linux-kernel.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
	stable@vger.kernel.org,
	Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
	"Peter Zijlstra (Intel)" <peterz@infradead.org>
Subject: [PATCH 5.4 41/58] selftests/rseq: Uplift rseq selftests for compatibility with glibc-2.35
Date: Tue,  5 Jul 2022 13:58:17 +0200	[thread overview]
Message-ID: <20220705115611.452786249@linuxfoundation.org> (raw)
In-Reply-To: <20220705115610.236040773@linuxfoundation.org>

From: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>

commit 233e667e1ae3e348686bd9dd0172e62a09d852e1 upstream.

glibc-2.35 (upcoming release date 2022-02-01) exposes the rseq per-thread
data in the TCB, accessible at an offset from the thread pointer, rather
than through an actual Thread-Local Storage (TLS) variable, as the
Linux kernel selftests initially expected.

The __rseq_abi TLS and glibc-2.35's ABI for per-thread data cannot
actively coexist in a process, because the kernel supports only a single
rseq registration per thread.

Here is the scheme introduced to ensure selftests can work both with an
older glibc and with glibc-2.35+:

- librseq exposes its own "rseq_offset, rseq_size, rseq_flags" ABI.

- librseq queries for glibc rseq ABI (__rseq_offset, __rseq_size,
  __rseq_flags) using dlsym() in a librseq library constructor. If those
  are found, copy their values into rseq_offset, rseq_size, and
  rseq_flags.

- Else, if those glibc symbols are not found, handle rseq registration
  from librseq and use its own IE-model TLS to implement the rseq ABI
  per-thread storage.

Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20220124171253.22072-8-mathieu.desnoyers@efficios.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 tools/testing/selftests/rseq/Makefile |    2 
 tools/testing/selftests/rseq/rseq.c   |  161 ++++++++++++++++------------------
 tools/testing/selftests/rseq/rseq.h   |   13 ++
 3 files changed, 88 insertions(+), 88 deletions(-)

--- a/tools/testing/selftests/rseq/Makefile
+++ b/tools/testing/selftests/rseq/Makefile
@@ -6,7 +6,7 @@ endif
 
 CFLAGS += -O2 -Wall -g -I./ -I../../../../usr/include/ -L$(OUTPUT) -Wl,-rpath=./ \
 	  $(CLANG_FLAGS)
-LDLIBS += -lpthread
+LDLIBS += -lpthread -ldl
 
 # Own dependencies because we only want to build against 1st prerequisite, but
 # still track changes to header files and depend on shared object.
--- a/tools/testing/selftests/rseq/rseq.c
+++ b/tools/testing/selftests/rseq/rseq.c
@@ -26,103 +26,113 @@
 #include <assert.h>
 #include <signal.h>
 #include <limits.h>
+#include <dlfcn.h>
 
 #include "../kselftest.h"
 #include "rseq.h"
 
-__thread struct rseq_abi __rseq_abi = {
-	.cpu_id = RSEQ_ABI_CPU_ID_UNINITIALIZED,
-};
+static const int *libc_rseq_offset_p;
+static const unsigned int *libc_rseq_size_p;
+static const unsigned int *libc_rseq_flags_p;
+
+/* Offset from the thread pointer to the rseq area.  */
+int rseq_offset;
+
+/* Size of the registered rseq area.  0 if the registration was
+   unsuccessful.  */
+unsigned int rseq_size = -1U;
 
-/*
- * Shared with other libraries. This library may take rseq ownership if it is
- * still 0 when executing the library constructor. Set to 1 by library
- * constructor when handling rseq. Set to 0 in destructor if handling rseq.
- */
-int __rseq_handled;
+/* Flags used during rseq registration.  */
+unsigned int rseq_flags;
 
-/* Whether this library have ownership of rseq registration. */
 static int rseq_ownership;
 
