[Search for users]
[Overall Top Noters]
[List of all Conferences]
[Download this site]
Title: | DECmcc user notes file. Does not replace IPMT. |
Notice: | Use IPMT for problems. Newsletter location in note 6187 |
Moderator: | TAEC::BEROUD |
|
Created: | Mon Aug 21 1989 |
Last Modified: | Wed Jun 04 1997 |
Last Successful Update: | Fri Jun 06 1997 |
Number of topics: | 6497 |
Total number of notes: | 27359 |
2726.0. "Phase V AM ACCVIO bugs and workaround question" by SWORD1::ES (Eugene Shvartsman) Wed Apr 08 1992 15:43
In our functional module we are calling Phase 5 AM module through
mcc_call_access.
When we have called it with out_p passed as static descriptor:
mcc_w_maxstrlen 6
mcc_b_dtype: 14
mcc_b_class: 1
mcc_a_pointer: 2150740
mcc_w_curlen: 6
mcc_b_flags: 0
mcc_b_ver: 1
mcc_l_id: 0
mcc_l_dt: 127
mcc_a_link: 0
the result has been:
%DEBUG-I-DYNIMGSET, setting image MCC_KERNEL_SHR
%DEBUG-I-DYNMODSET, setting module MCC_AES
%SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=03268065, PC
=000AB774, PSL=03C00000
break on unhandled exception at MCC_AES\mcc_aes_copy\%LINE 11214+4
and
DBG> SHOW CALLS
*MCC_AES mcc_aes_copy 11214 000000C0 000AB774
00000000 004257B2
00000000 00425D46
00000000 00422F20
MCC_FW_THREADS 00000000 000AA795
SHARE$CMA$RTL 00000000 00083FC1
SHARE$CMA$RTL 00000000 0007E9B8
SHARE$CMA$OPEN_RTL
00000000 0008DFB1
MCC_FW_THREADS 00000000 000AA89D
MCC_DISPATCH 00000000 000F1B2C
MCC_IM 00000000 000B3483
MCC_IM 00000000 000B31E3
MCC_IM 00000000 000B3139
MCC_IM 00000000 000B2B7B
MCC_IM 00000000 000B262C
MCC_IM 00000000 000B2453
SHARE$MCC_PTT ......
--------------------------------------------------------------------------------
When we have called it with out_p passed as dynamic descriptor:
mcc_w_maxstrlen 6
mcc_b_dtype: 14
mcc_b_class: 2
mcc_a_pointer: 2150740
mcc_w_curlen: 6
mcc_b_flags: 0
mcc_b_ver: 1
mcc_l_id: 0
mcc_l_dt: 127
mcc_a_link: 0
the result has been:
%SYSTEM-F-ACCVIO, access violation, reason mask=01, virtual address=020E000E, PC
=0000FEBC, PSL=03C00000
break on unhandled exception at SHARE$VAXCRTL+39100
and
DBG> SHOW CALLS
SHARE$VAXCRTL 00000000 0000FEBC
SHARE$CMA$LIB_SHR
00000000 00070A7E
SHARE$CMA$OPEN_LIB_SHR
00000000 0008F319
SHARE$MCC_KERNEL_SHR
00000000 000B2114
00000000 004256D1
00000000 00425D46
00000000 00422F20
SHARE$MCC_KERNEL_SHR
00000000 000AA795
SHARE$CMA$RTL 00000000 00083FC1
SHARE$CMA$RTL 00000000 0007E9B8
SHARE$CMA$OPEN_RTL
00000000 0008DFB1
SHARE$MCC_KERNEL_SHR
00000000 000AA89D
SHARE$MCC_KERNEL_SHR
00000000 000F1B2C
SHARE$MCC_KERNEL_SHR
00000000 000B3483
SHARE$MCC_KERNEL_SHR
00000000 000B31E3
SHARE$MCC_KERNEL_SHR
00000000 000B3139
SHARE$MCC_KERNEL_SHR
00000000 000B2B7B
SHARE$MCC_KERNEL_SHR
00000000 000B262C
SHARE$MCC_KERNEL_SHR
00000000 000B2453
PIL_MCC_CALL 00000000 00321BCF
.....
On another hand we have managed to call successfully this module with
the dynamic or static descriptors with mcc_w_maxstrlen = 512.
Both examples above are clearly bugs. What we need now until these bugs will be
fixed the workaround with the safe value for mcc_w_maxstrlen. What value should
be safe?
BTW, we are running DECmcc (X1.2.18 - cma BL9).
You advise on this matter will be greatly appreciated.
Thank you,
Gene
T.R | Title | User | Personal Name | Date | Lines |
---|
2726.1 | workaround | TOOK::PURRETTA | | Wed Apr 08 1992 18:54 | 14 |
| I received mail from someone in your group regarding this problem
and have sent back a reply. For the benefit of other readers I'll
summarize my response.
It's not clear from your SHOW CALLS where the accvio is coming from.
If you think it's in the DNA5 AM you should:
DBG> Set image mcc_dna5_am
DBG> show calls
There is a problem in the DNA5 AM out_p validation routine which may be
related to this and I've generated a QAR against the DNA5 AM so the problem
gets tracked. For a workaround, please call the AM using a static descriptor
with at least 1024 bytes of memory allocated. We know that works.
|
2726.2 | | SWORD1::ES | Eugene Shvartsman | Thu Apr 09 1992 15:06 | 6 |
| Thank you.
I'd like to mention that it works with dynamic descriptor with length
1024, because that how FCL is calling it now.
Gene
|