Casting in 64 bit Solaris 28 Forte 6

I have the following piece of code
char crap[10000];
if (rand() % 2 == 0) // really try fooling daemon
crap[0] = 0;
crap[1] = 0;
crap[2] = 0;
crap[3] = size;
crap[4] = 0;
crap[5] = 0;
crap[6] = 0;
crap[7] = rand() % 4;
i = 8;
cout << " set size " << size << " " << crap[3] << endl;
srand(time(0));
for (; i < 10000; i++)
crap[i] = rand() % 256;
//cores here
cout << " sending the daemon len " << *(unsigned long*) crap << " type " << *(long*)(crap + 4) << endl;
The second cout cores on the casting part of *(unsigned long*) crap .
Error for the core is -->terminated by signal BUS (invalid address alignment)
Now this doesnt happen in 32 bit, so why does it happen in 64 bit?? Any ideas or places where i can read up on it and find a fix?
Thanks
T

According to the C and C++ standards, if you cast a pointer to a different type and dereference, the results are undefined. If your code ever works, it is only by accident.
An array of char can be allocated at any location. On SPARC in particular, larger types have stricter alignment requirements. You are depending on an element of a char array being aligned properly for a type that has stricter alignment requirements, a dependecy that is certain to fail on one or more platforms.
In this particular case, the char array is not local to a function, and so is likely to be aligned on a 4-byte boundary, so the address crap+4 is also likely to be 4-byte aligned.
In 32-bit mode, type long is 32 bits, and must be aligned on a 4-byte boundary. If crap is on a 4-byte boundary, the code *(long*)(crap+4) will work.
In 64-bit mode, type long is 64 bits, and must be aligned on an 8-byte boundary. The address crap+4 has at best a 50% chance of being 8-byte aligned, and if it is not, you get a runtime failure trying to execute *(long*)(crap+4).
The best solution is not to use a char array to store integer data; use an array of the appropriate type. If you have mixed data types, use a struct, an array of structs, or an array of unions. The compiler will ensure that everything is properly aligned, no matter what compiler or platform you use.
If you must store integer data in a char array, you can use
#pragma align
to force the array to an appropriate alignment. If you need to compile in 64-bit mode, force the array to 8-byte alignment, and use an offset that is a multiple of 8 for the location of the long.
If you can't do any of these things, you have a few other options.
Compilier option -misalign tells the compiler to assume that data might not be aligned correctly, and to load one byte at a time if it can't be sure about alignment. This option kills performance, sometimes to an unacceptable degree. If you must use it, confine all the misaligned access to one module that does little besides access the data, and compile only that module with -misalign. This module must return data by value, not by pointer or reference, because a pointer or reference might point to misaligned data.
You are using an outdated and unsupported compiler. If you upgrade to a recent version, you can also use the -xmemalign option, which allows the compiler to be less pessimistic about alignment.
I recommend upgrading to Sun Studio 11, which is free. You can get it here:
http://developers.sun.com/sunstudio/

