The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm
tree, then please email the backport, including the original git commit
id to <stable(a)vger.kernel.org>.
thanks,
greg k-h
------------------ original commit in Linus's tree ------------------
From 64620e0a1e712a778095bd35cbb277dc2259281f Mon Sep 17 00:00:00 2001
From: Daniel Borkmann <daniel(a)iogearbox.net>
Date: Tue, 11 Jan 2022 14:43:41 +0000
Subject: [PATCH] bpf: Fix out of bounds access for ringbuf helpers
Both bpf_ringbuf_submit() and bpf_ringbuf_discard() have ARG_PTR_TO_ALLOC_MEM
in their bpf_func_proto definition as their first argument. They both expect
the result from a prior bpf_ringbuf_reserve() call which has a return type of
RET_PTR_TO_ALLOC_MEM_OR_NULL.
Meaning, after a NULL check in the code, the verifier will promote the register
type in the non-NULL branch to a PTR_TO_MEM and in the NULL branch to a known
zero scalar. Generally, pointer arithmetic on PTR_TO_MEM is allowed, so the
latter could have an offset.
The ARG_PTR_TO_ALLOC_MEM expects a PTR_TO_MEM register type. However, the non-
zero result from bpf_ringbuf_reserve() must be fed into either bpf_ringbuf_submit()
or bpf_ringbuf_discard() but with the original offset given it will then read
out the struct bpf_ringbuf_hdr mapping.
The verifier missed to enforce a zero offset, so that out of bounds access
can be triggered which could be used to escalate privileges if unprivileged
BPF was enabled (disabled by default in kernel).
Fixes: 457f44363a88 ("bpf: Implement BPF ring buffer and verifier support for it")
Reported-by: <tr3e.wang(a)gmail.com> (SecCoder Security Lab)
Signed-off-by: Daniel Borkmann <daniel(a)iogearbox.net>
Acked-by: John Fastabend <john.fastabend(a)gmail.com>
Acked-by: Alexei Starovoitov <ast(a)kernel.org>
diff --git a/kernel/bpf/verifier.c b/kernel/bpf/verifier.c
index e0b3f4d683eb..c72c57a6684f 100644
--- a/kernel/bpf/verifier.c
+++ b/kernel/bpf/verifier.c
@@ -5318,9 +5318,15 @@ static int check_func_arg(struct bpf_verifier_env *env, u32 arg,
case PTR_TO_BUF:
case PTR_TO_BUF | MEM_RDONLY:
case PTR_TO_STACK:
+ /* Some of the argument types nevertheless require a
+ * zero register offset.
+ */
+ if (arg_type == ARG_PTR_TO_ALLOC_MEM)
+ goto force_off_check;
break;
/* All the rest must be rejected: */
default:
+force_off_check:
err = __check_ptr_off_reg(env, reg, regno,
type == PTR_TO_BTF_ID);
if (err < 0)
Currently, due to the sequential use of min_t() and clamp_t() macros,
in cdc_ncm_check_tx_max(), if dwNtbOutMaxSize is not set, the logic
sets tx_max to 0. This is then used to allocate the data area of the
SKB requested later in cdc_ncm_fill_tx_frame().
This does not cause an issue presently because when memory is
allocated during initialisation phase of SKB creation, more memory
(512b) is allocated than is required for the SKB headers alone (320b),
leaving some space (512b - 320b = 192b) for CDC data (172b).
However, if more elements (for example 3 x u64 = [24b]) were added to
one of the SKB header structs, say 'struct skb_shared_info',
increasing its original size (320b [320b aligned]) to something larger
(344b [384b aligned]), then suddenly the CDC data (172b) no longer
fits in the spare SKB data area (512b - 384b = 128b).
Consequently the SKB bounds checking semantics fails and panics:
skbuff: skb_over_panic: text:ffffffff830a5b5f len:184 put:172 \
head:ffff888119227c00 data:ffff888119227c00 tail:0xb8 end:0x80 dev:<NULL>
------------[ cut here ]------------
kernel BUG at net/core/skbuff.c:110!
