| 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
| ||
|
Message-ID: <c3c2aa0e-845f-f141-1737-4253e405ab66@arm.com>
Date: Mon, 3 Apr 2023 12:11:59 +0100
From: Suzuki K Poulose <suzuki.poulose@....com>
To: James Clark <james.clark@....com>, coresight@...ts.linaro.org,
quic_jinlmao@...cinc.com, mike.leach@...aro.org
Cc: Mathieu Poirier <mathieu.poirier@...aro.org>,
Leo Yan <leo.yan@...aro.org>,
Alexander Shishkin <alexander.shishkin@...ux.intel.com>,
Maxime Coquelin <mcoquelin.stm32@...il.com>,
Alexandre Torgue <alexandre.torgue@...s.st.com>,
linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
linux-stm32@...md-mailman.stormreply.com
Subject: Re: [PATCH v3 07/13] coresight: Store pointers to connections rather
than an array of them
On 03/04/2023 11:16, James Clark wrote:
>
>
> On 03/04/2023 09:46, Suzuki K Poulose wrote:
>> On 29/03/2023 12:53, James Clark wrote:
>>> This will allow the same connection object to be referenced via the
>>> input connection list in a later commit rather than duplicating them.
>>>
>>> Signed-off-by: James Clark <james.clark@....com>
>>> ---
>>> drivers/hwtracing/coresight/coresight-core.c | 45 ++++++++++---------
>>> .../hwtracing/coresight/coresight-platform.c | 8 ++--
>>> drivers/hwtracing/coresight/coresight-priv.h | 1 +
>>> .../hwtracing/coresight/coresight-tmc-etr.c | 2 +-
>>> include/linux/coresight.h | 2 +-
>>> 5 files changed, 31 insertions(+), 27 deletions(-)
>>>
>>> diff --git a/drivers/hwtracing/coresight/coresight-core.c
>>> b/drivers/hwtracing/coresight/coresight-core.c
>>> index 12cbb68e8e1c..389f6203c8f0 100644
>>> --- a/drivers/hwtracing/coresight/coresight-core.c
>>> +++ b/drivers/hwtracing/coresight/coresight-core.c
>>> @@ -119,7 +119,7 @@ static int coresight_find_link_inport(struct
>>> coresight_device *csdev,
>>> struct coresight_connection *conn;
>>> for (i = 0; i < parent->pdata->nr_outconns; i++) {
>>> - conn = &parent->pdata->out_conns[i];
>>> + conn = parent->pdata->out_conns[i];
>>> if (conn->dest_dev == csdev)
>>> return conn->dest_port;
>>> }
>>> @@ -137,7 +137,7 @@ static int coresight_find_link_outport(struct
>>> coresight_device *csdev,
>>> struct coresight_connection *conn;
>>> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>>> - conn = &csdev->pdata->out_conns[i];
>>> + conn = csdev->pdata->out_conns[i];
>>> if (conn->dest_dev == child)
>>> return conn->src_port;
>>> }
>>> @@ -606,7 +606,7 @@ coresight_find_enabled_sink(struct
>>> coresight_device *csdev)
>>> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>>> struct coresight_device *child_dev;
>>> - child_dev = csdev->pdata->out_conns[i].dest_dev;
>>> + child_dev = csdev->pdata->out_conns[i]->dest_dev;
>>> if (child_dev)
>>> sink = coresight_find_enabled_sink(child_dev);
>>> if (sink)
>>> @@ -722,7 +722,7 @@ static int coresight_grab_device(struct
>>> coresight_device *csdev)
>>> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>>> struct coresight_device *child;
>>> - child = csdev->pdata->out_conns[i].dest_dev;
>>> + child = csdev->pdata->out_conns[i]->dest_dev;
>>> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
>>> if (!coresight_get_ref(child))
