From: Willy Tarreau <w@1wt.eu>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: Helge Deller <deller@gmx.de>,
Mathieu Desnoyers <mathieu.desnoyers@efficios.com>,
Ben Hutchings <ben@decadent.org.uk>, Willy Tarreau <w@1wt.eu>
Subject: [PATCH 2.6.32 38/42] parisc: Fix syscall restarts
Date: Sat, 23 Jan 2016 15:12:59 +0100 [thread overview]
Message-ID: <20160123141223.631979008@1wt.eu> (raw)
In-Reply-To: <aa387f55227cb730b41e3d621bf460ff@local>
2.6.32-longterm review patch. If anyone has any objections, please let me know.
------------------
From: Helge Deller <deller@gmx.de>
commit 71a71fb5374a23be36a91981b5614590b9e722c3 upstream.
On parisc syscalls which are interrupted by signals sometimes failed to
restart and instead returned -ENOSYS which in the worst case lead to
userspace crashes.
A similiar problem existed on MIPS and was fixed by commit e967ef02
("MIPS: Fix restart of indirect syscalls").
On parisc the current syscall restart code assumes that all syscall
callers load the syscall number in the delay slot of the ble
instruction. That's how it is e.g. done in the unistd.h header file:
ble 0x100(%sr2, %r0)
ldi #syscall_nr, %r20
Because of that assumption the current code never restored %r20 before
returning to userspace.
This assumption is at least not true for code which uses the glibc
syscall() function, which instead uses this syntax:
ble 0x100(%sr2, %r0)
copy regX, %r20
where regX depend on how the compiler optimizes the code and register
usage.
This patch fixes this problem by adding code to analyze how the syscall
number is loaded in the delay branch and - if needed - copy the syscall
number to regX prior returning to userspace for the syscall restart.
Signed-off-by: Helge Deller <deller@gmx.de>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
[bwh: Backported to 3.2: adjust context]
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
(cherry picked from commit 9f2dcffefc599c424ad1dd402e3a96da60639308)
Signed-off-by: Willy Tarreau <w@1wt.eu>
---
arch/parisc/kernel/signal.c | 64 ++++++++++++++++++++++++++++++++++++---------
1 file changed, 52 insertions(+), 12 deletions(-)
diff --git a/arch/parisc/kernel/signal.c b/arch/parisc/kernel/signal.c
index e8467e4..9b39674 100644
--- a/arch/parisc/kernel/signal.c
+++ b/arch/parisc/kernel/signal.c
@@ -474,6 +474,55 @@ handle_signal(unsigned long sig, siginfo_t *info, struct k_sigaction *ka,
return 1;
}
+/*
+ * Check how the syscall number gets loaded into %r20 within
+ * the delay branch in userspace and adjust as needed.
+ */
+
+static void check_syscallno_in_delay_branch(struct pt_regs *regs)
+{
+ u32 opcode, source_reg;
+ u32 __user *uaddr;
+ int err;
+
+ /* Usually we don't have to restore %r20 (the system call number)
+ * because it gets loaded in the delay slot of the branch external
+ * instruction via the ldi instruction.
+ * In some cases a register-to-register copy instruction might have
+ * been used instead, in which case we need to copy the syscall
+ * number into the source register before returning to userspace.
+ */
+
+ /* A syscall is just a branch, so all we have to do is fiddle the
+ * return pointer so that the ble instruction gets executed again.
+ */
+ regs->gr[31] -= 8; /* delayed branching */
+
+ /* Get assembler opcode of code in delay branch */
+ uaddr = (unsigned int *) ((regs->gr[31] & ~3) + 4);
+ err = get_user(opcode, uaddr);
+ if (err)
+ return;
+
+ /* Check if delay branch uses "ldi int,%r20" */
+ if ((opcode & 0xffff0000) == 0x34140000)
+ return; /* everything ok, just return */
+
+ /* Check if delay branch uses "nop" */
+ if (opcode == INSN_NOP)
+ return;
+
+ /* Check if delay branch uses "copy %rX,%r20" */
+ if ((opcode & 0xffe0ffff) == 0x08000254) {
+ source_reg = (opcode >> 16) & 31;
+ regs->gr[source_reg] = regs->gr[20];
+ return;
+ }
+
+ pr_warn("syscall restart: %s (pid %d): unexpected opcode 0x%08x\n",
+ current->comm, task_pid_nr(current), opcode);
+}
+
static inline void
syscall_restart(struct pt_regs *regs, struct k_sigaction *ka)
{
@@ -495,10 +544,7 @@ syscall_restart(struct pt_regs *regs, struct k_sigaction *ka)
}
/* fallthrough */
case -ERESTARTNOINTR:
- /* A syscall is just a branch, so all
- * we have to do is fiddle the return pointer.
