]> git.apps.os.sepia.ceph.com Git - ceph-client.git/log
ceph-client.git
5 months agohandling static and dynamic arrays better
Alex Markuze [Tue, 15 Apr 2025 12:35:23 +0000 (12:35 +0000)]
handling static and dynamic arrays better

5 months agoserializing -- need to modify sizof macro to handle arrays -- also serdes
Alex Markuze [Mon, 7 Apr 2025 20:10:08 +0000 (20:10 +0000)]
serializing -- need to modify sizof macro to handle arrays -- also serdes

5 months agoarray handling
Alex Markuze [Mon, 7 Apr 2025 19:57:20 +0000 (19:57 +0000)]
array handling

5 months agoresolving const and macro expansion issues -- need to handle arrays and variable...
Alex Markuze [Mon, 7 Apr 2025 19:46:20 +0000 (19:46 +0000)]
resolving const and macro expansion issues -- need to handle arrays and variable strings!

5 months agomacro fixups
Alex Markuze [Mon, 7 Apr 2025 18:58:54 +0000 (18:58 +0000)]
macro  fixups

5 months agotmp
Alex Markuze [Mon, 7 Apr 2025 16:34:16 +0000 (16:34 +0000)]
tmp

5 months agowtf
Alex Markuze [Mon, 7 Apr 2025 16:34:06 +0000 (16:34 +0000)]
wtf

5 months agoredundnat
Alex Markuze [Mon, 7 Apr 2025 16:00:02 +0000 (16:00 +0000)]
redundnat

5 months agoredundant
Alex Markuze [Mon, 7 Apr 2025 15:56:54 +0000 (15:56 +0000)]
redundant

5 months agoredundancy
Alex Markuze [Mon, 7 Apr 2025 15:50:49 +0000 (15:50 +0000)]
redundancy

5 months agoredundant
Alex Markuze [Mon, 7 Apr 2025 15:44:25 +0000 (15:44 +0000)]
redundant

5 months agoremove redundancy
Alex Markuze [Mon, 7 Apr 2025 15:43:22 +0000 (15:43 +0000)]
remove redundancy

5 months agoredundancy
Alex Markuze [Mon, 7 Apr 2025 15:41:30 +0000 (15:41 +0000)]
redundancy

5 months agomore rm redundancy
Alex Markuze [Mon, 7 Apr 2025 15:39:59 +0000 (15:39 +0000)]
more rm redundancy

5 months agomore rm redundancy
Alex Markuze [Mon, 7 Apr 2025 15:38:12 +0000 (15:38 +0000)]
more rm redundancy

5 months agomore rm redundancy
Alex Markuze [Mon, 7 Apr 2025 15:37:47 +0000 (15:37 +0000)]
more rm redundancy

5 months agomore rm redundancy
Alex Markuze [Mon, 7 Apr 2025 15:35:41 +0000 (15:35 +0000)]
more rm redundancy

5 months agomore rm redundancy
Alex Markuze [Mon, 7 Apr 2025 15:33:38 +0000 (15:33 +0000)]
more rm redundancy

5 months agomore rm redundancy
Alex Markuze [Mon, 7 Apr 2025 15:33:02 +0000 (15:33 +0000)]
more rm redundancy

5 months agomore rm redundancy
Alex Markuze [Mon, 7 Apr 2025 15:31:55 +0000 (15:31 +0000)]
more rm redundancy

5 months agoremoving redundant bits
Alex Markuze [Mon, 7 Apr 2025 15:20:51 +0000 (15:20 +0000)]
removing redundant bits

5 months agoremoveing redundant __func__ calls
Alex Markuze [Mon, 7 Apr 2025 14:51:41 +0000 (14:51 +0000)]
removeing redundant __func__ calls

5 months agoserializing input parameters
Alex Markuze [Mon, 7 Apr 2025 12:00:57 +0000 (12:00 +0000)]
serializing input parameters

5 months agoadding des code
Alex Markuze [Mon, 7 Apr 2025 11:49:37 +0000 (11:49 +0000)]
adding des code

5 months agofixup
Alex Markuze [Mon, 7 Apr 2025 11:37:37 +0000 (11:37 +0000)]
fixup

5 months agosizeof fixup
Alex Markuze [Mon, 7 Apr 2025 11:18:51 +0000 (11:18 +0000)]
sizeof fixup

5 months agoadding global id to store static info
Alex Markuze [Mon, 7 Apr 2025 10:31:20 +0000 (10:31 +0000)]
adding global id to store static info

5 months agoceph_san ser -- prep for binary tracing
Alex Markuze [Mon, 7 Apr 2025 09:59:38 +0000 (09:59 +0000)]
ceph_san ser -- prep for binary tracing

5 months agoRevert "ceph_san: binary logging"
Alex Markuze [Mon, 7 Apr 2025 09:33:49 +0000 (09:33 +0000)]
Revert "ceph_san: binary logging"

This reverts commit 3d957afa4285ed4deaaf42d200ba7ee1f3092f8d.

