| Title: | VAX and Alpha VMS |
| Notice: | This is a new VMSnotes, please read note 2.1 |
| Moderator: | VAXAXP::BERNARDO |
| Created: | Wed Jan 22 1997 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 703 |
| Total number of notes: | 3722 |
OpenVMS V6.2, alpha, $crmpsc system service Could someone explain exactly what the sec$m_no_overmap flag does for the $crmpsc system service. The reason I ask has to do with a customer issue. They are creating large global sections, and mapping them to pre-defined psects. The customer claims that if they link the image with the page alignment forced on the psects, calling $crmpsc results in overmapping the specified psects and corrupting local data. I have already told them that this is expected, and to use the solitary attribute instead. Using the solitary attribute does resolve the problem, but results in a larger virtual memory size of the image (see note preceding this one), so the customer tried going back to forcing page alignment and using the sec$m_no_overmap flag on the call to $crmpsc. He claims that this resolves the problem. Should it? Is using this flag a valid alternative to specifying the solitary attribute? Thanks, Jerry
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 518.1 | use expreg and pointers, if available | XDELTA::HOFFMAN | Steve, OpenVMS Engineering | Wed Apr 23 1997 18:34 | 4 |
The customer appears to be trying to (re)implement a COMMON. What is the language in use? (There are certainly various good alternatives for languages that support pointers.) | |||||
| 518.2 | AUSS::GARSON | DECcharity Program Office | Wed Apr 23 1997 23:49 | 14 | |
re .0
Trying to land a global section over the top of the right variables as
allocated by a compiler is fraught with danger. As .1 implies, the
customer should reconsider their approach.
If this is what the customer is trying to do then SEC$M_NO_OVERMAP
won't help i.e. they really are trying to overmap.
64K alignment on the PSECT should be workable.
Extreme care must be taken in computing the inadr if one is attempting
to overmap compiler allocated variables (as well as using the right
compiler and linker incantations).
| |||||
| 518.3 | Need alignment "at both ends" | WIBBIN::NOYCE | Pulling weeds, pickin' stones | Thu Apr 24 1997 18:16 | 3 |
> 64K alignment on the PSECT should be workable.
... as long as you also pad following the used data to the following
64K boundary. (This is what SOLITARY is supposed to do automatically.)
| |||||