Similar Messages

  • Re: Crystal Reports 5.0 (32-Bit) and Forte 2.0H16 on NT3.51

    Dear Silke,
    Crystal's default for identifying a table in the report design is
    <schema>.<table>. You can change that by going to the Database Location menu
    item and cutting the schema prefix off. You should do that for every table in
    the report. The PESetNthTableLogOnInfo function will, I believe, not change
    the schema prefix of a table, since it assumes you really want that schema,
    i.e., as an override. Also, if you are using the propagateAcrossAllTables
    feature of the function make sure that every table has the schema prefix cut
    off, otherwise those tables will be skipped, since the propagate feature applies
    to all other similar tables only. For example, some of the tables could be
    Sybase, etc.
    I believe you can change the default behaviour of the CR Report Designer in the
    File->Preferences menu, so that when reports are designed the schema prefix is
    not automatically attached.
    I have compiled some guidelines on designing reports to make them work properly
    with Forte. I will try to send them to you next week. Crystal Reports and its
    C API are a good product, and you will get good performance once the report
    design is set up properly.
    Yours very sincerely, John Hodgson.
    Schulte, Silke (DVD) wrote:
    The problem seems to be, that Crystal stores the database name in the
    report. If you want to change the database you have to develop your
    report new, although you give the connection-information via the API.
    The hotline from Crystal told us, when using subreports you have to use
    the PESetNthTableLogOnInfo and to go through all the subreports and
    change the connect-string.
    Our experience is, you have to do both: change the report and call the
    other function for the subreports.
    Thank you very much for all the tips.
    Regards
    Silke Schulte
    Union-Investment, Frankfurt / Main, Germany
    e-mail: [email protected]
    From: Ajith Kallambella M[SMTP:[email protected]]
    Sent: Mittwoch, 23. September 1998 15:17
    To: [email protected]; ForteUsers; [email protected]
    Subject: RE: Crystal Reports 5.0 (32-Bit) and Forte 2.0H16 on NT
    3.51
    Hi There,
    Crystal reports stores Database name as a hard-coded
    value in the
    report template. So if you have to use an other
    database, you will
    have to open the report template and modify the
    database name. I
    think this is why you are getting Crystal API error.
    Hope this helps!
    Ajith Kallambella M
    Forte systems Engineer,
    International Business Corporation.
    Hello,
    I wonder if we have a similar problem.
    We are using Crystal Reports 5 and Forte 3.0.F.2. For
    the interface
    between Forte and Crystal we have in Forte a Lib using
    C-wrappering to
    the C-API of Crystal.
    We have developed a report on a database in our
    development-environment. This report comes up with data
    and works
    correct. It's no different if we have saved the report
    in Crystal with
    data or without.
    If we want to use the same report on another database
    in our
    test-environment the window from crystal reports comes
    up and when
    starting there is an error from the Crystal API
    function.
    Both environments use the same version of an Oracle
    Database and are on
    Open VMS. The client for the test is in this case
    always the same.
    From the C-API we use the method PEStartPrintJob after
    we have specified
    the rest for the report.
    Is there a possibility that this is the same error you
    mentioned? Or do
    you have any other ideas?
    Any help is appreciated.
    Regards
    Silke Schulte
    Union-Investment Frankfurt (Germany)
    email: [email protected]
    From: Katie Carty[SMTP:[email protected]]
    Reply To: Katie Carty
    Sent: Montag, 21. September 1998 19:14
    To: 'Tim Hagemann'; [email protected]
    Subject: RE: Crystal Reports 5.0 (32-Bit) and Forte2.0H16 on NT
    3.51
    Tim,
    You need to upgrade to CrystalReports 6.0. There isa bug in the
    interaction between Forte and CrystalReports 5.0 (Idon't know on
    which side the bug resides, but I think that it is onthe Crystal end,
    since the problem occurs both through C-wrapperingand OLE
    interaction), which does not allow you to invoke thePrintReport
    method unless you have saved data with the report.
    As far as I know, there is no workaround, and theonly solution is to
    upgrade to CrystalReports 6.0.
    FYI, there is no ActiveX control supplied withCrystalReports 5.0 -
    you are actually using standard OLE integration.
    Hope this helps...
    Regards,
    -Katie
    -----Original Message-----
    From: Tim Hagemann [SMTP:[email protected]]
    Sent: Monday, September 21, 1998 11:27 AM
    To: [email protected]
    Subject: Crystal Reports 5.0 (32-Bit) and Forte2.0H16 on NT 3.51
    Hi Forte-User !
    We got a big problem here using the crystal active-xcontrol. On my
    windows-nt 4.0 machine (with
    crystal reports installed completely) all works well.On our test
    machine (with only the crystal ActiveX
    installed), the Forte runtime shuts down afterinvoking a special
    method of the ocx (PrintReport).
    The cause for shutting down (no stacktrace orsomething like that) is
    a generated c++ exception in crystal -
    but we don't know the cause of this exception.
    Has anyone made the control work under nt 3.51 ? Doesanybody know
    (exactly) what dll's we'll have
    to use ?
    TIA,
    Tim Hagemann
    Tim Hagemann
    Ascom GmbH Email:
    [email protected]
    Charlottenburger Allee 61 Phone: +49241 96806 273
    D-52068 Aachen Fax: +49241 96806 225
    >
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive
    <URL:http://pinehurst.sageit.com/listarchive/>
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive
    <URL:http://pinehurst.sageit.com/listarchive/>
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive
    <URL:http://pinehurst.sageit.com/listarchive/>
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive <URL:http://pinehurst.sageit.com/listarchive/>-
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive <URL:http://pinehurst.sageit.com/listarchive/>