- */
- regs->gr[31] -= 8; /* delayed branching */
+ check_syscallno_in_delay_branch(regs);
/* Preserve original r28. */
regs->gr[28] = regs->orig_r28;
break;
@@ -549,18 +595,12 @@ insert_restart_trampoline(struct pt_regs *regs)
}
case -ERESTARTNOHAND:
case -ERESTARTSYS:
- case -ERESTARTNOINTR: {
- /* Hooray for delayed branching. We don't
- * have to restore %r20 (the system call
- * number) because it gets loaded in the delay
- * slot of the branch external instruction.
- */
- regs->gr[31] -= 8;
+ case -ERESTARTNOINTR:
+ check_syscallno_in_delay_branch(regs);
/* Preserve original r28. */
regs->gr[28] = regs->orig_r28;
return;
- }
default:
break;
}
--
1.7.12.2.21.g234cd45.dirty
next prev parent reply other threads:[~2016-01-23 14:12 UTC|newest]
Thread overview: 43+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <aa387f55227cb730b41e3d621bf460ff@local>
2016-01-23 14:12 ` [PATCH 2.6.32 01/42] ip6mr: call del_timer_sync() in ip6mr_free_table() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 02/42] isdn_ppp: Add checks for allocation failure in isdn_ppp_open() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 04/42] RDS: fix race condition when sending a message on unbound socket Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 05/42] unix: avoid use-after-free in ep_remove_wait_queue Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 06/42] ext4: Fix null dereference in ext4_fill_super() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 07/42] Revert "net: add length argument to skb_copy_and_csum_datagram_iovec" Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 08/42] udp: properly support MSG_PEEK with truncated buffers Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 09/42] KEYS: Fix race between read and revoke Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 11/42] net: fix warnings in make htmldocs by moving macro definition out of field declaration Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 12/42] bluetooth: Validate socket address length in sco_sock_bind() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 13/42] sctp: translate host order to network order when setting a hmacid Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 14/42] fuse: break infinite loop in fuse_fill_write_pages() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 15/42] fix sysvfs symlinks Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 16/42] vfs: Avoid softlockups with sendfile(2) Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 17/42] ext4: Fix handling of extended tv_sec Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 18/42] nfs: if we have no valid attrs, then dont declare the attribute cache valid Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 19/42] wan/x25: Fix use-after-free in x25_asy_open_tty() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 20/42] ipv4: igmp: Allow removing groups from a removed interface Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 21/42] sched/core: Remove false-positive warning from wake_up_process() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 22/42] ipmi: move timer init to before irq is setup Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 23/42] tcp: initialize tp->copied_seq in case of cross SYN connection Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 24/42] net, scm: fix PaX detected msg_controllen overflow in scm_detach_fds Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 25/42] sctp: update the netstamp_needed counter when copying sockets Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 26/42] af_unix: fix a fatal race with bit fields Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 27/42] rfkill: copy the name into the rfkill struct Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 28/42] ses: Fix problems with simple enclosures Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 29/42] ses: fix additional element traversal bug Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 30/42] tty: Fix GPF in flush_to_ldisc() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 31/42] mISDN: fix a loop count Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 32/42] ser_gigaset: fix deallocation of platform device structure Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 33/42] spi: fix parent-device reference leak Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 34/42] s390/dis: Fix handling of format specifiers Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 35/42] USB: ipaq.c: fix a timeout loop Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 36/42] USB: fix invalid memory access in hub_activate() Willy Tarreau
2016-01-23 14:12 ` [PATCH 2.6.32 37/42] MIPS: Fix restart of indirect syscalls Willy Tarreau
2016-01-23 14:12 ` Willy Tarreau [this message]
2016-01-23 14:13 ` [PATCH 2.6.32 39/42] ipv6/addrlabel: fix ip6addrlbl_get() Willy Tarreau
2016-01-23 14:13 ` [PATCH 2.6.32 40/42] mm/memory_hotplug.c: check for missing sections in test_pages_in_a_zone() Willy Tarreau
2016-01-23 18:13 ` Ben Hutchings
2016-01-23 18:29 ` Willy Tarreau
2016-01-23 19:05 ` Willy Tarreau
2016-01-23 14:13 ` [PATCH 2.6.32 41/42] KVM: x86: Reload pit counters for all channels when restoring state Willy Tarreau
2016-01-23 14:13 ` [PATCH 2.6.32 42/42] kvm: x86: only channel 0 of the i8254 is linked to the HPET Willy Tarreau
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=20160123141223.631979008@1wt.eu \
--to=w@1wt.eu \
--cc=ben@decadent.org.uk \
--cc=deller@gmx.de \
--cc=linux-kernel@vger.kernel.org \
--cc=mathieu.desnoyers@efficios.com \
--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).