-static __thread volatile uint32_t __rseq_refcount;
+static
+__thread struct rseq_abi __rseq_abi __attribute__((tls_model("initial-exec"))) = {
+	.cpu_id = RSEQ_ABI_CPU_ID_UNINITIALIZED,
+};
 
-static void signal_off_save(sigset_t *oldset)
+static int sys_rseq(struct rseq_abi *rseq_abi, uint32_t rseq_len,
+		    int flags, uint32_t sig)
 {
-	sigset_t set;
-	int ret;
-
-	sigfillset(&set);
-	ret = pthread_sigmask(SIG_BLOCK, &set, oldset);
-	if (ret)
-		abort();
+	return syscall(__NR_rseq, rseq_abi, rseq_len, flags, sig);
 }
 
-static void signal_restore(sigset_t oldset)
+int rseq_available(void)
 {
-	int ret;
+	int rc;
 
-	ret = pthread_sigmask(SIG_SETMASK, &oldset, NULL);
-	if (ret)
+	rc = sys_rseq(NULL, 0, 0, 0);
+	if (rc != -1)
 		abort();
-}
-
-static int sys_rseq(volatile struct rseq_abi *rseq_abi, uint32_t rseq_len,
-		    int flags, uint32_t sig)
-{
-	return syscall(__NR_rseq, rseq_abi, rseq_len, flags, sig);
+	switch (errno) {
+	case ENOSYS:
+		return 0;
+	case EINVAL:
+		return 1;
+	default:
+		abort();
+	}
 }
 
 int rseq_register_current_thread(void)
 {
-	int rc, ret = 0;
-	sigset_t oldset;
+	int rc;
 
-	if (!rseq_ownership)
+	if (!rseq_ownership) {
+		/* Treat libc's ownership as a successful registration. */
 		return 0;
-	signal_off_save(&oldset);
-	if (__rseq_refcount == UINT_MAX) {
-		ret = -1;
-		goto end;
 	}
-	if (__rseq_refcount++)
-		goto end;
 	rc = sys_rseq(&__rseq_abi, sizeof(struct rseq_abi), 0, RSEQ_SIG);
-	if (!rc) {
-		assert(rseq_current_cpu_raw() >= 0);
-		goto end;
-	}
-	if (errno != EBUSY)
-		RSEQ_WRITE_ONCE(__rseq_abi.cpu_id, RSEQ_ABI_CPU_ID_REGISTRATION_FAILED);
-	ret = -1;
-	__rseq_refcount--;
-end:
-	signal_restore(oldset);
-	return ret;
+	if (rc)
+		return -1;
+	assert(rseq_current_cpu_raw() >= 0);
+	return 0;
 }
 
 int rseq_unregister_current_thread(void)
 {
-	int rc, ret = 0;
-	sigset_t oldset;
+	int rc;
 
-	if (!rseq_ownership)
+	if (!rseq_ownership) {
+		/* Treat libc's ownership as a successful unregistration. */
 		return 0;
-	signal_off_save(&oldset);
-	if (!__rseq_refcount) {
-		ret = -1;
-		goto end;
 	}
-	if (--__rseq_refcount)
-		goto end;
-	rc = sys_rseq(&__rseq_abi, sizeof(struct rseq_abi),
-		      RSEQ_ABI_FLAG_UNREGISTER, RSEQ_SIG);
-	if (!rc)
-		goto end;
-	__rseq_refcount = 1;
-	ret = -1;
-end:
-	signal_restore(oldset);
-	return ret;
+	rc = sys_rseq(&__rseq_abi, sizeof(struct rseq_abi), RSEQ_ABI_FLAG_UNREGISTER, RSEQ_SIG);
+	if (rc)
+		return -1;
+	return 0;
+}
+
+static __attribute__((constructor))
+void rseq_init(void)
+{
+	libc_rseq_offset_p = dlsym(RTLD_NEXT, "__rseq_offset");
+	libc_rseq_size_p = dlsym(RTLD_NEXT, "__rseq_size");
+	libc_rseq_flags_p = dlsym(RTLD_NEXT, "__rseq_flags");
+	if (libc_rseq_size_p && libc_rseq_offset_p && libc_rseq_flags_p) {
+		/* rseq registration owned by glibc */
+		rseq_offset = *libc_rseq_offset_p;
+		rseq_size = *libc_rseq_size_p;
+		rseq_flags = *libc_rseq_flags_p;
+		return;
+	}
+	if (!rseq_available())
+		return;
+	rseq_ownership = 1;
+	rseq_offset = (void *)&__rseq_abi - rseq_thread_pointer();
+	rseq_size = sizeof(struct rseq_abi);
+	rseq_flags = 0;
+}
+
+static __attribute__((destructor))
+void rseq_exit(void)
+{
+	if (!rseq_ownership)
+		return;
+	rseq_offset = 0;
+	rseq_size = -1U;
+	rseq_ownership = 0;
 }
 