6 months agoceph_san: binary logging
Alex Markuze [Tue, 25 Mar 2025 15:00:44 +0000 (15:00 +0000)]
ceph_san: binary logging

6 months agofixup warnings
Alex Markuze [Wed, 19 Mar 2025 14:28:25 +0000 (14:28 +0000)]
fixup warnings

6 months agodebugging TLS
Alex Markuze [Wed, 19 Mar 2025 14:25:00 +0000 (14:25 +0000)]
debugging TLS

6 months agodebug
Alex Markuze [Wed, 19 Mar 2025 13:33:37 +0000 (13:33 +0000)]
debug

6 months agouse spinlock bh?
Alex Markuze [Wed, 19 Mar 2025 12:17:17 +0000 (12:17 +0000)]
use spinlock bh?

6 months agomore debug info
Alex Markuze [Wed, 19 Mar 2025 11:43:49 +0000 (11:43 +0000)]
more debug info

6 months agopf: debugging pf corruption
Alex Markuze [Tue, 18 Mar 2025 20:44:27 +0000 (20:44 +0000)]
pf: debugging pf corruption

6 months agomore stats on pf
Alex Markuze [Tue, 18 Mar 2025 20:34:03 +0000 (20:34 +0000)]
more stats on pf

6 months agoformatting warnings
Alex Markuze [Tue, 18 Mar 2025 20:19:12 +0000 (20:19 +0000)]
formatting warnings

6 months agopf iterator debug
Alex Markuze [Tue, 18 Mar 2025 20:04:31 +0000 (20:04 +0000)]
pf iterator debug

6 months agoverbose debug
Alex Markuze [Tue, 18 Mar 2025 19:46:35 +0000 (19:46 +0000)]
verbose debug

6 months agoebug allocation
Alex Markuze [Tue, 18 Mar 2025 19:45:07 +0000 (19:45 +0000)]
ebug allocation

6 months agobetter wrapparound handling
Alex Markuze [Tue, 18 Mar 2025 19:35:02 +0000 (19:35 +0000)]
better wrapparound handling

6 months agoupdated datetime attempt
Alex Markuze [Tue, 18 Mar 2025 19:34:27 +0000 (19:34 +0000)]
updated datetime attempt

6 months agoAdd task status
Alex Markuze [Tue, 18 Mar 2025 16:46:06 +0000 (16:46 +0000)]
Add task status

6 months agofixups
Alex Markuze [Tue, 18 Mar 2025 16:43:15 +0000 (16:43 +0000)]
fixups

6 months agoformatting improvments
Alex Markuze [Tue, 18 Mar 2025 16:08:51 +0000 (16:08 +0000)]
formatting improvments

6 months agoanother fix
Alex Markuze [Tue, 18 Mar 2025 15:59:18 +0000 (15:59 +0000)]
another fix

6 months agocgroup: export
Alex Markuze [Tue, 18 Mar 2025 15:57:56 +0000 (15:57 +0000)]
cgroup: export

6 months agodebugfs: fixups
Alex Markuze [Tue, 18 Mar 2025 15:57:39 +0000 (15:57 +0000)]
debugfs: fixups

6 months agofixups
Alex Markuze [Tue, 18 Mar 2025 15:50:01 +0000 (15:50 +0000)]
fixups

6 months agocephsan new logger
Alex Markuze [Mon, 17 Mar 2025 16:19:00 +0000 (16:19 +0000)]
cephsan new logger

6 months agominor fixup
Alex Markuze [Mon, 17 Mar 2025 15:41:04 +0000 (15:41 +0000)]
minor fixup

6 months agopf spin lock
Alex Markuze [Mon, 17 Mar 2025 15:30:31 +0000 (15:30 +0000)]
pf spin lock

6 months agocephsan pf iterator
Alex Markuze [Mon, 17 Mar 2025 15:09:13 +0000 (15:09 +0000)]
cephsan pf iterator

6 months agopagefrag allocation fixups
Alex Markuze [Mon, 17 Mar 2025 15:08:26 +0000 (15:08 +0000)]
pagefrag allocation fixups

6 months agobug fix
Alex Markuze [Sun, 16 Mar 2025 19:08:26 +0000 (19:08 +0000)]
bug fix

6 months agolog entry_poison
Alex Markuze [Sun, 16 Mar 2025 19:06:18 +0000 (19:06 +0000)]
log entry_poison

6 months agofixup buffer allocation
Alex Markuze [Sun, 16 Mar 2025 19:03:51 +0000 (19:03 +0000)]
fixup buffer allocation

6 months agominor fixups
Alex Markuze [Sun, 16 Mar 2025 18:38:49 +0000 (18:38 +0000)]
minor fixups

6 months agoceph_san more fixups
Alex Markuze [Sun, 16 Mar 2025 16:55:19 +0000 (16:55 +0000)]
ceph_san more fixups

6 months agotls logger: removing alloc/free elem clallbacks'
Alex Markuze [Sun, 16 Mar 2025 16:51:11 +0000 (16:51 +0000)]
tls logger: removing alloc/free elem clallbacks'

6 months agocephsan logger: WIP
Alex Markuze [Sun, 16 Mar 2025 16:29:21 +0000 (16:29 +0000)]
cephsan logger: WIP

6 months agoceph_san: moving to magzaines
Alex Markuze [Sun, 16 Mar 2025 15:27:09 +0000 (15:27 +0000)]
ceph_san: moving to magzaines