    Dear Silke,
    Crystal's default for identifying a table in the report design is
    <schema>.<table>. You can change that by going to the Database Location menu
    item and cutting the schema prefix off. You should do that for every table in
    the report. The PESetNthTableLogOnInfo function will, I believe, not change
    the schema prefix of a table, since it assumes you really want that schema,
    i.e., as an override. Also, if you are using the propagateAcrossAllTables
    feature of the function make sure that every table has the schema prefix cut
    off, otherwise those tables will be skipped, since the propagate feature applies
    to all other similar tables only. For example, some of the tables could be
    Sybase, etc.
    I believe you can change the default behaviour of the CR Report Designer in the
    File->Preferences menu, so that when reports are designed the schema prefix is
    not automatically attached.
    I have compiled some guidelines on designing reports to make them work properly
    with Forte. I will try to send them to you next week. Crystal Reports and its
    C API are a good product, and you will get good performance once the report
    design is set up properly.
    Yours very sincerely, John Hodgson.
    Schulte, Silke (DVD) wrote:
    The problem seems to be, that Crystal stores the database name in the
    report. If you want to change the database you have to develop your
    report new, although you give the connection-information via the API.
    The hotline from Crystal told us, when using subreports you have to use
    the PESetNthTableLogOnInfo and to go through all the subreports and
    change the connect-string.
    Our experience is, you have to do both: change the report and call the
    other function for the subreports.
    Thank you very much for all the tips.
    Regards
    Silke Schulte
    Union-Investment, Frankfurt / Main, Germany
    e-mail: [email protected]
    From: Ajith Kallambella M[SMTP:[email protected]]
    Sent: Mittwoch, 23. September 1998 15:17
    To: [email protected]; ForteUsers; [email protected]
    Subject: RE: Crystal Reports 5.0 (32-Bit) and Forte 2.0H16 on NT
    3.51
    Hi There,
    Crystal reports stores Database name as a hard-coded
    value in the
    report template. So if you have to use an other
    database, you will
    have to open the report template and modify the
    database name. I
    think this is why you are getting Crystal API error.
    Hope this helps!
    Ajith Kallambella M
    Forte systems Engineer,
    International Business Corporation.
    Hello,
    I wonder if we have a similar problem.
    We are using Crystal Reports 5 and Forte 3.0.F.2. For
    the interface
    between Forte and Crystal we have in Forte a Lib using
    C-wrappering to
    the C-API of Crystal.
    We have developed a report on a database in our
    development-environment. This report comes up with data
    and works
    correct. It's no different if we have saved the report
    in Crystal with
    data or without.
    If we want to use the same report on another database
    in our
    test-environment the window from crystal reports comes
    up and when
    starting there is an error from the Crystal API
    function.
    Both environments use the same version of an Oracle
    Database and are on
    Open VMS. The client for the test is in this case
    always the same.
    From the C-API we use the method PEStartPrintJob after
    we have specified
    the rest for the report.
    Is there a possibility that this is the same error you
    mentioned? Or do
    you have any other ideas?
    Any help is appreciated.
    Regards
    Silke Schulte
    Union-Investment Frankfurt (Germany)
    email: [email protected]
    From: Katie Carty[SMTP:[email protected]]
    Reply To: Katie Carty
    Sent: Montag, 21. September 1998 19:14
    To: 'Tim Hagemann'; [email protected]
    Subject: RE: Crystal Reports 5.0 (32-Bit) and Forte2.0H16 on NT
    3.51
    Tim,
    You need to upgrade to CrystalReports 6.0. There isa bug in the
    interaction between Forte and CrystalReports 5.0 (Idon't know on
    which side the bug resides, but I think that it is onthe Crystal end,
    since the problem occurs both through C-wrapperingand OLE
    interaction), which does not allow you to invoke thePrintReport
    method unless you have saved data with the report.
    As far as I know, there is no workaround, and theonly solution is to
    upgrade to CrystalReports 6.0.
    FYI, there is no ActiveX control supplied withCrystalReports 5.0 -
    you are actually using standard OLE integration.
    Hope this helps...
    Regards,
    -Katie
    -----Original Message-----
    From: Tim Hagemann [SMTP:[email protected]]
    Sent: Monday, September 21, 1998 11:27 AM
    To: [email protected]
    Subject: Crystal Reports 5.0 (32-Bit) and Forte2.0H16 on NT 3.51
    Hi Forte-User !
    We got a big problem here using the crystal active-xcontrol. On my
    windows-nt 4.0 machine (with
    crystal reports installed completely) all works well.On our test
    machine (with only the crystal ActiveX
    installed), the Forte runtime shuts down afterinvoking a special
    method of the ocx (PrintReport).
    The cause for shutting down (no stacktrace orsomething like that) is
    a generated c++ exception in crystal -
    but we don't know the cause of this exception.
    Has anyone made the control work under nt 3.51 ? Doesanybody know
    (exactly) what dll's we'll have
    to use ?
    TIA,
    Tim Hagemann
    Tim Hagemann
    Ascom GmbH Email:
    [email protected]
    Charlottenburger Allee 61 Phone: +49241 96806 273
    D-52068 Aachen Fax: +49241 96806 225
    >
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive
    <URL:http://pinehurst.sageit.com/listarchive/>
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive
    <URL:http://pinehurst.sageit.com/listarchive/>
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive
    <URL:http://pinehurst.sageit.com/listarchive/>
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive <URL:http://pinehurst.sageit.com/listarchive/>-
    To unsubscribe, email '[email protected]' with
    'unsubscribe forte-users' as the body of the message.
    Searchable thread archive <URL:http://pinehurst.sageit.com/listarchive/>

