[parisc-linux] Status of hp-ux binary compatibility?

[parisc-linux] Status of hp-ux binary compatibility?

[Nathan Neulinger]

I just started looking at PA-RISC linux 0.9.2 on a J210 (which installed 
great on a narrow scsi drive, but didn't see the internal fw). Only 
problem I've seen so far is that it is spitting out ip_local_deliver bad 
loopback skb: PRE_ROUTING LOCAL_IN messages on the console for the lo 
device.

My main question though - what's the status of hpux binary support? The 
FAQ seems to indicate that some parisc binaries are already functional.

What is necessary to be done on the machine before I can run an HP 
executable?

-- Nathan

------------------------------------------------------------
Nathan Neulinger                       EMail:  nneul@umr.edu
University of Missouri - Rolla         Phone: (573) 341-4841
Computing Services                       Fax: (573) 341-4216

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Richard Allen]

On Fri, Sep 28, 2001 at 12:35:26PM -0500, Nathan Neulinger wrote:
> I just started looking at PA-RISC linux 0.9.2 on a J210 (which installed 
> great on a narrow scsi drive, but didn't see the internal fw). Only 
> problem I've seen so far is that it is spitting out ip_local_deliver bad 
> loopback skb: PRE_ROUTING LOCAL_IN messages on the console for the lo 
> device.
> 

I got rid of that message by killing the dhcp client daemon.


-- 
Rikki.         --  HP Technical Support, RHCE, RHCX, HP-UX Certified Admin.
               --  Solaris 7 Certified Systems and Network Administrator.
Bell Labs Unix --  Reach out and grep someone.
Those who do not understand Unix are condemned to reinvent it, poorly.

RE: [parisc-linux] Status of hp-ux binary compatibility?

[Neulinger, Nathan]

Ah, thanks!... wonder why it got enabled since I told it no on dhcp...
Weird. (I'm not a debian user, so some of it's ways of doing things are a
bit odd to me.)

-- Nathan

> -----Original Message-----
> From: Richard Allen [mailto:ra@hp.is]
> Sent: Friday, September 28, 2001 1:31 PM
> To: Nathan Neulinger
> Cc: parisc-linux@lists.parisc-linux.org
> Subject: Re: [parisc-linux] Status of hp-ux binary compatibility?
> 
> 
> On Fri, Sep 28, 2001 at 12:35:26PM -0500, Nathan Neulinger wrote:
> > I just started looking at PA-RISC linux 0.9.2 on a J210 
> (which installed 
> > great on a narrow scsi drive, but didn't see the internal fw). Only 
> > problem I've seen so far is that it is spitting out 
> ip_local_deliver bad 
> > loopback skb: PRE_ROUTING LOCAL_IN messages on the console 
> for the lo 
> > device.
> > 
> 
> I got rid of that message by killing the dhcp client daemon.
> 
> 
> -- 
> Rikki.         --  HP Technical Support, RHCE, RHCX, HP-UX 
> Certified Admin.
>                --  Solaris 7 Certified Systems and Network 
> Administrator.
> Bell Labs Unix --  Reach out and grep someone.
> Those who do not understand Unix are condemned to reinvent it, poorly.
> 

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Richard Hirst]

On Fri, Sep 28, 2001 at 02:02:09PM -0500, Neulinger, Nathan wrote:
> Ah, thanks!... wonder why it got enabled since I told it no on dhcp...
> Weird. (I'm not a debian user, so some of it's ways of doing things are a
> bit odd to me.)

Caused by a broken dhcp-client package.  Should be invoked via ifup/ifdown,
but was also being started via init scripts.  rm /etc/rc?.d/*dhcp-client
(if I rememeber the name right) to fix it.

Richard

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Matthew Wilcox]

On Fri, Sep 28, 2001 at 12:35:26PM -0500, Nathan Neulinger wrote:
> My main question though - what's the status of hpux binary support? The 
> FAQ seems to indicate that some parisc binaries are already functional.
> 
> What is necessary to be done on the machine before I can run an HP 
> executable?

It might work, but it's probably bitrotten since nobody's tried it in
over a year.