RIP: 0010:skb_panic+0x14f/0x160 net/core/skbuff.c:106
<snip>
Call Trace:
<IRQ>
skb_over_panic+0x2c/0x30 net/core/skbuff.c:115
skb_put+0x205/0x210 net/core/skbuff.c:1877
skb_put_zero include/linux/skbuff.h:2270 [inline]
cdc_ncm_ndp16 drivers/net/usb/cdc_ncm.c:1116 [inline]
cdc_ncm_fill_tx_frame+0x127f/0x3d50 drivers/net/usb/cdc_ncm.c:1293
cdc_ncm_tx_fixup+0x98/0xf0 drivers/net/usb/cdc_ncm.c:1514
By overriding the max value with the default CDC_NCM_NTB_MAX_SIZE_TX
when not offered through the system provided params, we ensure enough
data space is allocated to handle the CDC data, meaning no crash will
occur.
Cc: stable(a)vger.kernel.org
Cc: Oliver Neukum <oliver(a)neukum.org>
Cc: "David S. Miller" <davem(a)davemloft.net>
Cc: Jakub Kicinski <kuba(a)kernel.org>
Cc: linux-usb(a)vger.kernel.org
Cc: netdev(a)vger.kernel.org
Cc: linux-kernel(a)vger.kernel.org
Fixes: 289507d3364f9 ("net: cdc_ncm: use sysfs for rx/tx aggregation tuning")
Signed-off-by: Lee Jones <lee.jones(a)linaro.org>
---
drivers/net/usb/cdc_ncm.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/net/usb/cdc_ncm.c b/drivers/net/usb/cdc_ncm.c
index 24753a4da7e60..e303b522efb50 100644
--- a/drivers/net/usb/cdc_ncm.c
+++ b/drivers/net/usb/cdc_ncm.c
@@ -181,6 +181,8 @@ static u32 cdc_ncm_check_tx_max(struct usbnet *dev, u32 new_tx)
min = ctx->max_datagram_size + ctx->max_ndp_size + sizeof(struct usb_cdc_ncm_nth32);
max = min_t(u32, CDC_NCM_NTB_MAX_SIZE_TX, le32_to_cpu(ctx->ncm_parm.dwNtbOutMaxSize));
+ if (max == 0)
+ max = CDC_NCM_NTB_MAX_SIZE_TX; /* dwNtbOutMaxSize not set */
/* some devices set dwNtbOutMaxSize too low for the above default */
min = min(min, max);
--
2.34.0.384.gca35af8252-goog
From: Alexander Sverdlin <alexander.sverdlin(a)nokia.com>
Erase can be zeroed in spi_nor_parse_4bait() or
spi_nor_init_non_uniform_erase_map(). In practice it happened with
mt25qu256a, which supports 4K, 32K, 64K erases with 3b address commands,
but only 4K and 64K erase with 4b address commands.
Fixes: dc92843159a7 ("mtd: spi-nor: fix erase_type array to indicate current map conf")
Cc: stable(a)vger.kernel.org
Signed-off-by: Alexander Sverdlin <alexander.sverdlin(a)nokia.com>
---
Changes in v2:
erase->opcode -> erase->size
drivers/mtd/spi-nor/core.c | 2 ++
1 file changed, 2 insertions(+)
diff --git a/drivers/mtd/spi-nor/core.c b/drivers/mtd/spi-nor/core.c
index 88dd090..183ea9d 100644
--- a/drivers/mtd/spi-nor/core.c
+++ b/drivers/mtd/spi-nor/core.c
@@ -1400,6 +1400,8 @@ spi_nor_find_best_erase_type(const struct spi_nor_erase_map *map,
continue;
erase = &map->erase_type[i];
+ if (!erase->size)
+ continue;
/* Alignment is not mandatory for overlaid regions */
if (region->offset & SNOR_OVERLAID_REGION &&
--
2.10.2
This reverts commit 2dc016599cfa9672a147528ca26d70c3654a5423.
Users are reporting regressions in regulatory domain detection and
channel availability.
