yevgeny

Hello, I am getting strange messages on logs and I cannot find any reason. Could someone point me on what's wrong with my dovecot configuration? ~# dovecot -n # 2.3.4.1 (f79e8e7e4): /etc/dovecot/dovecot.conf # Pigeonhole version 0.5.4 () # OS: Linux 4.9.0-8-amd64 x86_64 Debian 9 ext4 ~# grep 'full mail' //var/log/dovecot.log Apr 16 19:25:43 imap(phaoost at

All, Is there any particular reason we don't use cl::ZeroOrMore more pervasively with the cl::opt class of command-line options? 1. Some of llvm's optimizations are enabled/disabled by llvm command-line options. Passing these options to our build system via CFF and CXXF result in redundant CL options. If the CL option is a cl::opt which defaults to cl::Optional, a warning is emitted.

I installed VS Express and included the IDE in $LIB. It looks like wine is able to run 'cl.exe' but 'cl.exe' itself has a problem. We've tried a number of variations of setting %TMP% and %TEMP% to no avail. Any ideas. TIA.

I'm attempting to build Firefox on Windows using clang-cl with -fprofile-instr-generate (*), but I'm running into some linker errors that I haven't been able to make sense of. lld-link.exe: error: duplicate symbol: __profc_?width at ios_base@std@@QEBA_JXZ in minidump-analyzer.obj and in ..\google-breakpad\src\processor\basic_code_modules.obj lld-link.exe: error: duplicate symbol:

Hi, I'm using Kubuntu and have a problem with the ISOLINUX CDROM booter with versions used since (K)Ubuntu 7.04. The ISOLINUX booter versions 3.36 and 3.53 do not boot my motherboard. The version shipped with Ubuntu 7.04 is I believe ISOLINUX 3.31 but it displays for such a short time I could be wrong. I'm writing to this mail list in case the original Ubuntu bug report never made it

comboot.txt says to set CL=9 "to get a valid return in CL for all versions". The code in __syslinux_get_derivative_info() doesn't set it. Is this comment obsolete? Looking at comboot.inc suggest this because P_CL is set to SECTOR_SHIFT for all derivatives but pxelinux which does set P_ECX to MyIP. - Sebastian

I've added some debugging flags to my register allocator here to turn on and off specific debug messages. I did this with llvm::cl::bits which allows me to turn on multiple options. But its kind of ugly because each option is independent as far as llvm is concerned and I have to prefix each option with something like RegallocDebug so users know what it's about. For example:

Hi, Just wanted to point this out here, in case it hasn't been run into yet. I've filed a bug with MS already, but it might be a warning to anyone thinking about uninstalling their older versions. Workaround was to build lib/TableGen (well, all of LLVM really due to a case of laziness) at /O1 and utils/TableGen at /Od. Turning inline expansion to 0 apparently works as well.

hi everyone, have you ever tried the argument -cl-kernel-arg-info for clang? can you share the usage? thank you -- View this message in context: http://llvm.1065342.n5.nabble.com/how-to-use-the-argument-cl-kernel-arg-info-for-clang-tp52524.html Sent from the LLVM - Dev mailing list archive at Nabble.com.