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Nathan Neulinger]

Matthew Wilcox wrote:
> 
> On Fri, Sep 28, 2001 at 12:35:26PM -0500, Nathan Neulinger wrote:
> > My main question though - what's the status of hpux binary support? The
> > FAQ seems to indicate that some parisc binaries are already functional.
> >
> > What is necessary to be done on the machine before I can run an HP
> > executable?
> 
> It might work, but it's probably bitrotten since nobody's tried it in
> over a year.

Yuck... That'd be most unfortunate since that is the primary reason I'm
looking at parisc-linux at the moment.

-- Nathan

------------------------------------------------------------
Nathan Neulinger                       EMail:  nneul@umr.edu
University of Missouri - Rolla         Phone: (573) 341-4841
Computing Services                       Fax: (573) 341-4216

RE: [parisc-linux] Status of hp-ux binary compatibility?

[Neulinger, Nathan]

I just built and upgrade to 2.4.9-pa27 and with it, as well as the 2.4.0
base install, I get nothing but segmentation faults whenever trying to
execute any HP-UX executable.

-- Nathan

> -----Original Message-----
> From: Matthew Wilcox [mailto:willy@debian.org]
> Sent: Sunday, September 30, 2001 11:19 PM
> To: Nathan Neulinger
> Cc: parisc-linux@lists.parisc-linux.org
> Subject: Re: [parisc-linux] Status of hp-ux binary compatibility?
> 
> 
> On Fri, Sep 28, 2001 at 12:35:26PM -0500, Nathan Neulinger wrote:
> > My main question though - what's the status of hpux binary 
> support? The 
> > FAQ seems to indicate that some parisc binaries are already 
> functional.
> > 
> > What is necessary to be done on the machine before I can run an HP 
> > executable?
> 
> It might work, but it's probably bitrotten since nobody's tried it in
> over a year.
> 
> -- 
> Revolutions do not require corporate support.
> 

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Matthew Wilcox]

On Mon, Oct 01, 2001 at 07:47:45AM -0500, Nathan Neulinger wrote:
> Yuck... That'd be most unfortunate since that is the primary reason I'm
> looking at parisc-linux at the moment.

what application specifically are you looking at running?  We only ever
implemented a few of the hundreds of hpux syscalls.

-- 
Revolutions do not require corporate support.

RE: [parisc-linux] Status of hp-ux binary compatibility?

[Neulinger, Nathan]

In particular - flexlm. 

We have a ton of licensed products served off an old 715, and relicensing
them to move the license server will be a nightmare. What I'd like to do is
run the lmgrd+vendor-daemon on a parisc-linux box and modload a syscall
replacement that lies about the hostid to this process.

But for testing, I've just tried very simple programs - cal, yes, echo, and
a staticly linked hello world. All of them segfault right after execution. 

As far as I know, the only system calls out of the ordinary the flexlm stuff
should use are file locking and perhaps some of the ioctl stuff for getting
interface address, but I figured on tracing to find out exactly what it's
doing. 

-- Nathan

> -----Original Message-----
> From: Matthew Wilcox [mailto:willy@debian.org]
> Sent: Monday, October 01, 2001 8:26 AM
> To: Nathan Neulinger
> Cc: Matthew Wilcox; parisc-linux@lists.parisc-linux.org
> Subject: Re: [parisc-linux] Status of hp-ux binary compatibility?
> 
> 
> On Mon, Oct 01, 2001 at 07:47:45AM -0500, Nathan Neulinger wrote:
> > Yuck... That'd be most unfortunate since that is the 
> primary reason I'm
> > looking at parisc-linux at the moment.
> 
> what application specifically are you looking at running?  We 
> only ever
> implemented a few of the hundreds of hpux syscalls.
> 
> -- 
> Revolutions do not require corporate support.
> 

RE: [parisc-linux] Status of hp-ux binary compatibility?