The problem this was trying to resolve was fixed in firmware anyway:
QCA6174 hw3.0: sdio-4.4.1: add firmware.bin_WLAN.RMH.4.4.1-00042
https://github.com/kvalo/ath10k-firmware/commit/4d382787f0efa77dba40394e0bc…
Link: https://bbs.archlinux.org/viewtopic.php?id=254535
Link: http://lists.infradead.org/pipermail/ath10k/2020-April/014871.html
Link: http://lists.infradead.org/pipermail/ath10k/2020-May/015152.html
Fixes: 2dc016599cfa ("ath: add support for special 0x0 regulatory domain")
Cc: <stable(a)vger.kernel.org>
Cc: Wen Gong <wgong(a)codeaurora.org>
Signed-off-by: Brian Norris <briannorris(a)chromium.org>
---
drivers/net/wireless/ath/regd.c | 10 +++++-----
1 file changed, 5 insertions(+), 5 deletions(-)
diff --git a/drivers/net/wireless/ath/regd.c b/drivers/net/wireless/ath/regd.c
index bee9110b91f3..20f4f8ea9f89 100644
--- a/drivers/net/wireless/ath/regd.c
+++ b/drivers/net/wireless/ath/regd.c
@@ -666,14 +666,14 @@ ath_regd_init_wiphy(struct ath_regulatory *reg,
/*
* Some users have reported their EEPROM programmed with
- * 0x8000 or 0x0 set, this is not a supported regulatory
- * domain but since we have more than one user with it we
- * need a solution for them. We default to 0x64, which is
- * the default Atheros world regulatory domain.
+ * 0x8000 set, this is not a supported regulatory domain
+ * but since we have more than one user with it we need
+ * a solution for them. We default to 0x64, which is the
+ * default Atheros world regulatory domain.
*/
static void ath_regd_sanitize(struct ath_regulatory *reg)
{
- if (reg->current_rd != COUNTRY_ERD_FLAG && reg->current_rd != 0)
+ if (reg->current_rd != COUNTRY_ERD_FLAG)
return;
printk(KERN_DEBUG "ath: EEPROM regdomain sanitized\n");
reg->current_rd = 0x64;
--
2.27.0.rc0.183.gde8f92d652-goog
The i801 controller provides a locking mechanism that the OS is supposed
to use to safely share the SMBus with ACPI AML or other firmware.
Previously, Linux attempted to get out of the way of ACPI AML entirely,
but left the bus locked if it used it before the first AML access. This
causes AML implementations that *do* attempt to safely share the bus
to time out if Linux uses it first; notably, this regressed ACPI video
backlight controls on 2015 iMacs after 01590f361e started instantiating
SPD EEPROMs on boot.
Commit 065b6211a8 fixed the immediate problem of leaving the bus locked,
but we can do better. The controller does have a proper locking mechanism,
so let's use it as intended. Since we can't rely on the BIOS doing this
properly, we implement the following logic:
- If ACPI AML uses the bus at all, we make a note and disable power
management. The latter matches already existing behavior.
- When we want to use the bus, we attempt to lock it first. If the
locking attempt times out, *and* ACPI hasn't tried to use the bus at
all yet, we cautiously go ahead and assume the BIOS forgot to unlock
the bus after boot. This preserves existing behavior.
- We always unlock the bus after a transfer.
- If ACPI AML tries to use the bus (except trying to lock it) while
we're in the middle of a transfer, or after we've determined
locking is broken, we know we cannot safely share the bus and give up.
Upon first usage of SMBus by ACPI AML, if nothing has gone horribly
wrong so far, users will see:
i801_smbus 0000:00:1f.4: SMBus controller is shared with ACPI AML. This seems safe so far.
If locking the SMBus times out, users will see:
i801_smbus 0000:00:1f.4: BIOS left SMBus locked
And if ACPI AML tries to use the bus concurrently with Linux, or it
previously used the bus and we failed to subsequently lock it as
above, the driver will give up and users will get:
i801_smbus 0000:00:1f.4: BIOS uses SMBus unsafely
i801_smbus 0000:00:1f.4: Driver SMBus register access inhibited
This fixes the regression introduced by 01590f361e, and further allows
safely sharing the SMBus on 2015 iMacs. Tested by running `i2cdump` in a
loop while changing backlight levels via the ACPI video device.
