- 16 Mar, 2016 7 commits
-
-
Jiri Slaby authored
-
Konrad Rzeszutek Wilk authored
commit 8d47065f upstream. Commit 408fb0e5 (xen/pciback: Don't allow MSI-X ops if PCI_COMMAND_MEMORY is not set) prevented enabling MSI-X on passed-through virtual functions, because it checked the VF for PCI_COMMAND_MEMORY but this is not a valid bit for VFs. Instead, check the physical function for PCI_COMMAND_MEMORY. Signed-off-by:
Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Reviewed-by:
Jan Beulich <jbeulich@suse.com> Signed-off-by:
David Vrabel <david.vrabel@citrix.com> Signed-off-by:
Jiri Slaby <jslaby@suse.cz>
-
Konrad Rzeszutek Wilk authored
commit 408fb0e5 upstream. commit f598282f ("PCI: Fix the NIU MSI-X problem in a better way") teaches us that dealing with MSI-X can be troublesome. Further checks in the MSI-X architecture shows that if the PCI_COMMAND_MEMORY bit is turned of in the PCI_COMMAND we may not be able to access the BAR (since they are memory regions). Since the MSI-X tables are located in there.. that can lead to us causing PCIe errors. Inhibit us performing any operation on the MSI-X unless the MEMORY bit is set. Note that Xen hypervisor with: "x86/MSI-X: access MSI-X table only after having enabled MSI-X" will return: xen_pciback: 0000:0a:00.1: error -6 enabling MSI-X for guest 3! When the generic MSI code tries to setup the PIRQ without MEMORY bit set. Which means with later versions of Xen (4.6) this patch is not neccessary. This is part of XSA-157 Reviewed-by:
-
Konrad Rzeszutek Wilk authored
commit 7cfb905b upstream. Otherwise just continue on, returning the same values as previously (return of 0, and op->result has the PIRQ value). This does not change the behavior of XEN_PCI_OP_disable_msi[|x]. The pci_disable_msi or pci_disable_msix have the checks for msi_enabled or msix_enabled so they will error out immediately. However the guest can still call these operations and cause us to disable the 'ack_intr'. That means the backend IRQ handler for the legacy interrupt will not respond to interrupts anymore. This will lead to (if the device is causing an interrupt storm) for the Linux generic code to disable the interrupt line. Naturally this will only happen if the device in question is plugged in on the motherboard on shared level interrupt GSI. This is part of XSA-157 Reviewed-by:
-
Konrad Rzeszutek Wilk authored
commit a396f3a2 upstream. Otherwise an guest can subvert the generic MSI code to trigger an BUG_ON condition during MSI interrupt freeing: for (i = 0; i < entry->nvec_used; i++) BUG_ON(irq_has_action(entry->irq + i)); Xen PCI backed installs an IRQ handler (request_irq) for the dev->irq whenever the guest writes PCI_COMMAND_MEMORY (or PCI_COMMAND_IO) to the PCI_COMMAND register. This is done in case the device has legacy interrupts the GSI line is shared by the backend devices. To subvert the backend the guest needs to make the backend to change the dev->irq from the GSI to the MSI interrupt line, make the backend allocate an interrupt handler, and then command the backend to free the MSI interrupt and hit the BUG_ON. Since the backend only calls 'request_irq' when the guest writes to the PCI_COMMAND register the guest needs to call XEN_PCI_OP_enable_msi before any other operation. This will cause the generic MSI code to setup an MSI entry and populate dev->irq with the new PIRQ value. Then the guest can write to PCI_COMMAND PCI_COMMAND_MEMORY and cause the backend to setup an IRQ handler for dev->irq (which instead of the GSI value has the MSI pirq). See 'xen_pcibk_control_isr'. Then the guest disables the MSI: XEN_PCI_OP_disable_msi which ends up triggering the BUG_ON condition in 'free_msi_irqs' as there is an IRQ handler for the entry->irq (dev->irq). Note that this cannot be done using MSI-X as the generic code does not over-write dev->irq with the MSI-X PIRQ values. The patch inhibits setting up the IRQ handler if MSI or MSI-X (for symmetry reasons) code had been called successfully. P.S. Xen PCIBack when it sets up the device for the guest consumption ends up writting 0 to the PCI_COMMAND (see xen_pcibk_reset_device). XSA-120 addendum patch removed that - however when upstreaming said addendum we found that it caused issues with qemu upstream. That has now been fixed in qemu upstream. This is part of XSA-157 Reviewed-by:
-
Konrad Rzeszutek Wilk authored
