Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-21 Thread Joerg Roedel
On Mon, Jul 20, 2020 at 06:48:03PM +0200, Thomas Gleixner wrote: > But with explicit sections which store only page aligned objects there > is an implicit guarantee that the object is alone in the page in which > it is placed. That works for all objects except the last one. That's > inconsistent.

Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-20 Thread Thomas Gleixner
Joerg Roedel writes: > On Sun, Jul 19, 2020 at 12:39:44PM +0200, Thomas Gleixner wrote: >> *(.bss..page_aligned) \ >> +. = ALIGN(PAGE_SIZE); \ >> *(.dynbss)

Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-20 Thread Joerg Roedel
On Sun, Jul 19, 2020 at 12:39:44PM +0200, Thomas Gleixner wrote: > The right fix is trivial. See below. > > Thanks, > > tglx > > arch/x86/kernel/vmlinux.lds.S |1 + > include/asm-generic/vmlinux.lds.h |1 + > 2 files changed, 2 insertions(+) > > ---

Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-19 Thread Thomas Gleixner
Arvind Sankar writes: > To repeat the commit message, the problem is not misaligned > bss..page_aligned objects, but symbols in _other_ bss sections, which > can get allocated in the last page of bss..page_aligned, because its end > isn't page-aligned (maybe it should be?) That's the real and

Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-18 Thread Arvind Sankar
On Sat, Jul 18, 2020 at 06:15:26PM -0700, h...@zytor.com wrote: > On July 18, 2020 12:25:46 PM PDT, Andy Lutomirski wrote: > > > >> On Jul 18, 2020, at 10:57 AM, h...@zytor.com wrote: > >> > >> On July 9, 2020 3:33:55 AM PDT, Joerg Roedel > >wrote: > >>> From: Joerg Roedel > >>> > >>> On

Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-18 Thread hpa
On July 18, 2020 12:25:46 PM PDT, Andy Lutomirski wrote: > >> On Jul 18, 2020, at 10:57 AM, h...@zytor.com wrote: >> >> On July 9, 2020 3:33:55 AM PDT, Joerg Roedel >wrote: >>> From: Joerg Roedel >>> >>> On x86-32 the idt_table with 256 entries needs only 2048 bytes. It >is >>> page-aligned,

Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-18 Thread Andy Lutomirski
> On Jul 18, 2020, at 10:57 AM, h...@zytor.com wrote: > > On July 9, 2020 3:33:55 AM PDT, Joerg Roedel wrote: >> From: Joerg Roedel >> >> On x86-32 the idt_table with 256 entries needs only 2048 bytes. It is >> page-aligned, but the end of the .bss..page_aligned section is not >> guaranteed

Re: [PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-18 Thread hpa
On July 9, 2020 3:33:55 AM PDT, Joerg Roedel wrote: >From: Joerg Roedel > >On x86-32 the idt_table with 256 entries needs only 2048 bytes. It is >page-aligned, but the end of the .bss..page_aligned section is not >guaranteed to be page-aligned. > >As a result, symbols from other .bss sections

[PATCH] x86/idt: Make sure idt_table takes a whole page

2020-07-09 Thread Joerg Roedel
From: Joerg Roedel On x86-32 the idt_table with 256 entries needs only 2048 bytes. It is page-aligned, but the end of the .bss..page_aligned section is not guaranteed to be page-aligned. As a result, symbols from other .bss sections may end up on the same 4k page as the idt_table, and will