aboutsummaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* MAINTAINERS: Add myself as a maintainer for altera_uart/altera_jtaguartTobias Klauser2011-02-171-0/+10
| | | | | Signed-off-by: Tobias Klauser <tklauser@distanz.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty: serial: altera_uart: Use port->regshift to store bus shiftTobias Klauser2011-02-171-7/+7
| | | | | | | | | | | | | Use the regshift member of struct uart_port to store the address stride from platform data. This way we can save one dereference per call of altera_uart_readl and altera_uart_writel. This also allows us to use the driver without platform data, which is needed for device tree support in the Nios2 port. Acked-by: Anton Vorontsov <cbouatmailru@gmail.com> Signed-off-by: Tobias Klauser <tklauser@distanz.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty: serial: altera_uart: Handle pdev->id == -1 in altera_uart_removeTobias Klauser2011-02-171-1/+7
| | | | | | | | | | Commit 6b5756f176568a710d008d3b478128fafb6707f0 introduced the possibility for pdev->id being -1 but the change was not done equally in altera_uart_remove. This patch fixes this. Acked-by: Anton Vorontsov <cbouatmailru@gmail.com> Signed-off-by: Tobias Klauser <tklauser@distanz.ch> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty,vt: fix VT_SETACTIVATE console switchJiri Olsa2011-02-171-1/+2
| | | | | | | | | | | using VT_SETACTIVATE ioctl for console switch did not work, since it put wrong param to the set_console function. Also ioctl returned misleading error, because of the missing break statement. I wonder anyone has ever used this one :). Signed-off-by: Jiri Olsa <jolsa@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* atmel_serial: enable PPS supportViktar Palstsiuk2011-02-171-0/+16
| | | | | | | | | Enables PPS support in atmel serial driver to make PPS API working. Signed-off-by: Viktar Palstsiuk <viktar.palstsiuk@promwad.com> Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com> Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty: Add msm_smd_tty driverNiranjana Vishwanathapura2011-02-173-0/+246
| | | | | | | | | | msm_smd_tty driver provides tty device interface to 'DS' and 'GPSNMEA' streaming SMD ports. Cc: Brian Swetland <swetland@google.com> Signed-off-by: Niranjana Vishwanathapura <nvishwan@codeaurora.org> Acked-by: Alan Cox <alan@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: change the divisor latch only when prescalar actually changedYin Kangkai2011-02-171-12/+20
| | | | | | | | | | | | | | | | | | | | In 8250.c original ns16550 autoconfig code, we change the divisor latch when we goto to high speed mode, we're assuming the previous speed is legacy. This some times is not true. For example in a system with both CONFIG_SERIAL_8250 and CONFIG_SERIAL_8250_PNP set, in this case, the code (autoconfig) will be called twice, one in serial8250_init/probe() and the other is from serial_pnp_probe. When serial_pnp_probe calls the autoconfig for NS16550A, it's already in high speed mode, change the divisor latch (quot << 3) in this case will make the UART console garbled. CC: Greg Kroah-Hartman <greg@kroah.com> CC: David Woodhouse <dwmw2@infradead.org> CC: linux-kernel@vger.kernel.org CC: stable@kernel.org Signed-off-by: Yin Kangkai <kangkai.yin@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: also set the uartclk value in resume after goes to highspeedYin Kangkai2011-02-171-0/+1
| | | | | | | | | | | | | | For any reason if the NS16550A was not work in high speed mode (e.g. we hold NS16550A from going to high speed mode in autoconfig_16550a()), now we are resume from suspend, we should also set the uartclk to the correct value. Otherwise it is still the old 1843200 and that will bring issues. CC: Greg Kroah-Hartman <greg@kroah.com> CC: David Woodhouse <dwmw2@infradead.org> CC: linux-kernel@vger.kernel.org CC: stable@kernel.org Signed-off-by: Yin Kangkai <kangkai.yin@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: ifx6x60: minor cleanupRuss Gorby2011-02-172-6/+4
| | | | | | | | | renamed spi_driver variable to not be h/w specific set driver name to use DRVNAME define removed commented-out define Signed-off-by: Russ Gorby <russ.gorby@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: ifx6x60: probe routine needs to call spi_setupRuss Gorby2011-02-171-0/+9
