This is the mail archive of the mailing list for the GDB project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: SEGV in dwarf2read.c -- gdb-7.2

On 11/03/2011 10:22 AM, Jan Kratochvil wrote:
On Thu, 03 Nov 2011 18:13:04 +0100, Michael Eager wrote:
With the addition of the call to age_cached_comp_units() in
dwarf2_fetch_die_location_block() and dw2_do_instantiate_symtab(),
I think that the cache is aged very aggressively.

With dwarf2_max_cache_age == 5, it looks to me that it is likely that CU
data will not be in the cache when needed.

There was for example this fix:

which is not yet even in 7.3.1 as I see now, it is only in FSF GDB HEAD.

Thanks. The patch looks similar to the one I came up with. I added a similar check higher in the call tree.

wrt CUs aging therefore only FSF GDB HEAD matters, older releases are known to
be buggy - at least due to this bug.

With FSF GDB HEAD I am not aware of any problems because any call to
age_cached_comp_units is done when CUs are no longer needed anywhere.

I do not discuss here about efficiency of the FSF GDB HEAD solution but I am
not aware of how it could crash any way.

I'll see if it works with my application.

-- Michael Eager 1960 Park Blvd., Palo Alto, CA 94306 650-325-8077

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]