Skip to contentSkip to conversation

s390x: uv-host: Fixups and extensions part 1

PATCH Superseded Janosch Frank 2 participants 18 messages May 2, 2023 – Jun 2, 2023

Subsystems (2)

v3 of 4 Superseded Latest is v4 →

lore
Share & tools
lore
#0 lore raw

v3 cover s390x: uv-host: Fixups and extensions part 1

The uv-host test has a lot of historical growth problems which have largely been overlooked since running it is harder than running a KVM (guest 2) based test.

This series fixes up smaller problems but still leaves the test with fails when running create config base and variable storage tests. Those problems will either be fixed up with the second series or with a firmware fix since I'm unsure on which side of the os/fw fence the problem exists.

The series is based on my other series that introduces pv-ipl and pv-icpt. The memory allocation fix will be added to the new version of that series so all G1 tests are fixed.

v3:
- Re-based on the ipl/icpt series
- Added review-bys

v2:
- Added patch that exchanges sigp_retry with the smp variant
- Re-worked the create config test handling
- Minor fixups

Janosch Frank (9):
s390x: uv-host: Fix UV init test memory allocation
s390x: uv-host: Check for sufficient amount of memory
s390x: uv-host: Beautify code
s390x: uv-host: Add cpu number check
s390x: uv-host: Fix create guest variable storage prefix check
s390x: uv-host: Switch to smp_sigp
s390x: uv-host: Properly handle config creation errors
s390x: uv-host: Fence access checks when UV debug is enabled
s390x: uv-host: Add the test to unittests.conf

 lib/s390x/asm/uv.h  |   1 +
 s390x/unittests.cfg |   5 ++
 s390x/uv-host.c     | 137 ++++++++++++++++++++++++++++++++++----------
 3 files changed, 114 insertions(+), 29 deletions(-)
Signature · 1 line
2.34.1
#1 lore raw

PATCH v3 1/9 s390x: uv-host: Fix UV init test memory allocation

The init memory has to be above 2G and 1M aligned but we're currently aligning on 2G which means the allocations need a lot of unused memory.

Also the second block of memory was never actually used for the double init test since its address is never put into the uvcb.

Let's fix that.

---
 s390x/uv-host.c | 23 ++++++++++++++---------
 1 file changed, 14 insertions(+), 9 deletions(-)
s390x/uv-host.c +14 -9
index 33e6eec6..9dfaebd7 100644
@@ -500,14 +500,17 @@ static void test_config_create(void)
 static void test_init(void)
 {
 	int rc;
-	uint64_t mem;
+	uint64_t tmp;
 
-	/* Donated storage needs to be over 2GB */
-	mem = (uint64_t)memalign_pages_flags(SZ_1M, uvcb_qui.uv_base_stor_len, AREA_NORMAL);
+	/*
+	 * Donated storage needs to be over 2GB, AREA_NORMAL does that
+	 * on s390x.
+	 */
+	tmp = (uint64_t)memalign_pages_flags(SZ_1M, uvcb_qui.uv_base_stor_len, AREA_NORMAL);
 
 	uvcb_init.header.len = sizeof(uvcb_init);
 	uvcb_init.header.cmd = UVC_CMD_INIT_UV;
-	uvcb_init.stor_origin = mem;
+	uvcb_init.stor_origin = tmp;
 	uvcb_init.stor_len = uvcb_qui.uv_base_stor_len;
 
 	report_prefix_push("init");
@@ -528,14 +531,14 @@ static void test_init(void)
 	rc = uv_call(0, (uint64_t)&uvcb_init);
 	report(rc == 1 && (uvcb_init.header.rc == 0x104 || uvcb_init.header.rc == 0x105),
 	       "storage origin invalid");
-	uvcb_init.stor_origin = mem;
+	uvcb_init.stor_origin = tmp;
 
 	if (uvcb_init.stor_len >= HPAGE_SIZE) {
 		uvcb_init.stor_origin = get_max_ram_size() - HPAGE_SIZE;
 		rc = uv_call(0, (uint64_t)&uvcb_init);
 		report(rc == 1 && uvcb_init.header.rc == 0x105,
 		       "storage + length invalid");
-		uvcb_init.stor_origin = mem;
+		uvcb_init.stor_origin = tmp;
 	} else {
 		report_skip("storage + length invalid, stor_len < HPAGE_SIZE");
 	}
@@ -544,7 +547,7 @@ static void test_init(void)
 	rc = uv_call(0, (uint64_t)&uvcb_init);
 	report(rc == 1 && uvcb_init.header.rc == 0x108,
 	       "storage below 2GB");
-	uvcb_init.stor_origin = mem;
+	uvcb_init.stor_origin = tmp;
 