Fixes: 01590f361e ("i2c: i801: Instantiate SPD EEPROMs automatically")
Cc: <stable(a)vger.kernel.org>
Signed-off-by: Hector Martin <marcan(a)marcan.st>
---
drivers/i2c/busses/i2c-i801.c | 96 ++++++++++++++++++++++++++++-------
1 file changed, 79 insertions(+), 17 deletions(-)
diff --git a/drivers/i2c/busses/i2c-i801.c b/drivers/i2c/busses/i2c-i801.c
index 04a1e38f2a6f..03be6310d6d7 100644
--- a/drivers/i2c/busses/i2c-i801.c
+++ b/drivers/i2c/busses/i2c-i801.c
@@ -287,11 +287,18 @@ struct i801_priv {
#endif
struct platform_device *tco_pdev;
+ /* BIOS left the controller marked busy. */
+ bool inuse_stuck;
/*
- * If set to true the host controller registers are reserved for
- * ACPI AML use. Protected by acpi_lock.
+ * If set to true, ACPI AML uses the host controller registers.
+ * Protected by acpi_lock.
*/
- bool acpi_reserved;
+ bool acpi_usage;
+ /*
+ * If set to true, ACPI AML uses the host controller registers in an
+ * unsafe way. Protected by acpi_lock.
+ */
+ bool acpi_unsafe;
struct mutex acpi_lock;
};
@@ -854,10 +861,37 @@ static s32 i801_access(struct i2c_adapter *adap, u16 addr,
int hwpec;
int block = 0;
int ret = 0, xact = 0;
+ int timeout = 0;
struct i801_priv *priv = i2c_get_adapdata(adap);
+ /*
+ * The controller provides a bit that implements a mutex mechanism
+ * between users of the bus. First, try to lock the hardware mutex.
+ * If this doesn't work, we give up trying to do this, but then
+ * bail if ACPI uses SMBus at all.
+ */
+ if (!priv->inuse_stuck) {
+ while (inb_p(SMBHSTSTS(priv)) & SMBHSTSTS_INUSE_STS) {
+ if (++timeout >= MAX_RETRIES) {
+ dev_warn(&priv->pci_dev->dev,
+ "BIOS left SMBus locked\n");
+ priv->inuse_stuck = true;
+ break;
+ }
+ usleep_range(250, 500);
+ }
+ }
+
mutex_lock(&priv->acpi_lock);
- if (priv->acpi_reserved) {
+ if (priv->acpi_usage && priv->inuse_stuck && !priv->acpi_unsafe) {
+ priv->acpi_unsafe = true;
+
+ dev_warn(&priv->pci_dev->dev, "BIOS uses SMBus unsafely\n");
+ dev_warn(&priv->pci_dev->dev,
+ "Driver SMBus register access inhibited\n");
+ }
+
+ if (priv->acpi_unsafe) {
mutex_unlock(&priv->acpi_lock);
return -EBUSY;
}
@@ -1639,6 +1673,16 @@ static bool i801_acpi_is_smbus_ioport(const struct i801_priv *priv,
address <= pci_resource_end(priv->pci_dev, SMBBAR);
}
+static acpi_status
+i801_acpi_do_access(u32 function, acpi_physical_address address,
+ u32 bits, u64 *value)
+{
+ if ((function & ACPI_IO_MASK) == ACPI_READ)
+ return acpi_os_read_port(address, (u32 *)value, bits);
+ else
+ return acpi_os_write_port(address, (u32)*value, bits);
+}
+
static acpi_status
i801_acpi_io_handler(u32 function, acpi_physical_address address, u32 bits,
u64 *value, void *handler_context, void *region_context)
@@ -1648,17 +1692,38 @@ i801_acpi_io_handler(u32 function, acpi_physical_address address, u32 bits,
acpi_status status;
/*
- * Once BIOS AML code touches the OpRegion we warn and inhibit any
- * further access from the driver itself. This device is now owned
- * by the system firmware.
+ * Non-i801 accesses pass through.