6 months agocephsan: bug fixes
Alex Markuze [Mon, 3 Mar 2025 16:51:54 +0000 (16:51 +0000)]
cephsan: bug fixes

7 months agosome alloc errors
Alex Markuze [Wed, 26 Feb 2025 20:10:39 +0000 (20:10 +0000)]
some alloc errors

7 months agoTLS logger
Alex Markuze [Wed, 26 Feb 2025 16:49:21 +0000 (16:49 +0000)]
TLS logger

7 months agotls: adding an allocation histogram
Alex Markuze [Wed, 26 Feb 2025 15:04:52 +0000 (15:04 +0000)]
tls: adding an allocation histogram

7 months agoAdding a tls storage struct
Alex Markuze [Tue, 25 Feb 2025 16:57:11 +0000 (16:57 +0000)]
Adding a tls storage struct

7 months agoAdding ceph_san
Alex Markuze [Mon, 10 Feb 2025 16:05:50 +0000 (16:05 +0000)]
Adding ceph_san

8 months agoceph: streamline request head structures in MDS client
Liang Jie [Fri, 10 Jan 2025 10:05:24 +0000 (18:05 +0800)]
ceph: streamline request head structures in MDS client

The existence of the ceph_mds_request_head_old structure in the MDS
client code is no longer required due to improvements in handling
different MDS request header versions. This patch removes the now
redundant ceph_mds_request_head_old structure and replaces its usage
with the flexible and extensible ceph_mds_request_head structure.

Changes include:
- Modification of find_legacy_request_head to directly cast the
  pointer to ceph_mds_request_head_legacy without going through the
  old structure.
- Update sizeof calculations in create_request_message to use
  offsetofend for consistency and future-proofing, rather than
  referencing the old structure.
- Use of the structured ceph_mds_request_head directly instead of the
  old one.

Additionally, this consolidation normalizes the handling of
request_head_version v1 to align with versions v2 and v3, leading to
a more consistent and maintainable codebase.

These changes simplify the codebase and reduce potential confusion
stemming from the existence of an obsolete structure.

Signed-off-by: Liang Jie <liangjie@lixiang.com>
Reviewed-by: Viacheslav Dubeyko <Slava.Dubeyko@ibm.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
8 months agoceph: fix memory leak in ceph_mds_auth_match()
Antoine Viallon [Tue, 14 Jan 2025 22:45:14 +0000 (23:45 +0100)]
ceph: fix memory leak in ceph_mds_auth_match()

We now free the temporary target path substring allocation on every
possible branch, instead of omitting the default branch.  In some
cases, a memory leak occured, which could rapidly crash the system
(depending on how many file accesses were attempted).

This was detected in production because it caused a continuous memory
growth, eventually triggering kernel OOM and completely hard-locking
the kernel.

Relevant kmemleak stacktrace:

    unreferenced object 0xffff888131e69900 (size 128):
      comm "git", pid 66104, jiffies 4295435999
      hex dump (first 32 bytes):
        76 6f 6c 75 6d 65 73 2f 63 6f 6e 74 61 69 6e 65  volumes/containe
        72 73 2f 67 69 74 65 61 2f 67 69 74 65 61 2f 67  rs/gitea/gitea/g
      backtrace (crc 2f3bb450):
        [<ffffffffaa68fb49>] __kmalloc_noprof+0x359/0x510
        [<ffffffffc32bf1df>] ceph_mds_check_access+0x5bf/0x14e0 [ceph]
        [<ffffffffc3235722>] ceph_open+0x312/0xd80 [ceph]
        [<ffffffffaa7dd786>] do_dentry_open+0x456/0x1120
        [<ffffffffaa7e3729>] vfs_open+0x79/0x360
        [<ffffffffaa832875>] path_openat+0x1de5/0x4390
        [<ffffffffaa834fcc>] do_filp_open+0x19c/0x3c0
        [<ffffffffaa7e44a1>] do_sys_openat2+0x141/0x180
        [<ffffffffaa7e4945>] __x64_sys_open+0xe5/0x1a0
        [<ffffffffac2cc2f7>] do_syscall_64+0xb7/0x210
        [<ffffffffac400130>] entry_SYSCALL_64_after_hwframe+0x77/0x7f

It can be triggered by mouting a subdirectory of a CephFS filesystem,
and then trying to access files on this subdirectory with an auth token
using a path-scoped capability:

    $ ceph auth get client.services
    [client.services]
            key = REDACTED
            caps mds = "allow rw fsname=cephfs path=/volumes/"
            caps mon = "allow r fsname=cephfs"
            caps osd = "allow rw tag cephfs data=cephfs"

    $ cat /proc/self/mounts
    services@[REDACTED].cephfs=/volumes/containers /ceph/containers ceph rw,noatime,name=services,secret=<hidden>,ms_mode=prefer-crc,mount_timeout=300,acl,mon_addr=[REDACTED]:3300,recover_session=clean 0 0

    $ seq 1 1000000 | xargs -P32 --replace={} touch /ceph/containers/file-{} && \
    seq 1 1000000 | xargs -P32 --replace={} cat /ceph/containers/file-{}

[ idryomov: combine if statements, rename rc to path_matched and make
            it a bool, formatting ]

