From: Greg KH <gregkh@linuxfoundation.org>
To: linux-kernel@vger.kernel.org, stable@vger.kernel.org
Cc: torvalds@linux-foundation.org, akpm@linux-foundation.org,
alan@lxorguk.ukuu.org.uk, Simon Arlott <simon@fire.lp0.eu>
Subject: [ 46/75] USB: ftdi_sio: fix race condition in TIOCMIWAIT, and abort of TIOCMIWAIT when the device is removed
Date: Thu, 19 Apr 2012 14:03:49 -0700 [thread overview]
Message-ID: <20120419210307.059005582@linuxfoundation.org> (raw)
In-Reply-To: <20120419210322.GA6478@kroah.com>
[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 3815 bytes --]
3.3-stable review patch. If anyone has any objections, please let me know.
------------------
From: Simon Arlott <simon@fire.lp0.eu>
commit 876ae50d94b02f3f523aa451b45ec5fb9c25d221 upstream.
There are two issues here, one is that the device is generating
spurious very fast modem status line changes somewhere:
CTS becomes high then low 18µs later:
[121226.924373] ftdi_process_packet: prev rng=0 dsr=10 dcd=0 cts=6
[121226.924378] ftdi_process_packet: status=10 prev=00 diff=10
[121226.924382] ftdi_process_packet: now rng=0 dsr=10 dcd=0 cts=7
(wake_up_interruptible is called)
[121226.924391] ftdi_process_packet: prev rng=0 dsr=10 dcd=0 cts=7
[121226.924394] ftdi_process_packet: status=00 prev=10 diff=10
[121226.924397] ftdi_process_packet: now rng=0 dsr=10 dcd=0 cts=8
(wake_up_interruptible is called)
This wakes up the task in TIOCMIWAIT:
[121226.924405] ftdi_ioctl: 19451 rng=0->0 dsr=10->10 dcd=0->0 cts=6->8
(wait from 20:51:46 returns and observes both changes)
Which then calls TIOCMIWAIT again:
20:51:46.400239 ioctl(3, TIOCMIWAIT, 0x20) = 0
22:11:09.441818 ioctl(3, TIOCMGET, [TIOCM_DTR|TIOCM_RTS]) = 0
22:11:09.442812 ioctl(3, TIOCMIWAIT, 0x20) = -1 EIO (Input/output error)
(the second wake_up_interruptible takes effect and an I/O error occurs)
The other issue is that TIOCMIWAIT will wait forever (unless the task is
interrupted) if the device is removed.
This change removes the -EIO return that occurs if the counts don't
appear to have changed. Multiple counts may have been processed as
one or the waiting task may have started waiting after recording the
current count.
It adds a bool to indicate that the device has been removed so that
TIOCMIWAIT doesn't wait forever, and wakes up any tasks so that they can
return -EIO.
