Skip to contentSkip to conversation

NFSv4.1 handle ERR_DELAY error reclaiming locking state on delegation recall

PATCH In review Sasha Levin 3 participants 19 messages Sep 14, 2020 – Sep 20, 2020

Subsystems (11)
lore
Share & tools
lore
author maintainer
#13 lore raw

PATCH 10/15 nvme-fc: cancel async events before freeing event struct

In reply to Sasha Levin

From: David Milburn <[email protected]>

[ Upstream commit e126e8210e950bb83414c4f57b3120ddb8450742 ]

Cancel async event work in case async event has been queued up, and nvme_fc_submit_async_event() runs after event has been freed.

---
 drivers/nvme/host/fc.c | 1 +
 1 file changed, 1 insertion(+)
drivers/nvme/host/fc.c +1 -0
index 13c89cc9d10cf..e95d2f75713e1 100644
@@ -1566,6 +1566,7 @@ nvme_fc_term_aen_ops(struct nvme_fc_ctrl *ctrl)
 	struct nvme_fc_fcp_op *aen_op;
 	int i;
 
+	cancel_work_sync(&ctrl->ctrl.async_event_work);
 	aen_op = ctrl->aen_ops;
 	for (i = 0; i < NVME_FC_NR_AEN_COMMANDS; i++, aen_op++) {
 		if (!aen_op->fcp_req.private)
Signature · 7 lines
2.25.1


_______________________________________________
Linux-nvme mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-nvme
author maintainer
#14 lore raw

PATCH 11/15 f2fs: fix indefinite loop scanning for free nid

In reply to Sasha Levin

From: Sahitya Tummala <[email protected]>

[ Upstream commit e2cab031ba7b5003cd12185b3ef38f1a75e3dae8 ]

If the sbi->ckpt->next_free_nid is not NAT block aligned and if there are free nids in that NAT block between the start of the block and next_free_nid, then those free nids will not be scanned in scan_nat_page(). This results into mismatch between nm_i->available_nids and the sum of nm_i->free_nid_count of all NAT blocks scanned. And nm_i->available_nids will always be greater than the sum of free nids in all the blocks. Under this condition, if we use all the currently scanned free nids, then it will loop forever in f2fs_alloc_nid() as nm_i->available_nids is still not zero but nm_i->free_nid_count of that partially scanned NAT block is zero.

Fix this to align the nm_i->next_scan_nid to the first nid of the corresponding NAT block.

---
 fs/f2fs/node.c | 3 +++
 1 file changed, 3 insertions(+)
fs/f2fs/node.c +3 -0
index e7b8e2b35e226..f8006f62c5467 100644
@@ -2011,6 +2011,9 @@ static void __build_free_nids(struct f2fs_sb_info *sbi, bool sync, bool mount)
 	if (unlikely(nid >= nm_i->max_nid))
 		nid = 0;
 
+	if (unlikely(nid % NAT_ENTRY_PER_BLOCK))
+		nid = NAT_BLOCK_OFFSET(nid) * NAT_ENTRY_PER_BLOCK;
+
 	/* Enough entries */
 	if (nm_i->nid_cnt[FREE_NID_LIST] >= NAT_ENTRY_PER_BLOCK)
 		return;
Signature · 8 lines
2.25.1



_______________________________________________
Linux-f2fs-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/linux-f2fs-devel
author maintainer
#9 lore raw

PATCH 8/15 rapidio: Replace 'select' DMAENGINES 'with depends on'

In reply to Sasha Levin

From: Laurent Pinchart <[email protected]>

[ Upstream commit d2b86100245080cfdf1e95e9e07477474c1be2bd ]

Enabling a whole subsystem from a single driver 'select' is frowned upon and won't be accepted in new drivers, that need to use 'depends on' instead. Existing selection of DMAENGINES will then cause circular dependencies. Replace them with a dependency.

