Hi Rod, Thanks for explanation, however I don't understand why you writing offset 0xb4 while there is something at 0xd4 in the object. In fact the original libapr from the distribution (OpenSolaris) also has offset 0xd4 in the PT_SUNWCAP section. Is that a mistype?
On Fri, Oct 29, 2010 at 6:47 PM, Rod Evans <[email protected]> wrote: > On 10/29/10 08:17 AM, Piotr Jasiukajtis wrote: > >> It's not working since b135. >> .... >> >> Build 135: >> >> ldd libapr-0.so.0.9.17 >> ./libapr-0.so.0.9.17: unknown capability: 0 libsendfile.so.1 => >> /lib/libsendfile.so.1 >> librt.so.1 => /lib/librt.so.1 >> ..... > > If you have no idea how you came by this file, or who built it, then > I think your only recourse is to edit the file into submission. > > The ELF file is very suspicious, as there are so many errors generated > from elfdump(). Prior to 135, capabilities processing was very simple. > But with 135 we added symbol capabilities (6918143). This expanded on > the capabilities definitions, and what we can now do with them. It seems > that prior to 135, our simple capabilities processing was not adversely > affected by the information in this file, but since 135 we are falling foul > of the invalid information in the file. > > You could edit the files PT_SUNWCAP to point to a "null" series of bytes. > This can be done by pointing to the NULL program header entry. So try > this: > > % elfedit libapr-0.so.0.9.17 >> phdr:p_offset PT_SUNWCAP 0xb4 > ... >> phdr:p_vaddr PT_SUNWCAP 0xb4 > ... >> phdr:p_filesz PT_SUNWCAP 0x10 > ... >> phdr:p_memsz PT_SUNWCAP 0x10 > ... >> write >> quit > > You'll still get a complaint from elfdump(), because we assume that if a > PT_SUNWCAP exists, there should also be an underlying .SUNW_cap section. > > % elfdump -H libapr-0.so.0.9.17 > libapr-0.so.0.9.17: PT_SUNWCAP header has no associated section > > But, using ldd() or executing the library gets past the capabilities error, > > % ldd libapr-0.so.0.9.17 > libsendfile.so.1 => /lib/libsendfile.so.1 > librt.so.1 => /lib/librt.so.1 > libm.so.2 => /lib/libm.so.2 > libsocket.so.1 => /lib/libsocket.so.1 > libnsl.so.1 => /lib/libnsl.so.1 > libresolv.so.2 => /lib/libresolv.so.2 > libpthread.so.1 => /lib/libpthread.so.1 > libc.so.1 => /lib/libc.so.1 > libmp.so.2 => /lib/libmp.so.2 > libmd.so.1 => /lib/libmd.so.1 > > Given how many error diagnostics are generated from elfdump() for this file, > I'd still be a little suspicious of its use. I'd try and find a better > copy. > > If you can find how the file was created then please pass use the > information > as it would be enlightening to get to the bottom of this. > > -- > > Rod. > -- Piotr Jasiukajtis | estibi | SCA OS0072 http://estseg.blogspot.com _______________________________________________ tools-linking mailing list [email protected]