[Neulinger, Nathan]

FYI, here's one of the oops:

do_page_fault() pid=481 command='cal' type=6 address=0xc0000007
vm_start = 0x40001000, vm_end = 0x40002000

     YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001000000000000001111
r0-3     00000000 c0000000 0000274f faf00000
r4-7     00000001 faf001e8 faf00194 00000000
r8-11    faf00280 000aa810 000b8910 000b86f0
r12-15   00000000 ffffffff 000b81b0 00000000
r16-19   00000000 00017800 00017800 4024c57c
r20-23   0000000b 401b071c 00000039 00000000
r24-27   00000000 00000000 faf02290 40001448
r28-31   00000000 00000000 faf02300 0000286f
sr0-3    00000725 00000000 00000000 00000725
sr4-7    00000725 00000725 00000725 00000725

IASQ: 00000725 00000725 IAOQ: c0000007 c000000b
 IIR: e02025b2    ISR: 00000725  IOR: faf0226c
 CPU:        0   CR30: 2734c000 CR31: 10388000
 ORIG_R28: 000b8950

do_page_fault() pid=482 command='echo' type=6 address=0xc0000007
vm_start = 0x40001000, vm_end = 0x40002000

     YZrvWESTHLNXBCVMcbcbcbcbOGFRQPDI
PSW: 00000000000001000000000000001111
r0-3     00000000 c0000000 000024df faf00000
r4-7     00000001 faf001ec faf00198 00000000
r8-11    faf00280 000aa810 000b8910 000b8850
r12-15   00000000 ffffffff 000b8370 00000000
r16-19   00000000 00017800 00017800 4024c57c
r20-23   0000000b 401b071c 00000039 00000000
r24-27   00000000 00000000 faf02290 40001150
r28-31   00000000 00000000 faf02300 000025ff
sr0-3    00000004 00000000 00000000 00000004
sr4-7    00000004 00000004 00000004 00000004

IASQ: 00000004 00000004 IAOQ: c0000007 c000000b
 IIR: e02025b2    ISR: 00000004  IOR: faf0226c
 CPU:        0   CR30: 2734c000 CR31: 10388000
 ORIG_R28: 000b84b0

-- Nathan



> -----Original Message-----
> From: Neulinger, Nathan 
> Sent: Monday, October 01, 2001 8:31 AM
> To: 'Matthew Wilcox'
> Cc: parisc-linux@lists.parisc-linux.org
> Subject: RE: [parisc-linux] Status of hp-ux binary compatibility?
> 
> 
> In particular - flexlm. 
> 
> We have a ton of licensed products served off an old 715, and 
> relicensing them to move the license server will be a 
> nightmare. What I'd like to do is run the lmgrd+vendor-daemon 
> on a parisc-linux box and modload a syscall replacement that 
> lies about the hostid to this process.
> 
> But for testing, I've just tried very simple programs - cal, 
> yes, echo, and a staticly linked hello world. All of them 
> segfault right after execution. 
> 
> As far as I know, the only system calls out of the ordinary 
> the flexlm stuff should use are file locking and perhaps some 
> of the ioctl stuff for getting interface address, but I 
> figured on tracing to find out exactly what it's doing. 
> 
> -- Nathan
> 
> > -----Original Message-----
> > From: Matthew Wilcox [mailto:willy@debian.org]
> > Sent: Monday, October 01, 2001 8:26 AM
> > To: Nathan Neulinger
> > Cc: Matthew Wilcox; parisc-linux@lists.parisc-linux.org
> > Subject: Re: [parisc-linux] Status of hp-ux binary compatibility?
> > 
> > 
> > On Mon, Oct 01, 2001 at 07:47:45AM -0500, Nathan Neulinger wrote:
> > > Yuck... That'd be most unfortunate since that is the 
> > primary reason I'm
> > > looking at parisc-linux at the moment.
> > 
> > what application specifically are you looking at running?  We 
> > only ever
> > implemented a few of the hundreds of hpux syscalls.
> > 
> > -- 
> > Revolutions do not require corporate support.
> > 
> 

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Matthew Wilcox]

On Mon, Oct 01, 2001 at 08:31:07AM -0500, Neulinger, Nathan wrote:
> But for testing, I've just tried very simple programs - cal, yes, echo, and
> a staticly linked hello world. All of them segfault right after execution. 

It'd be useful to know what address they segfault at.  My current
suspicion is the stack position changes have broken it, but it may be
something completely different.

