Tuesday Mar 30, 2010

nmap 5.30 (BETA1) Compile on OpenSolaris

The formal OpenSolaris offering of nmap is out of date. I used the following to compile nmap 5.30BETA1 on OpenSolaris build 134.

$ ac_cv_dnet_linux_pf_packet=no CPPFLAGS="-I/usr/include/pcre" \\ ./configure \\ --with-libpcap=/usr \\ --with-libpcre=/usr $ make $ pfexec make install

This assumes you already have the libpcap and libpcre packages installed. You can check with:

$ pkg list |egrep "pcap|pcre" library/pcre (opensolaris.org) 7.8-0.134 installed ----- system/library/libpcap (opensolaris.org) 1.0.0-0.134 installed -----

And repeating from my last nmap blog:

Alas, there are still issues lingering with later versions of nmap on OpenSolaris. See this networking-discuss thread as well as this (yet unanswered) nmap-dev posting. In the latter, it's suggested that non-root scans have better success - that's been my experience thus far. The obvious downside being scans needing raw Ethernet access can't be used (or use the --unprivileged option to nmap).

:wq

Friday Mar 26, 2010

Intel CPU/Memory HotPlug for OpenSolaris

Cool integration yesterday. Hotplug of CPUs and memory for Intel systems. Last summer, I worked with Intel to ensure that newly added resources are fault managed just as those present at start-of-day are fault managed. And gladly, that functionality is included in this integration.

When resources are added, the FMA topology is updated to reflect the new CPUs/memory, as are the #MC handlers. There's a gap in FRU identification with newly added resources (component labels and serial numbers are sourced from SMBIOS, which is static) but otherwise hotplugged components are handled in FMA.

:wq

Thursday Mar 25, 2010

skipfish on OpenSolaris

There's been several articles about skipfish, a web application recon tool from Google. Ignoring the documentation that called for glibc to be present, I tried the compile anyway. Aside from environment variable adjustments and an include file or two, the primary obstacle is the use of glibc's malloc_usable_size() - its primary apparent usage is to assist in zeroing out malloc'd memory.

I'd done some testing with pulling out malloc_usable_size and the tool would run, but bail when writing out final results (see the comments). I've since went the route of modifying skipfish to use Doug Lea's malloc library, which includes an implementation of malloc_usable_size(). And that's been working great.

initial allocations). I've uploaded a patch for the necessary changes. The patch is based on v1.26b of skipfish.

The basic steps:

Ensure you have the IDN library installed $ pkg list |grep idn library/idnkit (opensolaris.org) 0.5.11-0.134 installed ----- library/libidn (opensolaris.org) 1.9-0.134 installed ----- Download and unpack skipfish Copy the patch into the unpacked skipfish directory. $ cd skipfish $ patch -p1 < skipfish.1.26b.solaris.dmalloc.patch patching file alloc-inl.h patching file dlmalloc.c patching file dlmalloc.h patching file Makefile patching file report.c $ CC="/usr/bin/gcc" CFLAGS="-I/usr/include/idn -DUSE_LOCKS" \\ LIBS="-lsocket -lnsl -lpthread" make

From there, read the skipfish wiki for details on running the tool. Also check out the known issues. The item I hit immediately was not having a terminal size of at least 100x35. Things run fine, but the output to the terminal can get munged at smaller sizes.

If you do any testing of skipfish running on OpenSolaris, let me know how it goes. May look at getting this into the SourceJuicer. And by the way, I'm not a web-security expert. While I'm happy to (try to) answer questions about why skipfish may not compile on OpenSolaris, I cannot answer questions about the results the tool returns.

:wq

About

user9148476

Search

Archives
« April 2014
SunMonTueWedThuFriSat
  
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
   
       
Today