---
 drivers/rapidio/Kconfig | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/rapidio/Kconfig +1 -1
index d6d2f20c45977..21df2816def76 100644
@@ -25,7 +25,7 @@ config RAPIDIO_ENABLE_RX_TX_PORTS
 config RAPIDIO_DMA_ENGINE
 	bool "DMA Engine support for RapidIO"
 	depends on RAPIDIO
-	select DMADEVICES
+	depends on DMADEVICES
 	select DMA_ENGINE
 	help
 	  Say Y here if you want to use DMA Engine frameork for RapidIO data
Signature · 1 line
2.25.1
author maintainer
#18 lore raw

PATCH 15/15 KVM: MIPS: Change the definition of kvm type

In reply to Sasha Levin

From: Huacai Chen <[email protected]>

[ Upstream commit 15e9e35cd1dec2bc138464de6bf8ef828df19235 ]

MIPS defines two kvm types:

#define KVM_VM_MIPS_TE 0
#define KVM_VM_MIPS_VZ 1

In Documentation/virt/kvm/api.rst it is said that "You probably want to use 0 as machine type", which implies that type 0 be the "automatic" or "default" type. And, in user-space libvirt use the null-machine (with type 0) to detect the kvm capability, which returns "KVM not supported" on a VZ platform.

I try to fix it in QEMU but it is ugly:
https://lists.nongnu.org/archive/html/qemu-devel/2020-08/msg05629.html

And Thomas Huth suggests me to change the definition of kvm type:
https://lists.nongnu.org/archive/html/qemu-devel/2020-09/msg03281.html

So I define like this:

 #define KVM_VM_MIPS_AUTO        0
 #define KVM_VM_MIPS_VZ          1
 #define KVM_VM_MIPS_TE          2

Since VZ and TE cannot co-exists, using type 0 on a TE platform will still return success (so old user-space tools have no problems on new kernels); the advantage is that using type 0 on a VZ platform will not return failure. So, the only problem is "new user-space tools use type 2 on old kernels", but if we treat this as a kernel bug, we can backport this patch to old stable kernels.

Message-Id: <[email protected]>

---
 arch/mips/kvm/mips.c     | 2 ++
 include/uapi/linux/kvm.h | 5 +++--
 2 files changed, 5 insertions(+), 2 deletions(-)
arch/mips/kvm/mips.c +2 -0
index aa6c365f25591..8614225e92eb5 100644
@@ -131,6 +131,8 @@ void kvm_arch_check_processor_compat(void *rtn)
 int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
 {
 	switch (type) {
+	case KVM_VM_MIPS_AUTO:
+		break;
 #ifdef CONFIG_KVM_MIPS_VZ
 	case KVM_VM_MIPS_VZ:
 #else
include/uapi/linux/kvm.h +3 -2
index efe8873943f66..62f5e47aed160 100644
@@ -735,9 +735,10 @@ struct kvm_ppc_resize_hpt {
 #define KVM_VM_PPC_HV 1
 #define KVM_VM_PPC_PR 2
 
-/* on MIPS, 0 forces trap & emulate, 1 forces VZ ASE */
-#define KVM_VM_MIPS_TE		0
+/* on MIPS, 0 indicates auto, 1 forces VZ ASE, 2 forces trap & emulate */
+#define KVM_VM_MIPS_AUTO	0
 #define KVM_VM_MIPS_VZ		1
+#define KVM_VM_MIPS_TE		2
 
 #define KVM_S390_SIE_PAGE_OFFSET 1
 
Signature · 1 line
2.25.1
#11 lore raw

Re: [PATCH AUTOSEL 4.14 09/15] kobject: Drop unneeded conditional in __kobject_del()

In reply to Sasha Levin
Quoted 34 lines · Sasha Levin

From: Andy Shevchenko <[email protected]>

[ Upstream commit 07ecc6693f9157cf293da5d165c73fb28fd69bf4 ]

__kobject_del() is called from two places, in one where kobj is dereferenced before and thus can't be NULL, and in the other the NULL check is done before call. Drop unneeded conditional in __kobject_del().

---
 lib/kobject.c | 3 ---
 1 file changed, 3 deletions(-)
lib/kobject.c +0 -3
index bbbb067de8ecd..e02f1bb67c99f 100644
@@ -568,9 +568,6 @@ void kobject_del(struct kobject *kobj)
 {
 	struct kernfs_node *sd;
 
-	if (!kobj)
-		return;
-
 	sd = kobj->sd;
 	sysfs_remove_dir(kobj);
 	sysfs_put(sd);
Signature · 1 line
2.25.1

Ouch, no, this patch should not be backported ANYWHERE or bad things will happen.

