On Mon, 18 Sept 2023 at 16:15, Kumar Kartikeya Dwivedi <memxor@xxxxxxxxx> wrote:
>
> On Mon, 18 Sept 2023 at 15:56, Alexei Starovoitov
> <alexei.starovoitov@xxxxxxxxx> wrote:
> >
> > On Mon, Sep 18, 2023 at 6:54 AM Eric Dumazet <edumazet@xxxxxxxxxx> wrote:
> > >
> > > On Mon, Sep 18, 2023 at 3:41 PM Alexei Starovoitov
> > > <alexei.starovoitov@xxxxxxxxx> wrote:
> > > >
> > > > On Mon, Sep 18, 2023 at 6:25 AM Eric Dumazet <edumazet@xxxxxxxxxx> wrote:
> > > > >
> > > > > On Sat, Sep 16, 2023 at 6:59 PM Alexei Starovoitov
> > > > > <alexei.starovoitov@xxxxxxxxx> wrote:
> > > > > >
> > > > > > Hi David, hi Jakub, hi Paolo, hi Eric,
> > > > > >
> > > > > > The following pull-request contains BPF updates for your *net-next* tree.
> > > > > >
> > > > > > We've added 73 non-merge commits during the last 9 day(s) which contain
> > > > > > a total of 79 files changed, 5275 insertions(+), 600 deletions(-).
> > > > > >
> > > > > > The main changes are:
> > > > > >
> > > > > > 1) Basic BTF validation in libbpf, from Andrii Nakryiko.
> > > > > >
> > > > > > 2) bpf_assert(), bpf_throw(), exceptions in bpf progs, from Kumar Kartikeya Dwivedi.
> > > > > >
> > > > > > 3) next_thread cleanups, from Oleg Nesterov.
> > > > > >
> > > > > > 4) Add mcpu=v4 support to arm32, from Puranjay Mohan.
> > > > > >
> > > > > > 5) Add support for __percpu pointers in bpf progs, from Yonghong Song.
> > > > > >
> > > > > > 6) Fix bpf tailcall interaction with bpf trampoline, from Leon Hwang.
> > > > > >
> > > > > > 7) Raise irq_work in bpf_mem_alloc while irqs are disabled to improve refill probabablity, from Hou Tao.
> > > > > >
> > > > > > Please consider pulling these changes from:
> > > > > >
> > > > > > git://git.kernel.org/pub/scm/linux/kernel/git/bpf/bpf-next.git
> > > > > >
> > > > >
> > > > > This might have been raised already, but bpf on x86 now depends on
> > > > > CONFIG_UNWINDER_ORC ?
> > > > >
> > > > > $ grep CONFIG_UNWINDER_ORC .config
> > > > > # CONFIG_UNWINDER_ORC is not set
> > > > >
> > > > > $ make ...
> > > > > arch/x86/net/bpf_jit_comp.c:3022:58: error: no member named 'sp' in
> > > > > 'struct unwind_state'
> > > > > if (!addr || !consume_fn(cookie, (u64)addr,
> > > > > (u64)state.sp, (u64)state.bp))
> > > > > ~~~~~ ^
> > > > > 1 error generated.
> > > >
> > > > Kumar,
> > > > can probably explain better,
> > > > but no the bpf as whole doesn't depend.
> > > > One feature needs either ORC or frame unwinder.
> > > > It won't work with unwinder_guess.
> > > > The build error is a separate issue.
> > > > It hasn't been reported before.
> > >
> > > In my builds, I do have CONFIG_UNWINDER_FRAME_POINTER=y
> > >
> > > $ grep UNWIND .config
> > > # CONFIG_UNWINDER_ORC is not set
> > > CONFIG_UNWINDER_FRAME_POINTER=y
> > >
> > >
> > > I note state.sp is only available to CONFIG_UNWINDER_ORC
> > >
> > > arch/x86/include/asm/unwind.h
> > >
> > > #if defined(CONFIG_UNWINDER_ORC)
> > > bool signal, full_regs;
> > > unsigned long sp, bp, ip;
> > > struct pt_regs *regs, *prev_regs;
> > > #elif defined(CONFIG_UNWINDER_FRAME_POINTER)
> > > bool got_irq;
> > > unsigned long *bp, *orig_sp, ip; // this is orig_sp , not sp.
> >
> > Right. Our replies crossed.
> > Please ignore this PR. We need to fix this first.
>
> Hello,
> This is my bad. I totally missed it since I initially wrote this patch
> and never looked at it again.
> I suggest that I send a fix to disable this feature with
> CONFIG_UNWINDER_FRAME_POINTER=y, while I work on reenabling it again
> for it with a follow up.
Hi, I've attached a fix that should disable it for now. I'll work on a
follow up to reenable it for this config option.
Really sorry about this, I'll try to be more careful going forward.
From e95240ede0cbe3aa35bbb12bf7eb39e93c7c6e61 Mon Sep 17 00:00:00 2001
From: Kumar Kartikeya Dwivedi <memxor@xxxxxxxxx>
Date: Mon, 18 Sep 2023 16:07:27 +0200
Subject: [PATCH] bpf: Disable exceptions when CONFIG_UNWINDER_FRAME_POINTER=y
The build with CONFIG_UNWINDER_FRAME_POINTER=y is broken for
current exceptions feature as it assumes ORC unwinder specific fields in
the unwind_state. Disable exceptions when frame_pointer unwinder is
enabled for now.
Fixes: fd5d27b70188 ("arch/x86: Implement arch_bpf_stack_walk")
Reported-by: Eric Dumazet <edumazet@xxxxxxxxxx>
Signed-off-by: Kumar Kartikeya Dwivedi <memxor@xxxxxxxxx>
---
arch/x86/net/bpf_jit_comp.c | 9 ++++-----
1 file changed, 4 insertions(+), 5 deletions(-)
diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
index 84005f2114e0..8c10d9abc239 100644
--- a/arch/x86/net/bpf_jit_comp.c
+++ b/arch/x86/net/bpf_jit_comp.c
@@ -3003,16 +3003,15 @@ void bpf_jit_free(struct bpf_prog *prog)
bool bpf_jit_supports_exceptions(void)
{
/* We unwind through both kernel frames (starting from within bpf_throw
- * call) and BPF frames. Therefore we require one of ORC or FP unwinder
- * to be enabled to walk kernel frames and reach BPF frames in the stack
- * trace.
+ * call) and BPF frames. Therefore we require ORC unwinder to be enabled
+ * to walk kernel frames and reach BPF frames in the stack trace.
*/
- return IS_ENABLED(CONFIG_UNWINDER_ORC) || IS_ENABLED(CONFIG_UNWINDER_FRAME_POINTER);
+ return IS_ENABLED(CONFIG_UNWINDER_ORC);
}
void arch_bpf_stack_walk(bool (*consume_fn)(void *cookie, u64 ip, u64 sp, u64 bp), void *cookie)
{
-#if defined(CONFIG_UNWINDER_ORC) || defined(CONFIG_UNWINDER_FRAME_POINTER)
+#if defined(CONFIG_UNWINDER_ORC)
struct unwind_state state;
unsigned long addr;
--
2.41.0