>>> goto err;
>>> @@ -733,7 +733,7 @@ static int coresight_grab_device(struct
>>> coresight_device *csdev)
>>> for (i--; i >= 0; i--) {
>>> struct coresight_device *child;
>>> - child = csdev->pdata->out_conns[i].dest_dev;
>>> + child = csdev->pdata->out_conns[i]->dest_dev;
>>> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
>>> coresight_put_ref(child);
>>> }
>>> @@ -752,7 +752,7 @@ static void coresight_drop_device(struct
>>> coresight_device *csdev)
>>> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>>> struct coresight_device *child;
>>> - child = csdev->pdata->out_conns[i].dest_dev;
>>> + child = csdev->pdata->out_conns[i]->dest_dev;
>>> if (child && child->type == CORESIGHT_DEV_TYPE_HELPER)
>>> coresight_put_ref(child);
>>> }
>>> @@ -794,7 +794,7 @@ static int _coresight_build_path(struct
>>> coresight_device *csdev,
>>> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>>> struct coresight_device *child_dev;
>>> - child_dev = csdev->pdata->out_conns[i].dest_dev;
>>> + child_dev = csdev->pdata->out_conns[i]->dest_dev;
>>> if (child_dev &&
>>> _coresight_build_path(child_dev, sink, path) == 0) {
>>> found = true;
>>> @@ -964,7 +964,7 @@ coresight_find_sink(struct coresight_device
>>> *csdev, int *depth)
>>> struct coresight_device *child_dev, *sink = NULL;
>>> int child_depth = curr_depth;
>>> - child_dev = csdev->pdata->out_conns[i].dest_dev;
>>> + child_dev = csdev->pdata->out_conns[i]->dest_dev;
>>> if (child_dev)
>>> sink = coresight_find_sink(child_dev, &child_depth);
>>> @@ -1334,7 +1334,7 @@ static int coresight_orphan_match(struct
>>> device *dev, void *data)
>>> * an orphan connection whose name matches @csdev, link it.
>>> */
>>> for (i = 0; i < i_csdev->pdata->nr_outconns; i++) {
>>> - conn = &i_csdev->pdata->out_conns[i];
>>> + conn = i_csdev->pdata->out_conns[i];
>>> /* We have found at least one orphan connection */
>>> if (conn->dest_dev == NULL) {
>>> @@ -1372,7 +1372,7 @@ static int coresight_fixup_device_conns(struct
>>> coresight_device *csdev)
>>> int i, ret = 0;
>>> for (i = 0; i < csdev->pdata->nr_outconns; i++) {
>>> - struct coresight_connection *conn = &csdev->pdata->out_conns[i];
>>> + struct coresight_connection *conn = csdev->pdata->out_conns[i];
>>> conn->dest_dev =
>>> coresight_find_csdev_by_fwnode(conn->dest_fwnode);
>>> @@ -1406,15 +1406,12 @@ static int coresight_remove_match(struct
>>> device *dev, void *data)
>>> * a connection whose name matches @csdev, remove it.
>>> */
>>> for (i = 0; i < iterator->pdata->nr_outconns; i++) {
>>> - conn = &iterator->pdata->out_conns[i];
>>> + conn = iterator->pdata->out_conns[i];
>>> - if (conn->dest_dev == NULL)
>>> - continue;
>>> -
>>> - if (csdev->dev.fwnode == conn->dest_fwnode) {
>>> + /* Child_dev being set signifies that the links were made */
>>> + if (csdev->dev.fwnode == conn->dest_fwnode && conn->dest_dev) {
>>> iterator->orphan = true;
>>> coresight_remove_links(iterator, conn);
>>> -
>>> conn->dest_dev = NULL;
>>> /* No need to continue */
>>> break;
>>> @@ -1534,21 +1531,25 @@ void coresight_write64(struct coresight_device
>>> *csdev, u64 val, u32 offset)
>>> * to the output port of this device.
