Re: [clamav-users] Version 0.98.3 fails on Solaris
On May 13, 2014, at 4:24 AM, Al Varnell alvarn...@mac.commailto:alvarn...@mac.com wrote: On Tue, May 13, 2014 at 01:04 AM, James Lee wrote: (Please don't top post.) Please leave moderation functions to the moderators. There could possibly be a rule preventing it, but I’m unaware of any and there are examples in this thread of Sourcefire contributors top posting. For technical lists, it’s often preferred in order to retain all details. There is no requirement to bottom post (even though I think it to be better). I say, if a threat is already in bottom-post mode, keep doing it. Otherwise, whatever. This is not something to argue about I guess. -- Joel Esler Open Source Manager Threat Intelligence Team Lead Vulnerability Research Team ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 12/05/2014 21:33, Lawrence K. Chen, P.Eng. wrote: So, what's the definitive answer. To what question? (Please don't top post.) The OP's question relating to Can't allocate memory has been answered. James. ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On Tue, May 13, 2014 at 01:04 AM, James Lee wrote: (Please don't top post.) Please leave moderation functions to the moderators. There could possibly be a rule preventing it, but I’m unaware of any and there are examples in this thread of Sourcefire contributors top posting. For technical lists, it’s often preferred in order to retain all details. -Al- -- Al Varnell Mountain View, CA ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 13/05/2014 09:24, Al Varnell wrote: On Tue, May 13, 2014 at 01:04 AM, James Lee wrote: (Please don't top post.) Please leave moderation functions to the moderators. There could possibly be a rule preventing it, but I’m unaware of any and there are examples in this thread of Sourcefire contributors top posting. For technical lists, it’s often preferred in order to retain all details. OK, so what is the question? James. ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
James Lee wrote: On 10/05/2014 17:34, Shawn Webb wrote: Hello, The attached two patches will make building (with gcc) I'm not but.. and running on Solaris work. when using gcc configure sets HAVE_ATTRIB_PACKED=1 and the problem doesn't exist, ie, no patch is needed for gcc. In fact no patches need for cc if the value of HAVE_ATTRIB_PACKED is forced post configure. Any patch/fix should be for configure. Thats interesting. Without Shawns patches, but with HAVE_ATTRIB_PACKED and using Sun-CC it works (freshclam clamscan tested so far). But including the posted patches, the errors are back again. So it looks like a problem with the build system and the patch is (at least partly) compiler specific. Regards, Martin -- Martin Preen, Universität Freiburg, Institut für Informatik Georges-Koehler-Allee 52, Raum EG-006, 79110 Freiburg, Germany phone: ++49 761 203-8250pr...@informatik.uni-freiburg.de fax: ++49 761 203-8242 swt.informatik.uni-freiburg.de/staff/preen smime.p7s Description: S/MIME Cryptographic Signature ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
So, what's the definitive answer. I gave up late Friday night on getting it working on Solaris for the weekend maintenance window. Though the last thing I tried was to force -D__EXTENSIONS__ in...and caught a glimpse of it finally finishing its compile as I was securing my workstation. I also had to do some massaging, because the openssl we build is installed in a non-standard location. Before I had resorted to setting -D__EXTENSIONS__ from reading the headers to find out why definitions weren't being seen I had done some fiddling with some of the source files, but reverting those changes and trying again with just -D__EXTENSIONS__ has compiled all the way through. Now I suppose its to figure out if I need to test it before deploying into production. Had mentioned in the past, that