Ticket #356 (closed bug: fixed)

Opened 13 years ago

Last modified 11 years ago

better detect cpuarch with possible cmdline overrides

Reported by: rurban Owned by: jkeenan
Priority: normal Milestone:
Component: configure Version:
Severity: medium Keywords: 64bit
Cc: doughera Language:
Patch status: Platform: all

Description (last modified by rurban) (diff)

amd64 vs i386:

config::auto::arch just checks the perl5 config for the cpu type, esp. if 32-bit or 64-bit is used.

But it is easily possible to override that on the cmdline, and generate the wrong cpuarch then.

  • We could check for known arch overrides in ccflags/ldflags, like -m64, -m32 -Wl,-melf_x86_64 and such.
  • Or we could just check the ptr size with cc_build check in the compiler check and adjust the cpuarch later.

e.g. On solaris 10 with a 32-bit perl I use:

--ccflags='-m64 -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_TS_ER
RNO' --ldflags='-m64' --linkflags='-m64'
--ccflags='-m64 -D_LARGEFILE_SOURCE -D_FILE_OFFSET_BITS=64 -D_TS_ER
RNO' --ldflags='-m64' --linkflags='-m64' --floatval='long double'

Change History

  Changed 13 years ago by rurban

  • status changed from new to assigned
  • description modified (diff)

  Changed 13 years ago by rurban

  • keywords 64bit added

  Changed 12 years ago by bacek

  • status changed from assigned to closed
  • resolution set to wontfix

Current JIT is deprecated.

follow-ups: ↓ 5 ↓ 8   Changed 12 years ago by doughera

  • status changed from closed to reopened
  • resolution wontfix deleted
On Thu, 10 Sep 2009, Parrot wrote:

> #356: better detect jit cpuarch with possible cmdline overrides


> Changes (by bacek):
> 
>   * status:  assigned => closed
>   * resolution:  => wontfix
> 
> 
> Comment:
> 
>  Current JIT is deprecated.

Hang on a moment.  This ticket isn't really just about 'jit cpuarch', 
despite the slightly misleading subject line.  Parrot uses the 'cpuarch' 
setting for more than just JIT.

For OpenSolaris/x86, there's a patch in TT #423 for just this problem.
I'm unsure if it's an issue elsewhere.

-- 
    Andy Dougherty		doughera@lafayette.edu


in reply to: ↑ 4   Changed 12 years ago by jkeenan

Replying to doughera:

Hang on a moment. This ticket isn't really just about 'jit cpuarch', despite the slightly misleading subject line.

Agreed. I'm modifying the ticket's Summary to delete 'jit.

Other issues raised by rurban remain outstanding.

Thank you very much.

kid51

  Changed 12 years ago by jkeenan

  • status changed from reopened to new
  • owner changed from rurban to jkeenan

Hmm, the description didn't change. I suspect I have to take ownership of the ticket to change the description. Am trying that.

kid51

  Changed 12 years ago by jkeenan

  • owner jkeenan deleted
  • summary changed from better detect jit cpuarch with possible cmdline overrides to better detect cpuarch with possible cmdline overrides

in reply to: ↑ 4 ; follow-up: ↓ 9   Changed 11 years ago by jkeenan

  • cc doughera added

Replying to doughera:

{{{ On Thu, 10 Sep 2009, Parrot wrote: For OpenSolaris/x86, there's a patch in TT #423 for just this problem. I'm unsure if it's an issue elsewhere.

Andy,

FBOW, TT #423 has been closed. Could you evaluate the issues raised in this ticket and perhaps point a direction for us to explore?

Thank you very much.

kid51

in reply to: ↑ 8 ; follow-up: ↓ 11   Changed 11 years ago by doughera

Replying to jkeenan:

Replying to doughera:

{{{ On Thu, 10 Sep 2009, Parrot wrote: For OpenSolaris/x86, there's a patch in TT #423 for just this problem. I'm unsure if it's an issue elsewhere.

Andy, FBOW, TT #423 has been closed. Could you evaluate the issues raised in this ticket and perhaps point a direction for us to explore?

It's been fixed for Solaris, which was the main subject of this ticket. I am unaware of this being a practical issue anywhere else, though I can easily imagine it's possible it could be an issue.

More generally, without a clear sense of what, precisely, parrot intends to do with the information about "cpuarch", it's rather pointless to speculate on how to "improve" the detection.

For now, there doesn't seem to be an issue. I'd close the ticket.

  Changed 11 years ago by jkeenan

  • status changed from new to assigned
  • owner set to jkeenan

in reply to: ↑ 9   Changed 11 years ago by jkeenan

  • status changed from assigned to closed
  • resolution set to fixed

Replying to doughera:

More generally, without a clear sense of what, precisely, parrot intends to do with the information about "cpuarch", it's rather pointless to speculate on how to "improve" the detection. For now, there doesn't seem to be an issue. I'd close the ticket.

Thanks for the review. Closing the ticket.

Note: See TracTickets for help on using tickets.