| | | | | | | | | | | The probe routine should call spi_setup() to configure the SPI bus so it can properly communicate with the device. E.g. the device operates in SPI mode 1. Called spi_setup to configure SPI mode, max_speed_hz, and bpw Signed-off-by: Russ Gorby <russ.gorby@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: ifx6x60: set SPI max_speed_hz based on platform typeRuss Gorby2011-02-171-1/+2
| | | | | | | | | Platforms containing the 6260 can run up to 25Mhz. For these platforms set max_speed_hz to 25Mhz. Signed-off-by: Russ Gorby <russ.gorby@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: ifx6x60: changed internal bpw from boolean to intRuss Gorby2011-02-171-2/+2
| | | | | | | | | | | driver should support 32bit SPI transfers. The boolean variable only allowed 8/16. Changed to support 8/16/32 for future enabling of 32 bpw. Signed-off-by: Russ Gorby <russ.gorby@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: ifx6x60: dma_alloc_coherent must use parent devRuss Gorby2011-02-171-2/+2
| | | | | | | | | | | This driver is a SPI protocol driver and has no DMA ops associated with the device so the call will fail. Furthermore, the DMA allocation made here will be used by the SPI controller driver (parent dev) so it makes sense to pass that device instead. Signed-off-by: Russ Gorby <russ.gorby@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: ifx6x60: fixed call to tty_port_initRuss Gorby2011-02-171-1/+1
| | | | | | | | The port ops must be set AFTER calling port init as that function zeroes the structure Signed-off-by: Russ Gorby <russ.gorby@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty,vcs removing con_buf/conf_buf_mtxJiri Olsa2011-02-173-66/+52
| | | | | | | | | | | | | | | | | | | | | | | | | | | | seems there's no longer need for using con_buf/conf_buf_mtx as vcs_read/vcs_write buffer for user's data. The do_con_write function, that was the other user of this, is currently using its own kmalloc-ed buffer. Not sure when this got changed, as I was able to find this code in 2.6.9, but it's already gone as far as current git history goes - 2.6.12-rc2. AFAICS there's a behaviour change with the current change. The lseek is not completely mutually exclusive with the vcs_read/vcs_write - the file->f_pos might get updated via lseek callback during the vcs_read/vcs_write processing. I tried to find out if the prefered behaviour is to keep this in sync within read/write/lseek functions, but I did not find any pattern on different places. I guess if user end up calling write/lseek from different threads she should know what she's doing. If needed we could use dedicated fd mutex/buffer. Signed-off-by: Jiri Olsa <jolsa@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty,vcs: lseek/VC-release race fixJiri Olsa2011-02-171-0/+18
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | there's a race between vcs's lseek handler and VC release. The lseek handler does not hold console_lock and touches VC's size info. If during this the VC got released, there's an access violation. Following program triggers the issue for me: [SNIP] #define _BSD_SOURCE #include <stdio.h> #include <sys/types.h> #include <sys/stat.h> #include <fcntl.h> #include <sys/ioctl.h> #include <linux/vt.h> #include <unistd.h> #include <errno.h> static int run_seek(void) { while(1) { int fd; fd = open("./vcs30", O_RDWR); while(lseek(fd, 0, 0) != -1); close(fd); } } static int open_ioctl_tty(void) { return open("/dev/tty1", O_RDWR); } static int do_ioctl(int fd, int req, int i) { return ioctl(fd, req, i); } #define INIT(i) do_ioctl(ioctl_fd, VT_ACTIVATE, i) #define SHUT(i) do_ioctl(ioctl_fd, VT_DISALLOCATE, i) int main(int argc, char **argv) { int ioctl_fd = open_ioctl_tty(); if (ioctl < 0) { perror("open tty1 failed\n"); return -1; } if ((-1 == mknod("vcs30", S_IFCHR|0666, makedev(7, 30))) && (errno != EEXIST)) { printf("errno %d\n", errno); perror("failed to create vcs30"); return -1; } do_ioctl(ioctl_fd, VT_LOCKSWITCH, 0); if (!fork()) run_seek(); while(1) { INIT(30); SHUT(30); } return 0; } [SNIP] Signed-off-by: Jiri Olsa <jolsa@redhat.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* TTY: use appropriate printk priority levelMandeep Singh Baines2011-02-171-7/+8
| | | | | | | | | | printk()s without a priority level default to KERN_WARNING. To reduce noise at KERN_WARNING, this patch set the priority level appriopriately for unleveled printks()s. This should be useful to folks that look at dmesg warnings closely. Signed-off-by: Mandeep Singh Baines <msb@chromium.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* hvc: add Blackfin JTAG console supportMike Frysinger2011-02-173-0/+115