with our Solaris package CM system, I would often just update clamav project's source and go straight to making a release package and immediately distribute it. This is the first time that I'm not sure its release package worthy. Not sure I have somewhere to install a scratch package anymore FWIW, our build server is Solaris/x64 10 Generic_138889-07 (update 6) with gcc 4.1.1, ours is using gnu-as (/usr/sfw/bin/gas) and the system linker (/usr/ccs/bin/ld). On 05/10/14 11:13, James Lee wrote: On 10/05/2014 11:22, James Lee wrote: Indeed rantit's the usual brain dead pile-O-rubbish that is configure and friends/rant. *** simple workaround: CPPFLAGS=... -DHAVE_ATTRIB_PACKED My complete CPPFLAGS are: CPPFLAGS=-D__EXTENSIONS__ -DHAVE_ATTRIB_PACKED On 05/12/14 06:28, Martin Preen wrote: James Lee wrote: On 10/05/2014 17:34, Shawn Webb wrote: Hello, The attached two patches will make building (with gcc) I'm not but.. and running on Solaris work. when using gcc configure sets HAVE_ATTRIB_PACKED=1 and the problem doesn't exist, ie, no patch is needed for gcc. In fact no patches need for cc if the value of HAVE_ATTRIB_PACKED is forced post configure. Any patch/fix should be for configure. Thats interesting. Without Shawns patches, but with HAVE_ATTRIB_PACKED and using Sun-CC it works (freshclam clamscan tested so far). But including the posted patches, the errors are back again. So it looks like a problem with the build system and the patch is (at least partly) compiler specific. Regards, Martin -- Martin Preen, Universität Freiburg, Institut für Informatik Georges-Koehler-Allee 52, Raum EG-006, 79110 Freiburg, Germany phone: ++49 761 203-8250pr...@informatik.uni-freiburg.de fax: ++49 761 203-8242 swt.informatik.uni-freiburg.de/staff/preen ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml -- Who: Lawrence K. Chen, P.Eng. - W0LKC - Sr. Unix Systems Administrator For: Enterprise Server Technologies (EST) -- SafeZone Ally ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
May 8 07:44:31 mailhost freshclam[3924]: [ID 702911 mail.error] Verification: Can't allocate memory Can't allocate memory is the wrong message; it's not short of memory. The problem is failure to initialise openssl by calling OpenSSL_add_all_digests() resulting in EVP_get_digestbyname() returning null. I think it is intended to call this via cl_initialize_crypto(), except it's not called. __attribute__((constructor)) works, compiler tested, but not here. I've changed the #ifdef _WIN32 around cl_initialize_crypto() to add __SunOS and freshclam works. clamd appears to work too. Patch attached. Please don't take this as final, it's just a hint towards the problem. James. --- ../original/clamd/clamd.c 2014-05-06 19:39:56.0 +0100 +++ clamd/clamd.c 2014-05-09 19:27:33.196293494 +0100 @@ -139,7 +139,8 @@ sa.sa_handler = SIG_IGN; sigaction(SIGHUP, sa, NULL); sigaction(SIGUSR2, sa, NULL); -#else +#endif +#if defined(_WIN32) || defined(__SunOS) cl_initialize_crypto(); #endif --- ../original/clamdscan/clamdscan.c 2014-05-06 19:39:56.0 +0100 +++ clamdscan/clamdscan.c 2014-05-09 19:27:33.196546750 +0100 @@ -69,7 +69,7 @@ struct sigaction sigact; #endif -#if defined(_WIN32) +#if defined(_WIN32) || defined(__SunOS) cl_initialize_crypto(); #endif --- ../original/freshclam/freshclam.c 2014-05-06 19:39:56.0 +0100 +++ freshclam/freshclam.c 2014-05-09 19:27:33.196891959 +0100 @@ -302,7 +302,7 @@ if (check_flevel ()) exit (FCE_INIT); -#if defined(_WIN32) +#if defined(_WIN32) || defined(__SunOS) cl_initialize_crypto(); #endif --- ../original/libclamav/others.c 2014-05-06 19:42:41.0 +0100 +++ libclamav/others.c 2014-05-09 19:27:53.246477188 +0100 @@ -1293,7 +1293,7 @@ engine-cb_meta = callback; } -#if !defined(_WIN32) +#if !defined(_WIN32) !defined(__SunOS) __attribute__((constructor)) void init(void) { cl_initialize_crypto(); --- ../original/sigtool/sigtool.c 2014-05-06 19:39:56.0 +0100 +++ sigtool/sigtool.c 2014-05-09 19:27:33.197957626 +0100 @@ -2995,7 +2995,7 @@ if(check_flevel()) exit(1); -#if defined(_WIN32) +#if defined(_WIN32) || defined(__SunOS) cl_initialize_crypto(); #endif ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 10/05/2014 11:22, James Lee wrote: I think it is intended to call this via cl_initialize_crypto(), except it's not called. __attribute__((constructor)) works, compiler tested, but not here. ... Please don't take this as final, it's just a hint towards the problem. Indeed rantit's the usual brain dead pile-O-rubbish that is configure and friends/rant. *** simple workaround: CPPFLAGS=... -DHAVE_ATTRIB_PACKED My complete CPPFLAGS are: CPPFLAGS=-D__EXTENSIONS__ -DHAVE_ATTRIB_PACKED *** Fix Fix configure. configure checks for __attribute__((packed)) and make the wrong decision. configure assumes that if the compiler is not GNU it can't have the extension so doesn't bother checking: #ifdef __GNUC__ struct { int i __attribute__((packed)); } s; #else #error Only checking for packed attribute on gcc-like compilers #endif which is silly because if it's the gnu cc then the answer is already known, it's the non gnu cc that need checking. However moving on, based on the packed check this add to the headers: #ifndef HAVE_ATTRIB_PACKED #define __attribute__(x) #endif which renders all __attributes__ ineffective including the __attribute__((constructor)) which is needed to initialise openssl. For it to work in the general case __attribute__((constructor)) should be checked. James. ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
Hey All, The attached two patches will make building (with gcc) and running on Solaris work. I've also pasted them to the below linked sites in case the attachments don't go through. The patches ought to be applied in order. Patch 1: http://ix.io/ceV (001-clamav-solaris.patch) Patch 2: http://ix.io/ceW (002-clamav-solaris.patch) Thanks, Shawn On Sat, May 10, 2014 at 12:13 PM, James Lee li...@oxdrove.co.uk wrote: On 10/05/2014 11:22, James Lee wrote: I think it is intended to call this via cl_initialize_crypto(), except it's not called. __attribute__((constructor)) works, compiler tested, but not here. ... Please don't take this as final, it's just a hint towards the problem. Indeed rantit's the usual brain dead pile-O-rubbish that is configure and friends/rant. *** simple workaround: CPPFLAGS=... -DHAVE_ATTRIB_PACKED My complete CPPFLAGS are: CPPFLAGS=-D__EXTENSIONS__ -DHAVE_ATTRIB_PACKED *** Fix Fix configure. configure checks for __attribute__((packed)) and make the wrong decision. configure assumes that if the compiler is not GNU it can't have the extension so doesn't bother checking: #ifdef __GNUC__ struct { int i __attribute__((packed)); } s; #else #error Only checking for packed attribute on gcc-like compilers #endif which is silly because if it's the gnu cc then the answer is already known, it's the non gnu cc that need checking. However moving on, based on the packed check this add to the headers: #ifndef HAVE_ATTRIB_PACKED #define __attribute__(x) #endif which renders all __attributes__ ineffective including the __attribute__((constructor)) which is needed to initialise openssl. For it to work in the general case __attribute__((constructor)) should be checked. James. ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 10/05/2014 17:34, Shawn Webb wrote: Hello, The attached two patches will make building (with gcc) I'm not but.. and running on Solaris work. when using gcc configure sets HAVE_ATTRIB_PACKED=1 and the problem doesn't exist, ie, no patch is needed for gcc. In fact no patches need for cc if the value of HAVE_ATTRIB_PACKED is forced post configure. Any patch/fix should be for configure. James. ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On May 8, 2014, at 12:50 PM, Dennis Peterson denni...@inetnw.commailto:denni...