Signed-off-by: Simon Arlott <simon@fire.lp0.eu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
drivers/usb/serial/ftdi_sio.c | 12 +++++++-----
1 file changed, 7 insertions(+), 5 deletions(-)
--- a/drivers/usb/serial/ftdi_sio.c
+++ b/drivers/usb/serial/ftdi_sio.c
@@ -76,6 +76,7 @@ struct ftdi_private {
struct async_icount icount;
wait_queue_head_t delta_msr_wait; /* Used for TIOCMIWAIT */
char prev_status; /* Used for TIOCMIWAIT */
+ bool dev_gone; /* Used to abort TIOCMIWAIT */
char transmit_empty; /* If transmitter is empty or not */
struct usb_serial_port *port;
__u16 interface; /* FT2232C, FT2232H or FT4232H port interface
@@ -1679,6 +1680,7 @@ static int ftdi_sio_port_probe(struct us
init_waitqueue_head(&priv->delta_msr_wait);
priv->flags = ASYNC_LOW_LATENCY;
+ priv->dev_gone = false;
if (quirk && quirk->port_probe)
quirk->port_probe(priv);
@@ -1836,6 +1838,9 @@ static int ftdi_sio_port_remove(struct u
dbg("%s", __func__);
+ priv->dev_gone = true;
+ wake_up_interruptible_all(&priv->delta_msr_wait);
+
remove_sysfs_attrs(port);
kref_put(&priv->kref, ftdi_sio_priv_release);
@@ -2394,15 +2399,12 @@ static int ftdi_ioctl(struct tty_struct
*/
case TIOCMIWAIT:
cprev = priv->icount;
- while (1) {
+ while (!priv->dev_gone) {
interruptible_sleep_on(&priv->delta_msr_wait);
/* see if a signal did it */
if (signal_pending(current))
return -ERESTARTSYS;
cnow = priv->icount;
- if (cnow.rng == cprev.rng && cnow.dsr == cprev.dsr &&
- cnow.dcd == cprev.dcd && cnow.cts == cprev.cts)
- return -EIO; /* no change => error */
if (((arg & TIOCM_RNG) && (cnow.rng != cprev.rng)) ||
((arg & TIOCM_DSR) && (cnow.dsr != cprev.dsr)) ||
((arg & TIOCM_CD) && (cnow.dcd != cprev.dcd)) ||
@@ -2411,7 +2413,7 @@ static int ftdi_ioctl(struct tty_struct
}
cprev = cnow;
}
- /* not reached */
+ return -EIO;
break;
case TIOCSERGETLSR:
return get_lsr_info(port, (struct serial_struct __user *)arg);
next prev parent reply other threads:[~2012-04-19 21:03 UTC|newest]
Thread overview: 76+ messages / expand[flat|nested] mbox.gz Atom feed top
2012-04-19 21:03 [ 00/75] 3.3.3-stable review Greg KH
2012-04-19 21:03 ` [ 01/75] Btrfs: fix regression in scrub path resolving Greg KH
2012-04-19 21:03 ` [ 02/75] drm/radeon/kms: fix DVO setup on some r4xx chips Greg KH
2012-04-19 21:03 ` [ 03/75] drm/i915: Removed IVB forced enable of sprite dest key Greg KH
2012-04-19 21:03 ` [ 04/75] drm/i915/ringbuffer: Exclude last 2 cachlines of ring on 845g Greg KH
2012-04-19 21:03 ` [ 05/75] drm/radeon: only add the mm i2c bus if the hw_i2c module param is set Greg KH
2012-04-19 21:03 ` [ 06/75] drm/i915: properly compute dp dithering for user-created modes Greg KH
2012-04-19 21:03 ` [ 07/75] drm/i915: make rc6 module parameter read-only Greg KH
2012-04-19 21:03 ` [ 08/75] rtlwifi: Preallocate USB read buffers and eliminate kalloc in read routine Greg KH
2012-04-19 21:03 ` [ 09/75] rtlwifi: Add missing DMA buffer unmapping for PCI drivers Greg KH
2012-04-19 21:03 ` [ 10/75] ARM: 7379/1: DT: fix atags_to_fdt() second call site Greg KH
2012-04-19 21:03 ` [ 11/75] ARM: 7384/1: ThumbEE: Disable userspace TEEHBR access for !CONFIG_ARM_THUMBEE Greg KH