>>> */
>>> void coresight_release_platform_data(struct coresight_device *csdev,
>>> + struct device *dev,
>>> struct coresight_platform_data *pdata)
>>> {
>>> int i;
>>> - struct coresight_connection *conns = pdata->out_conns;
>>> + struct coresight_connection **conns = pdata->out_conns;
>>> for (i = 0; i < pdata->nr_outconns; i++) {
>>> /* If we have made the links, remove them now */
>>> - if (csdev && conns[i].dest_dev)
>>> - coresight_remove_links(csdev, &conns[i]);
>>> + if (csdev && conns[i]->dest_dev)
>>> + coresight_remove_links(csdev, conns[i]);
>>> /*
>>> * Drop the refcount and clear the handle as this device
>>> * is going away
>>> */
>>> - fwnode_handle_put(conns[i].dest_fwnode);
>>> + fwnode_handle_put(conns[i]->dest_fwnode);
>>> + devm_kfree(dev, conns[i]);
>>> }
>>> + devm_kfree(dev, pdata->out_conns);
>>> + devm_kfree(dev, pdata);
>>
>> Is there any particular reason, why we need to do this ? This should
>> be done automatically at device teardown, which is bound to happen
>> right after this call ?
>>
>
> I think they're not actually freed because connections are added before
> the coresight device is created using the 'parent' of the device which
> never goes away:
>
> struct coresight_device *coresight_register(struct coresight_desc *desc) {
> ...
> csdev->dev.parent = desc->dev;
>
> Another reason is that's is a good way to catch errors. There are so
> many different devices floating around that without an explicit free,
> it's very easy to pass the wrong one and be inconsistent. With the free
> it gives you a warning if it's mismatched.
>
> But either way, now that we know that they're not freed maybe we should
> fix that isntead? Although apart from doing a full copy of pdata into
> the coresight device I'm not sure how it can be done. Then I suppose we
> could remove these frees, although I still think the error catching is
> good. It's also not strictly related to this fix so maybe it can be a
> second fix.
Good point. May be that explains the leak we have been seeing in our
long running stress tests ? May be should extend this to the other
allocations too (e.g., drvdata for all devices ?) This can be done
separately. Ideally any allocation that we do, after the coresight
device has been created, must be allocated on the child device
and make use of what the devm_ provides us with.
Suzuki
>
>>
>>> if (csdev)
>>> coresight_remove_conns_sysfs_group(csdev);
>>> }
>>> @@ -1665,7 +1666,7 @@ struct coresight_device
>>> *coresight_register(struct coresight_desc *desc)
>>> err_out:
>>> /* Cleanup the connection information */
>>> - coresight_release_platform_data(NULL, desc->pdata);
>>> + coresight_release_platform_data(NULL, desc->dev, desc->pdata);
>>> return ERR_PTR(ret);
>>> }
>>> EXPORT_SYMBOL_GPL(coresight_register);
>>> @@ -1678,7 +1679,7 @@ void coresight_unregister(struct
>>> coresight_device *csdev)
>>> cti_assoc_ops->remove(csdev);
>>> coresight_remove_conns(csdev);
>>> coresight_clear_default_sink(csdev);
>>> - coresight_release_platform_data(csdev, csdev->pdata);
>>> + coresight_release_platform_data(csdev, csdev->dev.parent,
>>> csdev->pdata);
>>> device_unregister(&csdev->dev);
>>> }
>>> EXPORT_SYMBOL_GPL(coresight_unregister);
>>> diff --git a/drivers/hwtracing/coresight/coresight-platform.c
>>> b/drivers/hwtracing/coresight/coresight-platform.c
>>> index 80ed2e74620b..bea8f1ba309a 100644
>>> --- a/drivers/hwtracing/coresight/coresight-platform.c
>>> +++ b/drivers/hwtracing/coresight/coresight-platform.c
>>> @@ -36,7 +36,7 @@ int coresight_add_out_conn(struct device *dev,
>>> * Warn on any existing duplicate output port.
