[Crash-utility] [RFC PATCH 04/15] Fix verify_module() and next_module_vaddr()

[lijiang]

On Fri, May 26, 2023 at 8:50 AM HAGIO KAZUHITO(萩尾 一仁) <k-hagio-ab at nec.com>
wrote:

> (btw, isn't it possible for you to send emails only with plain text


I sent emails with plain text.


>
> part?  Your emails have two parts, html and plain text, but the html one
> is shown with proportional font and indents are fuzzy.  The plain text
> one looks auto-generated(?) and is wrapped excessively..  Or is my
> Thunderbird setting not good?)
>
>
Sorry about it. I checked again, and noticed that I marked the text with
the color/font. Maybe it is a compatibility issue, I'm using gmail.

>> +module_memory: /* TODO: test this */
> >> +       for (i = 0; i < st->mods_installed; i++) {
> >> +               lm = &st->load_modules[i];
> >> +
> >> +               for (j = MOD_TEXT; j < MOD_MEM_NUM_TYPES; j++) {
> >> +                       start = lm->mem[j].base;
> >> +                       end = start + lm->mem[j].size;
> >> +
> >> +                       if (vaddr >= end)
> >> +                               continue;
> >> +
> >> +                       if (vaddr < start) {
> >> +                               if (start < min) /* replace candidate */
> >> +                                       min = start;
> >>
> >
> > If so,  wouldn't it be overwritten in each module memory type loop? And
> > later it only checked them one time(see the code at the end of this
> > function).
> >
> > In addition, why does it need to deal with this one here? Could you
> please
> > explain more details?
>
> sorry, I don't understand your questions.
>
> This function searches for the next lowest module address from the
> vaddr.  Now the regions of "a module" are not one block and even not
> sorted, this function needs to search all regions of all modules.  min
> is a candidate, so when a lower start address than min is found, min is
> replaced with the start address.
>

Thank you for the explanation, Kazu. That is my question.

Thanks.
Lianbo
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/crash-utility/attachments/20230530/c9d4fce6/attachment.htm>