2012-04-19 21:03 ` [ 12/75] md/raid1,raid10: Fix calculation of vcnt when processing error recovery Greg KH
2012-04-19 21:03 ` [ 13/75] md/bitmap: prevent bitmap_daemon_work running while initialising bitmap Greg KH
2012-04-19 21:03 ` [ 14/75] [PATCH] Bluetooth: uart-ldisc: Fix memory leak Greg KH
2012-04-19 21:03 ` [ 15/75] Bluetooth: hci_ldisc: fix NULL-pointer dereference on tty_close Greg KH
2012-04-19 21:03 ` [ 16/75] Bluetooth: hci_core: fix NULL-pointer dereference at unregister Greg KH
2012-04-19 21:03 ` [ 17/75] Bluetooth: Remove unneeded locking Greg KH
2012-04-19 21:03 ` [ 18/75] Revert "Btrfs: increase the global block reserve estimates" Greg KH
2012-04-19 21:03 ` [ 19/75] ALSA: hda/realtek - Add a fixup entry for Acer Aspire 8940G Greg KH
2012-04-19 21:03 ` [ 20/75] ext4: address scalability issue by removing extent cache statistics Greg KH
2012-04-19 21:03 ` [ 21/75] ia64: fix futex_atomic_cmpxchg_inatomic() Greg KH
2012-04-19 21:03 ` [ 22/75] panic: fix stack dump print on direct call to panic() Greg KH
2012-04-19 21:03 ` [ 23/75] drivers/rtc/rtc-pl031.c: enable clock on all ST variants Greg KH
2012-04-19 21:03 ` [ 24/75] hugetlb: fix race condition in hugetlb_fault() Greg KH
2012-04-19 21:03 ` [ 25/75] staging: iio: hmc5843: Fix crash in probe function Greg KH
2012-04-19 21:03 ` [ 26/75] Revert "serial/8250_pci: init-quirk msi support for kt serial controller" Greg KH
2012-04-19 21:03 ` [ 27/75] serial: samsung: fix omission initialize ulcon in reset port fn() Greg KH
2012-04-19 21:03 ` [ 28/75] Revert "serial/8250_pci: setup-quirk workaround for the kt serial controller" Greg KH
2012-04-19 21:03 ` [ 29/75] serial/8250_pci: add a "force background timer" flag and use it for the "kt" serial port Greg KH
2012-04-19 21:03 ` [ 30/75] tty: serial: altera_uart: Check for NULL platform_data in probe Greg KH
2012-04-19 21:03 ` [ 31/75] sparc64: Eliminate obsolete __handle_softirq() function Greg KH
2012-04-19 21:03 ` [ 32/75] sparc64: Fix bootup crash on sun4v Greg KH
2012-04-19 21:03 ` [ 33/75] cciss: Initialize scsi host max_sectors for tape drive support Greg KH
2012-04-19 21:03 ` [ 34/75] cciss: Fix scsi tape io with more than 255 scatter gather elements Greg KH
2012-04-19 21:03 ` [ 35/75] perf hists: Catch and handle out-of-date hist entry maps Greg KH
2012-04-19 21:03 ` [ 36/75] video:uvesafb: Fix oops that uvesafb try to execute NX-protected page Greg KH
2012-04-19 21:03 ` [ 37/75] IB/srpt: Set srq_type to IB_SRQT_BASIC Greg KH
2012-04-19 21:03 ` [ 38/75] nohz: Fix stale jiffies update in tick_nohz_restart() Greg KH
2012-04-19 21:03 ` [ 39/75] pch_uart: Fix MSI setting issue Greg KH
2012-04-19 21:03 ` [ 40/75] x86: Use correct byte-sized register constraint in __xchg_op() Greg KH
2012-04-19 21:03 ` [ 41/75] x86: Use correct byte-sized register constraint in __add() Greg KH
2012-04-19 21:03 ` [ 42/75] USB: serial: fix race between probe and open Greg KH
2012-04-19 21:03 ` [ 43/75] USB: pl2303: fix DTR/RTS being raised on baud rate change Greg KH