Please drop from all AUTOSEL trees.

thanks,

greg k-h

author maintainer
#15 lore raw

PATCH 12/15 i2c: algo: pca: Reapply i2c bus settings after reset

In reply to Sasha Levin

From: Evan Nimmo <[email protected]>

[ Upstream commit 0a355aeb24081e4538d4d424cd189f16c0bbd983 ]

If something goes wrong (such as the SCL being stuck low) then we need to reset the PCA chip. The issue with this is that on reset we lose all config settings and the chip ends up in a disabled state which results in a lock up/high CPU usage. We need to re-apply any configuration that had previously been set and re-enable the chip.

---
 drivers/i2c/algos/i2c-algo-pca.c | 35 +++++++++++++++++++++-----------
 include/linux/i2c-algo-pca.h     | 15 ++++++++++++++
 2 files changed, 38 insertions(+), 12 deletions(-)
drivers/i2c/algos/i2c-algo-pca.c +23 -12
index 3a9db4626cb60..1886588b9ea3e 100644
@@ -50,8 +50,22 @@ static void pca_reset(struct i2c_algo_pca_data *adap)
 		pca_outw(adap, I2C_PCA_INDPTR, I2C_PCA_IPRESET);
 		pca_outw(adap, I2C_PCA_IND, 0xA5);
 		pca_outw(adap, I2C_PCA_IND, 0x5A);
+
+		/*
+		 * After a reset we need to re-apply any configuration
+		 * (calculated in pca_init) to get the bus in a working state.
+		 */
+		pca_outw(adap, I2C_PCA_INDPTR, I2C_PCA_IMODE);
+		pca_outw(adap, I2C_PCA_IND, adap->bus_settings.mode);
+		pca_outw(adap, I2C_PCA_INDPTR, I2C_PCA_ISCLL);
+		pca_outw(adap, I2C_PCA_IND, adap->bus_settings.tlow);
+		pca_outw(adap, I2C_PCA_INDPTR, I2C_PCA_ISCLH);
+		pca_outw(adap, I2C_PCA_IND, adap->bus_settings.thi);
+
+		pca_set_con(adap, I2C_PCA_CON_ENSIO);
 	} else {
 		adap->reset_chip(adap->data);
+		pca_set_con(adap, I2C_PCA_CON_ENSIO | adap->bus_settings.clock_freq);
 	}
 }
 
@@ -435,13 +449,14 @@ static int pca_init(struct i2c_adapter *adap)
 				" Use the nominal frequency.\n", adap->name);
 		}
 
-		pca_reset(pca_data);
-
 		clock = pca_clock(pca_data);
 		printk(KERN_INFO "%s: Clock frequency is %dkHz\n",
 		     adap->name, freqs[clock]);
 
-		pca_set_con(pca_data, I2C_PCA_CON_ENSIO | clock);
+		/* Store settings as these will be needed when the PCA chip is reset */
+		pca_data->bus_settings.clock_freq = clock;
+
+		pca_reset(pca_data);
 	} else {
 		int clock;
 		int mode;
@@ -508,19 +523,15 @@ static int pca_init(struct i2c_adapter *adap)
 			thi = tlow * min_thi / min_tlow;
 		}
 
+		/* Store settings as these will be needed when the PCA chip is reset */
+		pca_data->bus_settings.mode = mode;
+		pca_data->bus_settings.tlow = tlow;
+		pca_data->bus_settings.thi = thi;
+
 		pca_reset(pca_data);
 
 		printk(KERN_INFO
 		     "%s: Clock frequency is %dHz\n", adap->name, clock * 100);
-
-		pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_IMODE);
-		pca_outw(pca_data, I2C_PCA_IND, mode);
-		pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_ISCLL);
-		pca_outw(pca_data, I2C_PCA_IND, tlow);
-		pca_outw(pca_data, I2C_PCA_INDPTR, I2C_PCA_ISCLH);
-		pca_outw(pca_data, I2C_PCA_IND, thi);
-
-		pca_set_con(pca_data, I2C_PCA_CON_ENSIO);
 	}
 	udelay(500); /* 500 us for oscillator to stabilise */
 
include/linux/i2c-algo-pca.h +15 -0
index d03071732db4a..7c522fdd9ea73 100644
@@ -53,6 +53,20 @@
 #define I2C_PCA_CON_SI		0x08 /* Serial Interrupt */
 #define I2C_PCA_CON_CR		0x07 /* Clock Rate (MASK) */
 