commit 5e0ce145 upstream. The guest sequence of: a) XEN_PCI_OP_enable_msix b) XEN_PCI_OP_enable_msix results in hitting an NULL pointer due to using freed pointers. The device passed in the guest MUST have MSI-X capability. The a) constructs and SysFS representation of MSI and MSI groups. The b) adds a second set of them but adding in to SysFS fails (duplicate entry). 'populate_msi_sysfs' frees the newly allocated msi_irq_groups (note that in a) pdev->msi_irq_groups is still set) and also free's ALL of the MSI-X entries of the device (the ones allocated in step a) and b)). The unwind code: 'free_msi_irqs' deletes all the entries and tries to delete the pdev->msi_irq_groups (which hasn't been set to NULL). However the pointers in the SysFS are already freed and we hit an NULL pointer further on when 'strlen' is attempted on a freed pointer. The patch adds a simple check in the XEN_PCI_OP_enable_msix to guard against that. The check for msi_enabled is not stricly neccessary. This is part of XSA-157 Reviewed-by:
-
Konrad Rzeszutek Wilk authored
commit 56441f3c upstream. The guest sequence of: a) XEN_PCI_OP_enable_msi b) XEN_PCI_OP_enable_msi c) XEN_PCI_OP_disable_msi results in hitting an BUG_ON condition in the msi.c code. The MSI code uses an dev->msi_list to which it adds MSI entries. Under the above conditions an BUG_ON() can be hit. The device passed in the guest MUST have MSI capability. The a) adds the entry to the dev->msi_list and sets msi_enabled. The b) adds a second entry but adding in to SysFS fails (duplicate entry) and deletes all of the entries from msi_list and returns (with msi_enabled is still set). c) pci_disable_msi passes the msi_enabled checks and hits: BUG_ON(list_empty(dev_to_msi_list(&dev->dev))); and blows up. The patch adds a simple check in the XEN_PCI_OP_enable_msi to guard against that. The check for msix_enabled is not stricly neccessary. This is part of XSA-157. Reviewed-by:
-
- 14 Mar, 2016 22 commits
-
-
Rusty Russell authored
commit 8244062e upstream. For CONFIG_KALLSYMS, we keep two symbol tables and two string tables. There's one full copy, marked SHF_ALLOC and laid out at the end of the module's init section. There's also a cut-down version that only contains core symbols and strings, and lives in the module's core section. After module init (and before we free the module memory), we switch the mod->symtab, mod->num_symtab and mod->strtab to point to the core versions. We do this under the module_mutex. However, kallsyms doesn't take the module_mutex: it uses preempt_disable() and rcu tricks to walk through the modules, because it's used in the oops path. It's also used in /proc/kallsyms. There's nothing atomic about the change of these variables, so we can get the old (larger!) num_symtab and the new symtab pointer; in fact this is what I saw when trying to reproduce. By grouping these variables together, we can use a carefully-dereferenced pointer to ensure we always get one or the other (the free of the module init section is already done in an RCU callback, so that's safe). We allocate the init one at the end of the module init section, and keep the core one inside the struct module itself (it could also have been allocated at the end of the module core, but that's probably overkill). [ Rebased for 4.4-stable and older, because the following changes aren't in the older trees: - e0224418: adds arg to is_core_symbol - 7523e4dc: module_init/module_core/init_size/core_size become init_layout.base/core_layout.base/init_layout.size/core_layout.size. Original commit: 8244062e ] Reported-by:
Weilong Chen <chenweilong@huawei.com> Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=111541Signed-off-by:
Rusty Russell <rusty@rustcorp.com.au> Signed-off-by:
-
Jason Andryuk authored
commit a6807590 upstream. The comparisons should be >= since 0x800 and 0x80 require an additional bit to store. For the 3 byte case, the existing shift would drop off 2 more bits than intended. For the 2 byte case, there should be 5 bits bits in byte 1, and 6 bits in byte 2. Signed-off-by:
Jason Andryuk <jandryuk@gmail.com> Reviewed-by:
Laszlo Ersek <lersek@redhat.com> Cc: Peter Jones <pjones@redhat.com> Cc: Matthew Garrett <mjg59@coreos.com> Cc: "Lee, Chun-Yi" <jlee@suse.com> Signed-off-by:
Matt Fleming <matt@codeblueprint.co.uk> Signed-off-by:
-
Matt Fleming authored