  • VXIMove returns error 0x8000 only in 64-bit Solaris 8

    We are looking to migrate our Solaris CPU to a faster Blade 150 (or better) PCI-based workstation. The downside is that these latest workstations all contain 64-bit  processors so we are forced to run a 64-bit Solaris kernel. According to the NI-VXI documentation, the driver works on 32-bit and 64-bit kernels so I was not too concerned. Unfortunately our applications are not working properly when using the VXImove API call to move a chunk of data from VXI to a local memory address. The error is always 0x8000.
    We eliminated Solaris 8 as the problem by duplicating the setup on an Ultra 10 workstation which is also PCI-based but runs 32-bit Solaris 8. The same drivers were used in both cases. The VXImove works properly when running in Solaris 8 32-bit.
    I have attached my resman output  files for reference. Thanks in advance.
    Attachments:
    resman.err.txt ‏1 KB
    resman.out.txt ‏6 KB
    resman.tbl.txt ‏1 KB

    Hello All,
    I would like to clarify one of the statements made earlier in this forum.  NI VXI is not supported as a 64-bit application, under any circumstances.  One can user NI VXI on a 64-bit system, but they must run it as a 32-bit application in order for it to function correctly, and therefore lose the 64-bit processing capability. 
    It is true that NI-VISA is simply an added layer between the driver and your application.  As KnowledgeBase 2QDASEPG: Supported NI-VISA Operating Systems and Features points out, NI-VISA only supports VXI functionality on 32-bit operating systems.  This is because the NI VXI driver itself is only supported as a 32-bit application.
    Chris_G
    Sr Test Engineer
    Medtronic, Inc.

