|
From: Ivo R. <iv...@iv...> - 2017-01-31 04:12:52
|
Please have a look at the proposed patch for bug 375772 [1]: +1 error in get_elf_symbol_info() when computing value of 'hi' address for ML_(find_rx_mapping)() Kind regards, I. [1] https://bugs.kde.org/show_bug.cgi?id=375772 |
|
From: Mark W. <mj...@re...> - 2017-01-31 12:20:04
|
On Tue, 2017-01-31 at 05:12 +0100, Ivo Raisr wrote: > Please have a look at the proposed patch for bug 375772 [1]: > +1 error in get_elf_symbol_info() when computing value of 'hi' address for > ML_(find_rx_mapping)() > Symbol sizes could be zero, which would cause this to make hi < lo. I cannot immediately see if the code guards against that. Otherwise this looks correct. Cheers, Mark |
|
From: Ivo R. <iv...@iv...> - 2017-01-31 12:29:16
|
2017-01-31 13:19 GMT+01:00 Mark Wielaard <mj...@re...>: > On Tue, 2017-01-31 at 05:12 +0100, Ivo Raisr wrote: > > Please have a look at the proposed patch for bug 375772 [1]: > > +1 error in get_elf_symbol_info() when computing value of 'hi' address > for > > ML_(find_rx_mapping)() > > > > Symbol sizes could be zero, which would cause this to make hi < lo. I > cannot immediately see if the code guards against that. Otherwise this > looks correct. > Indeed, that was my worry as well. Fortunately get_elf_symbol_info() guards against this problem at the beginning of this function. Thank you for looking into it, Mark! I. |