| | | | | | | | | This converts the existing bfin_jtag_comm TTY driver to the HVC layer so that the common HVC code can worry about all of the TTY/polling crap and leave the Blackfin code to worry about the Blackfin bits. Signed-off-by: Mike Frysinger <vapier@gentoo.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* vt: Add virtual console keyboard mode OFFArthur Taylor2011-02-174-3/+9
| | | | | | | | | | | | virtual console: add keyboard mode OFF Add a new mode for the virtual console keyboard OFF in which all input other than shift keys is ignored. Prevents vt input buffers from overflowing when a program opens but doesn't read from a tty, like X11 using evdev for input. Signed-off-by: Arthur Taylor <art@ified.ca> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* hvc_dcc: Simplify assembly for v6 and v7 ARMStephen Boyd2011-02-031-24/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | The inline assembly differences for v6 vs. v7 in the hvc_dcc driver are purely optimizations. On a v7 processor, an mrc with the pc sets the condition codes to the 28-31 bits of the register being read. It just so happens that the TX/RX full bits the DCC driver is testing for are high enough in the register to be put into the condition codes. On a v6 processor, this "feature" isn't implemented and thus we have to do the usual read, mask, test operations to check for TX/RX full. Since we already test the RX/TX full bits before calling __dcc_getchar() and __dcc_putchar() we don't actually need to do anything special for v7 over v6. The only difference is in hvc_dcc_get_chars(). We would test RX full, poll RX full, and then read a character from the buffer, whereas now we will test RX full, read a character from the buffer, and then test RX full again for the second iteration of the loop. It doesn't seem possible for the buffer to go from full to empty between testing the RX full and reading a character. Therefore, replace the v7 versions with the v6 versions and everything works the same. Acked-by: Tony Lindgren <tony@atomide.com> Cc: Arnd Bergmann <arnd@arndb.de> Acked-by: Nicolas Pitre <nicolas.pitre@linaro.org> Cc: Daniel Walker <dwalker@codeaurora.org> Signed-off-by: Stephen Boyd <sboyd@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* hvc_dcc: Simplify put_chars()/get_chars() loopsStephen Boyd2011-02-031-8/+4
| | | | | | | | | | | | Casting and anding with 0xff is unnecessary in hvc_dcc_put_chars() since buf is already a char[]. __dcc_get_char() can't return an int less than 0 since it only returns a char. Simplify the if statement in hvc_dcc_get_chars() to take this into account. Cc: Daniel Walker <dwalker@codeaurora.org> Signed-off-by: Stephen Boyd <sboyd@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* hvc_dcc: Fix bad code generation by marking assembly volatileStephen Boyd2011-02-031-6/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Without marking the asm __dcc_getstatus() volatile my compiler decides it can cache the value of __ret in a register and then check the value of it continually in hvc_dcc_put_chars() (I had to replace get_wait/put_wait with 1 and fixup the branch otherwise my disassembler barfed on __dcc_(get|put)char). 00000000 <hvc_dcc_put_chars>: 0: ee103e11 mrc 14, 0, r3, cr0, cr1, {0} 4: e3a0c000 mov ip, #0 ; 0x0 8: e2033202 and r3, r3, #536870912 ; 0x20000000 c: ea000006 b 2c <hvc_dcc_put_chars+0x2c> 10: e3530000 cmp r3, #0 ; 0x0 14: 1afffffd bne 10 <hvc_dcc_put_chars+0x10> 18: e7d1000c ldrb r0, [r1, ip] 1c: ee10fe11 mrc 14, 0, pc, cr0, cr1, {0} 20: 2afffffd bcs 1c <hvc_dcc_put_chars+0x1c> 24: ee000e15 mcr 14, 0, r0, cr0, cr5, {0} 28: e28cc001 add ip, ip, #1 ; 0x1 2c: e15c0002 cmp ip, r2 30: bafffff6 blt 10 <hvc_dcc_put_chars+0x10> 34: e1a00002 mov r0, r2 38: e12fff1e bx lr As you can see, the value of the mrc is checked against DCC_STATUS_TX (bit 29) and then stored in r3 for later use. Marking the asm volatile produces the following: 00000000 <hvc_dcc_put_chars>: 0: e3a03000 mov r3, #0 ; 0x0 4: ea000007 b 28 <hvc_dcc_put_chars+0x28> 8: ee100e11 mrc 14, 0, r0, cr0, cr1, {0} c: e3100202 tst r0, #536870912 ; 0x20000000 10: 1afffffc bne 8 <hvc_dcc_put_chars+0x8> 14: e7d10003 ldrb r0, [r1, r3] 18: ee10fe11 mrc 14, 0, pc, cr0, cr1, {0} 1c: 2afffffd bcs 18 <hvc_dcc_put_chars+0x18> 20: ee000e15 mcr 14, 0, r0, cr0, cr5, {0} 24: e2833001 add r3, r3, #1 ; 0x1 28: e1530002 cmp r3, r2 2c: bafffff5 blt 8 <hvc_dcc_put_chars+0x8> 30: e1a00002 mov r0, r2 34: e12fff1e bx lr which looks better and actually works. Mark all the inline assembly in this file as volatile since we don't want the compiler to optimize away these statements or move them around in any way. Acked-by: Tony Lindgren <tony@atomide.com> Cc: Arnd Bergmann <arnd@arndb.de> Acked-by: Nicolas Pitre <nicolas.pitre@linaro.org> Cc: Daniel Walker <dwalker@codeaurora.org> Signed-off-by: Stephen Boyd <sboyd@codeaurora.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: pch_uart: revert Kconfig for non-DMA modeTomoya MORINAGA2011-02-031-3/+2