@inetnw.com wrote: On 5/8/14, 9:00 AM, Dennis Peterson wrote: On 5/8/14, 8:23 AM, Shawn Webb wrote: Hey Martin, Is there a way you can get to me main.cvd.broken? I'm wondering if the change to OpenSSL for hashing has somehow changed parsing CVDs and CLDs on big-endian machines running Solaris. I thoroughly tested the code on a sparc64 machine (an old SunFire 280r) running FreeBSD 9.2 successfully. To help me debug the issue: what version of OpenSSL do you have installed? Can you give me the output of the clamdconf command (preferably to a pastebin service)? Can you give me (again, pastebin) the output of your config.log? I can install Solaris on this sparc64 machine as early as next week. Thanks, Shawn Are we to understand Sourcefire does not have a proper Solaris Sparc environment for testing ClamAV products? dp My point is asking is I have three Sparc systems (Ultra 10, Ultra 2 dual proc, and Netra X1) that are going to a landfill if I don't find a home for them. All are in perfect working condition. I'll split the shipping cost (US only) if you're interested. OS not included. Understood. I’ll talk to the guys to see if we have a home for them in the lab. We’re currently planning on the expansion of our lab, so it’s a possibility. -- Joel Esler Open Source Manager Threat Intelligence Team Lead Vulnerability Research Team ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
Lars Hecking wrote: I've been building with static openssl for a while as well, and am still using gcc 3.4.6 as I couldn't get newer versions to compile - although it seems possible, and I'll try again; maybe using gcc 4.7 or 4.6. Well, 4.7.3 doesn't build for me. Trying 4.6.4 now. Using 3.4.6 and latest binutils. gcc -c -g -DIN_GCC -W -Wall -Wwrite-strings -Wcast-qual -Wstrict-prototypes -Wmissing-prototypes -Wold-style-definition -fno-common -DHAVE_CONFIG_H -DGENERATOR_FILE -I. -Ibuild -I../../gcc -I../../gcc/build -I../../gcc/../include -I./../intl -I../../gcc/../libcpp/include -I../../gcc/../libdecnumber -I../../gcc/../libdecnumber/dpd -I../libdecnumber -I/opt/local/include \ -o build/gencondmd.o build/gencondmd.c In file included from build/gencondmd.c:40: ../../gcc/config/sparc/constraints.md: In function `satisfies_constraint_T': ../../gcc/config/sparc/constraints.md:137: error: `TARGET_64BIT' undeclared (first use in this function) ../../gcc/config/sparc/constraints.md:137: error: (Each undeclared identifier is reported only once ../../gcc/config/sparc/constraints.md:137: error: for each function it appears in.) I've done some additional tests today. With Sun-CC the binaries won't work, even without optimization. But I managed to compile it with gcc 4.8.2 and that works (-O4 -mcpu=ultrasparc3 -I/opt/openssl/include -D__EXTENSIONS__). Thats good, but I don't know if I should worry about some compiler specific code in ClamAV (since 0.98) causing this bug. Maybe there is some other bigger problem. Regards, Martin -- Martin Preen, Universität Freiburg, Institut für Informatik Georges-Koehler-Allee 52, Raum EG-006, 79110 Freiburg, Germany phone: ++49 761 203-8250pr...@informatik.uni-freiburg.de fax: ++49 761 203-8242 swt.informatik.uni-freiburg.de/staff/preen smime.p7s Description: S/MIME Cryptographic Signature ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 09/05/2014 14:56, Joel Esler (jesler) wrote: Hello, Don't get over excited about Sparc, freshclam has the same problem on i386 Solaris. ... May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.info] freshclam daemon 0.98.3 (OS: solaris2.10, ARCH: i386, CPU: i386) May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.debug] Max retries == 3 May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.info] ClamAV update process started at Thu May 8 07:41:13 2014 May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.debug] Using IPv6 aware code May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.debug] Querying