+/**
+ * struct pca_i2c_bus_settings - The configured PCA i2c bus settings
+ * @mode: Configured i2c bus mode
+ * @tlow: Configured SCL LOW period
+ * @thi: Configured SCL HIGH period
+ * @clock_freq: The configured clock frequency
+ */
+struct pca_i2c_bus_settings {
+	int mode;
+	int tlow;
+	int thi;
+	int clock_freq;
+};
+
 struct i2c_algo_pca_data {
 	void 				*data;	/* private low level data */
 	void (*write_byte)		(void *data, int reg, int val);
@@ -64,6 +78,7 @@ struct i2c_algo_pca_data {
 	 * For PCA9665, use the frequency you want here. */
 	unsigned int			i2c_clock;
 	unsigned int			chip;
+	struct pca_i2c_bus_settings		bus_settings;
 };
 
 int i2c_pca_add_bus(struct i2c_adapter *);
Signature · 1 line
2.25.1
author maintainer
#17 lore raw

PATCH 14/15 gcov: add support for GCC 10.1

In reply to Sasha Levin

From: Peter Oberparleiter <[email protected]>

[ Upstream commit 40249c6962075c040fd071339acae524f18bfac9 ]

Using gcov to collect coverage data for kernels compiled with GCC 10.1 causes random malfunctions and kernel crashes. This is the result of a changed GCOV_COUNTERS value in GCC 10.1 that causes a mismatch between the layout of the gcov_info structure created by GCC profiling code and the related structure used by the kernel.

Fix this by updating the in-kernel GCOV_COUNTERS value. Also re-enable config GCOV_KERNEL for use with GCC 10.

---
 kernel/gcov/gcc_4_7.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
kernel/gcov/gcc_4_7.c +3 -1
index ca5e5c0ef8536..5b9e76117ded1 100644
@@ -19,7 +19,9 @@
 #include <linux/vmalloc.h>
 #include "gcov.h"
 
-#if (__GNUC__ >= 7)
+#if (__GNUC__ >= 10)
+#define GCOV_COUNTERS			8
+#elif (__GNUC__ >= 7)
 #define GCOV_COUNTERS			9
 #elif (__GNUC__ > 5) || (__GNUC__ == 5 && __GNUC_MINOR__ >= 1)
 #define GCOV_COUNTERS			10
Signature · 1 line
2.25.1
author maintainer
#10 lore raw

PATCH 9/15 kobject: Drop unneeded conditional in __kobject_del()

In reply to Sasha Levin

From: Andy Shevchenko <[email protected]>

[ Upstream commit 07ecc6693f9157cf293da5d165c73fb28fd69bf4 ]

__kobject_del() is called from two places, in one where kobj is dereferenced before and thus can't be NULL, and in the other the NULL check is done before call. Drop unneeded conditional in __kobject_del().

---
 lib/kobject.c | 3 ---
 1 file changed, 3 deletions(-)
lib/kobject.c +0 -3
index bbbb067de8ecd..e02f1bb67c99f 100644
@@ -568,9 +568,6 @@ void kobject_del(struct kobject *kobj)
 {
 	struct kernfs_node *sd;
 
-	if (!kobj)
-		return;
-
 	sd = kobj->sd;
 	sysfs_remove_dir(kobj);
 	sysfs_put(sd);
Signature · 1 line
2.25.1
author maintainer
#16 lore raw

PATCH 13/15 spi: Fix memory leak on splited transfers

In reply to Sasha Levin

From: Gustav Wiklander <[email protected]>

[ Upstream commit b59a7ca15464c78ea1ba3b280cfc5ac5ece11ade ]

In the prepare_message callback the bus driver has the opportunity to split a transfer into smaller chunks. spi_map_msg is done after prepare_message.

Function spi_res_release releases the splited transfers in the message. Therefore spi_res_release should be called after spi_map_msg.

The previous try at this was commit c9ba7a16d0f1 which released the splited transfers after spi_finalize_current_message had been called. This introduced a race since the message struct could be out of scope because the spi_sync call got completed.