| | | | | | | | | | PCH_DMA is not always enabled when a user uses PCH_UART. Since overhead of DMA is not small, in case of low frequent communication, without DMA is better. Thus, "select PCH_DMA" and DMADEVICES are unnecessary Signed-off-by: Tomoya MORINAGA <tomoya-linux@dsn.okisemi.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: pch_uart: support new device ML7213Tomoya MORINAGA2011-02-032-8/+24
| | | | | | | | | Support ML7213 device of OKI SEMICONDUCTOR. ML7213 is companion chip of Intel Atom E6xx series for IVI(In-Vehicle Infotainment). ML7213 is completely compatible for Intel EG20T PCH. Signed-off-by: Tomoya MORINAGA <tomoya-linux@dsn.okisemi.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* 68328serial: remove unsed m68k_serial->tqueue_hangupTejun Heo2011-02-032-24/+0
| | | | | | | | m68k_serial->tqueue_hangup is unused. Remove it. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Greg Ungerer <gerg@uclinux.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial-core: reset the console speed on resumeYin Kangkai2011-02-031-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On some platforms, we need to restore the console speed on resume even it was not suspended (no_console_suspend), and on others we don't have to do that. So don't care about the "console_suspend_enabled" and unconditionally reset the console speed if it is a console. This is actually a redo of ba15ab0 (Set proper console speed on resume if console suspend is disabled) from Deepak Saxena. I also tried to investigate more to find out if this change will break others, here is what I've found out: commit 891b9dd10764352926e1e107756aa229dfa2c210 Author: Jason Wang <jason77.wang@gmail.com> serial-core: restore termios settings when resume console ports commit ca2e71aa8cfb0056ce720f3fd53f59f5fac4a3e1 Author: Jason Wang <jason77.wang@gmail.com> serial-core: skip call set_termios/console_start when no_console_suspend commit 4547be7809a3b775ce750ec7f8b5748954741523 Author: Stanislav Brabec <sbrabec@suse.cz> serial-core: resume serial hardware with no_console_suspend commit ba15ab0e8de0d4439a91342ad52d55ca9e313f3d Author: Deepak Saxena <dsaxena@laptop.org> Set proper console speed on resume if console suspend is disabled from ba15ab0, we learned that, even if the console suspend is disabled (when no_console_suspend is set), we may still need to "reset the port to the state it was in before we suspended." Then with 4547be7, this piece of code is removed. And then Jason Wang added that back in ca2e71a and 891b9dd, to fix some breakage on OMAP3EVM platform. From ca2e71a we learned that the "set_termios" things is actually needed by both console is suspended and not suspended. That's why I removed the console_suspended_enabled condition, and only call console_start() when we actually suspeneded it. I also noticed in this thread: http://marc.info/?t=129079257100004&r=1&w=2, which talked about on some platforms, UART HW will be cut power whether or not we set no_console_suspend, and then on resume it does not work quite well. I have a similar HW, and this patch fixed this issue, don't know if this patch also works on their platforms. [Update: Stanislav tested this patch on Zaurus and reported it improves the situation. Thanks.] CC: Greg KH <greg@kroah.com> CC: Deepak Saxena <dsaxena@laptop.org> CC: Jason Wang <jason77.wang@gmail.com> CC: Stanislav Brabec <sbrabec@suse.cz> CC: Daniel Drake <dsd@laptop.org> Signed-off-by: Yin Kangkai <kangkai.yin@linux.intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: mrst_max3110: make buffer largerDan Carpenter2011-02-031-1/+1
| | | | | | | | | | This is used to store the spi_device ->modalias so they have to be the same size. SPI_NAME_SIZE is 32. Signed-off-by: Dan Carpenter <error27@gmail.com> Signed-off-by: Alan Cox <alan@linux.intel.com> Cc: stable@kernel.org Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty_ldisc: don't use flush_scheduled_work()Tejun Heo2011-02-031-2/+15