  • How can I use syscall() from 32-bit application in 64-bit Solaris (v. 5.8)

    Solaris 5.8
    Sun Blade 150 (64-bit)
    I have got SYGSYS when call syscall(...) from my 32-bit application.
    Is it possible to use syscall() in 32-bit applications under 64-bit Solaris?

    Hello again.
    You may use the debugger (mdb) to open the core file generated by the SIGSYS trap:mdb coreThen you type::regs to list the register values at the time of the trap (in hexadecimal form).
    For Sparc machines (it is more complicated on x86 machines) g1 holds the system call number, o0 holds the first argument, o1 the second one ... o5 the 6th one. (For system calls with more than 6 arguments the 7th, 8th... argument are stored on the stack.)
    If g1 is 0 you know that the system call is done indirectly using SYS_syscall so the arguments are shifted (see above) and o0 holds the system call number.
    You can exit mdb using::quitMartin

  • Where is the 64-bit download for oracle_proxy.so for 64-bit Solaris ?

    Hi, Experts,
    We tried to start the 64-bit Sun ONE web server in 64-bit Solaris env. with the Oracle AS plugin,
    oracle_proxy.so library. But then hit below error,
    [06/Jul/2011:15:44:13] failure (12345): CORE2253: Error running Init
    function load-modules: dlopen of /xyz/proxy/oracle_proxy.so failed (ld.so.1:
    webservd: fatal: /xyz/proxy/oracle_proxy.so:
    wrong ELF class: ELFCLASS32)
    We tried to search from
    http://www.oracle.com/technetwork/middleware/ias/downloads/index.html
    Oracle 10gR3 Companion (10.1.3.x) CD
    And downloaded AS101330_companionCD_sparc64_disk1.cpio however the file
    is still 32 BIT file :
    xxxyyy$ file oracle_proxy.so
    oracle_proxy.so: ELF 32-bit MSB dynamic lib SPARC Version 1,
    dynamically linked, not stripped
    Where can I get a 64-bit version of this above file, oracle_proxy.so ?
    Walter

    Here's a download location:
    http://support.apple.com/kb/DL1577

  • Where is the 64-bit download for AS oracle_proxy.so for 64-bit Solaris ?

    Hi, Experts,
    This is a cross-posting of
    Where is the 64-bit download for oracle_proxy.so for 64-bit Solaris ?
    (Forum Home » Application Server » Oracle Application Server - General )
    in the download forum.
    Where can I get a 64-bit version of this above file, oracle_proxy.so in downloads ?
    Walter

    Here's a download location:
    http://support.apple.com/kb/DL1577

  • 64 BIT SOLARIS ON INTEL

    When is the 64 BIT SOLARIS ON INTEL IS EXPECTED to be released
    Does Solaris on INtel Support clustering.If yes how how many
    machines in a cluster
    Thanks
    Noby([email protected])

    When is the 64 BIT SOLARIS ON INTEL IS EXPECTED to be released
    Does Solaris on INtel Support clustering.If yes how how many
    machines in a cluster
    Thanks
    Noby([email protected])

  • Hyperion Planning on 64 bit Solaris. ODI on 32 bit Windows.

    Hi,
    I have Hyperion planning on 64 bit Solaris and ODI installed on 32 bit windows server. Could not reverse a planning app in ODI.
    The error message says - Could not connect to Planning instance.
    I have checked the RMI service at port 11333 on solaris and its running. I can telnet to this port from the windows server successfully.
    Has any one come across such an issue before or knows of any issues regarding this?
    Any help will be greatly appreciated.
    Thanks.

    I had a problem with ODI being on linux and planning on windows.
    I could telnet the ping / telnet across between boxes. When reversing it could connect to planning but could not return, it turned out to be an issue with loopback on linux that had an extra ip of 10.0.0.0
    I removed the loopback from the host and it started reversing without a problem.
    If you telnet then maybe it is some sort of routing host file issue.
    Cheers
    John
    http://john-goodwin.blogspot.com/

  • Display Postscript Toolkit (dpstk library) for 64 bit Solaris

    Does anyone know whether the
    DPS Toolkit (dpstk library) is available for 64 bit Solaris operating systems?

