linux-perf-users.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Oleg Nesterov <oleg@redhat.com>
To: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Sven Schnelle <svens@linux.ibm.com>,
	Michael Ellerman <mpe@ellerman.id.au>,
	Masami Hiramatsu <mhiramat@kernel.org>,
	Peter Zijlstra <peterz@infradead.org>,
	Ingo Molnar <mingo@redhat.com>,
	Arnaldo Carvalho de Melo <acme@kernel.org>,
	Namhyung Kim <namhyung@kernel.org>,
	Mark Rutland <mark.rutland@arm.com>,
	Alexander Shishkin <alexander.shishkin@linux.intel.com>,
	Jiri Olsa <jolsa@kernel.org>, Ian Rogers <irogers@google.com>,
	Adrian Hunter <adrian.hunter@intel.com>,
	"Liang, Kan" <kan.liang@linux.intel.com>,
	linux-kernel@vger.kernel.org, linux-trace-kernel@vger.kernel.org,
	linux-perf-users@vger.kernel.org
Subject: Re: [PATCH] uprobes: use vm_special_mapping close() functionality
Date: Wed, 4 Sep 2024 11:56:17 +0200	[thread overview]
Message-ID: <20240904095449.GA28781@redhat.com> (raw)
In-Reply-To: <CAHk-=wi=qJr4r2DTLDMDh=ryK-x9sciGEeL+ZaWExpiHGyPhiQ@mail.gmail.com>

I didn't have time to write a full reply yesterday.

On 09/03, Linus Torvalds wrote:
>
> On Tue, 3 Sept 2024 at 02:09, Oleg Nesterov <oleg@redhat.com> wrote:
> >
> > but with or without this fix __create_xol_area() also needs
> >
> >         area->xol_mapping.mremap = NULL;
>
> I think the whole thing needs to be zeroed out.

Again, this is what Sven did and I agree with this fix for now.

> It was always horribly buggy. The close thing just made it more
> *obviously* buggy, because closing a vma is a lot more common than
> mremap'ing it.

Well, this code is very old, I don't think it was "always" buggy.
But at least today it is horribly ugly, and

> Either use kzalloc(), or do a proper initializer something like this:
>
> -       area->xol_mapping.name = "[uprobes]";
> -       area->xol_mapping.fault = NULL;
> -       area->xol_mapping.pages = area->pages;
> +       area->xol_mapping = (struct vm_special_mapping) {
> +               .name = "[uprobes]",
> +               .pages = area->pages,
> +               .close = uprobe_clear_state,
> +       };

either way the code is still ugly, imo.

How about the (untested) patch below?

I am not going to send this patch right now, it conflicts with the ongoing
close/mremap changes, but what do you think?

We do not need to add the instance of vm_special_mapping into mm_struct,
a single instance (xol_mapping below) with .fault = xol_fault() is enough.
And, with this change xol_area no longer needs "struct page *pages[2]", it
can be turned into "struct page *page".

Oleg.
---

--- a/kernel/events/uprobes.c
+++ b/kernel/events/uprobes.c
@@ -101,7 +101,6 @@ struct xol_area {
 	atomic_t 			slot_count;	/* number of in-use slots */
 	unsigned long 			*bitmap;	/* 0 = free slot */
 
-	struct vm_special_mapping	xol_mapping;
 	struct page 			*pages[2];
 	/*
 	 * We keep the vma's vm_start rather than a pointer to the vma
@@ -1453,6 +1452,21 @@ void uprobe_munmap(struct vm_area_struct *vma, unsigned long start, unsigned lon
 		set_bit(MMF_RECALC_UPROBES, &vma->vm_mm->flags);
 }
 
+static vm_fault_t xol_fault(const struct vm_special_mapping *sm,
+			    struct vm_area_struct *vma, struct vm_fault *vmf)
+{
+	struct xol_area *area = vma->vm_mm->uprobes_state.xol_area;
+
+	vmf->page = area->pages[0];
+	get_page(vmf->page);
+	return 0;
+}
+
+static const struct vm_special_mapping xol_mapping = {
+	.name = "[uprobes]",
+	.fault = xol_fault,
+};
+
 /* Slot allocation for XOL */
 static int xol_add_vma(struct mm_struct *mm, struct xol_area *area)
 {
@@ -1479,7 +1493,7 @@ static int xol_add_vma(struct mm_struct *mm, struct xol_area *area)
 
 	vma = _install_special_mapping(mm, area->vaddr, PAGE_SIZE,
 				VM_EXEC|VM_MAYEXEC|VM_DONTCOPY|VM_IO,
-				&area->xol_mapping);
+				&xol_mapping);
 	if (IS_ERR(vma)) {
 		ret = PTR_ERR(vma);
 		goto fail;
@@ -1518,9 +1532,6 @@ static struct xol_area *__create_xol_area(unsigned long vaddr)
 	if (!area->bitmap)
 		goto free_area;
 
-	area->xol_mapping.name = "[uprobes]";
-	area->xol_mapping.fault = NULL;
-	area->xol_mapping.pages = area->pages;
 	area->pages[0] = alloc_page(GFP_HIGHUSER);
 	if (!area->pages[0])
 		goto free_bitmap;


  parent reply	other threads:[~2024-09-04  9:56 UTC|newest]

Thread overview: 18+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
     [not found] <CAHk-=wjD0XLhkzou89J-TK=L6B88pFoNYxN1uTWRQB3U5Czywg@mail.gmail.com>
2024-09-03  7:36 ` [PATCH] uprobes: use vm_special_mapping close() functionality Sven Schnelle
2024-09-03  7:49   ` Sven Schnelle
2024-09-04  3:57     ` Michael Ellerman
2024-09-04 21:26       ` Andrew Morton
2024-09-03  9:08   ` Oleg Nesterov
2024-09-03  9:32     ` Sven Schnelle
2024-09-03 19:12     ` Linus Torvalds
2024-09-03 19:31       ` Sven Schnelle
2024-09-03 19:34         ` Linus Torvalds
2024-09-03 19:32       ` Oleg Nesterov
2024-09-04  9:56       ` Oleg Nesterov [this message]
2024-09-04 10:03       ` Oleg Nesterov
2024-09-11  9:44   ` Oleg Nesterov
2024-09-11  9:57     ` Oleg Nesterov
2024-09-11 10:12       ` Oleg Nesterov
2024-09-11 13:13   ` [PATCH -mm 1/3] Revert "uprobes: use vm_special_mapping close() functionality" Oleg Nesterov
2024-09-11 13:14     ` [PATCH -mm 2/3] uprobes: introduce the global struct vm_special_mapping xol_mapping Oleg Nesterov
2024-09-11 13:14     ` [PATCH -mm 3/3] uprobes: turn xol_area->pages[2] into xol_area->page Oleg Nesterov

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=20240904095449.GA28781@redhat.com \
    --to=oleg@redhat.com \
    --cc=acme@kernel.org \
    --cc=adrian.hunter@intel.com \
    --cc=alexander.shishkin@linux.intel.com \
    --cc=irogers@google.com \
    --cc=jolsa@kernel.org \
    --cc=kan.liang@linux.intel.com \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-perf-users@vger.kernel.org \
    --cc=linux-trace-kernel@vger.kernel.org \
    --cc=mark.rutland@arm.com \
    --cc=mhiramat@kernel.org \
    --cc=mingo@redhat.com \
    --cc=mpe@ellerman.id.au \
    --cc=namhyung@kernel.org \
    --cc=peterz@infradead.org \
    --cc=svens@linux.ibm.com \
    --cc=torvalds@linux-foundation.org \
    /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;
as well as URLs for NNTP newsgroup(s).