From: Kris Van Hees <kris.van.hees@oracle.com>
To: Kris Van Hees <kris.van.hees@oracle.com>
Cc: Nick Alcock <nick.alcock@oracle.com>,
Eugene Loh <eugene.loh@oracle.com>,
dtrace@lists.linux.dev, dtrace-devel@oss.oracle.com
Subject: Re: [PATCH 01/14] Fix stack-skip counts for caller and stackdepth
Date: Thu, 19 Jun 2025 12:32:13 -0400 [thread overview]
Message-ID: <aFQ7jfnA5eKnDMRc@oracle.com> (raw)
In-Reply-To: <aFQ4xujz8hJRUGjB@oracle.com>
On Thu, Jun 19, 2025 at 12:20:22PM -0400, Kris Van Hees wrote:
> On Thu, Jun 19, 2025 at 02:03:56PM +0100, Nick Alcock wrote:
> > On 16 Jun 2025, Eugene Loh verbalised:
> >
> > > On 6/13/25 10:33, Nick Alcock wrote:
> > >
> > >>> Note that we declare the skip count volatile. The compiler might
> > >>> optimize code that uses the STACK_SKIP value, but we will subsequently
> > >>> perform relocations that adjust this value.
> > >> ... why doesn't this apply to every other extern global variable in
> > >> get_bvar()? They're all similarly relocated...
> > >
> > > Right. There is potentially a broader problem. But we simply do not have evidence of misbehavior in other cases. Ruggedizing
> > > other cases could be the subject of a different patch.
> >
> > Aha, OK. I was just wondering if there was some extra reason.
> >
> > > The problem in this case is that the compiler seems to assume &symbol!=0, which is reasonable except that we violate that behavior
> > > for our relocation tricks.
> >
> > I wonder where the code for that is... plenty of symbols have value
> > zero.
> >
> > But, really...
> >
> > > Consider the C code:
> > >
> > > uint64_t dt_bvar_stackdepth(const dt_dctx_t *dctx)
> > > {
> > > uint32_t bufsiz = (uint32_t) (uint64_t) (&STKSIZ);
> > > char *buf = dctx->mem + (uint64_t)(&STACK_OFF);
> >
> > Hm...
> >
> > extern uint64_t STACK_SKIP;
> >
> > So we encode information about the stack size and skip value by encoding
> > it in the *address* of the variable? Is there some reason we don't use
> > its value? unlike the stack offset, we're *using* it as a value, not an
> > address...
>
> The issue seems to be (and perhaps this is a cross compiler problem) that e.g.
>
> extern uint64_t PC;
>
> and then code accessing the value of PC (e.g. foo(PC) as a call argument) will
> yield:
>
> 50: 18 02 00 00 00 00 00 00 lddw %r2,0
> 58: 00 00 00 00 00 00 00 00
> 50: R_BPF_INSN_64 PC
> 60: 79 22 00 00 00 00 00 00 ldxdw %r2,[%r2+0]
>
> which shows that it is interpreting PC as an address to a symbol, because it
> loads the address of the symbol and then dereferences it with offset 0. So,
> we cannot plug in the value during relocation because the only value we can
> put there would be an address where the vlaue can be found. To get around
> this, we "use" Tthe address as the entity to store the value in, knowing that
> we *never* will interpret it as an address for these specific externs.
Not a compiler error - since PC is extern uint64_t PC it *is* a variable and
so it is present (and accessible) as an address in .data in the location where
it is actually defined. Since we never define it, we don't have a .data (which
is fine because we only use this for constants known at link time) BUT the
compiler of course is free to assume that we *do* have an address to the
storage location for PC and thus that we get the value that way. It does not
know that the value is constant. So, the trick I use is needed to make this
work.
> > > uint64_t retv;
> > > volatile uint64_t skip = (uint64_t)(&STACK_SKIP);
> > >
> > > retv = bpf_get_stack(dctx->ctx,
> > > buf,
> > > bufsiz,
> > > skip & BPF_F_SKIP_FIELD_MASK);
> > >
> > > if (skip)
> > > return retv / sizeof(uint64_t) - 1; // branch "A"
> > > return retv / sizeof(uint64_t); // branch "B"
> > > }
> > >
> > > If you omit "volatile", the compiler assumes &STACK_SKIP!=0. The emitted code has:
> >
> > (which is a reasonable assumption if not freestanding, I'd say. Why
> > don't we compile BPF code with -ffreestanding? BPF is almost the
> > *definition* of a freestanding environment...)