Fixes this leak on spi bus driver spi-bcm2835.c when transfer size is greater than 65532:

Kmemleak:

sg_alloc_table+0x28/0xc8
spi_map_buf+0xa4/0x300
__spi_pump_messages+0x370/0x748
__spi_sync+0x1d4/0x270
spi_sync+0x34/0x58
spi_test_execute_msg+0x60/0x340 [spi_loopback_test]
spi_test_run_iter+0x548/0x578 [spi_loopback_test]
spi_test_run_test+0x94/0x140 [spi_loopback_test]
spi_test_run_tests+0x150/0x180 [spi_loopback_test]
spi_loopback_test_probe+0x50/0xd0 [spi_loopback_test]
spi_drv_probe+0x84/0xe0
---
 drivers/spi/spi.c | 9 +++++++--
 1 file changed, 7 insertions(+), 2 deletions(-)
drivers/spi/spi.c +7 -2
index ab6a4f85bcde7..acc8eeed73f07 100644
@@ -1125,8 +1125,6 @@ static int spi_transfer_one_message(struct spi_controller *ctlr,
 	if (msg->status && ctlr->handle_err)
 		ctlr->handle_err(ctlr, msg);
 
-	spi_res_release(ctlr, msg);
-
 	spi_finalize_current_message(ctlr);
 
 	return ret;
@@ -1384,6 +1382,13 @@ void spi_finalize_current_message(struct spi_controller *ctlr)
 
 	spi_unmap_msg(ctlr, mesg);
 
+	/* In the prepare_messages callback the spi bus has the opportunity to
+	 * split a transfer to smaller chunks.
+	 * Release splited transfers here since spi_map_msg is done on the
+	 * splited transfers.
+	 */
+	spi_res_release(ctlr, mesg);
+
 	if (ctlr->cur_msg_prepared && ctlr->unprepare_message) {
 		ret = ctlr->unprepare_message(ctlr, mesg);
 		if (ret) {
Signature · 1 line
2.25.1
author maintainer
#8 lore raw

PATCH 7/15 SUNRPC: stop printk reading past end of string

In reply to Sasha Levin

From: "J. Bruce Fields" <[email protected]>

[ Upstream commit 8c6b6c793ed32b8f9770ebcdf1ba99af423c303b ]

Since p points at raw xdr data, there's no guarantee that it's NULL terminated, so we should give a length. And probably escape any special characters too.

---
 net/sunrpc/rpcb_clnt.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
net/sunrpc/rpcb_clnt.c +2 -2
index da21efac80f4a..7c0b1feb36299 100644
@@ -981,8 +981,8 @@ static int rpcb_dec_getaddr(struct rpc_rqst *req, struct xdr_stream *xdr,
 	p = xdr_inline_decode(xdr, len);
 	if (unlikely(p == NULL))
 		goto out_fail;
-	dprintk("RPC: %5u RPCB_%s reply: %s\n", req->rq_task->tk_pid,
-			req->rq_task->tk_msg.rpc_proc->p_name, (char *)p);
+	dprintk("RPC: %5u RPCB_%s reply: %*pE\n", req->rq_task->tk_pid,
+			req->rq_task->tk_msg.rpc_proc->p_name, len, (char *)p);
 
 	if (rpc_uaddr2sockaddr(req->rq_xprt->xprt_net, (char *)p, len,
 				sap, sizeof(address)) == 0)
Signature · 1 line
2.25.1
author maintainer
#7 lore raw

PATCH 6/15 spi: spi-loopback-test: Fix out-of-bounds read

In reply to Sasha Levin

From: Vincent Whitchurch <[email protected]>

[ Upstream commit 837ba18dfcd4db21ad58107c65bfe89753aa56d7 ]

The "tx/rx-transfer - crossing PAGE_SIZE" test always fails when len=131071 and rx_offset >= 5:

spi-loopback-test spi0.0: Running test tx/rx-transfer - crossing PAGE_SIZE
...
with iteration values: len = 131071, tx_off = 0, rx_off = 3 with iteration values: len = 131071, tx_off = 0, rx_off = 4 with iteration values: len = 131071, tx_off = 0, rx_off = 5
loopback strangeness - rx changed outside of allowed range at: ...a4321000
spi_msg@ffffffd5a4157690

     frame_length:  131071
     actual_length: 131071
     spi_transfer@ffffffd5a41576f8
       len:    131071
       tx_buf: ffffffd5a4340ffc

Note that rx_offset > 3 can only occur if the SPI controller driver sets ->dma_alignment to a higher value than 4, so most SPI controller drivers are not affect.