| | | | | | | | | | flush_scheduled_work() is scheduled to be deprecated. Explicitly sync flush the used work items instead. Note that before this change, flush_scheduled_work() wouldn't have properly flushed tty->buf.work if it were on timer. Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: omap-serial: Enable the UART wake-up bits alwaysJarkko Nikula2011-02-031-3/+3
| | | | | | | | | | | | | | | | | | | OMAP can do also dynamic idling so wake-up enable register should be set also while system is running. If UART_OMAP_WER is not set, then for instance the RX activity cannot wake up the UART port that is sleeping. This RX wake-up feature was working when the 8250 driver was used instead of omap-serial. Reason for this is that the 8250 doesn't set the UART_OMAP_WER and then arch/arm/mach-omap2/pm34xx.c ends up saving and restoring the reset default which is the same than value OMAP_UART_WER_MOD_WKUP here. Fix this by moving the conditional UART_OMAP_WER write from serial_omap_pm into serial_omap_startup where wake-up bits are set unconditionally. Signed-off-by: Jarkko Nikula <jhnikula@gmail.com> Cc: Govindraj.R <govindraj.raja@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* OMAP: Enable Magic SysRq on serial console ttyOxThomas Weber2011-02-031-1/+4
| | | | | | | | | | | | | | | | | | Magic SysRq key is not working for OMAP on new serial console ttyOx because SUPPORT_SYSRQ is not defined for omap-serial. This patch defines SUPPORT_SYSRQ in omap-serial and enables handling of Magic SysRq character. Further there is an issue of losing first break character. Removing the reset of the lsr_break_flag fixes this issue. Signed-off-by: Thomas Weber <weber@corscience.de> Acked-by: Govindraj.R <govindraj.raja@ti.com> Tested-by: Manjunath G Kondaiah <manjugk@ti.com> Acked-by: Kevin Hilman <khilman@ti.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: ifx6x60: expanded info available from platform dataRuss Gorby2011-02-033-24/+32
| | | | | | | | | | | | Some platform attributes (e.g. max_hz, use_dma) were being intuited from the modem type. These things should be specified by the platform data. Added max_hz, use_dma to ifx_modem_platform_data definition, replaced is_6160 w/ modem_type, and changed clients accordingly Signed-off-by: Russ Gorby <russ.gorby@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* serial: mfd: remove the timeout workaround for A0Feng Tang2011-02-031-38/+0
| | | | | | | | | | | This is kind of a revert for commit 669b7a0938e "hsu: add a periodic timer to check dma rx channel", which is a workaround for a bug in A0 stepping silicon, where a dma rx data timeout is missing for some case. Since new silicon has fixed it and the old version is phasing out, no need to carry on it any more. Signed-off-by: Feng Tang <feng.tang@intel.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* tty: serial: bfin_sport_uart: fix signedness errorVasiliy Kulikov2011-02-031-1/+1
| | | | | | | | sport->port.irq is unsigned, check for <0 doesn't make sense. Explicitly cast it to int to check for error. Signed-off-by: Vasiliy Kulikov <segoon@openwall.com> Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
* Linux 2.6.38-rc3Linus Torvalds2011-02-011-1/+1
|
* Merge branch 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6Linus Torvalds2011-02-0125-210/+174
|\ | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | * 'bugfixes' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6: NFS: NFSv4 readdir loses entries NFS: Micro-optimize nfs4_decode_dirent() NFS: Fix an NFS client lockdep issue NFS construct consistent co_ownerid for v4.1 NFS: nfs_wcc_update_inode() should set nfsi->attr_gencount NFS improve pnfs_put_deviceid_cache debug print NFS fix cb_sequence error processing NFS do not find client in NFSv4 pg_authenticate NLM: Fix "kernel BUG at fs/lockd/host.c:417!" or ".../host.c:283!" NFS: Prevent memory allocation failure in nfsacl_encode() NFS: nfsacl_{encode,decode} should return signed integer NFS: Fix "kernel BUG at fs/nfs/nfs3xdr.c:1338!" NFS: Fix "kernel BUG at fs/aio.c:554!" NFS4: Avoid potential NULL pointer dereference in decode_and_add_ds(). NFS: fix handling of malloc failure during nfs_flush_multi()
| * NFS: NFSv4 readdir loses entriesChuck Lever2011-01-281-3/+0