 	smp_cpu_setup(1, PSW_WITH_CUR_MASK(cpu_loop));
 	rc = uv_call(0, (uint64_t)&uvcb_init);
@@ -555,10 +558,12 @@ static void test_init(void)
 	rc = uv_call(0, (uint64_t)&uvcb_init);
 	report(rc == 0 && uvcb_init.header.rc == UVC_RC_EXECUTED, "successful");
 
-	mem = (uint64_t)memalign(1UL << 31, uvcb_qui.uv_base_stor_len);
+	tmp = uvcb_init.stor_origin;
+	uvcb_init.stor_origin =	(uint64_t)memalign_pages_flags(HPAGE_SIZE, uvcb_qui.uv_base_stor_len, AREA_NORMAL);
 	rc = uv_call(0, (uint64_t)&uvcb_init);
 	report(rc == 1 && uvcb_init.header.rc == 0x101, "double init");
-	free((void *)mem);
+	free((void *)uvcb_init.stor_origin);
+	uvcb_init.stor_origin = tmp;
 
 	report_prefix_pop();
 }
Signature · 1 line
2.34.1
#2 lore raw

Re: [kvm-unit-tests PATCH v3 1/9] s390x: uv-host: Fix UV init test memory allocation

Quoting Janosch Frank (2023-05-02 15:07:24)

Quoted 3 lines

The init memory has to be above 2G and 1M aligned but we're currently aligning on 2G which means the allocations need a lot of unused memory.

I know I already gave my R-b here, but...

s390x/uv-host.c static void test_config_create(void) · line 506

Quoted 13 lines
s390x/uv-host.c +1 -3
index 33e6eec6..9dfaebd7 100644
@@ -500,14 +500,17 @@ static void test_config_create(void)
 static void test_init(void)
 {
        int rc;
-       uint64_t mem;
+       uint64_t tmp;
 
-       /* Donated storage needs to be over 2GB */
-       mem = (uint64_t)memalign_pages_flags(SZ_1M, uvcb_qui.uv_base_stor_len, AREA_NORMAL);

...maybe out of coffee, but can you point me to the place where we're aligning to 2G here? I only see alignment to 1M and your change only seems to rename mem to tmp:

Quoted 5 lines
s390x/uv-host.c +5 -0
+       /*
+        * Donated storage needs to be over 2GB, AREA_NORMAL does that
+        * on s390x.
+        */
+       tmp = (uint64_t)memalign_pages_flags(SZ_1M, uvcb_qui.uv_base_stor_len, AREA_NORMAL);
#3 lore raw

Re: [kvm-unit-tests PATCH v3 1/9] s390x: uv-host: Fix UV init test memory allocation

Quoted 18 lines · Nico Boehr

Quoting Janosch Frank (2023-05-02 15:07:24)

Quoted 3 lines

The init memory has to be above 2G and 1M aligned but we're currently aligning on 2G which means the allocations need a lot of unused memory.

I know I already gave my R-b here, but...

Quoted 11 lines
s390x/uv-host.c +1 -1
index 33e6eec6..9dfaebd7 100644
@@ -500,14 +500,17 @@ static void test_config_create(void)
  static void test_init(void)
  {
         int rc;
-       uint64_t mem;
+       uint64_t tmp;
 
Quoted 11 lines

s390x/uv-host.c static void test_config_create(void) · line 506

Quoted 2 lines
s390x/uv-host.c +0 -2
-       /* Donated storage needs to be over 2GB */
-       mem = (uint64_t)memalign_pages_flags(SZ_1M, uvcb_qui.uv_base_stor_len, AREA_NORMAL);

...maybe out of coffee, but can you point me to the place where we're aligning to 2G here? I only see alignment to 1M and your change only seems to rename mem to tmp:

Quoted 4 lines
s390x/uv-host.c +4 -0
+       /*
+        * Donated storage needs to be over 2GB, AREA_NORMAL does that
+        * on s390x.
+        */

This comment explains it :)
Its a re-name of mem to tmp and an extension of this comment so it makes more sense.