The allocated Rx buffer is of size SPI_TEST_MAX_SIZE_PLUS, which is 132 KiB (assuming 4 KiB pages). This test uses an initial offset into the rx_buf of PAGE_SIZE - 4, and a len of 131071, so the range expected to be written in this transfer ends at (4096 - 4) + 5 + 131071 == 132 KiB, which is also the end of the allocated buffer. But the code which verifies the content of the buffer reads a byte beyond the allocated buffer and spuriously fails because this out-of-bounds read doesn't return the expected value.

Fix this by using ITERATE_LEN instead of ITERATE_MAX_LEN to avoid testing sizes which cause out-of-bounds reads.

---
 drivers/spi/spi-loopback-test.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/spi/spi-loopback-test.c +1 -1
index bed7403bb6b3a..b9a7117b6dce3 100644
@@ -99,7 +99,7 @@ static struct spi_test spi_tests[] = {
 	{
 		.description	= "tx/rx-transfer - crossing PAGE_SIZE",
 		.fill_option	= FILL_COUNT_8,
-		.iterate_len    = { ITERATE_MAX_LEN },
+		.iterate_len    = { ITERATE_LEN },
 		.iterate_tx_align = ITERATE_ALIGN,
 		.iterate_rx_align = ITERATE_ALIGN,
 		.transfer_count = 1,
Signature · 1 line
2.25.1
author maintainer
#4 lore raw

PATCH 5/15 regulator: pwm: Fix machine constraints application

In reply to Sasha Levin

From: Vincent Whitchurch <[email protected]>

[ Upstream commit 59ae97a7a9e1499c2070e29841d1c4be4ae2994a ]

If the zero duty cycle doesn't correspond to any voltage in the voltage table, the PWM regulator returns an -EINVAL from get_voltage_sel() which results in the core erroring out with a "failed to get the current voltage" and ending up not applying the machine constraints.

Instead, return -ENOTRECOVERABLE which makes the core set the voltage since it's at an unknown value.

For example, with this device tree:

	fooregulator {
		compatible = "pwm-regulator";
		pwms = <&foopwm 0 100000>;
		regulator-min-microvolt = <2250000>;
		regulator-max-microvolt = <2250000>;
		regulator-name = "fooregulator";
		regulator-always-on;
		regulator-boot-on;
		voltage-table = <2250000 30>;
	};

Before this patch:

fooregulator: failed to get the current voltage(-22)

After this patch:

fooregulator: Setting 2250000-2250000uV
fooregulator: 2250 mV

---
 drivers/regulator/pwm-regulator.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/regulator/pwm-regulator.c +1 -1
index a2fd140eff81a..34f3b9778ffa1 100644
@@ -285,7 +285,7 @@ static int pwm_regulator_init_table(struct platform_device *pdev,
 		return ret;
 	}
 
-	drvdata->state			= -EINVAL;
+	drvdata->state			= -ENOTRECOVERABLE;
 	drvdata->duty_cycle_table	= duty_cycle_table;
 	memcpy(&drvdata->ops, &pwm_regulator_voltage_table_ops,
 	       sizeof(drvdata->ops));
Signature · 1 line
2.25.1
author maintainer
#3 lore raw

PATCH 4/15 scsi: lpfc: Fix FLOGI/PLOGI receive race condition in pt2pt discovery

In reply to Sasha Levin

From: James Smart <[email protected]>

[ Upstream commit 7b08e89f98cee9907895fabb64cf437bc505ce9a ]

The driver is unable to successfully login with remote device. During pt2pt login, the driver completes its FLOGI request with the remote device having WWN precedence. The remote device issues its own (delayed) FLOGI after accepting the driver's and, upon transmitting the FLOGI, immediately recognizes it has already processed the driver's FLOGI thus it transitions to sending a PLOGI before waiting for an ACC to its FLOGI.

In the driver, the FLOGI is received and an ACC sent, followed by the PLOGI being received and an ACC sent. The issue is that the PLOGI reception occurs before the response from the adapter from the FLOGI ACC is received. Processing of the PLOGI sets state flags to perform the REG_RPI mailbox command and proceed with the rest of discovery on the port. The same completion routine used by both FLOGI and PLOGI is generic in nature. One of the things it does is clear flags, and those flags happen to drive the rest of discovery. So what happened was the PLOGI processing set the flags, the FLOGI ACC completion cleared them, thus when the PLOGI ACC completes it doesn't see the flags and stops.