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On recent 2.6.38-rc kernels, connectathon basic test 6 fails on NFSv4 mounts of OpenSolaris with something like: > ./test6: readdir > ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.12' dir entry, pass 0 > ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.82' dir entry, pass 0 > ./test6: (/mnt/klimt/matisse.test) didn't read expected 'file.164' dir entry, pass 0 > ./test6: (/mnt/klimt/matisse.test) Test failed with 3 errors > basic tests failed > Tests failed, leaving /mnt/klimt mounted > [cel@matisse cthon04]$ I narrowed the problem down to nfs4_decode_dirent() reporting that the decode buffer had overflowed while decoding the entries for those missing files. verify_attr_len() assumes both it's pointer arguments reside on the same page. When these arguments point to locations on two different pages, verify_attr_len() can report false errors. This can happen now that a large NFSv4 readdir result can span pages. We have reasonably good checking in nfs4_decode_dirent() anyway, so it should be safe to simply remove the extra checking. At a guess, this was introduced by commit 6650239a, "NFS: Don't use vm_map_ram() in readdir". Cc: stable@kernel.org [2.6.37] Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: Micro-optimize nfs4_decode_dirent()Chuck Lever2011-01-281-3/+3
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Make the decoding of NFSv4 directory entries slightly more efficient by: 1. Avoiding unnecessary byte swapping when checking XDR booleans, and 2. Not bumping "p" when its value will be immediately replaced by xdr_inline_decode() This commit makes nfs4_decode_dirent() consistent with similar logic in the other two decode_dirent() functions. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: Fix an NFS client lockdep issueTrond Myklebust2011-01-281-2/+4
| | | | | | | | | | | | | | | | | | | | There is no reason to be freeing the delegation cred in the rcu callback, and doing so is resulting in a lockdep complaint that rpc_credcache_lock is being called from both softirq and non-softirq contexts. Reported-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com> Cc: stable@kernel.org
| * NFS construct consistent co_ownerid for v4.1Andy Adamson2011-01-251-20/+10
| | | | | | | | | | | | | | | | | | As stated in section 2.4 of RFC 5661, subsequent instances of the client need to present the same co_ownerid. Concatinate the client's IP dot address, host name, and the rpc_auth pseudoflavor to form the co_ownerid. Signed-off-by: Andy Adamson <andros@netapp.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: nfs_wcc_update_inode() should set nfsi->attr_gencountTrond Myklebust2011-01-251-9/+17
| | | | | | | | | | | | | | | | If the call to nfs_wcc_update_inode() results in an attribute update, we need to ensure that the inode's attr_gencount gets bumped too, otherwise we are not protected against races with other GETATTR calls. Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS improve pnfs_put_deviceid_cache debug printAndy Adamson2011-01-251-1/+1
| | | | | | | | | | | | | | What we really want to know is the ref count. Signed-off-by: Andy Adamson <andros@netapp.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS fix cb_sequence error processingAndy Adamson2011-01-251-1/+1
| | | | | | | | | | | | | | | | | | | | Always assign the cb_process_state nfs_client pointer so a processing error in cb_sequence after the nfs_client is found and referenced returns a non-NULL cb_process_state nfs_client and the matching nfs_put_client in nfs4_callback_compound dereferences the client. Signed-off-by: Andy Adamson <andros@netapp.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS do not find client in NFSv4 pg_authenticateAndy Adamson2011-01-2510-128/+42
| | | | | | | | | | | | | | | | | | | | The information required to find the nfs_client cooresponding to the incoming back channel request is contained in the NFS layer. Perform minimal checking in the RPC layer pg_authenticate method, and push more detailed checking into the NFS layer where the nfs_client can be found. Signed-off-by: Andy Adamson <andros@netapp.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NLM: Fix "kernel BUG at fs/lockd/host.c:417!" or ".../host.c:283!"Chuck Lever2011-01-251-4/+5