Cc: stable@vger.kernel.org
Fixes: 596afb0b8933 ("ceph: add ceph_mds_check_access() helper")
Signed-off-by: Antoine Viallon <antoine@lesviallon.fr>
Reviewed-by: Viacheslav Dubeyko <Slava.Dubeyko@ibm.com>
Signed-off-by: Ilya Dryomov <idryomov@gmail.com>
8 months agoLinux 6.13-rc7
Linus Torvalds [Sun, 12 Jan 2025 22:37:56 +0000 (14:37 -0800)]
Linux 6.13-rc7

8 months agoMerge tag 'char-misc-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregk...
Linus Torvalds [Sun, 12 Jan 2025 22:34:00 +0000 (14:34 -0800)]
Merge tag 'char-misc-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc

Pull char/misc/IIO driver fixes from Greg KH:
 "Here are a bunch of small IIO and interconnect and other driver fixes
  to resolve reported issues. Included in here are:

   - loads of iio driver fixes as a result of an audit of places where
    uninitialized data would leak to userspace.

   - other smaller, and normal, iio driver fixes.

   - mhi driver fix

   - interconnect driver fixes

   - pci1xxxx driver fix

  All of these have been in linux-next for a while with no reported
  issues"

* tag 'char-misc-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/char-misc: (32 commits)
  misc: microchip: pci1xxxx: Resolve return code mismatch during GPIO set config
  misc: microchip: pci1xxxx: Resolve kernel panic during GPIO IRQ handling
  interconnect: icc-clk: check return values of devm_kasprintf()
  interconnect: qcom: icc-rpm: Set the count member before accessing the flex array
  iio: adc: ti-ads1119: fix sample size in scan struct for triggered buffer
  iio: temperature: tmp006: fix information leak in triggered buffer
  iio: inkern: call iio_device_put() only on mapped devices
  iio: adc: ad9467: Fix the "don't allow reading vref if not available" case
  iio: adc: at91: call input_free_device() on allocated iio_dev
  iio: adc: ad7173: fix using shared static info struct
  iio: adc: ti-ads124s08: Use gpiod_set_value_cansleep()
  iio: adc: ti-ads1119: fix information leak in triggered buffer
  iio: pressure: zpa2326: fix information leak in triggered buffer
  iio: adc: rockchip_saradc: fix information leak in triggered buffer
  iio: imu: kmx61: fix information leak in triggered buffer
  iio: light: vcnl4035: fix information leak in triggered buffer
  iio: light: bh1745: fix information leak in triggered buffer
  iio: adc: ti-ads8688: fix information leak in triggered buffer
  iio: dummy: iio_simply_dummy_buffer: fix information leak in triggered buffer
  iio: test: Fix GTS test config
  ...

8 months agoMerge tag 'driver-core-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 12 Jan 2025 22:26:31 +0000 (14:26 -0800)]
Merge tag 'driver-core-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core

Pull driver core and debugfs fixes from Greg KH:
 "Here are some small driver core and debugfs fixes that resolve some
  reported problems:

   - debugfs runtime error reporting fixes

   - topology cpumask race-condition fix

   - MAINTAINERS file email update

  All of these have been in linux-next this week with no reported
  issues"

* tag 'driver-core-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/driver-core:
  fs: debugfs: fix open proxy for unsafe files
  MAINTAINERS: align Danilo's maintainer entries
  topology: Keep the cpumask unchanged when printing cpumap
  debugfs: fix missing mutex_destroy() in short_fops case
  fs: debugfs: differentiate short fops with proxy ops

8 months agoMerge tag 'staging-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh...
Linus Torvalds [Sun, 12 Jan 2025 22:22:13 +0000 (14:22 -0800)]
Merge tag 'staging-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging

Pull staging driver fixes from Greg KH:
 "Here are some small staging driver fixes that resolve some reported
  issues and have been in my tree for too long due to the holiday break.
  They resolve the following issues:

   - lots of gpib build-time fixes as reported by testers and 0-day

   - gpib logical fixes

   - mailmap fix

  All of these have been in linux-next for a while, with no reported
  issues other than the duplicated change"

* tag 'staging-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/staging:
  staging: gpib: mite: remove unused global functions
  staging: gpib: refer to correct config symbol in tnt4882 Makefile
  mailmap: update Bingwu Zhang's email address
  staging: gpib: fix address space mixup
  staging: gpib: use ioport_map
  staging: gpib: fix pcmcia dependencies
  staging: gpib: add module author and description fields
  staging: gpib: fix Makefiles
  staging: gpib: make global 'usec_diff' functions static
  staging: gpib: Modify mismatched function name
  staging: gpib: Add lower bound check for secondary address
  staging: gpib: Fix erroneous removal of blank before newline

8 months agoMerge tag 'tty-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty
Linus Torvalds [Sun, 12 Jan 2025 21:27:15 +0000 (13:27 -0800)]
Merge tag 'tty-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty

Pull serial driver fixes from Greg KH:
 "Here are three small serial driver fixes tree. They resolve some
  reported issues:

   - stm32 break control fix

   - 8250 runtime pm usage counter fix

   - imx driver locking fix

  All have been in my tree and linux-next for three weeks now, with no
  reported issues"

* tag 'tty-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty:
  serial: stm32: use port lock wrappers for break control
  serial: imx: Use uart_port_lock_irq() instead of uart_port_lock()
  tty: serial: 8250: Fix another runtime PM usage counter underflow

8 months agoMerge tag 'usb-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb
Linus Torvalds [Sun, 12 Jan 2025 21:09:00 +0000 (13:09 -0800)]
Merge tag 'usb-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb

Pull USB fixes from Greg KH:
 "Here are some small USB driver fixes and new device ids for 6.13-rc7.
  Included in here are:

   - usb serial new device ids

   - typec bugfixes for reported issues

   - dwc3 driver fixes

   - chipidea driver fixes

   - gadget driver fixes

   - other minor fixes for reported problems.