Quoted 1 line
Quoted 1 line
s390x/uv-host.c +1 -0
+       tmp = (uint64_t)memalign_pages_flags(SZ_1M, uvcb_qui.uv_base_stor_len, AREA_NORMAL);
Reviewed-by
#4 lore raw

Re: [kvm-unit-tests PATCH v3 1/9] s390x: uv-host: Fix UV init test memory allocation

Quoting Janosch Frank (2023-05-10 11:03:54)

Quoted 26 lines · Nico Boehr
Quoted 25 lines

Quoting Janosch Frank (2023-05-02 15:07:24)

Quoted 3 lines

The init memory has to be above 2G and 1M aligned but we're currently aligning on 2G which means the allocations need a lot of unused memory.

I know I already gave my R-b here, but...

s390x/uv-host.c static void test_config_create(void) · line 506

Quoted 13 lines
s390x/uv-host.c +1 -3
index 33e6eec6..9dfaebd7 100644
@@ -500,14 +500,17 @@ static void test_config_create(void)
  static void test_init(void)
  {
         int rc;
-       uint64_t mem;
+       uint64_t tmp;
 
-       /* Donated storage needs to be over 2GB */
-       mem = (uint64_t)memalign_pages_flags(SZ_1M, uvcb_qui.uv_base_stor_len, AREA_NORMAL);

...maybe out of coffee, but can you point me to the place where we're aligning to 2G here? I only see alignment to 1M and your change only seems to rename mem to tmp:

Quoted 8 lines
Quoted 4 lines

s390x/uv-host.c static void test_config_create(void) · line 508

Quoted 4 lines
s390x/uv-host.c +4 -0
+       /*
+        * Donated storage needs to be over 2GB, AREA_NORMAL does that
+        * on s390x.
+        */

This comment explains it :)
Its a re-name of mem to tmp and an extension of this comment so it makes more sense.

Alright, thanks. I guess coffee level too low.

#5 lore raw

PATCH v3 2/9 s390x: uv-host: Check for sufficient amount of memory

In reply to Janosch Frank

The UV init storage needs to be above 2G so we need a little over 2G of memory when running the test.

---
 s390x/uv-host.c | 8 ++++++++
 1 file changed, 8 insertions(+)
s390x/uv-host.c +8 -0
index 9dfaebd7..34ca7148 100644
@@ -15,6 +15,7 @@
 #include <sclp.h>
 #include <smp.h>
 #include <uv.h>
+#include <snippet.h>
 #include <mmu.h>
 #include <asm/page.h>
 #include <asm/sigp.h>
@@ -720,6 +721,13 @@ int main(void)
 	test_invalid();
 	test_uv_uninitialized();
 	test_query();
+
+	if (get_ram_size() < SNIPPET_PV_MIN_MEM_SIZE) {
+		report_skip("Not enough memory. This test needs about %ld MB of memory",
+			    SNIPPET_PV_MIN_MEM_SIZE / 1024 / 1024);
+		goto done;
+	}
+
 	test_init();
 
 	setup_vmem();
Signature · 1 line
2.34.1
#6 lore raw

PATCH v3 3/9 s390x: uv-host: Beautify code

In reply to Janosch Frank

Fixup top comment and add missing space.

---
 s390x/uv-host.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
s390x/uv-host.c +2 -2
index 34ca7148..1c04d8ef 100644
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Guest Ultravisor Call tests
+ * Host Ultravisor Call tests
  *
  * Copyright (c) 2021 IBM Corp
  *
@@ -34,7 +34,7 @@ static struct uv_cb_csc uvcb_csc;
 
 extern int diag308_load_reset(u64 code);
 
-struct cmd_list{
+struct cmd_list {
 	const char *name;
 	uint16_t cmd;
 	uint16_t len;
Signature · 1 line
2.34.1
#7 lore raw

PATCH v3 4/9 s390x: uv-host: Add cpu number check

In reply to Janosch Frank

We should only run a test that needs more than one cpu if a sufficient number of cpus are available.