-- 
Revolutions do not require corporate support.

RE: [parisc-linux] Status of hp-ux binary compatibility?

[Neulinger, Nathan]

When you get down to it, so is making any change to how the software is
installed. I can't even remember the last time we installed a package on
unix the way the vendor insisted it be installed. 

-- Nathan

> -----Original Message-----
> From: Alan Cox [mailto:alan@lxorguk.ukuu.org.uk]
> Sent: Monday, October 01, 2001 10:34 AM
> To: nneul@umr.edu
> Cc: willy@debian.org; parisc-linux@lists.parisc-linux.org
> Subject: Re: [parisc-linux] Status of hp-ux binary compatibility?
> 
> 
> > them to move the license server will be a nightmare. What 
> I'd like to do is
> > run the lmgrd+vendor-daemon on a parisc-linux box and 
> modload a syscall
> > replacement that lies about the hostid to this process.
> 
> Thats now a criminal offence in the USA. You are circumventing digital
> rights management. The DMCA has no exclusion for actually 
> having the rights
> to do so, or exemption swhen you have the license
> 

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Alan Cox]

> them to move the license server will be a nightmare. What I'd like to do is
> run the lmgrd+vendor-daemon on a parisc-linux box and modload a syscall
> replacement that lies about the hostid to this process.

Thats now a criminal offence in the USA. You are circumventing digital
rights management. The DMCA has no exclusion for actually having the rights
to do so, or exemption swhen you have the license

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Matthew Wilcox]

On Mon, Oct 01, 2001 at 08:35:52AM -0500, Neulinger, Nathan wrote:
> IASQ: 00000725 00000725 IAOQ: c0000007 c000000b
>  IIR: e02025b2    ISR: 00000725  IOR: faf0226c
>  CPU:        0   CR30: 2734c000 CR31: 10388000
>  ORIG_R28: 000b8950

Interesting.  It's trying to execute code in userspace at 0xc000'0000 and
up.  This is very strange.  I'll investigate this today.

-- 
Revolutions do not require corporate support.

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Jes Sorensen]

>>>>> "Nathan" == Neulinger, Nathan <nneul@umr.edu> writes:

Nathan> But for testing, I've just tried very simple programs - cal,
Nathan> yes, echo, and a staticly linked hello world. All of them
Nathan> segfault right after execution.

Nathan> As far as I know, the only system calls out of the ordinary
Nathan> the flexlm stuff should use are file locking and perhaps some
Nathan> of the ioctl stuff for getting interface address, but I
Nathan> figured on tracing to find out exactly what it's doing.

I really don't think it should be that much work to get the basics
going, it might be a bit more to get more advanced applications up and
running.

I don't know anything really about the implementation of flexlm, but
I'd say it should be doable with a limited effort. Time to get out
that compiler and debugger and get hacking ;-)

Cheers
Jes

RE: [parisc-linux] Status of hp-ux binary compatibility?

[Neulinger, Nathan]

> I really don't think it should be that much work to get the basics
> going, it might be a bit more to get more advanced applications up and
> running.
> 
> I don't know anything really about the implementation of flexlm, but
> I'd say it should be doable with a limited effort. Time to get out
> that compiler and debugger and get hacking ;-)
> 
> Cheers
> Jes

As far as I know, flexlm should be a pretty basic socket server, nothing
fancy, other than it's retrieval of the HP host id, which I presume is just
a syscall of some sort. Unfortunately, it does require use of shared lib
libc since that's how it was linked. 

-- Nathan 

Re: [parisc-linux] Status of hp-ux binary compatibility?

[Jes Sorensen]

>>>>> "Nathan" == Neulinger, Nathan <nneul@umr.edu> writes:

Nathan> As far as I know, flexlm should be a pretty basic socket
Nathan> server, nothing fancy, other than it's retrieval of the HP
Nathan> host id, which I presume is just a syscall of some
Nathan> sort. Unfortunately, it does require use of shared lib libc
Nathan> since that's how it was linked.

Hmmm, admittedly I know little about the HPUX shared linker. However
even then I don't think it should be too big of a problem, provided
Linux/HPPA is able to map binaries at the same place as HPUX.

Jes