commit e246eb56 upstream. Laszlo explains why this is a good idea, 'This is because the pstore filesystem can be backed by UEFI variables, and (for example) a crash might dump the last kilobytes of the dmesg into a number of pstore entries, each entry backed by a separate UEFI variable in the above GUID namespace, and with a variable name according to the above pattern. Please see "drivers/firmware/efi/efi-pstore.c". While this patch series will not prevent the user from deleting those UEFI variables via the pstore filesystem (i.e., deleting a pstore fs entry will continue to delete the backing UEFI variable), I think it would be nice to preserve the possibility for the sysadmin to delete Linux-created UEFI variables that carry portions of the crash log, *without* having to mount the pstore filesystem.' There's also no chance of causing machines to become bricked by deleting these variables, which is the whole purpose of excluding things from the whitelist. Use the LINUX_EFI_CRASH_GUID guid and a wildcard '*' for the match so that we don't have to update the string in the future if new variable name formats are created for crash dump variables. Reported-by:
Peter Jones <pjones@redhat.com> Tested-by:
-
Peter Jones authored
commit ed8b0de5 upstream. "rm -rf" is bricking some peoples' laptops because of variables being used to store non-reinitializable firmware driver data that's required to POST the hardware. These are 100% bugs, and they need to be fixed, but in the mean time it shouldn't be easy to *accidentally* brick machines. We have to have delete working, and picking which variables do and don't work for deletion is quite intractable, so instead make everything immutable by default (except for a whitelist), and make tools that aren't quite so broad-spectrum unset the immutable flag. Signed-off-by:
Lee, Chun-Yi <jlee@suse.com> Acked-by:
Matthew Garrett <mjg59@coreos.com> Signed-off-by:
-
Peter Jones authored
commit 8282f5d9 upstream. All the variables in this list so far are defined to be in the global namespace in the UEFI spec, so this just further ensures we're validating the variables we think we are. Including the guid for entries will become more important in future patches when we decide whether or not to allow deletion of variables based on presence in this list. Signed-off-by:
-
Peter Jones authored
commit 3dcb1f55 upstream. Actually translate from ucs2 to utf8 before doing the test, and then test against our other utf8 data, instead of fudging it. Signed-off-by:
-
Peter Jones authored
commit e0d64e6a upstream. Translate EFI's UCS-2 variable names to UTF-8 instead of just assuming all variable names fit in ASCII. Signed-off-by:
-
Peter Jones authored
commit 73500267 upstream. This adds ucs2_utf8size(), which tells us how big our ucs2 string is in bytes, and ucs2_as_utf8, which translates from ucs2 to utf8.. Signed-off-by:
-
Marcelo Tosatti authored
commit 7cae2bed upstream. As reported at https://bugs.launchpad.net/qemu/+bug/1494350, it is possible to have vcpu->arch.st.last_steal initialized from a thread other than vcpu thread, say the iothread, via KVM_SET_MSRS. Which can cause an overflow later (when subtracting from vcpu threads sched_info.run_delay). To avoid that, move steal time accumulation to vcpu entry time, before copying steal time data to guest. Signed-off-by:
Marcelo Tosatti <mtosatti@redhat.com> Reviewed-by:
David Matlack <dmatlack@google.com> Signed-off-by:
Paolo Bonzini <pbonzini@redhat.com> Signed-off-by:
-
Andreas Schwab authored
commit f15838e9 upstream. Since binutils 2.26 BFD is doing suffix merging on STRTAB sections. But dedotify modifies the symbol names in place, which can also modify unrelated symbols with a name that matches a suffix of a dotted name. To remove the leading dot of a symbol name we can just increment the pointer into the STRTAB section instead. Backport to all stables to avoid breakage when people update their binutils - mpe. Signed-off-by:
Andreas Schwab <schwab@linux-m68k.org> Signed-off-by:
Michael Ellerman <mpe@ellerman.id.au> Signed-off-by:
-
Felix Fietkau authored