2012-04-19 21:03 ` [ 44/75] USB: option: re-add NOVATELWIRELESS_PRODUCT_HSPA_HIGHSPEED to option_id array Greg KH
2012-04-19 21:03 ` [ 45/75] USB: ftdi_sio: fix status line change handling for TIOCMIWAIT and TIOCGICOUNT Greg KH
2012-04-19 21:03 ` Greg KH [this message]
2012-04-19 21:03 ` [ 47/75] USB: sierra: add support for Sierra Wireless MC7710 Greg KH
2012-04-19 21:03 ` [ 48/75] USB: dont clear urb->dev in scatter-gather library Greg KH
2012-04-19 21:03 ` [ 49/75] USB: dont ignore suspend errors for root hubs Greg KH
2012-04-19 21:03 ` [ 50/75] xhci: dont re-enable IE constantly Greg KH
2012-04-19 21:03 ` [ 51/75] xhci: Dont write zeroed pointers to xHC registers Greg KH
2012-04-19 21:03 ` [ 52/75] xhci: Restore event ring dequeue pointer on resume Greg KH
2012-04-19 21:03 ` [ 53/75] USB: fix bug of device descriptor got from superspeed device Greg KH
2012-04-19 21:03 ` [ 54/75] xHCI: add XHCI_RESET_ON_RESUME quirk for VIA xHCI host Greg KH
2012-04-19 21:03 ` [ 55/75] xHCI: Correct the #define XHCI_LEGACY_DISABLE_SMI Greg KH
2012-04-19 21:03 ` [ 56/75] [S390] fix tlb flushing for page table pages Greg KH
2012-04-19 21:04 ` [ 57/75] memcg: fix Bad page state after replace_page_cache Greg KH
2012-04-19 21:04 ` [ 58/75] serial: PL011: clear pending interrupts Greg KH
2012-04-19 21:04 ` [ 59/75] serial: PL011: move interrupt clearing Greg KH
2012-04-19 21:04 ` [ 60/75] fcaps: clear the same personality flags as suid when fcaps are used Greg KH
2012-04-19 21:04 ` [ 61/75] xhci: Fix register save/restore order Greg KH
2012-04-19 21:04 ` [ 62/75] usb: gadget: pch_udc: Fix disconnect issue Greg KH
2012-04-19 21:04 ` [ 63/75] usb: gadget: pch_udc: Fix wrong return value Greg KH
2012-04-19 21:04 ` [ 64/75] usb: gadget: pch_udc: Fix USB suspend issue Greg KH
2012-04-19 21:04 ` [ 65/75] usb: gadget: pch_udc: Fix usb/gadget/pch_udc: Fix ether gadget connect/disconnect issue Greg KH
2012-04-19 21:04 ` [ 66/75] usb: gadget: pch_udc: Reduce redundant interrupt Greg KH
2012-04-19 21:04 ` [ 67/75] security: fix compile error in commoncap.c Greg KH
2012-04-19 21:04 ` [ 68/75] spi-topcliff-pch: fix -Wuninitialized warning Greg KH
2012-04-19 21:04 ` [ 69/75] Bluetooth: Adding USB device 13d3:3375 as an Atheros AR3012 Greg KH
2012-04-19 21:04 ` [ 70/75] Bluetooth: Add Atheros maryann PIDVID support Greg KH
2012-04-19 21:04 ` [ 71/75] futex: Do not leak robust list to unprivileged process Greg KH
2012-04-19 21:04 ` [ 72/75] drm/i915: Hold mode_config lock whilst changing mode for lastclose() Greg KH
2012-04-19 21:04 ` [ 73/75] drm/radeon/kms: fix the regression of DVI connector check Greg KH
2012-04-19 21:04 ` [ 74/75] drm/radeon: disable MSI on RV515 Greg KH
2012-04-19 21:04 ` [ 75/75] drm/radeon: fix load detect on rn50 with hardcoded EDIDs Greg KH
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=20120419210307.059005582@linuxfoundation.org \
--to=gregkh@linuxfoundation.org \
--cc=akpm@linux-foundation.org \
--cc=alan@lxorguk.ukuu.org.uk \
--cc=linux-kernel@vger.kernel.org \
--cc=simon@fire.lp0.eu \
--cc=stable@vger.kernel.org \
--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).