*/
- mutex_lock(&priv->acpi_lock);
+ if (!i801_acpi_is_smbus_ioport(priv, address))
+ return i801_acpi_do_access(function, address, bits, value);
- if (!priv->acpi_reserved && i801_acpi_is_smbus_ioport(priv, address)) {
- priv->acpi_reserved = true;
+ if (!mutex_trylock(&priv->acpi_lock)) {
+ mutex_lock(&priv->acpi_lock);
+ /*
+ * This better be a read of the status register to acquire
+ * the lock...
+ */
+ if (!priv->acpi_unsafe &&
+ !(address == SMBHSTSTS(priv) &&
+ (function & ACPI_IO_MASK) == ACPI_READ)) {
+ /*
+ * Uh-oh, ACPI AML is trying to do something with the
+ * controller without locking it properly.
+ */
+ priv->acpi_unsafe = true;
+
+ dev_warn(&pdev->dev, "BIOS uses SMBus unsafely\n");
+ dev_warn(&pdev->dev,
+ "Driver SMBus register access inhibited\n");
+ }
+ }
- dev_warn(&pdev->dev, "BIOS is accessing SMBus registers\n");
- dev_warn(&pdev->dev, "Driver SMBus register access inhibited\n");
+ if (!priv->acpi_usage) {
+ priv->acpi_usage = true;
+
+ if (!priv->acpi_unsafe)
+ dev_info(&pdev->dev,
+ "SMBus controller is shared with ACPI AML. This seems safe so far.\n");
/*
* BIOS is accessing the host controller so prevent it from
@@ -1667,10 +1732,7 @@ i801_acpi_io_handler(u32 function, acpi_physical_address address, u32 bits,
pm_runtime_get_sync(&pdev->dev);
}
- if ((function & ACPI_IO_MASK) == ACPI_READ)
- status = acpi_os_read_port(address, (u32 *)value, bits);
- else
- status = acpi_os_write_port(address, (u32)*value, bits);
+ status = i801_acpi_do_access(function, address, bits, value);
mutex_unlock(&priv->acpi_lock);
@@ -1706,7 +1768,7 @@ static void i801_acpi_remove(struct i801_priv *priv)
ACPI_ADR_SPACE_SYSTEM_IO, i801_acpi_io_handler);
mutex_lock(&priv->acpi_lock);
- if (priv->acpi_reserved)
+ if (priv->acpi_usage)
pm_runtime_put(&priv->pci_dev->dev);
mutex_unlock(&priv->acpi_lock);
}
--
2.32.0
When booting with ACPI unavailable or disabled, get_smp_config() ends up
calling MP_processor_info() for each CPU found in the MPS
table. Previously, this resulted in boot_cpu_physical_apicid getting
unconditionally overwritten by the apicid of whatever processor had the
CPU_BOOTPROCESSOR flag. This occurred even if boot_cpu_physical_apicid
had already been more reliably determined in register_lapic_address() by
calling read_apic_id() from the actual boot processor.
Ordinariliy, this is not a problem because the boot processor really is
the one with the CPU_BOOTPROCESSOR flag. However, kexec is an exception
in which the kernel may be booted from any processor regardless of the
MPS table contents. In this case, boot_cpu_physical_apicid may not
indicate the actual boot processor.
This was particularly problematic when the second kernel was booted with
NR_CPUS fewer than the number of physical processors. It's the job of
generic_processor_info() to decide which CPUs to bring up in this case.
That obviously must include the real boot processor which it takes care
to save a slot for. It relies upon the contents of
boot_cpu_physical_apicid to do this, which if incorrect, may result in
the boot processor getting left out.