> >
> > > *) no run-time "if (skip)" check
> > >
> > > *) no code for branch "B"
> > >
> > > *) only code for branch "A"
> > >
> > > If you include "volatile", however, the compiler caches &STACK_SKIP on
> > > the BPF stack and later performs a run-time check on its value to
> > > correctly execute either branch "A" or branch "B".
> >
> > This feels very mucvh like a workaround to me. Does compiling BPF with
> > -ffreestanding help?
> >
> > I mean it fixes a bug, so I suppose it should go in if nothing else
> > works, but using volatile is almost always a desperate sticking plaster
> > and this feels like one of those occasions to me.
> >
> > --
> > NULL && (void)
next prev parent reply other threads:[~2025-06-19 16:32 UTC|newest]
Thread overview: 41+ messages / expand[flat|nested] mbox.gz Atom feed top
2025-05-22 18:01 [PATCH 01/14] Fix stack-skip counts for caller and stackdepth eugene.loh
2025-05-22 18:01 ` [PATCH 02/14] Add stack-skip frame count for rawtp provider eugene.loh
2025-06-13 14:35 ` Nick Alcock
2025-05-22 18:01 ` [PATCH 03/14] Test: remove unnecessary "unstable" tag eugene.loh
2025-06-13 14:35 ` [DTrace-devel] " Nick Alcock
2025-05-22 18:01 ` [PATCH 04/14] Test: caller and stackdepth tests for fbt provider eugene.loh
2025-06-13 14:40 ` [DTrace-devel] " Nick Alcock
2025-06-16 19:43 ` Eugene Loh
2025-06-25 4:14 ` Kris Van Hees
2025-05-22 18:01 ` [PATCH 05/14] Test: caller and stackdepth tests for dtrace provider eugene.loh
2025-06-13 14:42 ` Nick Alcock
2025-05-22 18:01 ` [PATCH 06/14] Test: caller and stackdepth tests for rawtp provider eugene.loh
2025-06-13 14:45 ` Nick Alcock
2025-05-22 18:01 ` [PATCH 07/14] Test: caller and stackdepth tests for cpc provider eugene.loh
2025-06-13 14:48 ` [DTrace-devel] " Nick Alcock
2025-05-22 18:01 ` [PATCH 08/14] Test: caller and stackdepth tests for ip provider eugene.loh
2025-06-13 14:51 ` [DTrace-devel] " Nick Alcock
2025-06-17 3:38 ` Eugene Loh
2025-06-25 4:15 ` Kris Van Hees
2025-05-22 18:01 ` [PATCH 09/14] Test: caller and stackdepth tests for profile provider eugene.loh
2025-06-13 14:52 ` [DTrace-devel] " Nick Alcock
2025-05-22 18:01 ` [PATCH 10/14] Test: caller and stackdepth tests for sched provider eugene.loh
2025-06-13 14:52 ` Nick Alcock
2025-05-22 18:01 ` [PATCH 11/14] Test: caller and stackdepth tests for proc provider eugene.loh
2025-06-13 14:53 ` Nick Alcock
2025-06-15 17:50 ` Eugene Loh
2025-06-19 12:52 ` Nick Alcock
2025-06-25 4:18 ` Kris Van Hees
2025-05-22 18:01 ` [PATCH 12/14] Test: caller and stackdepth tests for rawfbt provider eugene.loh
2025-06-13 14:54 ` [DTrace-devel] " Nick Alcock
2025-06-17 23:17 ` Eugene Loh
2025-05-22 18:01 ` [PATCH 13/14] Test: caller and stackdepth tests for io provider eugene.loh
2025-06-13 14:56 ` [DTrace-devel] " Nick Alcock
2025-05-22 18:01 ` [PATCH 14/14] Test: caller and stackdepth tests for lockstat provider eugene.loh
2025-06-13 14:57 ` Nick Alcock
2025-06-13 14:33 ` [PATCH 01/14] Fix stack-skip counts for caller and stackdepth Nick Alcock
2025-06-16 19:21 ` Eugene Loh
2025-06-19 13:03 ` Nick Alcock
2025-06-19 16:20 ` Kris Van Hees
2025-06-19 16:32 ` Kris Van Hees [this message]
2025-06-23 14:04 ` Nick Alcock
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=aFQ7jfnA5eKnDMRc@oracle.com \
--to=kris.van.hees@oracle.com \
--cc=dtrace-devel@oss.oracle.com \
--cc=dtrace@lists.linux.dev \
--cc=eugene.loh@oracle.com \
--cc=nick.alcock@oracle.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox