From: Song Liu <song@kernel.org>
To: <bpf@vger.kernel.org>, <netdev@vger.kernel.org>,
<linux-kernel@vger.kernel.org>
Cc: <ast@kernel.org>, <daniel@iogearbox.net>, <andrii@kernel.org>,
<kernel-team@fb.com>, <peterz@infradead.org>, <x86@kernel.org>,
Song Liu <songliubraving@fb.com>
Subject: [PATCH v2 bpf-next 7/7] bpf, x86_64: use bpf_prog_pack allocator
Date: Tue, 14 Dec 2021 22:01:02 -0800 [thread overview]
Message-ID: <20211215060102.3793196-8-song@kernel.org> (raw)
In-Reply-To: <20211215060102.3793196-1-song@kernel.org>
From: Song Liu <songliubraving@fb.com>
Use bpf_prog_pack allocator in x86_64 jit.
The program header from bpf_prog_pack is read only during the jit process.
Therefore, the binary is first written to a temporary buffer, and later
copied to final location with text_poke_jit().
Similarly, jit_fill_hole() is updated to fill the hole with 0xcc using
text_poke_jit().
Signed-off-by: Song Liu <songliubraving@fb.com>
---
arch/x86/net/bpf_jit_comp.c | 91 +++++++++++++++++++++++++++++++------
1 file changed, 78 insertions(+), 13 deletions(-)
diff --git a/arch/x86/net/bpf_jit_comp.c b/arch/x86/net/bpf_jit_comp.c
index 2231d483f82c..631bfbb7b1aa 100644
--- a/arch/x86/net/bpf_jit_comp.c
+++ b/arch/x86/net/bpf_jit_comp.c
@@ -216,11 +216,33 @@ static u8 simple_alu_opcodes[] = {
[BPF_ARSH] = 0xF8,
};
+static char jit_hole_buffer[PAGE_SIZE] = {};
+
static void jit_fill_hole(void *area, unsigned int size)
+{
+ struct bpf_binary_header *hdr = area;
+ int i;
+
+ for (i = 0; i < roundup(size, PAGE_SIZE); i += PAGE_SIZE) {
+ int s;
+
+ s = min_t(int, PAGE_SIZE, size - i);
+ text_poke_jit(area + i, jit_hole_buffer, s);
+ }
+
+ /* bpf_jit_binary_alloc_pack cannot write size directly to the ro
+ * mapping. Write it here with text_poke_jit().
+ */
+ text_poke_jit(&hdr->size, &size, sizeof(size));
+}
+
+static int __init x86_jit_fill_hole_init(void)
{
/* Fill whole space with INT3 instructions */
- memset(area, 0xcc, size);
+ memset(jit_hole_buffer, 0xcc, PAGE_SIZE);
+ return 0;
}
+pure_initcall(x86_jit_fill_hole_init);
struct jit_context {
int cleanup_addr; /* Epilogue code offset */
@@ -867,7 +889,7 @@ static void emit_nops(u8 **pprog, int len)
#define INSN_SZ_DIFF (((addrs[i] - addrs[i - 1]) - (prog - temp)))
-static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image,
+static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image, u8 *tmp_image,
int oldproglen, struct jit_context *ctx, bool jmp_padding)
{
bool tail_call_reachable = bpf_prog->aux->tail_call_reachable;
@@ -894,8 +916,8 @@ static int do_jit(struct bpf_prog *bpf_prog, int *addrs, u8 *image,
push_callee_regs(&prog, callee_regs_used);
ilen = prog - temp;
- if (image)
- memcpy(image + proglen, temp, ilen);
+ if (tmp_image)
+ memcpy(tmp_image + proglen, temp, ilen);
proglen += ilen;
addrs[0] = proglen;
prog = temp;
@@ -1289,8 +1311,10 @@ st: if (is_imm8(insn->off))
pr_err("extable->insn doesn't fit into 32-bit\n");
return -EFAULT;
}
- ex->insn = delta;
+ /* switch ex to temporary buffer for writes */
+ ex = (void *)tmp_image + ((void *)ex - (void *)image);
+ ex->insn = delta;
ex->type = EX_TYPE_BPF;
if (dst_reg > BPF_REG_9) {
@@ -1671,7 +1695,7 @@ st: if (is_imm8(insn->off))
pr_err("bpf_jit: fatal error\n");
return -EFAULT;
}
- memcpy(image + proglen, temp, ilen);
+ memcpy(tmp_image + proglen, temp, ilen);
}
proglen += ilen;
addrs[i] = proglen;
@@ -2213,8 +2237,10 @@ int arch_prepare_bpf_dispatcher(void *image, s64 *funcs, int num_funcs)
struct x64_jit_data {
struct bpf_binary_header *header;
+ struct bpf_binary_header *tmp_header;
int *addrs;
u8 *image;
+ u8 *tmp_image;
int proglen;
struct jit_context ctx;
};
@@ -2224,6 +2250,7 @@ struct x64_jit_data {
struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
{
+ struct bpf_binary_header *tmp_header = NULL;
struct bpf_binary_header *header = NULL;
struct bpf_prog *tmp, *orig_prog = prog;
struct x64_jit_data *jit_data;
@@ -2232,6 +2259,7 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
bool tmp_blinded = false;
bool extra_pass = false;
bool padding = false;
+ u8 *tmp_image = NULL;
u8 *image = NULL;
int *addrs;
int pass;
@@ -2266,7 +2294,9 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
ctx = jit_data->ctx;
oldproglen = jit_data->proglen;
image = jit_data->image;
+ tmp_image = jit_data->tmp_image;
header = jit_data->header;
+ tmp_header = jit_data->tmp_header;
extra_pass = true;