current.cvd.clamav.net May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.debug] TTL: 391 May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.debug] Software version from DNS: 0.98.3 May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.debug] Retrieving http://database.clamav.net/main.cvd May 8 07:41:13 mailhost freshclam[3924]: [ID 702911 mail.debug] Trying to download http://database.clamav.net/main.cvd (IP: 193.1.193.64) May 8 07:44:31 mailhost freshclam[3924]: [ID 702911 mail.info] Downloading main.cvd [100%] May 8 07:44:31 mailhost freshclam[3924]: [ID 702911 mail.error] Verification: Can't allocate memory ... James. ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
[clamav-users] Version 0.98.3 fails on Solaris
Hello, after building 0.98.3 on Solaris 10 (Sparc) I got some error messages from freshclam. The first run: ERROR: Corrupted database file /var/clamav/main.cvd: Can't allocate memory Corrupted database file renamed to /var/clamav/main.cvd.broken ERROR: Corrupted database file /var/clamav/daily.cld: Malformed database Corrupted database file renamed to /var/clamav/daily.cld.broken ERROR: Corrupted database file /var/clamav/bytecode.cld: Malformed database Corrupted database file renamed to /var/clamav/bytecode.cld.broken Any subsequent run: Downloading main.cvd [100%] LibClamAV debug: Initialized 0.98.3 engine LibClamAV debug: in cli_cvdload() LibClamAV debug: cli_cvdverify: Cannot generate hash, out of memory LibClamAV debug: Cleaning up phishcheck LibClamAV debug: Phishcheck cleaned up ERROR: Verification: Can't allocate memory There is enough memory and up to version 0.98 there was never any memory problem or similar messages. How can one resolve this ? Regards, Martin -- Martin Preen, Universität Freiburg, Institut für Informatik Georges-Koehler-Allee 52, Raum EG-006, 79110 Freiburg, Germany phone: ++49 761 203-8250pr...@informatik.uni-freiburg.de fax: ++49 761 203-8242 swt.informatik.uni-freiburg.de/staff/preen smime.p7s Description: S/MIME Cryptographic Signature ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On Thu, May 8, 2014 at 11:13 AM, Martin Preen pr...@informatik.uni-freiburg.de wrote: Hello, after building 0.98.3 on Solaris 10 (Sparc) I got some error messages from freshclam. The first run: ERROR: Corrupted database file /var/clamav/main.cvd: Can't allocate memory Corrupted database file renamed to /var/clamav/main.cvd.broken ERROR: Corrupted database file /var/clamav/daily.cld: Malformed database Corrupted database file renamed to /var/clamav/daily.cld.broken ERROR: Corrupted database file /var/clamav/bytecode.cld: Malformed database Corrupted database file renamed to /var/clamav/bytecode.cld.broken Any subsequent run: Downloading main.cvd [100%] LibClamAV debug: Initialized 0.98.3 engine LibClamAV debug: in cli_cvdload() LibClamAV debug: cli_cvdverify: Cannot generate hash, out of memory LibClamAV debug: Cleaning up phishcheck LibClamAV debug: Phishcheck cleaned up ERROR: Verification: Can't allocate memory There is enough memory and up to version 0.98 there was never any memory problem or similar messages. How can one resolve this ? Regards, Martin Hey Martin, Is there a way you can get to me main.cvd.broken? I'm wondering if the change to OpenSSL for hashing has somehow changed parsing CVDs and CLDs on big-endian machines running Solaris. I thoroughly tested the code on a sparc64 machine (an old SunFire 280r) running FreeBSD 9.2 successfully. To help me debug the issue: what version of OpenSSL do you have installed? Can you give me the output of the clamdconf command (preferably to a pastebin service)? Can you give me (again, pastebin) the output of your config.log? I can install Solaris on this sparc64 machine as early as next week. Thanks, Shawn ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
Shawn Webb wrote: On Thu, May 8, 2014 at 11:13 AM, Martin Preen pr...@informatik.uni-freiburg.de wrote: Hello, after building 0.98.3 on Solaris 10 (Sparc) I got some error messages from freshclam. The first run: ERROR: Corrupted database file /var/clamav/main.cvd: Can't allocate memory Corrupted database file renamed to /var/clamav/main.cvd.broken ERROR: Corrupted database file /var/clamav/daily.cld: Malformed database Corrupted database file renamed to /var/clamav/daily.cld.broken ERROR: Corrupted database file /var/clamav/bytecode.cld: Malformed database Corrupted database file renamed to /var/clamav/bytecode.cld.broken Any subsequent run: Downloading main.cvd [100%] LibClamAV debug: Initialized 0.98.3 engine LibClamAV debug: in cli_cvdload() LibClamAV debug: cli_cvdverify: Cannot generate hash, out of memory LibClamAV debug: Cleaning up phishcheck LibClamAV debug: Phishcheck cleaned up ERROR: Verification: Can't allocate memory There is enough memory and up to version 0.98 there was never any memory problem or similar messages. How can one resolve this ? Regards, Martin Hey Martin, Is there a way you can get to me main.cvd.broken? I'm wondering if the change to OpenSSL for hashing has somehow changed parsing CVDs and CLDs on big-endian machines running Solaris. I thoroughly tested the code on a sparc64 machine (an old SunFire 280r) running FreeBSD 9.2 successfully. To help me debug the issue: what version of OpenSSL do you have installed? Can you give me the output of the clamdconf command (preferably to a pastebin service)? Can you give me (again, pastebin) the output of your config.log? I can install Solaris on this sparc64 machine as early as next week. I'll send you an URL for downloading main.cvd.broken. Martin -- Martin Preen, Universität Freiburg, Institut für Informatik Georges-Koehler-Allee 52, Raum EG-006, 79110 Freiburg, Germany phone: ++49 761 203-8250pr...@informatik.uni-freiburg.de fax: ++49 761 203-8242 swt.informatik.uni-freiburg.de/staff/preen smime.p7s Description: S/MIME Cryptographic Signature ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
Shawn Webb wrote: On Thu, May 8, 2014 at 11:13 AM, Martin Preen pr...@informatik.uni-freiburg.de wrote: Hello, after building 0.98.3 on Solaris 10 (Sparc) I got some error messages from freshclam. The first run: ERROR: Corrupted database file /var/clamav/main.cvd: Can't allocate memory Corrupted database file renamed to /var/clamav/main.cvd.broken ERROR: Corrupted database file /var/clamav/daily.cld: Malformed database Corrupted database file renamed to /var/clamav/daily.cld.broken ERROR: Corrupted database file /var/clamav/bytecode.cld: Malformed database Corrupted database file renamed to /var/clamav/bytecode.cld.broken Any subsequent run: Downloading main.cvd [100%] LibClamAV debug: Initialized 0.98.3 engine LibClamAV debug: in cli_cvdload() LibClamAV debug: cli_cvdverify: Cannot generate hash, out of memory LibClamAV debug: Cleaning up phishcheck LibClamAV debug: Phishcheck cleaned up ERROR: Verification: Can't allocate memory There is enough memory and up to version 0.98 there was never any memory problem or similar messages. How can one resolve this ? Regards, Martin Hey Martin, Is there a way you can get to me main.cvd.broken? I'm wondering if the change to OpenSSL for hashing has somehow changed parsing CVDs and CLDs on big-endian machines running Solaris. I thoroughly tested the code on a sparc64 machine (an old SunFire 280r) running FreeBSD 9.2 successfully. To help me debug the issue: what version of OpenSSL do you have installed? Can you give me the output of the clamdconf command (preferably to a pastebin service)? Can you give me (again, pastebin) the output of your config.log? I can install Solaris on this sparc64 machine as early as next week. It is OpenSSL 1.0.1g and Sun C 5.11 SunOS_sparc 2010/08/13 with CFLAGS='-xtarget=ultra -xO4'. Martin -- Martin Preen, Universität Freiburg, Institut für Informatik Georges-Koehler-Allee 52, Raum EG-006, 79110 Freiburg, Germany phone: ++49 761 203-8250pr...