  All of these have been in linux-next for a while, with no reported
  issues"

* tag 'usb-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/usb:
  USB: serial: option: add Neoway N723-EA support
  USB: serial: option: add MeiG Smart SRM815
  USB: serial: cp210x: add Phoenix Contact UPS Device
  usb: typec: fix pm usage counter imbalance in ucsi_ccg_sync_control()
  usb-storage: Add max sectors quirk for Nokia 208
  usb: gadget: midi2: Reverse-select at the right place
  usb: gadget: f_fs: Remove WARN_ON in functionfs_bind
  USB: core: Disable LPM only for non-suspended ports
  usb: fix reference leak in usb_new_device()
  usb: typec: tcpci: fix NULL pointer issue on shared irq case
  usb: gadget: u_serial: Disable ep before setting port to null to fix the crash caused by port being null
  usb: chipidea: ci_hdrc_imx: decrement device's refcount in .remove() and in the error path of .probe()
  usb: typec: ucsi: Set orientation as none when connector is unplugged
  usb: gadget: configfs: Ignore trailing LF for user strings to cdev
  USB: usblp: return error when setting unsupported protocol
  usb: gadget: f_uac2: Fix incorrect setting of bNumEndpoints
  usb: typec: tcpm/tcpci_maxim: fix error code in max_contaminant_read_resistance_kohm()
  usb: host: xhci-plat: set skip_phy_initialization if software node has XHCI_SKIP_PHY_INIT property
  usb: dwc3-am62: Disable autosuspend during remove
  usb: dwc3: gadget: fix writing NYET threshold

8 months agoMerge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm
Linus Torvalds [Sun, 12 Jan 2025 20:04:53 +0000 (12:04 -0800)]
Merge tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm

Pull kvm fixes from Paolo Bonzini:
 "The largest part here is for KVM/PPC, where a NULL pointer dereference
  was introduced in the 6.13 merge window and is now fixed.

  There's some "holiday-induced lateness", as the s390 submaintainer put
  it, but otherwise things looks fine.

  s390:

   - fix a latent bug when the kernel is compiled in debug mode

   - two small UCONTROL fixes and their selftests

  arm64:

   - always check page state in hyp_ack_unshare()

   - align set_id_regs selftest with the fact that ASIDBITS field is RO

   - various vPMU fixes for bugs that only affect nested virt

  PPC e500:

   - Fix a mostly impossible (but just wrong) case where IRQs were never
     re-enabled

   - Observe host permissions instead of mapping readonly host pages as
     guest-writable. This fixes a NULL-pointer dereference in 6.13

   - Replace brittle VMA-based attempts at building huge shadow TLB
     entries with PTE lookups"

* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
  KVM: e500: perform hugepage check after looking up the PFN
  KVM: e500: map readonly host pages for read
  KVM: e500: track host-writability of pages
  KVM: e500: use shadow TLB entry as witness for writability
  KVM: e500: always restore irqs
  KVM: s390: selftests: Add has device attr check to uc_attr_mem_limit selftest
  KVM: s390: selftests: Add ucontrol gis routing test
  KVM: s390: Reject KVM_SET_GSI_ROUTING on ucontrol VMs
  KVM: s390: selftests: Add ucontrol flic attr selftests
  KVM: s390: Reject setting flic pfault attributes on ucontrol VMs
  KVM: s390: vsie: fix virtual/physical address in unpin_scb()
  KVM: arm64: Only apply PMCR_EL0.P to the guest range of counters
  KVM: arm64: nv: Reload PMU events upon MDCR_EL2.HPME change
  KVM: arm64: Use KVM_REQ_RELOAD_PMU to handle PMCR_EL0.E change
  KVM: arm64: Add unified helper for reprogramming counters by mask
  KVM: arm64: Always check the state from hyp_ack_unshare()
  KVM: arm64: Fix set_id_regs selftest for ASIDBITS becoming unwritable

8 months agoMerge tag 'perf_urgent_for_v6.13_rc7' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 12 Jan 2025 19:57:45 +0000 (11:57 -0800)]
Merge tag 'perf_urgent_for_v6.13_rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull perf fix from Borislav Petkov:

 - Fix a #GP in the perf user callchain code caused by a race between
   uprobe freeing the task and the bpf profiler unwinding the task's
   user stack

* tag 'perf_urgent_for_v6.13_rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  uprobes: Fix race in uprobe_free_utask