Fix by modifying the generic completion routine to not clear the rest of discovery flag (NLP_ACC_REGLOGIN) unless the completion is also associated with performing a mailbox command as part of its handling. For things such as FLOGI ACC, there isn't a subsequent action to perform with the adapter, thus there is no mailbox cmd ptr. PLOGI ACC though will perform REG_RPI upon completion, thus there is a mailbox cmd ptr.

---
 drivers/scsi/lpfc/lpfc_els.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)
drivers/scsi/lpfc/lpfc_els.c +3 -1
index db1111f7e85ae..566e8d07cb058 100644
@@ -4104,7 +4104,9 @@ lpfc_cmpl_els_rsp(struct lpfc_hba *phba, struct lpfc_iocbq *cmdiocb,
 out:
 	if (ndlp && NLP_CHK_NODE_ACT(ndlp) && shost) {
 		spin_lock_irq(shost->host_lock);
-		ndlp->nlp_flag &= ~(NLP_ACC_REGLOGIN | NLP_RM_DFLT_RPI);
+		if (mbox)
+			ndlp->nlp_flag &= ~NLP_ACC_REGLOGIN;
+		ndlp->nlp_flag &= ~NLP_RM_DFLT_RPI;
 		spin_unlock_irq(shost->host_lock);
 
 		/* If the node is not being used by another discovery thread,
Signature · 1 line
2.25.1

Sasha Levin

author maintainer
#0 lore raw

PATCH 1/15 NFSv4.1 handle ERR_DELAY error reclaiming locking state on delegation recall

From: Olga Kornievskaia <[email protected]>

[ Upstream commit 3d7a9520f0c3e6a68b6de8c5812fc8b6d7a52626 ]

A client should be able to handle getting an ERR_DELAY error while doing a LOCK call to reclaim state due to delegation being recalled. This is a transient error that can happen due to server moving its volumes and invalidating its file location cache and upon reference to it during the LOCK call needing to do an expensive lookup (leading to an ERR_DELAY error on a PUTFH).

---
 fs/nfs/nfs4proc.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)
fs/nfs/nfs4proc.c +6 -1
index 4cfb84119e017..997b731ee19ab 100644
@@ -6687,7 +6687,12 @@ int nfs4_lock_delegation_recall(struct file_lock *fl, struct nfs4_state *state,
 	err = nfs4_set_lock_state(state, fl);
 	if (err != 0)
 		return err;
-	err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
+	do {
+		err = _nfs4_do_setlk(state, F_SETLK, fl, NFS_LOCK_NEW);
+		if (err != -NFS4ERR_DELAY)
+			break;
+		ssleep(1);
+	} while (err == -NFS4ERR_DELAY);
 	return nfs4_handle_delegation_recall_error(server, state, stateid, fl, err);
 }
 
Signature · 1 line
2.25.1
author maintainer
#2 lore raw

PATCH 3/15 scsi: libfc: Fix for double free()

From: Javed Hasan <[email protected]>

[ Upstream commit 5a5b80f98534416b3b253859897e2ba1dc241e70 ]

Fix for '&fp->skb' double free.

Link:
https://lore.kernel.org/r/[email protected]

---
 drivers/scsi/libfc/fc_disc.c | 2 --
 1 file changed, 2 deletions(-)
drivers/scsi/libfc/fc_disc.c +0 -2
index 62f83cc151b22..0db0d0ad9f966 100644
@@ -658,8 +658,6 @@ static void fc_disc_gpn_id_resp(struct fc_seq *sp, struct fc_frame *fp,
 	fc_frame_free(fp);
 out:
 	kref_put(&rdata->kref, fc_rport_destroy);
-	if (!IS_ERR(fp))
-		fc_frame_free(fp);
 }
 
 /**
Signature · 1 line
2.25.1
author maintainer
#1 lore raw

PATCH 2/15 scsi: pm8001: Fix memleak in pm8001_exec_internal_task_abort

In reply to Sasha Levin

From: Dinghao Liu <[email protected]>

[ Upstream commit ea403fde7552bd61bad6ea45e3feb99db77cb31e ]

When pm8001_tag_alloc() fails, task should be freed just like it is done in the subsequent error paths.