padding = true;
goto skip_init_addrs;
@@ -2297,14 +2327,18 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
for (pass = 0; pass < MAX_PASSES || image; pass++) {
if (!padding && pass >= PADDING_PASSES)
padding = true;
- proglen = do_jit(prog, addrs, image, oldproglen, &ctx, padding);
+ proglen = do_jit(prog, addrs, image, tmp_image, oldproglen, &ctx, padding);
if (proglen <= 0) {
out_image:
image = NULL;
- if (header)
- bpf_jit_binary_free(header);
+ tmp_image = NULL;
+ if (header) {
+ bpf_jit_binary_free_pack(header);
+ kfree(tmp_header);
+ }
prog = orig_prog;
header = NULL;
+ tmp_header = NULL;
goto out_addrs;
}
if (image) {
@@ -2327,13 +2361,22 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
sizeof(struct exception_table_entry);
/* allocate module memory for x86 insns and extable */
- header = bpf_jit_binary_alloc(roundup(proglen, align) + extable_size,
- &image, align, jit_fill_hole);
+ header = bpf_jit_binary_alloc_pack(roundup(proglen, align) + extable_size,
+ &image, align, jit_fill_hole);
if (!header) {
prog = orig_prog;
goto out_addrs;
}
- prog->aux->extable = (void *) image + roundup(proglen, align);
+ tmp_header = kzalloc(header->size, GFP_KERNEL);
+ if (!tmp_header) {
+ bpf_jit_binary_free_pack(header);
+ header = NULL;
+ prog = orig_prog;
+ goto out_addrs;
+ }
+ tmp_header->size = header->size;
+ tmp_image = (void *)tmp_header + ((void *)image - (void *)header);
+ prog->aux->extable = (void *)image + roundup(proglen, align);
}
oldproglen = proglen;
cond_resched();
@@ -2345,13 +2388,16 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
if (image) {
if (!prog->is_func || extra_pass) {
bpf_tail_call_direct_fixup(prog);
- bpf_jit_binary_lock_ro(header);
+ if (header->size > bpf_prog_pack_max_size())
+ bpf_jit_binary_lock_ro(header);
} else {
jit_data->addrs = addrs;
jit_data->ctx = ctx;
jit_data->proglen = proglen;
jit_data->image = image;
+ jit_data->tmp_image = tmp_image;
jit_data->header = header;
+ jit_data->tmp_header = tmp_header;
}
prog->bpf_func = (void *)image;
prog->jited = 1;
@@ -2367,6 +2413,11 @@ struct bpf_prog *bpf_int_jit_compile(struct bpf_prog *prog)
kvfree(addrs);
kfree(jit_data);
prog->aux->jit_data = NULL;
+ jit_data = NULL;
+ if (tmp_header) {
+ text_poke_jit(header, tmp_header, header->size);
+ kfree(tmp_header);
+ }
}
out:
if (tmp_blinded)
@@ -2380,3 +2431,17 @@ bool bpf_jit_supports_kfunc_call(void)
{
return true;
}
+
+void bpf_jit_free(struct bpf_prog *fp)
+{
+ if (fp->jited) {
+ struct bpf_binary_header *hdr = bpf_jit_binary_hdr(fp);
+
+ if (hdr->size > bpf_prog_pack_max_size())
+ bpf_jit_binary_free(hdr);
+ else
+ bpf_jit_binary_free_pack(hdr);
+ }
+
+ bpf_prog_unlock_free(fp);
+}
--
2.30.2
next prev parent reply other threads:[~2021-12-15 6:01 UTC|newest]
Thread overview: 19+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-12-15 6:00 [PATCH v2 bpf-next 0/7] bpf_prog_pack allocator Song Liu
2021-12-15 6:00 ` [PATCH v2 bpf-next 1/7] x86/Kconfig: select HAVE_ARCH_HUGE_VMALLOC with HAVE_ARCH_HUGE_VMAP Song Liu
2021-12-15 6:00 ` [PATCH v2 bpf-next 2/7] bpf: use bytes instead of pages for bpf_jit_[charge|uncharge]_modmem Song Liu
2021-12-15 8:56 ` Peter Zijlstra
2021-12-16 19:45 ` Song Liu
2021-12-15 6:00 ` [PATCH v2 bpf-next 3/7] bpf: use size instead of pages in bpf_binary_header Song Liu
2021-12-15 6:00 ` [PATCH v2 bpf-next 4/7] bpf: add a pointer of bpf_binary_header to bpf_prog Song Liu
2021-12-15 6:01 ` [PATCH v2 bpf-next 5/7] x86/alternative: introduce text_poke_jit Song Liu
2021-12-15 9:06 ` Peter Zijlstra
2021-12-15 9:17 ` Peter Zijlstra
2021-12-16 19:40 ` Song Liu
2021-12-15 6:01 ` [PATCH v2 bpf-next 6/7] bpf: introduce bpf_prog_pack allocator Song Liu
2021-12-15 6:01 ` Song Liu [this message]
2021-12-16 20:06 ` [PATCH v2 bpf-next 0/7] " Andrii Nakryiko
2021-12-17 1:53 ` Song Liu
2021-12-17 16:42 ` Andrii Nakryiko
2021-12-17 16:43 ` Andrii Nakryiko
2021-12-17 17:13 ` Song Liu
2021-12-17 17:16 ` Andrii Nakryiko
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=20211215060102.3793196-8-song@kernel.org \
--to=song@kernel.org \
--cc=andrii@kernel.org \
--cc=ast@kernel.org \
--cc=bpf@vger.kernel.org \
--cc=daniel@iogearbox.net \
--cc=kernel-team@fb.com \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=peterz@infradead.org \
--cc=songliubraving@fb.com \
--cc=x86@kernel.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.