 int32_t rseq_fallback_current_cpu(void)
@@ -136,20 +146,3 @@ int32_t rseq_fallback_current_cpu(void)
 	}
 	return cpu;
 }
-
-void __attribute__((constructor)) rseq_init(void)
-{
-	/* Check whether rseq is handled by another library. */
-	if (__rseq_handled)
-		return;
-	__rseq_handled = 1;
-	rseq_ownership = 1;
-}
-
-void __attribute__((destructor)) rseq_fini(void)
-{
-	if (!rseq_ownership)
-		return;
-	__rseq_handled = 0;
-	rseq_ownership = 0;
-}
--- a/tools/testing/selftests/rseq/rseq.h
+++ b/tools/testing/selftests/rseq/rseq.h
@@ -43,12 +43,19 @@
 #define RSEQ_INJECT_FAILED
 #endif
 
-extern __thread struct rseq_abi __rseq_abi;
-extern int __rseq_handled;
+#include "rseq-thread-pointer.h"
+
+/* Offset from the thread pointer to the rseq area.  */
+extern int rseq_offset;
+/* Size of the registered rseq area.  0 if the registration was
+   unsuccessful.  */
+extern unsigned int rseq_size;
+/* Flags used during rseq registration.  */
+extern unsigned int rseq_flags;
 
 static inline struct rseq_abi *rseq_get_abi(void)
 {
-	return &__rseq_abi;
+	return (struct rseq_abi *) ((uintptr_t) rseq_thread_pointer() + rseq_offset);
 }
 
 #define rseq_likely(x)		__builtin_expect(!!(x), 1)



  parent reply	other threads:[~2022-07-05 12:10 UTC|newest]