---
 drivers/scsi/pm8001/pm8001_sas.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
drivers/scsi/pm8001/pm8001_sas.c +1 -1
index e64a13f0bce17..61a2da30f94b7 100644
@@ -795,7 +795,7 @@ pm8001_exec_internal_task_abort(struct pm8001_hba_info *pm8001_ha,
 
 		res = pm8001_tag_alloc(pm8001_ha, &ccb_tag);
 		if (res)
-			return res;
+			goto ex_err;
 		ccb = &pm8001_ha->ccb_info[ccb_tag];
 		ccb->device = pm8001_dev;
 		ccb->ccb_tag = ccb_tag;
Signature · 1 line
2.25.1
#5 lore raw

Re: [PATCH AUTOSEL 4.14 05/15] regulator: pwm: Fix machine constraints application

In reply to Sasha Levin
Quoted 11 lines · Sasha Levin

From: Vincent Whitchurch <[email protected]>

[ Upstream commit 59ae97a7a9e1499c2070e29841d1c4be4ae2994a ]

If the zero duty cycle doesn't correspond to any voltage in the voltage table, the PWM regulator returns an -EINVAL from get_voltage_sel() which results in the core erroring out with a "failed to get the current voltage" and ending up not applying the machine constraints.

Instead, return -ENOTRECOVERABLE which makes the core set the voltage since it's at an unknown value.

For this patch to work it needs 84b3a7c9c6befe5ab4d49070fe7 ("regulator:
core: Allow for regulators that can't be read at bootup") which was merged in v4.18. Without that this patch is not going to have any effect so it probably shouldn't be backported to older kernels.

author maintainer
#6 lore raw

Re: [PATCH AUTOSEL 4.14 05/15] regulator: pwm: Fix machine constraints application

Quoted 10 lines · Vincent Whitchurch
Quoted 9 lines · Sasha Levin

From: Vincent Whitchurch <[email protected]>

[ Upstream commit 59ae97a7a9e1499c2070e29841d1c4be4ae2994a ]

If the zero duty cycle doesn't correspond to any voltage in the voltage table, the PWM regulator returns an -EINVAL from get_voltage_sel() which results in the core erroring out with a "failed to get the current voltage" and ending up not applying the machine constraints.

Quoted 7 lines
Quoted 2 lines

Instead, return -ENOTRECOVERABLE which makes the core set the voltage since it's at an unknown value.

For this patch to work it needs 84b3a7c9c6befe5ab4d49070fe7 ("regulator:
core: Allow for regulators that can't be read at bootup") which was merged in v4.18. Without that this patch is not going to have any effect so it probably shouldn't be backported to older kernels.

Dropped for those older kernels, thanks!

Signature · 2 lines
Thanks,
Sasha
author maintainer
#12 lore raw

Re: [PATCH AUTOSEL 4.14 09/15] kobject: Drop unneeded conditional in __kobject_del()

In reply to Greg Kroah-Hartman
Quoted 39 lines · Greg Kroah-Hartman
Quoted 32 lines · Sasha Levin

From: Andy Shevchenko <[email protected]>

[ Upstream commit 07ecc6693f9157cf293da5d165c73fb28fd69bf4 ]

__kobject_del() is called from two places, in one where kobj is dereferenced before and thus can't be NULL, and in the other the NULL check is done before call. Drop unneeded conditional in __kobject_del().

---
 lib/kobject.c | 3 ---
 1 file changed, 3 deletions(-)
lib/kobject.c +0 -4
index bbbb067de8ecd..e02f1bb67c99f 100644
@@ -568,9 +568,6 @@ void kobject_del(struct kobject *kobj)
 {
 	struct kernfs_node *sd;
 
-	if (!kobj)
-		return;
-
 	sd = kobj->sd;
 	sysfs_remove_dir(kobj);
 	sysfs_put(sd);
--
Quoted 2 lines

2.25.1

Ouch, no, this patch should not be backported ANYWHERE or bad things will happen.

Quoted 1 line

Please drop from all AUTOSEL trees.

Dropped, thanks!

Signature · 2 lines
Thanks,
Sasha