Files
kernel/drivers/firmware/arm_scmi/shmem.c
Cristian Marussi 9b5e1b93c8 firmware: arm_scmi: Check mailbox/SMT channel for consistency
commit 437a310b22 upstream.

On reception of a completion interrupt the shared memory area is accessed
to retrieve the message header at first and then, if the message sequence
number identifies a transaction which is still pending, the related
payload is fetched too.

When an SCMI command times out the channel ownership remains with the
platform until eventually a late reply is received and, as a consequence,
any further transmission attempt remains pending, waiting for the channel
to be relinquished by the platform.

Once that late reply is received the channel ownership is given back
to the agent and any pending request is then allowed to proceed and
overwrite the SMT area of the just delivered late reply; then the wait
for the reply to the new request starts.

It has been observed that the spurious IRQ related to the late reply can
be wrongly associated with the freshly enqueued request: when that happens
the SCMI stack in-flight lookup procedure is fooled by the fact that the
message header now present in the SMT area is related to the new pending
transaction, even though the real reply has still to arrive.

This race-condition on the A2P channel can be detected by looking at the
channel status bits: a genuine reply from the platform will have set the
channel free bit before triggering the completion IRQ.

Add a consistency check to validate such condition in the A2P ISR.

Reported-by: Xinglong Yang <xinglong.yang@cixtech.com>
Closes: https://lore.kernel.org/all/PUZPR06MB54981E6FA00D82BFDBB864FBF08DA@PUZPR06MB5498.apcprd06.prod.outlook.com/
Fixes: 5c8a47a5a9 ("firmware: arm_scmi: Make scmi core independent of the transport type")
Cc: stable@vger.kernel.org # 5.15+
Signed-off-by: Cristian Marussi <cristian.marussi@arm.com>
Tested-by: Xinglong Yang <xinglong.yang@cixtech.com>
Link: https://lore.kernel.org/r/20231220172112.763539-1-cristian.marussi@arm.com
Signed-off-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2024-01-31 16:19:07 -08:00

131 lines
3.8 KiB
C

// SPDX-License-Identifier: GPL-2.0
/*
* For transport using shared mem structure.
*
* Copyright (C) 2019 ARM Ltd.
*/
#include <linux/ktime.h>
#include <linux/io.h>
#include <linux/processor.h>
#include <linux/types.h>
#include <asm-generic/bug.h>
#include "common.h"
/*
* SCMI specification requires all parameters, message headers, return
* arguments or any protocol data to be expressed in little endian
* format only.
*/
struct scmi_shared_mem {
__le32 reserved;
__le32 channel_status;
#define SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR BIT(1)
#define SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE BIT(0)
__le32 reserved1[2];
__le32 flags;
#define SCMI_SHMEM_FLAG_INTR_ENABLED BIT(0)
__le32 length;
__le32 msg_header;
u8 msg_payload[];
};
void shmem_tx_prepare(struct scmi_shared_mem __iomem *shmem,
struct scmi_xfer *xfer, struct scmi_chan_info *cinfo)
{
ktime_t stop;
/*
* Ideally channel must be free by now unless OS timeout last
* request and platform continued to process the same, wait
* until it releases the shared memory, otherwise we may endup
* overwriting its response with new message payload or vice-versa.
* Giving up anyway after twice the expected channel timeout so as
* not to bail-out on intermittent issues where the platform is
* occasionally a bit slower to answer.
*
* Note that after a timeout is detected we bail-out and carry on but
* the transport functionality is probably permanently compromised:
* this is just to ease debugging and avoid complete hangs on boot
* due to a misbehaving SCMI firmware.
*/
stop = ktime_add_ms(ktime_get(), 2 * cinfo->rx_timeout_ms);
spin_until_cond((ioread32(&shmem->channel_status) &
SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE) ||
ktime_after(ktime_get(), stop));
if (!(ioread32(&shmem->channel_status) &
SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE)) {
WARN_ON_ONCE(1);
dev_err(cinfo->dev,
"Timeout waiting for a free TX channel !\n");
return;
}
/* Mark channel busy + clear error */
iowrite32(0x0, &shmem->channel_status);
iowrite32(xfer->hdr.poll_completion ? 0 : SCMI_SHMEM_FLAG_INTR_ENABLED,
&shmem->flags);
iowrite32(sizeof(shmem->msg_header) + xfer->tx.len, &shmem->length);
iowrite32(pack_scmi_header(&xfer->hdr), &shmem->msg_header);
if (xfer->tx.buf)
memcpy_toio(shmem->msg_payload, xfer->tx.buf, xfer->tx.len);
}
u32 shmem_read_header(struct scmi_shared_mem __iomem *shmem)
{
return ioread32(&shmem->msg_header);
}
void shmem_fetch_response(struct scmi_shared_mem __iomem *shmem,
struct scmi_xfer *xfer)
{
size_t len = ioread32(&shmem->length);
xfer->hdr.status = ioread32(shmem->msg_payload);
/* Skip the length of header and status in shmem area i.e 8 bytes */
xfer->rx.len = min_t(size_t, xfer->rx.len, len > 8 ? len - 8 : 0);
/* Take a copy to the rx buffer.. */
memcpy_fromio(xfer->rx.buf, shmem->msg_payload + 4, xfer->rx.len);
}
void shmem_fetch_notification(struct scmi_shared_mem __iomem *shmem,
size_t max_len, struct scmi_xfer *xfer)
{
size_t len = ioread32(&shmem->length);
/* Skip only the length of header in shmem area i.e 4 bytes */
xfer->rx.len = min_t(size_t, max_len, len > 4 ? len - 4 : 0);
/* Take a copy to the rx buffer.. */
memcpy_fromio(xfer->rx.buf, shmem->msg_payload, xfer->rx.len);
}
void shmem_clear_channel(struct scmi_shared_mem __iomem *shmem)
{
iowrite32(SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE, &shmem->channel_status);
}
bool shmem_poll_done(struct scmi_shared_mem __iomem *shmem,
struct scmi_xfer *xfer)
{
u16 xfer_id;
xfer_id = MSG_XTRACT_TOKEN(ioread32(&shmem->msg_header));
if (xfer->hdr.seq != xfer_id)
return false;
return ioread32(&shmem->channel_status) &
(SCMI_SHMEM_CHAN_STAT_CHANNEL_ERROR |
SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE);
}
bool shmem_channel_free(struct scmi_shared_mem __iomem *shmem)
{
return (ioread32(&shmem->channel_status) &
SCMI_SHMEM_CHAN_STAT_CHANNEL_FREE);
}