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Nick Bowler <nbowler@elliptictech.com> reports: > We were just having some NFS server troubles, and my client machine > running 2.6.38-rc1+ (specifically, commit 2b1caf6ed7b888c95) crashed > hard (syslog output appended to this mail). > > I'm not sure what the exact timeline was or how to reproduce this, > but the server was rebooted during all this. Since I've never seen > this happen before, it is possibly a regression from previous kernel > releases. However, I recently updated my nfs-utils (on the client) to > version 1.2.3, so that might be related as well. [ BUG output redacted ] When done searching, the for_each_host loop in next_host_state() falls through and returns the final host on the host chain without bumping it's reference count. Since the host's ref count is only one at that point, releasing the host in nlm_host_rebooted() attempts to destroy the host prematurely, and therefore hits a BUG(). Likely, the original intent of the for_each_host behavior in next_host_state() was to handle the case when the host chain is empty. Searching the chain and finding no suitable host to return needs to be handled as well. Defensively restructure next_host_state() always to return NULL when the loop falls through. Introduced by commit b10e30f6 "lockd: reorganize nlm_host_rebooted". Cc: J. Bruce Fields <bfields@fieldses.org> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: Prevent memory allocation failure in nfsacl_encode()Chuck Lever2011-01-254-13/+31
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | nfsacl_encode() allocates memory in certain cases. This of course is not guaranteed to work. Since commit 9f06c719 "SUNRPC: New xdr_streams XDR encoder API", the kernel's XDR encoders can't return a result indicating possibly a failure, so a memory allocation failure in nfsacl_encode() has become fatal (ie, the XDR code Oopses) in some cases. However, the allocated memory is a tiny fixed amount, on the order of 40-50 bytes. We can easily use a stack-allocated buffer for this, with only a wee bit of nose-holding. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: nfsacl_{encode,decode} should return signed integerChuck Lever2011-01-252-8/+28
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Clean up. The nfsacl_encode() and nfsacl_decode() functions return negative errno values, and each call site verifies that the returned value is not negative. Change the synopsis of both of these functions to reflect this usage. Document the synopsis and return values. Reported-by: Trond Myklebust <trond.myklebust@netapp.com> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: Fix "kernel BUG at fs/nfs/nfs3xdr.c:1338!"Chuck Lever2011-01-251-1/+4
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Milan Broz <mbroz@redhat.com> reports: > on today Linus' tree I get OOps if using nfs. > > server (2.6.36) exports dir: > /dir 172.16.1.0/24(rw,async,all_squash,no_subtree_check,anonuid=500,anongid=500) > > on client it is mounted in fstab > server:/dir /mnt/tst nfs rw,soft 0 0 > > and these commands OOpses it (simplified from a configure script): > > cd /dir > touch x > install x y > > [ 105.327701] ------------[ cut here ]------------ > [ 105.327979] kernel BUG at fs/nfs/nfs3xdr.c:1338! > [ 105.328075] invalid opcode: 0000 [#1] PREEMPT SMP > [ 105.328223] last sysfs file: /sys/devices/virtual/bdi/0:16/uevent > [ 105.328349] Modules linked in: usbcore dm_mod > [ 105.328553] > [ 105.328678] Pid: 3710, comm: install Not tainted 2.6.37+ #423 440BX Desktop Reference Platform/VMware Virtual Platform > [ 105.328853] EIP: 0060:[<c116c06c>] EFLAGS: 00010282 CPU: 0 > [ 105.329152] EIP is at nfs3_xdr_enc_setacl3args+0x61/0x98 > [ 105.329249] EAX: ffffffea EBX: ce941d98 ECX: 00000000 EDX: 00000004 > [ 105.329340] ESI: ce941cd0 EDI: 000000a4 EBP: ce941cc0 ESP: ce941cb4 > [ 105.329431] DS: 007b ES: 007b FS: 00d8 GS: 00e0 SS: 0068 > [ 105.329525] Process install (pid: 3710, ti=ce940000 task=ced36f20 task.ti=ce940000) > [ 105.336600] Stack: > [ 105.336693] ce941cd0 ce9dc000 00000000 ce941cf8 c12ecd02 c12f43e0 c116c00b cf754158 > [ 105.336982] ce9dc004 cf754284 ce9dc004 cf7ffee8 ceff9978 ce9dc000 cf7ffee8 ce9dc000 > [ 105.337182] ce9dc000 ce941d14 c12e698d cf75412c ce941d98 cf7ffee8 cf7fff20 00000000 > [ 105.337405] Call Trace: > [ 105.337695] [<c12ecd02>] rpcauth_wrap_req+0x75/0x7f > [ 105.337806] [<c12f43e0>] ? xdr_encode_opaque+0x12/0x15 > [ 105.337898] [<c116c00b>] ? nfs3_xdr_enc_setacl3args+0x0/0x98 > [ 105.337988] [<c12e698d>] call_transmit+0x17e/0x1e8 > [ 105.338072] [<c12ec307>] __rpc_execute+0x6d/0x1a6 > [ 105.338155] [<c12ec474>] rpc_execute+0x34/0x37 > [ 105.338235] [<c12e738d>] rpc_run_task+0xb5/0xbd > [ 105.338316] [<c12e7474>] rpc_call_sync+0x3d/0x58 > [ 105.338402] [<c116d0c6>] nfs3_proc_setacls+0x18e/0x24f > [ 105.338493] [<c10b3f76>] ? __kmalloc+0x148/0x1c4 > [ 105.338579] [<c10ecd01>] ? posix_acl_alloc+0x12/0x22 > [ 