---
 s390x/uv-host.c | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)
s390x/uv-host.c +9 -5
index 1c04d8ef..ed13bc26 100644
@@ -550,11 +550,15 @@ static void test_init(void)
 	       "storage below 2GB");
 	uvcb_init.stor_origin = tmp;
 
-	smp_cpu_setup(1, PSW_WITH_CUR_MASK(cpu_loop));
-	rc = uv_call(0, (uint64_t)&uvcb_init);
-	report(rc == 1 && uvcb_init.header.rc == 0x102,
-	       "too many running cpus");
-	smp_cpu_stop(1);
+	if (smp_query_num_cpus() > 1) {
+		smp_cpu_setup(1, PSW_WITH_CUR_MASK(cpu_loop));
+		rc = uv_call(0, (uint64_t)&uvcb_init);
+		report(rc == 1 && uvcb_init.header.rc == 0x102,
+		       "too many running cpus");
+		smp_cpu_stop(1);
+	} else {
+		report_skip("Not enough cpus for 0x102 test");
+	}
 
 	rc = uv_call(0, (uint64_t)&uvcb_init);
 	report(rc == 0 && uvcb_init.header.rc == UVC_RC_EXECUTED, "successful");
Signature · 1 line
2.34.1
#8 lore raw

PATCH v3 5/9 s390x: uv-host: Fix create guest variable storage prefix check

In reply to Janosch Frank

We want more than one cpu and the rc is 10B, not 109.

---
 s390x/uv-host.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)
s390x/uv-host.c +6 -2
index ed13bc26..0048a19f 100644
@@ -435,11 +435,15 @@ static void test_config_create(void)
 	       "base storage origin contains lowcore");
 	uvcb_cgc.conf_base_stor_origin = tmp;
 
-	if (smp_query_num_cpus() == 1) {
+	/*
+	 * Let's not make it too easy and use a second cpu to set a
+	 * non-zero prefix.
+	 */
+	if (smp_query_num_cpus() > 1) {
 		sigp_retry(1, SIGP_SET_PREFIX,
 			   uvcb_cgc.conf_var_stor_origin + PAGE_SIZE, NULL);
 		rc = uv_call(0, (uint64_t)&uvcb_cgc);
-		report(uvcb_cgc.header.rc == 0x10e && rc == 1 &&
+		report(uvcb_cgc.header.rc == 0x10b && rc == 1 &&
 		       !uvcb_cgc.guest_handle, "variable storage area contains lowcore");
 		sigp_retry(1, SIGP_SET_PREFIX, 0x0, NULL);
 	}
Signature · 1 line
2.34.1
#9 lore raw

PATCH v3 6/9 s390x: uv-host: Switch to smp_sigp

In reply to Janosch Frank

Let's move to the new smp_sigp() interface which abstracts cpu numbers.

---
 s390x/uv-host.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
s390x/uv-host.c +2 -2
index 0048a19f..a9543593 100644
@@ -440,12 +440,12 @@ static void test_config_create(void)
 	 * non-zero prefix.
 	 */
 	if (smp_query_num_cpus() > 1) {
-		sigp_retry(1, SIGP_SET_PREFIX,
+		smp_sigp(1, SIGP_SET_PREFIX,
 			   uvcb_cgc.conf_var_stor_origin + PAGE_SIZE, NULL);
 		rc = uv_call(0, (uint64_t)&uvcb_cgc);
 		report(uvcb_cgc.header.rc == 0x10b && rc == 1 &&
 		       !uvcb_cgc.guest_handle, "variable storage area contains lowcore");
-		sigp_retry(1, SIGP_SET_PREFIX, 0x0, NULL);
+		smp_sigp(1, SIGP_SET_PREFIX, 0x0, NULL);
 	}
 
 	tmp = uvcb_cgc.guest_sca;
Signature · 1 line
2.34.1
#11 lore raw

PATCH v3 7/9 s390x: uv-host: Properly handle config creation errors

In reply to Janosch Frank

If the first bit is set on a error rc, the hypervisor will need to destroy the config before trying again. Let's properly handle those cases so we're not using stale data.

---
 lib/s390x/asm/uv.h |  1 +
 s390x/uv-host.c    | 66 +++++++++++++++++++++++++++++++++++++++-------
 2 files changed, 58 insertions(+), 9 deletions(-)
lib/s390x/asm/uv.h +1 -0
index 38920461..e9fb19af 100644
@@ -24,6 +24,7 @@
 #define UVC_RC_NO_RESUME	0x0007
 #define UVC_RC_INV_GHANDLE	0x0020
 #define UVC_RC_INV_CHANDLE	0x0021
+#define UVC_RC_DSTR_NEEDED_FLG	0x8000
 
 #define UVC_CMD_QUI			0x0001
 #define UVC_CMD_INIT_UV			0x000f
s390x/uv-host.c +57 -9

140 lines not on this page · raw message

Signature · 1 line
2.34.1
Reviewed-by on v3 7/9
#12 lore raw

Re: [kvm-unit-tests PATCH v3 7/9] s390x: uv-host: Properly handle config creation errors

Quoting Janosch Frank (2023-05-02 15:07:30)

Quoted 5 lines

If the first bit is set on a error rc, the hypervisor will need to destroy the config before trying again. Let's properly handle those cases so we're not using stale data.