>>> */
>>> for (i = 0; i < pdata->nr_outconns; ++i) {
>>> - conn = &pdata->out_conns[i];
>>> + conn = pdata->out_conns[i];
>>> /* Output == -1 means ignore the port for example for
>>> helpers */
>>> if (conn->src_port != -1 &&
>>> conn->src_port == new_conn->src_port) {
>>> @@ -53,7 +53,9 @@ int coresight_add_out_conn(struct device *dev,
>>> if (!pdata->out_conns)
>>> return -ENOMEM;
>>> - pdata->out_conns[pdata->nr_outconns - 1] = *new_conn;
>>> + pdata->out_conns[pdata->nr_outconns - 1] = devm_kmalloc(
>>> + dev, sizeof(struct coresight_connection), GFP_KERNEL);
>>> + *pdata->out_conns[pdata->nr_outconns - 1] = *new_conn;
>>> return 0;
>>> }
>>> EXPORT_SYMBOL_GPL(coresight_add_out_conn);
>>> @@ -859,7 +861,7 @@ coresight_get_platform_data(struct device *dev)
>>> error:
>>> if (!IS_ERR_OR_NULL(pdata))
>>> /* Cleanup the connection information */
>>> - coresight_release_platform_data(NULL, pdata);
>>> + coresight_release_platform_data(NULL, dev, pdata);
>>> return ERR_PTR(ret);
>>> }
>>> EXPORT_SYMBOL_GPL(coresight_get_platform_data);
>>> diff --git a/drivers/hwtracing/coresight/coresight-priv.h
>>> b/drivers/hwtracing/coresight/coresight-priv.h
>>> index 788ff19c60f6..65ae6d161c57 100644
>>> --- a/drivers/hwtracing/coresight/coresight-priv.h
>>> +++ b/drivers/hwtracing/coresight/coresight-priv.h
>>> @@ -207,6 +207,7 @@ static inline void *coresight_get_uci_data(const
>>> struct amba_id *id)
>>> }
>>> void coresight_release_platform_data(struct coresight_device *csdev,
>>> + struct device *dev,
>>> struct coresight_platform_data *pdata);
>>> struct coresight_device *
>>> coresight_find_csdev_by_fwnode(struct fwnode_handle *r_fwnode);
>>> diff --git a/drivers/hwtracing/coresight/coresight-tmc-etr.c
>>> b/drivers/hwtracing/coresight/coresight-tmc-etr.c
>>> index 61234cb8052a..1bbe5410a23d 100644
>>> --- a/drivers/hwtracing/coresight/coresight-tmc-etr.c
>>> +++ b/drivers/hwtracing/coresight/coresight-tmc-etr.c
>>> @@ -782,7 +782,7 @@ tmc_etr_get_catu_device(struct tmc_drvdata *drvdata)
>>> return NULL;
>>> for (i = 0; i < etr->pdata->nr_outconns; i++) {
>>> - tmp = etr->pdata->out_conns[i].dest_dev;
>>> + tmp = etr->pdata->out_conns[i]->dest_dev;
>>> if (tmp && coresight_is_catu_device(tmp))
>>> return tmp;
>>> }
>>> diff --git a/include/linux/coresight.h b/include/linux/coresight.h
>>> index ccbc5eafcb6b..7197b07deede 100644
>>> --- a/include/linux/coresight.h
>>> +++ b/include/linux/coresight.h
>>> @@ -111,7 +111,7 @@ struct coresight_platform_data {
>>> int high_outport;
>>> int nr_inconns;
>>> int nr_outconns;
>>> - struct coresight_connection *out_conns;
>>> + struct coresight_connection **out_conns;
>>
>> minor nit: Do we need to update the comment too ?
>>
>
> Will do
>
>> Otherwise looks good to me
>> Suzuki
>>
>>> };
>>> /**
>>
Powered by blists - more mailing lists