Summary: | Limit crypt rounds to avoid denial of service | ||
---|---|---|---|
Product: | glibc | Reporter: | Hanno Boeck <hanno> |
Component: | crypt | Assignee: | Not yet assigned to anyone <unassigned> |
Status: | UNCONFIRMED --- | ||
Severity: | normal | CC: | adhemerval.zanella, drepper.fsp, fweimer |
Priority: | P2 | Flags: | fweimer:
security+
|
Version: | unspecified | ||
Target Milestone: | --- | ||
See Also: | https://sourceware.org/bugzilla/show_bug.cgi?id=25441 | ||
Host: | Target: | ||
Build: | Last reconfirmed: | ||
Attachments: | patch to avoid DoS in crypt function |
Description
Hanno Boeck
2016-09-17 08:07:13 UTC
I think this change is reasonable, the only details I am not sure is which ROUNDS_MAX would the most suitable, mainly on slower CPUs. On ARMv7 Cortex-A15 using 9999999 rounds it took about 20s, which is high but still manageable. Can you send this patch upstream? "Can you send this patch upstream?" What do you mean by this? Isn't this the official glibc bug tracker, thus upstream? (In reply to Hanno Boeck from comment #2) > "Can you send this patch upstream?" > > What do you mean by this? Isn't this the official glibc bug tracker, thus > upstream? Patch discussions are not done in bugtracker, but rather on libc-alpha [1]. By send this upstream I mean send a complete patch for libc mailist. [1] https://www.gnu.org/software/libc/involved.html |