commit 7a36b930 upstream. The value 5000 was put here with the addition of the timeout field to ieee80211_start_tx_ba_session. It was originally added in mac80211 to save resources for drivers like iwlwifi, which only supports a limited number of concurrent aggregation sessions. Since iwlwifi does not use minstrel_ht and other drivers don't need this, 0 is a better default - especially since there have been recent reports of aggregation setup related issues reproduced with ath9k. This should improve stability without causing any adverse effects. Acked-by:
Avery Pennarun <apenwarr@gmail.com> Signed-off-by:
Felix Fietkau <nbd@openwrt.org> Signed-off-by:
Johannes Berg <johannes.berg@intel.com> Signed-off-by:
-
Johannes Berg authored
commit 8bf86273 upstream. Beniamino reported that he was getting an RTM_NEWLINK message for a given interface, after the RTM_DELLINK for it. It turns out that the message is a wireless extensions message, which was sent because the interface had been connected and disconnection while it was deleted caused a wext message. For its netlink messages, wext uses RTM_NEWLINK, but the message is without all the regular rtnetlink attributes, so "ip monitor link" prints just rudimentary information: 5: wlan1: <BROADCAST,MULTICAST> mtu 1500 qdisc mq state DOWN group default link/ether 02:00:00:00:01:00 brd ff:ff:ff:ff:ff:ff Deleted 5: wlan1: <BROADCAST,MULTICAST> mtu 1500 qdisc noop state DOWN group default link/ether 02:00:00:00:01:00 brd ff:ff:ff:ff:ff:ff 5: wlan1: <BROADCAST,MULTICAST,UP> link/ether (from my hwsim reproduction) This can cause userspace to get confused since it doesn't expect an RTM_NEWLINK message after RTM_DELLINK. The reason for this is that wext schedules a worker to send out the messages, and the scheduling delay can cause the messages to get out to userspace in different order. To fix this, have wext register a netdevice notifier and flush out any pending messages when netdevice state changes. This fixes any ordering whenever the original message wasn't sent by a notifier itself. Reported-by:
Beniamino Galvani <bgalvani@redhat.com> Signed-off-by:
-
Takashi Iwai authored
commit d0784829 upstream. "MBC Mode", "VSS Mode", "VSS HPF Mode" and "Enhanced EQ Mode" ctls in wm8958 codec driver are enum, while the current driver accesses wrongly via value.integer.value[]. They have to be via value.enumerated.item[] instead. Signed-off-by:
Takashi Iwai <tiwai@suse.de> Signed-off-by:
Mark Brown <broonie@kernel.org> Signed-off-by:
-
Takashi Iwai authored
commit 8019c0b3 upstream. The DRC Mode like "AIF1DRC1 Mode" and EQ Mode like "AIF1.1 EQ Mode" in wm8994 codec driver are enum ctls, while the current driver accesses wrongly via value.integer.value[]. They have to be via value.enumerated.item[] instead. Signed-off-by:
-
Steven Rostedt (Red Hat) authored
commit dc17147d upstream. Commit f3775549 ("tracepoints: Do not trace when cpu is offline") added a check to make sure that tracepoints only get called when the cpu is online, as it uses rcu_read_lock_sched() for protection. Commit 3a630178 ("tracing: generate RCU warnings even when tracepoints are disabled") added lockdep checks (including rcu checks) for events that are not enabled to catch possible RCU issues that would only be triggered if a trace event was enabled. Commit f3775549 only stopped the warnings when the trace event was enabled but did not prevent warnings if the trace event was called when disabled. To fix this, the cpu online check is moved to where the condition is added to the trace event. This will place the cpu online check in all places that it may be used now and in the future. Fixes: f3775549 ("tracepoints: Do not trace when cpu is offline") Fixes: 3a630178 ("tracing: generate RCU warnings even when tracepoints are disabled") Reported-by:
Sudeep Holla <sudeep.holla@arm.com> Tested-by:
Steven Rostedt <rostedt@goodmis.org> Signed-off-by:
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
-
Radim Krčmář authored