    Thanks for getting me to the troublemaker ArielK. 
    I can confirm disabling APSDaemon.exe, aka, new Apple Push for WiFi Sync and renaming (or deleting if you are more risk oriented) APSDaemon.exe and APSDaemon_main.dll in this folder "C:\Program Files (x86)\Common Files\Apple\Apple Application Support" resolved my issue. 
    Alternatively, another slightly more annoying but less techy workaround for those nervous about changing filenames is to start iTunes in Safe Mode every time.  You must do it via Programs > iTunes > hold down Ctrl-Shift then click iTunes (does not work if you do it by clicking a short cut in programs list or taskbar, whatever...).  Maybe this disables WiFi Sync by default, dunno.  Don't care at this point, but it seemed to work.
    I can live without WiFi Sync for now, but its a nice feature I would really like to use in the future.
    Apple please fix!!!
    Running: Windows 7 Home Premium x64, iTunes 10.5.2.11.

  • Oracle 10g installation in an SPARC 64 bit Solaris 10 server

    Good morning,
    I'm trying to install Oracle 10g Enterprise Edition in a server with Solaris 10 as operating system. When I run the oui, the prerequisites part is passed but the installation gives me an error telling that the installer can't write in the /tmp directory. Nevertheless, I have writing grants and space enough in that directory, in fact the installer writes the installAction.log in this directory. So I don't understand which can be the problem?
    Any suggestions?
    Thanks in advance.
    Regards,
    Mabel

    It could be you don't have enough free space on the tmp dir.
    Check this point from the
    Oracle® Database Installation Guide. 10g Release 2 (10.2) for Solaris Operating System (SPARC 64-Bit). Part Number B15690-02
    If you determined that the /tmp directory has less than 400 MB of free disk space, then identify a file system with at least 400 MB of free space and set the TMP and TMPDIR environment variables to specify a temporary directory on this file system:
    1. Use the df -k command to identify a suitable file system with sufficient free space.
    2. If necessary, enter commands similar to the following to create a temporary directory on the file system that you identified, and set the appropriate permissions on the directory:
    $ su - root
    # mkdir /mount_point/tmp
    # chmod a+wr /mount_point/tmp
    # exit
    3. Enter commands similar to the following to set the TMP and TMPDIR environment variables:
    Bourne, Bash, or Korn shell:
    $ TMP=/mount_point/tmp
    $ TMPDIR=/mount_point/tmp
    $ export TMP TMPDIR
    C shell:
    % setenv TMP /mount_point/tmp
    % setenv TMPDIR /mount_point/tmp
    ~ Madrid.

  • How to get internals of FILE * on 64-bit solaris

    Hi,
    We are in the process of migrating our applications to 64-bit operating system. We have some code which was using the internal members of a FILE structure. These members were exposed in the 32-bit headers but they are opaque in 64-bit. Anyway of finding out the equivalent structure fields for the 64-bit headers.
    Here is how the structure is defined in stdio_impl.h on Solaris 7.
    32-bit:
    ========
    typedef struct __FILE_TAG FILE;
    struct __FILE_TAG {
    ssize_t _cnt;
    unsigned char _ptr;
    unsigned char _base;
    unsigned char _flag;
    unsigned char _file;
    unsigned __orientation:2;
    unsigned __filler:6;
    64-bit
    =======
    struct __FILE_TAG {
    long __pad[16];
    Thanks for any help.
    -Chandra

    I'm not sure about Solaris 7. On Solaris 8 you should be able to use __fbufsize() and related functions (see the man page).

  • Is there a way to run Windows programs on an Ultra SPARC 64-Bit Solaris 10?