105.338665] [<c116d5c8>] nfs3_proc_setacl+0xa0/0xca > [ 105.338748] [<c116d69c>] nfs3_setxattr+0x62/0x88 > [ 105.338834] [<c1317042>] ? sub_preempt_count+0x7c/0x89 > [ 105.338926] [<c116d63a>] ? nfs3_setxattr+0x0/0x88 > [ 105.339026] [<c10cfa79>] __vfs_setxattr_noperm+0x26/0x95 > [ 105.339114] [<c10cfb43>] vfs_setxattr+0x5b/0x76 > [ 105.339211] [<c10cfbfb>] setxattr+0x9d/0xc3 > [ 105.339298] [<c10a2ea8>] ? handle_pte_fault+0x258/0x5cb > [ 105.339428] [<c1091ff6>] ? __free_pages+0x1a/0x23 > [ 105.339517] [<c10498ea>] ? up_read+0x16/0x2c > [ 105.339599] [<c10b8365>] ? fget+0x0/0xa3 > [ 105.339677] [<c10b8365>] ? fget+0x0/0xa3 > [ 105.339760] [<c1025d23>] ? get_parent_ip+0xb/0x31 > [ 105.339843] [<c1317042>] ? sub_preempt_count+0x7c/0x89 > [ 105.339931] [<c10cfc72>] sys_fsetxattr+0x51/0x79 > [ 105.340014] [<c1002853>] sysenter_do_call+0x12/0x32 > [ 105.340133] Code: 2e 76 18 00 58 31 d2 8b 7f 28 f6 43 04 01 74 03 8b 53 08 6a 00 8b 46 04 6a 01 8b 0b 52 89 fa e8 85 10 f8 ff 83 c4 0c 85 c0 79 04 <0f> 0b eb fe 31 c9 f6 43 04 04 74 03 8b 4b 0c 68 00 10 00 00 8d > [ 105.350321] EIP: [<c116c06c>] nfs3_xdr_enc_setacl3args+0x61/0x98 SS:ESP 0068:ce941cb4 > [ 105.364385] ---[ end trace 01fcfe7f0f7f6e4a ]--- nfs3_xdr_enc_setacl3args() is not properly setting up the target buffer before nfsacl_encode() attempts to encode the ACL. Introduced by commit d9c407b1 "NFS: Introduce new-style XDR encoding functions for NFSv3." Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: Fix "kernel BUG at fs/aio.c:554!"Chuck Lever2011-01-251-14/+20
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Nick Piggin reports: > I'm getting use after frees in aio code in NFS > > [ 2703.396766] Call Trace: > [ 2703.396858] [<ffffffff8100b057>] ? native_sched_clock+0x27/0x80 > [ 2703.396959] [<ffffffff8108509e>] ? put_lock_stats+0xe/0x40 > [ 2703.397058] [<ffffffff81088348>] ? lock_release_holdtime+0xa8/0x140 > [ 2703.397159] [<ffffffff8108a2a5>] lock_acquire+0x95/0x1b0 > [ 2703.397260] [<ffffffff811627db>] ? aio_put_req+0x2b/0x60 > [ 2703.397361] [<ffffffff81039701>] ? get_parent_ip+0x11/0x50 > [ 2703.397464] [<ffffffff81612a31>] _raw_spin_lock_irq+0x41/0x80 > [ 2703.397564] [<ffffffff811627db>] ? aio_put_req+0x2b/0x60 > [ 2703.397662] [<ffffffff811627db>] aio_put_req+0x2b/0x60 > [ 2703.397761] [<ffffffff811647fe>] do_io_submit+0x2be/0x7c0 > [ 2703.397895] [<ffffffff81164d0b>] sys_io_submit+0xb/0x10 > [ 2703.397995] [<ffffffff8100307b>] system_call_fastpath+0x16/0x1b > > Adding some tracing, it is due to nfs completing the request then > returning something other than -EIOCBQUEUED, so aio.c > also completes the request. To address this, prevent the NFS direct I/O engine from completing async iocbs when the forward path returns an error without starting any I/O. This fix appears to survive ^C during both "xfstest no. 208" and "fsx -Z." It's likely this bug has existed for a very long while, as we are seeing very similar symptoms in OEL 5. Copying stable. Cc: Stable <stable@kernel.org> Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS4: Avoid potential NULL pointer dereference in decode_and_add_ds().Jesper Juhl2011-01-251-2/+7
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | On Mon, 17 Jan 2011, Mi Jinlong wrote: > > > Jesper Juhl: > > strrchr() can return NULL if nothing is found. If this happens we'll > > dereference a NULL pointer in > > fs/nfs/nfs4filelayoutdev.c::decode_and_add_ds(). > > > > I tried to find some other code that guarantees that this can never > > happen but I was unsuccessful. So, unless someone else can point to some > > code that ensures this can never be a problem, I believe this patch is > > needed. > > > > While I was changing this code I also noticed that all the dprintk() > > statements, except one, start with "%s:". The one missing the ":" I added > > it to. > > Maybe another one also should be changed at decode_and_add_ds() at line 243: > > 243 printk("%s Decoded address and port %s\n", __func__, buf); > Missed that one. Thanks. Signed-off-by: Jesper Juhl <jj@chaosbits.net> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
| * NFS: fix handling of malloc failure during nfs_flush_multi()Fred Isaman2011-01-191-1/+1
| | | | | | | | | | | | | | | | | | Cleanup of the allocated list entries should not call put_nfs_open_context() on each entry, as the context will always be NULL, causing an oops. Signed-off-by: Fred Isaman <iisaman@netapp.com> Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
generated by cgit v1.1 at 2025-01-15 15:58:37 +0000