8 months agoMerge tag 'x86_urgent_for_v6.13_rc7' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Sun, 12 Jan 2025 19:55:48 +0000 (11:55 -0800)]
Merge tag 'x86_urgent_for_v6.13_rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip

Pull x86 fixes from Borislav Petkov:

 - Check whether shadow stack is active before using the ptrace regset
   getter

 - Remove a wrong BUG_ON in the early static call code which breaks Xen
   PVH when booting as dom0

* tag 'x86_urgent_for_v6.13_rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip:
  x86/fpu: Ensure shadow stack is active before "getting" registers
  x86/static-call: Remove early_boot_irqs_disabled check to fix Xen PVH dom0

8 months agoMerge tag 'kvm-s390-master-6.13-1' of https://git.kernel.org/pub/scm/linux/kernel...
Paolo Bonzini [Sun, 12 Jan 2025 11:51:05 +0000 (12:51 +0100)]
Merge tag 'kvm-s390-master-6.13-1' of https://git.kernel.org/pub/scm/linux/kernel/git/kvms390/linux into HEAD

KVM: s390: three small bugfixes

Fix a latent bug when the kernel is compiled in debug mode.
Two small UCONTROL fixes and their selftests.

8 months agoMerge tag 'kvmarm-fixes-6.13-3' of https://git.kernel.org/pub/scm/linux/kernel/git...
Paolo Bonzini [Sun, 12 Jan 2025 11:50:39 +0000 (12:50 +0100)]
Merge tag 'kvmarm-fixes-6.13-3' of https://git.kernel.org/pub/scm/linux/kernel/git/kvmarm/kvmarm into HEAD

KVM/arm64 changes for 6.13, part #3

 - Always check page state in hyp_ack_unshare()

 - Align set_id_regs selftest with the fact that ASIDBITS field is RO

 - Various vPMU fixes for bugs that only affect nested virt

8 months agoMerge branch 'kvm-e500-check-writable-pfn' into HEAD
Paolo Bonzini [Sun, 12 Jan 2025 11:48:14 +0000 (12:48 +0100)]
Merge branch 'kvm-e500-check-writable-pfn' into HEAD

The new __kvm_faultin_pfn() function is upset by the fact that e500
KVM ignores host page permissions - __kvm_faultin requires a "writable"
outgoing argument, but e500 KVM is passing NULL.

While a simple fix would be possible that simply allows writable to
be NULL, it is quite ugly to have e500 KVM ignore completely the host
permissions and map readonly host pages as guest-writable.  Merge a more
complete fix and remove the VMA-based attempts at building huge shadow TLB
entries.  Using a PTE lookup, similar to what is done for x86, is better
and works with remap_pfn_range() because it does not assume that VM_PFNMAP
areas are contiguous.  Note that the same incorrect logic is there in
ARM's get_vma_page_shift() and RISC-V's kvm_riscv_gstage_ioremap().

Fortunately, for e500 most of the code is already there; it just has to
be changed to compute the range from find_linux_pte()'s output rather
than find_vma().  The new code works for both VM_PFNMAP and hugetlb
mappings, so the latter is removed.

Patches 2-5 were tested by the reporter, Christian Zigotzky.  Since
the difference with v1 is minimal, I am going to send it to Linus
today.

8 months agoKVM: e500: perform hugepage check after looking up the PFN
Paolo Bonzini [Wed, 8 Jan 2025 15:49:50 +0000 (16:49 +0100)]
KVM: e500: perform hugepage check after looking up the PFN

e500 KVM tries to bypass __kvm_faultin_pfn() in order to map VM_PFNMAP
VMAs as huge pages.  This is a Bad Idea because VM_PFNMAP VMAs could
become noncontiguous as a result of callsto remap_pfn_range().

Instead, use the already existing host PTE lookup to retrieve a
valid host-side mapping level after __kvm_faultin_pfn() has
returned.  Then find the largest size that will satisfy the
guest's request while staying within a single host PTE.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
8 months agoKVM: e500: map readonly host pages for read
Paolo Bonzini [Wed, 8 Jan 2025 15:14:55 +0000 (16:14 +0100)]
KVM: e500: map readonly host pages for read

The new __kvm_faultin_pfn() function is upset by the fact that e500 KVM
ignores host page permissions - __kvm_faultin requires a "writable"
outgoing argument, but e500 KVM is nonchalantly passing NULL.

If the host page permissions do not include writability, the shadow
TLB entry is forcibly mapped read-only.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
8 months agoKVM: e500: track host-writability of pages
Paolo Bonzini [Wed, 8 Jan 2025 15:21:38 +0000 (16:21 +0100)]
KVM: e500: track host-writability of pages

Add the possibility of marking a page so that the UW and SW bits are
force-cleared.  This is stored in the private info so that it persists
across multiple calls to kvmppc_e500_setup_stlbe.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
8 months agoKVM: e500: use shadow TLB entry as witness for writability
Paolo Bonzini [Wed, 8 Jan 2025 15:19:28 +0000 (16:19 +0100)]
KVM: e500: use shadow TLB entry as witness for writability

kvmppc_e500_ref_setup is returning whether the guest TLB entry is writable,
which is than passed to kvm_release_faultin_page.  This makes little sense
for two reasons: first, because the function sets up the private data for
the page and the return value feels like it has been bolted on the side;
second, because what really matters is whether the _shadow_ TLB entry is
writable.  If it is not writable, the page can be released as non-dirty.
Shift from using tlbe_is_writable(gtlbe) to doing the same check on
the shadow TLB entry.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
8 months agoKVM: e500: always restore irqs
Paolo Bonzini [Sun, 12 Jan 2025 09:34:44 +0000 (10:34 +0100)]
KVM: e500: always restore irqs

If find_linux_pte fails, IRQs will not be restored.  This is unlikely
to happen in practice since it would have been reported as hanging
hosts, but it should of course be fixed anyway.