Thread overview: 65+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-07-05 11:57 [PATCH 5.4 00/58] 5.4.204-rc1 review Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 01/58] ipv6: take care of disable_policy when restoring routes Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 02/58] nvdimm: Fix badblocks clear off-by-one error Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 03/58] powerpc/prom_init: Fix kernel config grep Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 04/58] powerpc/bpf: Fix use of user_pt_regs in uapi Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 05/58] dm raid: fix accesses beyond end of raid member array Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 06/58] dm raid: fix KASAN warning in raid5_add_disks Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 07/58] s390/archrandom: simplify back to earlier design and initialize earlier Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 08/58] SUNRPC: Fix READ_PLUS crasher Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 09/58] net: rose: fix UAF bugs caused by timer handler Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 10/58] net: usb: ax88179_178a: Fix packet receiving Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 11/58] virtio-net: fix race between ndo_open() and virtio_device_ready() Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 12/58] selftests/net: pass ipv6_args to udpgso_benchs IPv6 TCP test Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 13/58] net: tun: unlink NAPI from device on destruction Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 14/58] net: tun: stop NAPI when detaching queues Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 15/58] RDMA/qedr: Fix reporting QP timeout attribute Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 16/58] linux/dim: Fix divide by 0 in RDMA DIM Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 17/58] usbnet: fix memory allocation in helpers Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 18/58] net: ipv6: unexport __init-annotated seg6_hmac_net_init() Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 19/58] caif_virtio: fix race between virtio_device_ready() and ndo_open() Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 20/58] PM / devfreq: exynos-ppmu: Fix refcount leak in of_get_devfreq_events Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 21/58] s390: remove unneeded select BUILD_BIN2C Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 22/58] netfilter: nft_dynset: restore set element counter when failing to update Greg Kroah-Hartman
2022-07-05 11:57 ` [PATCH 5.4 23/58] net/sched: act_api: Notify user space if any actions were flushed before error Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 24/58] net: bonding: fix possible NULL deref in rlb code Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 25/58] net: bonding: fix use-after-free after 802.3ad slave unbind Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 26/58] nfc: nfcmrvl: Fix irq_of_parse_and_map() return value Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 27/58] NFC: nxp-nci: Dont issue a zero length i2c_master_read() Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 28/58] net: tun: avoid disabling NAPI twice Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 29/58] xen/gntdev: Avoid blocking in unmap_grant_pages() Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 30/58] hwmon: (ibmaem) dont call platform_device_del() if platform_device_add() fails Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 31/58] net: dsa: bcm_sf2: force pause link settings Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 32/58] sit: use min Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 33/58] ipv6/sit: fix ipip6_tunnel_get_prl return value Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 34/58] rseq/selftests,x86_64: Add rseq_offset_deref_addv() Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 35/58] selftests/rseq: remove ARRAY_SIZE define from individual tests Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 36/58] selftests/rseq: introduce own copy of rseq uapi header Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 37/58] selftests/rseq: Remove useless assignment to cpu variable Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 38/58] selftests/rseq: Remove volatile from __rseq_abi Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 39/58] selftests/rseq: Introduce rseq_get_abi() helper Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 40/58] selftests/rseq: Introduce thread pointer getters Greg Kroah-Hartman
2022-07-05 11:58 ` Greg Kroah-Hartman [this message]
2022-07-05 11:58 ` [PATCH 5.4 42/58] selftests/rseq: Fix ppc32: wrong rseq_cs 32-bit field pointer on big endian Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 43/58] selftests/rseq: Fix ppc32 missing instruction selection "u" and "x" for load/store Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 44/58] selftests/rseq: Fix ppc32 offsets by using long rather than off_t Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 45/58] selftests/rseq: Fix warnings about #if checks of undefined tokens Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 46/58] selftests/rseq: Remove arm/mips asm goto compiler work-around Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 47/58] selftests/rseq: Fix: work-around asm goto compiler bugs Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 48/58] selftests/rseq: x86-64: use %fs segment selector for accessing rseq thread area Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 49/58] selftests/rseq: x86-32: use %gs " Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 50/58] selftests/rseq: Change type of rseq_offset to ptrdiff_t Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 51/58] xen/blkfront: fix leaking data in shared pages Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 52/58] xen/netfront: " Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 53/58] xen/netfront: force data bouncing when backend is untrusted Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 54/58] xen/blkfront: " Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 55/58] xen/arm: Fix race in RB-tree based P2M accounting Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 56/58] net: usb: qmi_wwan: add Telit 0x1060 composition Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 57/58] net: usb: qmi_wwan: add Telit 0x1070 composition Greg Kroah-Hartman
2022-07-05 11:58 ` [PATCH 5.4 58/58] clocksource/drivers/ixp4xx: remove EXPORT_SYMBOL_GPL from ixp4xx_timer_setup() Greg Kroah-Hartman
2022-07-05 16:53 ` [PATCH 5.4 00/58] 5.4.204-rc1 review Florian Fainelli
2022-07-06  5:51 ` Samuel Zou
2022-07-06  6:55 ` Naresh Kamboju
2022-07-06 10:19 ` Sudip Mukherjee (Codethink)
2022-07-06 13:43 ` Guenter Roeck
2022-07-07  0:00 ` Shuah Khan

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=20220705115611.452786249@linuxfoundation.org \
    --to=gregkh@linuxfoundation.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=mathieu.desnoyers@efficios.com \
    --cc=peterz@infradead.org \
    --cc=stable@vger.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 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).