    I know the answer is more than likely going to be no, but I just thought I would ask...
    I have an Sparc Ultra-60 running Solaris 10 64-Bit sitting here that is a web / forum / ftp server, and I would like to run a Telnet BBS program on it that is a Windows x86 based software.
    I know WINE will run on OpenSolaris, but I don't know and I have searched high and low on the internet for the answer, so I came to the experts! :)
    Have a great weekend.!
    --- Charles!
    Edited by: 808111 on Nov 7, 2010 3:04 PM

    Hello Charles,
    I have an Sparc Ultra-60 running Solaris 10 64-Bit sitting here that is a web / forum / ftp server, and I would like to run a Telnet BBS program on it that is a Windows x86 based software.
    you might be able to get one of these SunPCi add-in boards. These are boards with thieir own CPU, memory, sound circuty and graphics chip (shared memory), later ones with network port but no own mass storage (the disk is emulated by a file on the Solaris filesystem). A full copy of the supported Windows operating system (the last generation supported Windows XP) is required (the installer depends on the layout of the retail copy). Depending on the ages/version of the board the CPU range from AMD K5 300/400 MHz, Celeron 600/700 to AMD K6 with over 1 GHz.
    Unfortunately the first 2 generations (SunPCi and SunPCi II/Pro) don't directly support Solaris 10 (a file from the SunWSPCi package had to be modified to get the software installed) and the last generation SunPCi III/Pro requires a patch to work in 2010. The SunWSPCi packages that came with the boards are outdated and Oracle removed the packages from the download center.
    The documentation for these boards is still available on docs.sun.com (enter the following in google "SunPCi +site:docs.sun.com").
    Michael

  • Porting 32-bit Applications on 64-bit Solaris 9i

    Hi All,
    I tried to port an existing and well set 32-bit application onto Solaris 9.While I was compling the C-code part of it,I encountered the error
    wrong ELF class: ELFCLASS64.
    which I realized was due to intermixing of 32 and 64-bit applications.
    I just wanted to know if any oracle 10g libraries called or linked with the application happen to be of 32-bit ELFCLASS as it's only while linking with them that the error is thrown.
    cc -L/u01/app/oracle/product/10.1.0.3/lib/ -L/u01/app/oracle/product/10.1.0.3/rdbms/lib -o cps_lib_test cps_lib_test.o -lclntsh `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lnro10 `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lclient10 -lnnetd10 -lvsn10 -lcommon10 -lgeneric10 -lmm -lcore10 -lxml10 -lunls10 -lsnls10 -lnls10 -lcore10 -lnls10 `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lnro10 `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lclient10 -lnnetd10 -lvsn10 -lcommon10 -lgeneric10 -lcore10 -lxml10 -lunls10 -lsnls10 -lnls10 -lcore10 -lnls10 -lclient10 -lnnetd10 -lvsn10 -lcommon10 -lgeneric10 -lcore10 -lxml10 -lunls10 -lsnls10 -lnls10 -lcore10 -lnls10 `cat /u01/app/oracle/product/10.1.0.3/lib/sysliblist` -R/u01/app/oracle/product/10.1.0.3/lib -laio -lposix4 -lkstat -lm -lthread cps_lib.so
    ld: fatal: file cps_lib_test.o: wrong ELF class: ELFCLASS64
    ld: fatal: File processing errors. No output written to cps_lib_test
    *** Error code 1
    make: Fatal error: Command failed for target `build'
    Can someone please help?

    Sorry for the i attached to Solaris 9.
    The output of file cps_lib_test.o is ELF 64-bit MSB relocatable SPARC Version 1.
    Almost all the resulting files are of the type ELFCLASS64.But probably some of the oracle libraries while linking are causing the problem.