Cc: stable@vger.kernel.org
Reported-by: Sean Christopherson <seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
8 months agoMerge tag 'probes-fixes-v6.13-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sun, 12 Jan 2025 04:34:12 +0000 (20:34 -0800)]
Merge tag 'probes-fixes-v6.13-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace

Pull probes fix from Masami Hiramatsu:
 "Fix to free trace_kprobe objects at a failure path in
  __trace_kprobe_create() function. This fixes a memory leak"

* tag 'probes-fixes-v6.13-rc6' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace:
  tracing/kprobes: Fix to free objects when failed to copy a symbol

8 months agoMerge tag 'hwmon-for-v6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Sat, 11 Jan 2025 19:42:48 +0000 (11:42 -0800)]
Merge tag 'hwmon-for-v6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging

Pull hwmon fix from Guenter Roeck:
 "One patch to fix error handling in drivetemp driver"

* tag 'hwmon-for-v6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/groeck/linux-staging:
  hwmon: (drivetemp) Fix driver producing garbage data when SCSI errors occur

8 months agoMerge tag 'block-6.13-20250111' of git://git.kernel.dk/linux
Linus Torvalds [Sat, 11 Jan 2025 19:17:08 +0000 (11:17 -0800)]
Merge tag 'block-6.13-20250111' of git://git.kernel.dk/linux

Pull block fix from Jens Axboe:
 "A single fix for a use-after-free in the BFQ IO scheduler"

* tag 'block-6.13-20250111' of git://git.kernel.dk/linux:
  block, bfq: fix waker_bfqq UAF after bfq_split_bfqq()

8 months agoMerge tag 'io_uring-6.13-20250111' of git://git.kernel.dk/linux
Linus Torvalds [Sat, 11 Jan 2025 18:59:43 +0000 (10:59 -0800)]
Merge tag 'io_uring-6.13-20250111' of git://git.kernel.dk/linux

Pull io_uring fixes from Jens Axboe:

 - Fix for multishot timeout updates only using the updated value for
   the first invocation, not subsequent ones

 - Silence a false positive lockdep warning

 - Fix the eventfd signaling and putting RCU logic

 - Fix fault injected SQPOLL setup not clearing the task pointer in the
   error path

 - Fix local task_work looking at the SQPOLL thread rather than just
   signaling the safe variant. Again one of those theoretical issues,
   which should be closed up none the less.

* tag 'io_uring-6.13-20250111' of git://git.kernel.dk/linux:
  io_uring: don't touch sqd->thread off tw add
  io_uring/sqpoll: zero sqd->thread on tctx errors
  io_uring/eventfd: ensure io_eventfd_signal() defers another RCU period
  io_uring: silence false positive warnings
  io_uring/timeout: fix multishot updates

8 months agoMerge tag '6.13-rc6-SMB3-client-fix' of git://git.samba.org/sfrench/cifs-2.6
Linus Torvalds [Sat, 11 Jan 2025 18:49:50 +0000 (10:49 -0800)]
Merge tag '6.13-rc6-SMB3-client-fix' of git://git.samba.org/sfrench/cifs-2.6

Pull smb client fix from Steve French:

 - fix unneeded session setup retry due to stale password e.g. for DFS
   automounts

* tag '6.13-rc6-SMB3-client-fix' of git://git.samba.org/sfrench/cifs-2.6:
  smb: client: sync the root session and superblock context passwords before automounting

8 months agoMerge tag 'soc-fixes-6.13-3' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc
Linus Torvalds [Sat, 11 Jan 2025 18:42:05 +0000 (10:42 -0800)]
Merge tag 'soc-fixes-6.13-3' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc

Pull SoC fixes from Arnd Bergmann:
 "Over the Christmas break a couple of devicetree fixes came in for
  Rockchips, Qualcomm and NXP/i.MX. These add some missing board
  specific properties, address build time warnings,

  The USB/TOG supoprt on X1 Elite regressed, so two earlier DT changes
  get reverted for now.