This condition can be discovered by smp_sanity_check() and rectified by
adding the boot processor to the phys_cpu_present_map with the warning
"weird, boot CPU (#%d) not listed by the BIOS". However, commit
3e730dad3b6da ("x86/apic: Unify interrupt mode setup for SMP-capable
system") caused setup_local_APIC() to be called before this could happen
resulting in a BUG_ON(!apic->apic_id_registered()):
[ 0.655452] ------------[ cut here ]------------
[ 0.660610] Kernel BUG at setup_local_APIC+0x74/0x280 [verbose debug info unavailable]
[ 0.669466] invalid opcode: 0000 [#1] SMP
[ 0.673948] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 4.19.109.Ar-16509018.eostrunkkernel419 #1
[ 0.683670] Hardware name: Quanta Quanta LY6 (1LY6UZZ0FBC), BIOS 1.0.6.0-e7d6a55 11/26/2015
[ 0.693007] RIP: 0010:setup_local_APIC+0x74/0x280
[ 0.698264] Code: 80 e4 fe bf f0 00 00 00 89 c6 48 8b 05 0f 1a 8e 00 ff 50 10 e8 12 53 fd ff 48 8b 05 00 1a 8e 00 ff 90 a0 00 00 00 85 c0 75 02 <0f> 0b 48 8b 05 ed 19 8e 00 41 be 00 02 00 00 ff 90 b0 00 00 00 48
[ 0.719251] RSP: 0000:ffffffff81a03e20 EFLAGS: 00010246
[ 0.725091] RAX: 0000000000000000 RBX: 0000000000000003 RCX: 0000000000000000
[ 0.733066] RDX: 0000000000000000 RSI: 000000000000000f RDI: 0000000000000020
[ 0.741041] RBP: ffffffff81a03e98 R08: 0000000000000002 R09: 0000000000000000
[ 0.749014] R10: ffffffff81a204e0 R11: ffffffff81b50ea7 R12: 0000000000000000
[ 0.756989] R13: ffffffff81aef920 R14: ffffffff81af60a0 R15: 0000000000000000
[ 0.764965] FS: 0000000000000000(0000) GS:ffff888036800000(0000) knlGS:0000000000000000
[ 0.774007] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[ 0.780427] CR2: ffff888035c01000 CR3: 0000000035a08000 CR4: 00000000000006b0
[ 0.788401] Call Trace:
[ 0.791137] ? amd_iommu_prepare+0x15/0x2a
[ 0.795717] apic_bsp_setup+0x55/0x75
[ 0.799808] apic_intr_mode_init+0x169/0x16e
[ 0.804579] x86_late_time_init+0x10/0x17
[ 0.809062] start_kernel+0x37e/0x3fe
[ 0.813154] x86_64_start_reservations+0x2a/0x2c
[ 0.818316] x86_64_start_kernel+0x72/0x75
[ 0.822886] secondary_startup_64+0xa4/0xb0
[ 0.827564] ---[ end trace 237b64da0fd9b22e ]---
This change avoids these issues by only setting boot_cpu_physical_apicid
from the MPS table if it is not already set, which can occur in the
construct_default_ISA_mptable() path. Otherwise,
boot_cpu_physical_apicid will already have been set in
register_lapic_address() and should therefore remain untouched.
Looking through all the places where boot_cpu_physical_apicid is
accessed, nearly all of them assume that boot_cpu_physical_apicid should
match read_apic_id() on the booting processor. The only place that might
intend to use the BSP apicid listed in the MPS table is amd_numa_init(),
which explicitly requires boot_cpu_physical_apicid to be the lowest
apicid of all processors. Ironically, due to the early exit short
circuit in early_get_smp_config(), it instead gets
boot_cpu_physical_apicid = read_apic_id() rather than the MPS table
BSP. The behaviour of amd_numa_init() is therefore unaffected by this
change.
Fixes: 3e730dad3b6da ("x86/apic: Unify interrupt mode setup for SMP-capable system")
Signed-off-by: Kevin Mitchell <kevmitch(a)arista.com>
Cc: <stable(a)vger.kernel.org>
---
arch/x86/kernel/mpparse.c | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
diff --git a/arch/x86/kernel/mpparse.c b/arch/x86/kernel/mpparse.c
index afac7ccce72f..6f22f09bfe11 100644
--- a/arch/x86/kernel/mpparse.c
+++ b/arch/x86/kernel/mpparse.c
@@ -64,7 +64,8 @@ static void __init MP_processor_info(struct mpc_cpu *m)
if (m->cpuflag & CPU_BOOTPROCESSOR) {
bootup_cpu = " (Bootup-CPU)";
- boot_cpu_physical_apicid = m->apicid;
+ if (boot_cpu_physical_apicid == -1U)
+ boot_cpu_physical_apicid = m->apicid;
}
pr_info("Processor #%d%s\n", m->apicid, bootup_cpu);
--
2.26.2