I am not a huge fan of the naming of cgc_check_data(), but since I can't really come up with a better suggestion let's avoid the bike shedding and be pragmatic:

#13 lore raw

PATCH v3 8/9 s390x: uv-host: Fence access checks when UV debug is enabled

In reply to Janosch Frank

The debug print directly accesses the UV header which will result in a second accesses exception which will abort the test. Let's fence the access tests instead.

---
 s390x/uv-host.c | 10 ++++++++++
 1 file changed, 10 insertions(+)
s390x/uv-host.c +10 -0
index 4418e9b9..19bdd465 100644
@@ -165,6 +165,15 @@ static void test_access(void)
 
 	report_prefix_push("access");
 
+	/*
+	 * If debug is enabled info from the uv header is printed
+	 * which would lead to a second exception and a test abort.
+	 */
+	if (UVC_ERR_DEBUG) {
+		report_skip("Debug doesn't work with access tests");
+		goto out;
+	}
+
 	report_prefix_push("non-crossing");
 	protect_page(uvcb, PAGE_ENTRY_I);
 	for (i = 0; cmds[i].name; i++) {
@@ -197,6 +206,7 @@ static void test_access(void)
 	uvcb += 1;
 	unprotect_page(uvcb, PAGE_ENTRY_I);
 
+out:
 	free_pages(pages);
 	report_prefix_pop();
 }
Signature · 1 line
2.34.1
#14 lore raw

PATCH v3 9/9 s390x: uv-host: Add the test to unittests.conf

In reply to Janosch Frank

Better to skip than to not run it at all.

---
 s390x/unittests.cfg | 5 +++++
 1 file changed, 5 insertions(+)
s390x/unittests.cfg +5 -0
index d6e7b170..3bba03ce 100644
@@ -231,3 +231,8 @@ extra_params = -m 2200
 [pv-diags]
 file = pv-diags.elf
 extra_params = -m 2200
+
+[uv-host]
+file = uv-host.elf
+smp = 2
+extra_params = -m 2200
Signature · 1 line
2.34.1
#16 lore raw
In reply to Janosch Frank

Quoting Janosch Frank (2023-05-02 15:07:23)

Quoted 10 lines

The uv-host test has a lot of historical growth problems which have largely been overlooked since running it is harder than running a KVM (guest 2) based test.

This series fixes up smaller problems but still leaves the test with fails when running create config base and variable storage tests. Those problems will either be fixed up with the second series or with a firmware fix since I'm unsure on which side of the os/fw fence the problem exists.

Quoted 3 lines

The series is based on my other series that introduces pv-ipl and pv-icpt. The memory allocation fix will be added to the new version of that series so all G1 tests are fixed.

I have also pushed this to our CI, thanks.

Also here, I took the liberty of adding
groups = pv-host
in the last patch. If you are OK with it, I can carry that when picking for the PR.

#17 lore raw
Quoted 15 lines · Nico Boehr

Quoting Janosch Frank (2023-05-02 15:07:23)

Quoted 10 lines

The uv-host test has a lot of historical growth problems which have largely been overlooked since running it is harder than running a KVM (guest 2) based test.

This series fixes up smaller problems but still leaves the test with fails when running create config base and variable storage tests. Those problems will either be fixed up with the second series or with a firmware fix since I'm unsure on which side of the os/fw fence the problem exists.

Quoted 3 lines

The series is based on my other series that introduces pv-ipl and pv-icpt. The memory allocation fix will be added to the new version of that series so all G1 tests are fixed.

Quoted 6 lines

I have also pushed this to our CI, thanks.

Also here, I took the liberty of adding
groups = pv-host
in the last patch. If you are OK with it, I can carry that when picking for the PR.

Sure, since Thomas reviewed the patch introducing the group I'm fine with that.