  Aside from the devicetree fixes, there is One build fix for the stm32
  firewall driver, and a defconfig change to enable SPDIF support for
  i.MX"

* tag 'soc-fixes-6.13-3' of git://git.kernel.org/pub/scm/linux/kernel/git/soc/soc:
  firewall: remove misplaced semicolon from stm32_firewall_get_firewall
  arm64: dts: rockchip: add hevc power domain clock to rk3328
  arm64: dts: rockchip: Fix the SD card detection on NanoPi R6C/R6S
  arm64: dts: qcom: sa8775p: fix the secure device bootup issue
  Revert "arm64: dts: qcom: x1e80100: enable OTG on USB-C controllers"
  Revert "arm64: dts: qcom: x1e80100-crd: enable otg on usb ports"
  arm64: dts: qcom: x1e80100: Fix up BAR space size for PCIe6a
  Revert "arm64: dts: qcom: x1e78100-t14s: enable otg on usb-c ports"
  ARM: dts: imxrt1050: Fix clocks for mmc
  ARM: imx_v6_v7_defconfig: enable SND_SOC_SPDIF
  arm64: dts: imx95: correct the address length of netcmix_blk_ctrl
  arm64: dts: imx8-ss-audio: add fallback compatible string fsl,imx6ull-esai for esai
  arm64: dts: rockchip: rename rfkill label for Radxa ROCK 5B
  arm64: dts: rockchip: add reset-names for combphy on rk3568
  arm64: dts: qcom: sa8775p: Fix the size of 'addr_space' regions

8 months agoMAINTAINERS: powerpc: Update my status
Michael Ellerman [Fri, 10 Jan 2025 23:57:38 +0000 (10:57 +1100)]
MAINTAINERS: powerpc: Update my status

Maddy is taking over the day-to-day maintenance of powerpc. I will still
be around to help, and as a backup.

Re-order the main POWERPC list to put Maddy first to reflect that.

KVM/powerpc patches will be handled by Maddy via the powerpc tree with
review from Nick, so replace myself with Maddy there.

Remove myself from BPF, leaving Hari & Christophe as maintainers.

Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
8 months agosmb: client: sync the root session and superblock context passwords before automounting
Meetakshi Setiya [Wed, 8 Jan 2025 10:10:34 +0000 (05:10 -0500)]
smb: client: sync the root session and superblock context passwords before automounting

In some cases, when password2 becomes the working password, the
client swaps the two password fields in the root session struct, but
not in the smb3_fs_context struct in cifs_sb. DFS automounts inherit
fs context from their parent mounts. Therefore, they might end up
getting the passwords in the stale order.
The automount should succeed, because the mount function will end up
retrying with the actual password anyway. But to reduce these
unnecessary session setup retries for automounts, we can sync the
parent context's passwords with the root session's passwords before
duplicating it to the child's fs context.

Cc: stable@vger.kernel.org
Signed-off-by: Meetakshi Setiya <msetiya@microsoft.com>
Reviewed-by: Shyam Prasad N <sprasad@microsoft.com>
Acked-by: Paulo Alcantara (Red Hat) <pc@manguebit.com>
Signed-off-by: Steve French <stfrench@microsoft.com>
8 months agoMerge tag 'sched_ext-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 10 Jan 2025 23:11:58 +0000 (15:11 -0800)]
Merge tag 'sched_ext-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext

Pull sched_ext fixes from Tejun Heo:

 - Fix corner case bug where ops.dispatch() couldn't extend the
   execution of the current task if SCX_OPS_ENQ_LAST is set.

 - Fix ops.cpu_release() not being called when a SCX task is preempted
   by a higher priority sched class task.

 - Fix buitin idle mask being incorrectly left as busy after an idle CPU
   is picked and kicked.

 - scx_ops_bypass() was unnecessarily using rq_lock() which comes with
   rq pinning related sanity checks which could trigger spuriously.
   Switch to raw_spin_rq_lock().

* tag 'sched_ext-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/sched_ext:
  sched_ext: idle: Refresh idle masks during idle-to-idle transitions
  sched_ext: switch class when preempted by higher priority scheduler
  sched_ext: Replace rq_lock() to raw_spin_rq_lock() in scx_ops_bypass()
  sched_ext: keep running prev when prev->scx.slice != 0

8 months agoMerge tag 'cgroup-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel...
Linus Torvalds [Fri, 10 Jan 2025 23:03:02 +0000 (15:03 -0800)]
Merge tag 'cgroup-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup

Pull cgroup fixes from Tejun Heo:
 "Cpuset fixes:

   - Fix isolated CPUs leaking into sched domains

   - Remove now unnecessary kernfs active break which can trigger a
     warning

   - Comment updates"

* tag 'cgroup-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup:
  cgroup/cpuset: remove kernfs active break
  cgroup/cpuset: Prevent leakage of isolated CPUs into sched domains
  cgroup/cpuset: Remove stale text

8 months agoMerge tag 'wq-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git...
Linus Torvalds [Fri, 10 Jan 2025 22:52:30 +0000 (14:52 -0800)]
Merge tag 'wq-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq

Pull workqueue fix from Tejun Heo:

 - Add a WARN_ON_ONCE() on queue_delayed_work_on() on an offline CPU as
   such work items won't get executed till the CPU comes back online

* tag 'wq-for-6.13-rc6-fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/wq:
  workqueue: warn if delayed_work is queued to an offlined cpu.

8 months agoMerge tag 'thermal-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael...
Linus Torvalds [Fri, 10 Jan 2025 22:46:49 +0000 (14:46 -0800)]
Merge tag 'thermal-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm

Pull thermal control fix from Rafael Wysocki:
 "Fix an OF node leak in the code parsing thermal zone DT properties
  (Joe Hattori)"

* tag 'thermal-6.13-rc7' of git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm:
  thermal: of: fix OF node leak in of_thermal_zone_find()