| lists.openwall.net | lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC | |
|
Open Source and information security mailing list archives
| ||
|
Date: Wed, 2 Aug 2023 19:33:02 -0500
From: Jassi Brar <jassisinghbrar@...il.com>
To: Elliot Berman <quic_eberman@...cinc.com>
Cc: Alex Elder <elder@...aro.org>,
Srinivas Kandagatla <srinivas.kandagatla@...aro.org>,
Prakruthi Deepak Heragu <quic_pheragu@...cinc.com>,
Jonathan Corbet <corbet@....net>,
Murali Nalajala <quic_mnalajal@...cinc.com>,
Trilok Soni <quic_tsoni@...cinc.com>,
Srivatsa Vaddagiri <quic_svaddagi@...cinc.com>,
Carl van Schaik <quic_cvanscha@...cinc.com>,
Dmitry Baryshkov <dmitry.baryshkov@...aro.org>,
Bjorn Andersson <andersson@...nel.org>,
Konrad Dybcio <konrad.dybcio@...aro.org>,
Arnd Bergmann <arnd@...db.de>,
Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
Rob Herring <robh+dt@...nel.org>,
Krzysztof Kozlowski <krzysztof.kozlowski+dt@...aro.org>,
Conor Dooley <conor+dt@...nel.org>,
Bagas Sanjaya <bagasdotme@...il.com>,
Will Deacon <will@...nel.org>, Andy Gross <agross@...nel.org>,
Catalin Marinas <catalin.marinas@....com>,
linux-arm-msm@...r.kernel.org, devicetree@...r.kernel.org,
linux-kernel@...r.kernel.org, linux-doc@...r.kernel.org,
linux-arm-kernel@...ts.infradead.org
Subject: Re: [PATCH v14 06/25] mailbox: Add Gunyah message queue mailbox
On Tue, Jun 13, 2023 at 12:21 PM Elliot Berman <quic_eberman@...cinc.com> wrote:
......
> diff --git a/drivers/mailbox/gunyah-msgq.c b/drivers/mailbox/gunyah-msgq.c
> new file mode 100644
> index 0000000000000..7f777339278eb
> --- /dev/null
> +++ b/drivers/mailbox/gunyah-msgq.c
> @@ -0,0 +1,219 @@
> +// SPDX-License-Identifier: GPL-2.0-only
> +/*
> + * Copyright (c) 2022-2023 Qualcomm Innovation Center, Inc. All rights reserved.
> + */
> +
> +#include <linux/mailbox_controller.h>
> +#include <linux/module.h>
> +#include <linux/interrupt.h>
> +#include <linux/gunyah.h>
> +#include <linux/printk.h>
> +#include <linux/init.h>
> +#include <linux/slab.h>
> +#include <linux/wait.h>
> +
I believe some includes could be removed without issue.
>
> +#define mbox_chan_to_msgq(chan) (container_of(chan->mbox, struct gh_msgq, mbox))
> +
container_of need not be in brackets.
> +static irqreturn_t gh_msgq_rx_irq_handler(int irq, void *data)
> +{
> + struct gh_msgq *msgq = data;
> + struct gh_msgq_rx_data rx_data;
> + enum gh_error gh_error;
> + bool ready = true;
> +
Please limit the scope of rx_data and gh_error by moving them inside
the while() body.
> + while (ready) {
> + gh_error = gh_hypercall_msgq_recv(msgq->rx_ghrsc->capid,
> + &rx_data.data, sizeof(rx_data.data),
> + &rx_data.length, &ready);
> + if (gh_error != GH_ERROR_OK) {
> + if (gh_error != GH_ERROR_MSGQUEUE_EMPTY)
> + dev_warn(msgq->mbox.dev, "Failed to receive data: %d\n", gh_error);
> + break;
> + }
> + if (likely(gh_msgq_chan(msgq)->cl))
> + mbox_chan_received_data(gh_msgq_chan(msgq), &rx_data);
> + }
> +
> + return IRQ_HANDLED;
> +}
> +
> +static int gh_msgq_send_data(struct mbox_chan *chan, void *data)
> +{
> + struct gh_msgq *msgq = mbox_chan_to_msgq(chan);
> + struct gh_msgq_tx_data *msgq_data = data;
> + u64 tx_flags = 0;
> + enum gh_error gh_error;
> + bool ready;
> +
> + if (!msgq->tx_ghrsc)
> + return -EOPNOTSUPP;
> +
If we hit this error, the fix will still be in the upper layer.
So please drop the check and, if needed, add one in the client driver.
> + if (msgq_data->push)
> + tx_flags |= GH_HYPERCALL_MSGQ_TX_FLAGS_PUSH;
> +
> + gh_error = gh_hypercall_msgq_send(msgq->tx_ghrsc->capid, msgq_data->length, msgq_data->data,
> + tx_flags, &ready);
> +
> + /**
> + * unlikely because Linux tracks state of msgq and should not try to
> + * send message when msgq is full.
> + */
> + if (unlikely(gh_error == GH_ERROR_MSGQUEUE_FULL))
> + return -EAGAIN;
> +
If it is not expected to hit, please remove the check.
If there can be a 'race' like situation, still remove this and try to
find an appropriate place to avoid the race.
> + /**
> + * Propagate all other errors to client. If we return error to mailbox
> + * framework, then no other messages can be sent and nobody will know
> + * to retry this message.
> + */
> + msgq->last_ret = gh_error_remap(gh_error);
> +
> + /**
> + * This message was successfully sent, but message queue isn't ready to
> + * accept more messages because it's now full. Mailbox framework
> + * requires that we only report that message was transmitted when
> + * we're ready to transmit another message. We'll get that in the form
> + * of tx IRQ once the other side starts to drain the msgq.
> + */
> + if (gh_error == GH_ERROR_OK) {
> + if (!ready)
> + return 0;
> + } else {
> + dev_err(msgq->mbox.dev, "Failed to send data: %d (%d)\n", gh_error, msgq->last_ret);
> + }
> +
> + /**
> + * We can send more messages.
>
... until we can not (when the platform specific queue is full).
> Mailbox framework requires that tx done
> + * happens asynchronously to sending the message.
>
hence the mailbox api needs to track each transfer's stage.
> Gunyah message queues
> + * tell us right away on the hypercall return whether we can send more
> + * messages. To work around this, defer the txdone to a tasklet.
> + */
>
If not here, you would still have to put the 'defer' somewhere in the
upper layer.
So it is not exactly a "workaround".
> + tasklet_schedule(&msgq->txdone_tasklet);
> +
> + return 0;
> +}
> +
> diff --git a/include/linux/gunyah.h b/include/linux/gunyah.h
> index 01a6f202d037e..982e27d10d57f 100644
> --- a/include/linux/gunyah.h
> +++ b/include/linux/gunyah.h
> @@ -8,11 +8,68 @@
>
> #include <linux/bitfield.h>
> #include <linux/errno.h>
> +#include <linux/interrupt.h>
> #include <linux/limits.h>
> +#include <linux/mailbox_controller.h>
> +#include <linux/mailbox_client.h>
> #include <linux/types.h>
>
controller.h and client.h aren't supposed to be in the same header.
The Client and the Controller driver could include them respectively
before this header.
Cheers.
Powered by blists - more mailing lists