@informatik.uni-freiburg.de fax: ++49 761 203-8242 swt.informatik.uni-freiburg.de/staff/preen smime.p7s Description: S/MIME Cryptographic Signature ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 5/8/14, 8:23 AM, Shawn Webb wrote: Hey Martin, Is there a way you can get to me main.cvd.broken? I'm wondering if the change to OpenSSL for hashing has somehow changed parsing CVDs and CLDs on big-endian machines running Solaris. I thoroughly tested the code on a sparc64 machine (an old SunFire 280r) running FreeBSD 9.2 successfully. To help me debug the issue: what version of OpenSSL do you have installed? Can you give me the output of the clamdconf command (preferably to a pastebin service)? Can you give me (again, pastebin) the output of your config.log? I can install Solaris on this sparc64 machine as early as next week. Thanks, Shawn Are we to understand Sourcefire does not have a proper Solaris Sparc environment for testing ClamAV products? dp ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On Thu, May 8, 2014 at 5:23 PM, Shawn Webb sw...@sourcefire.com wrote: I can install Solaris on this sparc64 machine as early as next week. OpenCSW provides a complete Solaris build farm if you're interested, we used to host automatic builds for ClamAV before Sourcefire aquired it. http://www.opencsw.org/extend-it/signup/to-upstream-maintainers/ ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 8 May 2014 at 11:23, Shawn Webb wrote: On Thu, May 8, 2014 at 11:13 AM, Martin Preen Is there a way you can get to me main.cvd.broken? I'm wondering if the change to OpenSSL for hashing has somehow changed parsing CVDs and CLDs on big-endian machines running Solaris. I thoroughly tested the code on a sparc64 machine (an old SunFire 280r) running FreeBSD 9.2 successfully. To help me debug the issue: what version of OpenSSL do you have installed? Can you give me the output of the clamdconf command (preferably to a pastebin service)? Can you give me (again, pastebin) the output of your config.log? I can install Solaris on this sparc64 machine as early as next week. If for Solaris 10 it is ok, I'm afraid you won't be able to test Solaris 11 on this machine: you need a sun4v or M-series sun4u, or a 64-bit x86 (machine or virtual machine). Pierre Note: I did not try 0.98.3 yet but I skipped 0.98.1 as well because it did not compile with my usual script neither (on Solaris 10 sparc) and I have had to time to debug the problem yet. ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml
Re: [clamav-users] Version 0.98.3 fails on Solaris
On 5/8/14, 9:00 AM, Dennis Peterson wrote: On 5/8/14, 8:23 AM, Shawn Webb wrote: Hey Martin, Is there a way you can get to me main.cvd.broken? I'm wondering if the change to OpenSSL for hashing has somehow changed parsing CVDs and CLDs on big-endian machines running Solaris. I thoroughly tested the code on a sparc64 machine (an old SunFire 280r) running FreeBSD 9.2 successfully. To help me debug the issue: what version of OpenSSL do you have installed? Can you give me the output of the clamdconf command (preferably to a pastebin service)? Can you give me (again, pastebin) the output of your config.log? I can install Solaris on this sparc64 machine as early as next week. Thanks, Shawn Are we to understand Sourcefire does not have a proper Solaris Sparc environment for testing ClamAV products? dp My point is asking is I have three Sparc systems (Ultra 10, Ultra 2 dual proc, and Netra X1) that are going to a landfill if I don't find a home for them. All are in perfect working condition. I'll split the shipping cost (US only) if you're interested. OS not included. dp ___ Help us build a comprehensive ClamAV guide: https://github.com/vrtadmin/clamav-faq http://www.clamav.net/support/ml