[rfc] [patch] commands to display x86 hardware structures
|[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]|
On Tue, 8 Jul 2003 13:50:42 +0530, "Vamsi Krishna S ." <vamsi at in.ibm.com> wrote: >Here is a patch (against 2.4.21+kdb-4.3 latest) to add the >following commands: > >- rdv: display all processor registers. Special system registers > are interpreted and their contents displayed verbosely. > This should have been "rd v", but as I wanted to keep > this seperate, I had to introduce a new command. >- gdt: display the contents of GDT >- idt: display the contents of IDT >- ldt: display the contents of the current LDT >- ldtp: display the contents of LDT of the kdb_current task > (All process-specific commands apply to the process > selected with the "pid" command.) >- ptex: display page table entries (pgd/pte) for the given address. > If a user address is given, it is looked up in > the context of the process previously selected with > "pid" command. I realise that "pte" command is > present in kdbm_vm.c, please take a look at this too. > Eventually, we will need to keep only one of them. > >All commands except "rdv" can be repeated. > >Please review and comment, especially on: > >- kdbm_x86.c file header, if you want any changes there >- whether "rdv" should be turned into "rd v" - needs moving > a lot of code from kdbm_x86.c to kdbasupport.c >- syntax of the commands and the format of their output >- whether "ptex" should be removed in favour of existing > "pte" >- exported symbols (kdb_current_task and kdba_dumpregs). Nice. I will include the patch as it is, to let people get some experience with the commands. Then we can decide if the commands should be rationalised or not. In previous versions of kdb I made the mistake of trying to get it exactly right before releasing the code. The pte command should become common code, with a few arch specific #ifdefs in it. Most of the pte scanmning code is the same for i386, ia64, etc.