commit 7099e2e1 upstream. Linux guests on Haswell (and also SandyBridge and Broadwell, at least) would crash if you decided to run a host command that uses PEBS, like perf record -e 'cpu/mem-stores/pp' -a This happens because KVM is using VMX MSR switching to disable PEBS, but SDM [2015-12] 18.4.4.4 Re-configuring PEBS Facilities explains why it isn't safe: When software needs to reconfigure PEBS facilities, it should allow a quiescent period between stopping the prior event counting and setting up a new PEBS event. The quiescent period is to allow any latent residual PEBS records to complete its capture at their previously specified buffer address (provided by IA32_DS_AREA). There might not be a quiescent period after the MSR switch, so a CPU ends up using host's MSR_IA32_DS_AREA to access an area in guest's memory. (Or MSR switching is just buggy on some models.) The guest can learn something about the host this way: If the guest doesn't map address pointed by MSR_IA32_DS_AREA, it results in #PF where we leak host's MSR_IA32_DS_AREA through CR2. After that, a malicious guest can map and configure memory where MSR_IA32_DS_AREA is pointing and can therefore get an output from host's tracing. This is not a critical leak as the host must initiate with PEBS tracing and I have not been able to get a record from more than one instruction before vmentry in vmx_vcpu_run() (that place has most registers already overwritten with guest's). We could disable PEBS just few instructions before vmentry, but disabling it earlier shouldn't affect host tracing too much. We also don't need to switch MSR_IA32_PEBS_ENABLE on VMENTRY, but that optimization isn't worth its code, IMO. (If you are implementing PEBS for guests, be sure to handle the case where both host and guest enable PEBS, because this patch doesn't.) Fixes: 26a4f3c0 ("perf/x86: disable PEBS on a guest entry.") Reported-by:
Jiří Olša <jolsa@redhat.com> Signed-off-by:
Radim Krčmář <rkrcmar@redhat.com> Signed-off-by:
-
Jiri Slaby authored
This reverts commit 40df18b4, commit f6ff4f67 upstream. It causes oopses: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 IP: [<ffffffffa010345d>] radeon_fence_ref+0xd/0x50 [radeon] Signed-off-by:
-
Richard Weinberger authored
commit e4f6daac upstream. ubi_start_leb_change() allocates too few bytes. ubi_more_leb_change_data() will write up to req->upd_bytes + ubi->min_io_size bytes. Signed-off-by:
Richard Weinberger <richard@nod.at> Reviewed-by:
Boris Brezillon <boris.brezillon@free-electrons.com> Signed-off-by:
-
Maciej W. Rozycki authored
commit e723e3f7 upstream. Avoid sending a partially initialised `siginfo_t' structure along SIGFPE signals issued from `do_ov' and `do_trap_or_bp', leading to information leaking from the kernel stack. Signed-off-by:
Maciej W. Rozycki <macro@imgtec.com> Signed-off-by:
Ralf Baechle <ralf@linux-mips.org> Signed-off-by:
-
Yegor Yefremov authored
commit c0992d0f upstream. Add support for Quectel UC20 and blacklist the QMI interface. Signed-off-by:
Yegor Yefremov <yegorslists@googlemail.com> [johan: amend commit message ] Signed-off-by:
Johan Hovold <johan@kernel.org> Signed-off-by:
-
Daniele Palmas authored
commit 5deef555 upstream. This patch adds support for 0x1045 PID of Telit LE922. Signed-off-by:
Daniele Palmas <dnlplm@gmail.com> Signed-off-by:
-
Vittorio Alfieri authored
commit 3c4c615d upstream. The Parrot NMEA GPS Flight Recorder is a USB composite device consisting of hub, flash storage, and cp210x usb to serial chip. It is an accessory to the mass-produced Parrot AR Drone 2. The device emits standard NMEA messages which make the it compatible with NMEA compatible software. It was tested using gpsd version 3.11-3 as an NMEA interpreter and using the official Parrot Flight Recorder. Signed-off-by:
Vittorio Alfieri <vittorio88@gmail.com> Signed-off-by:
-
- 07 Mar, 2016 11 commits
-
-
Takashi Iwai authored
commit 3a72494a upstream. The timer user status compat ioctl returned the bogus struct used for 64bit architectures instead of the 32bit one. This patch addresses it to return the proper struct. Signed-off-by:
-
Takashi Iwai authored
commit c1099c32 upstream. HDSPM driver contains a code issuing zero-division potentially in system sample rate ctl code. This patch fixes it by not processing a zero or invalid rate value as a divisor, as well as excluding the invalid value to be passed via the given ctl element. Signed-off-by:
-
Takashi Iwai authored
commit eab3c4db upstream. snd-hdsp driver accesses enum item values (int) instead of boolean values (long) wrongly for some ctl elements. This patch fixes them. Signed-off-by:
-
Takashi Iwai authored
commit 537e4813 upstream. snd-hdspm driver accesses enum item values (int) instead of boolean values (long) wrongly for some ctl elements. This patch fixes them. Signed-off-by:
-
Takashi Iwai authored