  • Porting 32-bit Applications on 64-bit Solaris 9

    Hi All,
    I tried to port an existing and well set 32-bit application onto Solaris 9.While I was compling the C-code part of it,I encountered the error
    wrong ELF class: ELFCLASS64.
    which I realized was due to intermixing of 32 and 64-bit applications.
    I just wanted to know if any oracle 10g libraries called or linked with the application happen to be of 32-bit ELFCLASS as it's only while linking with them that the error is thrown.
    cc -L/u01/app/oracle/product/10.1.0.3/lib/ -L/u01/app/oracle/product/10.1.0.3/rdbms/lib -o cps_lib_test cps_lib_test.o -lclntsh `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lnro10 `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lclient10 -lnnetd10 -lvsn10 -lcommon10 -lgeneric10 -lmm -lcore10 -lxml10 -lunls10 -lsnls10 -lnls10 -lcore10 -lnls10 `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lnro10 `cat /u01/app/oracle/product/10.1.0.3/lib/ldflags` -lnsslb10 -lncrypt10 -lnsgr10 -lnzjs10 -ln10 -lnnz10 -lnl10 -lclient10 -lnnetd10 -lvsn10 -lcommon10 -lgeneric10 -lcore10 -lxml10 -lunls10 -lsnls10 -lnls10 -lcore10 -lnls10 -lclient10 -lnnetd10 -lvsn10 -lcommon10 -lgeneric10 -lcore10 -lxml10 -lunls10 -lsnls10 -lnls10 -lcore10 -lnls10 `cat /u01/app/oracle/product/10.1.0.3/lib/sysliblist` -R/u01/app/oracle/product/10.1.0.3/lib -laio -lposix4 -lkstat -lm -lthread cps_lib.so
    ld: fatal: file cps_lib_test.o: wrong ELF class: ELFCLASS64
    ld: fatal: File processing errors. No output written to cps_lib_test
    *** Error code 1
    make: Fatal error: Command failed for target `build'
    Can someone please help?

    Check your client installation of oracle whether it is 64bit.
    One of your libraries you are linking is 32 bit instead of 64 bit

  • SUn CLuster 3.2 install - scinstall on x86 32 bit Solaris 5.10

    Ok - I have 2 machines with 32-bit x86 SOlaris 5.10 - I installed the Sun Cluster 3.2 software but everytime I try scinstall it says rebooting other node and the other node never brings sun cluster up -
    Questions:
    1. The private interface does not come up on reboot - should it - and should I have an entry for cluster-priv1 in /etc/hosts
    2. I know I am supposed to do a scvx -xv - I then try to enable the cluster service, but everything says disabled
    I have tried this 5 times no luck - I have lots of cluster experience - and can get Oracle CRS working fine
    Any thoughts

    Yeahh, guys!!!
    I was trying to establish a two-node cluster using VirtualBox + Solaris x86 + Sun Cluster 3.2. The node where I was running scinstall to configure my cluster environment was rebooting the other node in the end of the configuration process but it was hanging in the "Rebooting node01..." message just because it was not able to establish the cluster.
    After see your comments, I changed Solaris x86 to Solaris Express Community Edition and Sun Cluster to Cluster Express and now everything is working fine!
    Thanks!
    Jansen Sena <[email protected]>

Maybe you are looking for

  • Key mapping

    Hi All, I have 3 fields in my source file Id,Category (Key Mapping) ,Name. In Import Manager I have mapped Remote Key field with ID . Name with Name category with Category Qualified Range (A 1-10) I have source file like this Id name category 11  10 

  • WLC, Bonjour, Airprint in diffrent SSID

    Hi, I have two wlc 5508 version 7.5.102.0. one foreign and one anchor. the APs are connectet to the foreign wlc.foreign wlc has a tunnel to anchor wlc. On Anchor, I creat the SSIDs and there are breakout to internet. I use an apple iPad and HP MFP 27

  • Problem with Gmail Notifier

    I"m not sure if anyone here can help, but I figured it was worth a try. Most of the time now my Gmail Notifier icon is showing that it's not connected and when I click on it it says 'server certificate not yet valid' When I click this, it say page in

  • Safari sending error message, won't quit or reopen properly

    In the middle of browsing, Safari on my Macbook, running Yosemite, received the following error message. An option to close the application does not appear. When I receive the message, I can click on 'ok' or 'reopen.' 'OK' closes the message and 'reo

  • Do we have constant file in 7.5NW

    Hi All, Do we required to change our Account Dimension names in Constant file in BPC 7.5 NW version? If so where do we need change.  What is the path for that.  And when I am trying to move results to P&L Application it is not working it showing some