commit 197b958c upstream. The OSS sequencer client tries to drain the pending events at releasing. Unfortunately, as spotted by syzkaller fuzzer, this may lead to an unkillable process state when the event has been queued at the far future. Since the process being released can't be signaled any longer, it remains and waits for the echo-back event in that far future. Back to history, the draining feature was implemented at the time we misinterpreted POSIX definition for blocking file operation. Actually, such a behavior is superfluous at release, and we should just release the device as is instead of keeping it up forever. This patch just removes the draining call that may block the release for too long time unexpectedly. BugLink: http://lkml.kernel.org/r/CACT4Y+Y4kD-aBGj37rf-xBw9bH3GMU6P+MYg4W1e-s-paVD2pg@mail.gmail.comReported-by:
Dmitry Vyukov <dvyukov@google.com> Signed-off-by:
-
Takashi Iwai authored
commit b24e7ad1 upstream. X32 ABI takes the 64bit timespec, thus the timer user status ioctl becomes incompatible with IA32. This results in NOTTY error when the ioctl is issued. Meanwhile, this struct in X32 is essentially identical with the one in X86-64, so we can just bypassing to the existing code for this specific compat ioctl. Signed-off-by:
-
Takashi Iwai authored
commit 2251fbbc upstream. Like the previous fixes for ctl and PCM, we need a fix for incompatible X32 ABI regarding the rawmidi: namely, struct snd_rawmidi_status has the timespec, and the size and the alignment on X32 differ from IA32. This patch fixes the incompatible ioctl for X32. Signed-off-by:
-
Takashi Iwai authored
commit 6236d8bb upstream. The X32 ABI takes the same alignment like x86-64, and this may result in the incompatible struct size from ia32. Unfortunately, we hit this in some control ABI: struct snd_ctl_elem_value differs between them due to the position of 64bit variable array. This ends up with the unknown ioctl (ENOTTY) error. The fix is to add the compat entries for the new aligned struct. Reported-and-tested-by:
Steven Newbury <steve@snewbury.org.uk> Signed-off-by:
-
David Woodhouse authored
commit be629c62 upstream. When a directory is deleted, we don't take too much care about killing off all the dirents that belong to it — on the basis that on remount, the scan will conclude that the directory is dead anyway. This doesn't work though, when the deleted directory contained a child directory which was moved *out*. In the early stages of the fs build we can then end up with an apparent hard link, with the child directory appearing both in its true location, and as a child of the original directory which are this stage of the mount process we don't *yet* know is defunct. To resolve this, take out the early special-casing of the "directories shall not have hard links" rule in jffs2_build_inode_pass1(), and let the normal nlink processing happen for directories as well as other inodes. Then later in the build process we can set ic->pino_nlink to the parent inode#, as is required for directories during normal operaton, instead of the nlink. And complain only *then* about hard links which are still in evidence even after killing off all the unreachable paths. Reported-by:
Liu Song <liu.song11@zte.com.cn> Signed-off-by:
David Woodhouse <David.Woodhouse@intel.com> Signed-off-by:
-
David Woodhouse authored
commit 49e91e70 upstream. With this fix, all code paths should now be obtaining the page lock before f->sem. Reported-by:
Szabó Tamás <sztomi89@gmail.com> Tested-by:
Thomas Betker <thomas.betker@rohde-schwarz.com> Signed-off-by:
-
Thomas Betker authored
commit 157078f6 upstream. This reverts commit 5ffd3412 ("jffs2: Fix lock acquisition order bug in jffs2_write_begin"). The commit modified jffs2_write_begin() to remove a deadlock with jffs2_garbage_collect_live(), but this introduced new deadlocks found by multiple users. page_lock() actually has to be called before mutex_lock(&c->alloc_sem) or mutex_lock(&f->sem) because jffs2_write_end() and jffs2_readpage() are called with the page locked, and they acquire c->alloc_sem and f->sem, resp. In other words, the lock order in jffs2_write_begin() was correct, and it is the jffs2_garbage_collect_live() path that has to be changed. Revert the commit to get rid of the new deadlocks, and to clear the way for a better fix of the original deadlock. Reported-by:
Deng Chao <deng.chao1@zte.com.cn> Reported-by:
Ming Liu <liu.ming50@gmail.com> Reported-by:
wangzaiwei <wangzaiwei@top-vision.cn> Signed-off-by:
-
